R4.aqua Sixsigma Idesigner - En.v3

Building a Community of SPI* Knowledge
*SPI … System, Software, Service, Safety, Security Process, Product Improvement, Innovation and Infrastructure
https://www.eurospi.net https://soqrates.eurospi.net https://academy.eurospi.net
Conference and book series since 1994.

Working parties of leading suppliers and experts since 2003.
Online Campus with ECQA since 2008 and own academy
portals since 2020.
Organized by
2
Automotive Quality Manager
(Integrated)
Six Sigma Stream
AQUA JRCo Job Role Committee

Release 4, Version 1.0
© ISCN Ges.m.b.H 2020, the slides are copyrighted by ISCN Ges.m.b.H, no reproduction
or distribution is allowed without written consent by ISCN Ges.m.b.H.
3
Unit 1: Introduction
Element 1: Integration view and general part
Subcomponent: Six Sigma
4
Value and Foundations of Six Sigma
Six Sigma
Focus on reducing variability
The goal of Six Sigma is to increase business profits by eliminating the variability, defects and waste
that undermine customer loyalty
Six Sigma is a rigorous and systematic methodology that utilizes

information (management by facts) and statistical analysis to
measure and improve a company's operational performance by
preventing 'defects' i.e. meet and exceed stakeholders
expectations
Origin:
1986 Motorola – Bill Smith & Mikel Harry: ‘DMAIC’
1996 General Electric – Jack Welch: ‘The GE Way’
© ISCN Ges.m.b.H 5
Process Capability & Performance
What does Six Sigma Capability mean?
What does “out of specification” mean?
Listen carefully to what the customer wants

Look carefully at your process capability
LSL USL
Voice of the Process

Voice of the customer
© ISCN Ges.m.b.H 6
Capable or Not?
LSL USL
Not Capable:
Too much variation
LSL USL
Capable:
But not centred
© ISCN Ges.m.b.H 7
Defects Defects Defects Defects
LSL USL LSL USL
Centre
Centre the Process on Target Few Defects Few Defects
Spread
Reduce Variation
LSL USL
© ISCN Ges.m.b.H 8
Inherent Process Variation

Part of the process variation, resulting from “common causes”
This type of variation can be estimated by averaging standard deviations of
subgroups (pooling).
Total Process Variation
This is the variation, resulting from “common causes” as well as from “special
causes”.
This variation can be estimated by calculating the standard deviation over all
measurements
Process Capability
This is the 6sWithin range of the inherent variation of the process
Process Performance
This is the 6sOverall range of the total variation of the process
© ISCN Ges.m.b.H 9
Short Term vs. Long Term Capability
Short term Capability
Long term Performance
A Process has a tendency to shift by approximately 1.5

© ISCN Ges.m.b.H 10
Capability Indices
Statisticians have developed two key indices for measuring Short-Term
Capability: Cp and Cpk
USL  LSL
Cp 
6 Within
x  LSL USL  x
CpL  CpU 
3   Within 3   Within
Cpk  min CpL, CpU 
Performance Indices
Statisticians have also developed two key indices for measuring the
performance (also known as Long-term Capability):
USL  LSL
Pp 
6 Overall
x  LSL USL  x
PpL  PpU 
3   Overall 3   Overall
Ppk  min PpL, PpU 
Cp Pp
The tolerance divided by the process The tolerance divided by the process
capability performance
tolerance tolerance
process capability process performance
6Within 6Overall
Cpk Ppk
The distance from the process average The distance from the process average
to the nearest tolerance to the nearest tolerance
limit divided by half of the process limit divided by half of the process
capability performance
x x
USL-x USL-x
3Overall
3Within
Capability vs. Performance
Process Capability : The “Potential” Capability of the process

(Cp and Cpk)
Process Performance : The processes “Overall” performance
(Pp and Ppk)
What the process can do (Cp) and what the process does (Ppk)
Six Sigma Process
Capability : Cp = 2.0
Performance : Ppk = 1.5
A Ppk=1.5 is needed to get under 3.4 ppm
 Ppk PPM
2 0.2 308537
3 0.5 66807
4 0.8 6210
5 1.2 233
6 1.5 3.4
Six Sigma
Focus on the reduction of variation, hence variability
LSL USL Defects: outside specification
3σ 2700 ppm
defects
LSL USL
6σ
4.5σ 3.4 ppm

defects
Why Six Sigma Level?

Loss in
σ level ppm % of sales Status
7σ 0.02 <5%
6σ 3.4 5-10% World Class
5σ 233 10-15%
4σ 6.210 15-20% Average
3σ 66.810 20-30%
2σ 308.537 30-40% Bankrupt
1σ 690.000 40-50%
Design for Six Sigma
DfSS - Design for Six Sigma

Focus on robust design
The goal of DfSS is to design products/processes that exceed customer expectations i.e. flawless
product launch and predictable reliability
DfSS is a systematic and rigorous methodology using tools,

training, and measurements to enable the design of new
products and processes that meet customer expectations at Six
Sigma quality levels.
Design for Six Sigma
DfSS - Design for Six Sigma

Focus on robust design
Traditional development
Design for Six Sigma / QFD
Number of changes
• Robust design
• Reliability
• FMEA
• QFD
• ...
20-24 months 14-17 months 1-3 months 1st day of

production
Lean Six Sigma (LSS)
ISO 13053-1 DMAIC Methodology
Scope
This part of ISO 13053 recommends the preferred or best practice for each of
the phases of the DMAIC methodology used during the execution of a Six Sigma
project.
It also recommends how Six Sigma projects should be managed and describes
the roles, expertise and training of the personnel involved in such projects.
It is applicable to organizations using manufacturing processes as well as service

and transactional processes.
ISO 13053-2 Tools and Techniques
Define Measure Analyze Improve Control
Lean Principles SIPOC Brainstorming 5S Housekeeping Statistical process control

Project Selection Process Flow Mapping Multi-voting, Kaizen Visual Factory
Quality Function Deployment Types of data Regression analysis 8D Problem Solving Control Plan
Project Management Sampling Correlation coefficient Theory of Constraint OCAP
Team formation Basic statistics Analysis of variance Total Productive Maintenance Lessons learned
Stakeholder analysis Visualisation of data Multivariate studies VSM - Future State Standardized work
Risk Analysis Distributions Attributes data analysis Work and Line Balancing Training deployment
Tollgate reviews Normality test Practical significance Flow Auditing
Project Charter Descriptive statistics Sample size Pull Poka Yoke
Cost of Poor Quality Central limit theorem Hypothesis testing SMED Change Management
DPMO / PPM calculation Basic probability concepts Confidence Intervals First Time Right Effective communication
Voice of the customer Measurement systems analysis Tests for means DOE - Full Factorials Documentation
Kano Metrology Tests variances DOE - Fractional Factorials
Takt time Process performance metrics Tests for proportions Response Surface Modelling
Cycle time Process capability studies Paired-comparison OFAT
Critical to Quality OEE analysis Chi square tests Design for Six Sigma
CTQ Flowdown Contingency tables Design for X
Non-parametric tests Tolerance Analysis
FMEA Reliability
Root Cause analysis
Waste identification
VSM - Current State
Gap analysis
Transformation
Maturity Levels of Processes
Initial (Level 1): no description of any process in the

organization;
Managed (Level 2): reactive only on customer Level 5:

demand, the process to respond to the customer Optimized
has been formalized;
Level 4: Quantitatively
Managed
Defined (Level 3): the processes of the whole
organization are defined Level 3: Defined
Level 2: Managed
Quantitatively Managed (Level 4): all the processes
of Level 3 are quantitatively managed with Level 1: Initial
indicators
Optimized (Level 5): the processes can be optimized

with the use of indicators.
Summary
• Six Sigma is a holistic methodology aiming at reducing the variability of processes

based on a statistical approach.
• In the context of automotive quality engineering, Design for Six Sigma is of
particular relevance.
• (Lean) Six Sigma has several different qualification levels which are needed in
order to fulfill specific roles in Six Sigma projects and organizations.
• The ISO 13053 covers large parts of Six Sigma.
• In Six Sigma, the process quality is measured by the process’s short-term and
long-term capabilities.
• These capability figures (Cp, Cpk, Pp, Ppk) indicate the process’s capability with
respect to a Six Sigma process.
• For a Six Sigma process, Cp = 2, Pp = 1,5.
Unit 1: Introduction
Element 2: Organisational readiness
25
Roles and Responsibilities
Six Sigma Organisation

There are different roles and responsibilities involved in the successful
implementation of Six Sigma.
Top management must drive and support Six Sigma
Champion
MBB
Black Belt Top Down
Green Belt
Orange Belt
Bottom up
Yellow Belt
There should be a broad support system throughout the whole organization

Six Sigma Organization
CEO
Six Sigma
Champion
Training BB and GB Business Unit Manager
>5 years experience
>20 projects
Master
>2 years experience Black belt
>4 – 6 projects / year
175k$ p. project
Black belt Black belt Black belt
Green Green Green Green Green Green

Belt Belt Belt Belt Belt Belt
Yellow belts
Mikel Harry & Schroeder - 2000

Summary
• Six Sigma qualification levels are indicated by colored belts.

• Key roles in Six Sigma organizations are the Champions, Master Black Belts, Black
Belts, Green Belts, Orange and Yellow Belts.
• In order to teach a specific qualification (belt) level, the teacher needs to be
qualified for a superior level.
Unit 2: Product development
Element 1: Lifecycle
29
DMADV Roadmap
Design for Six Sigma:

DMADV roadmap
Verify design
Identify functions. Generate and
Define the project performance.
select concepts.
Implement design.
D M A D V
Define Measure Analyze Design Verify
Develop design, test/optimize

Establish Voice Of Customer design components and total
design.
DMAIC Roadmap
DMAIC Roadmap
Assure that
Define the problem Analyze the process. improvements will
and the objectives. Define factors of influence. sustain.
D M A I C
Define Measure Analyze Improve Control
What do we need to improve? Identify and implement

Can we measure this? improvements.
Summary
• DMADV describes the life cycle of products and processes that still have to be
designed.
• DMAIC describes the life cycle of existing products and processes that shall be
improved.
• In each of the life cycle phases, specific tools from the LSS toolbox are relevant
for application.
Unit 3: Quality and safety management
Element 2: Hazard & risk management
33
Failure Mode & Effect Analysis (FMEA)
FMEA: Failure Mode & Effect Analysis
• Identifies and evaluates the potential failures of a product or process and

their effects.
• Identifies actions that minimalises the probability of potential failures
occurring.
• Documents the product or process.
• Reduces the probability that failures reach the customer.
• Increases customer satisfaction.
• Is a proactive action, not a re-active action.
Failure Mode & Effect Analysis (FMEA) Types
Examines Identifies
Concept of a Failures in the

System FMEA
product concept
Product FMEA
Design of the Failures in the
Design FMEA
components design
Concept of a Failures in the

Process FMEA
process process
Customer Concept Design Product Process

System FMEA Start
Design FMEA Production
Process FMEA
Cost reduction through FMEA
Cost of Failure
Detection
Prevention of failures during Detection at
during
concept / design phase the customer
manufacturing
€1.000.000
QFD FMEA FTA SPC
€100.000
€10.000
€1.000
€1 €100
Designer / Manufacturer Customer
Main steps of the FMEA
Structuring Make a tree-structure of the

STEP 1 the system product/process.
Write down functions and describe the

STEP 2 functional relations.
Functional analysis
Write down failures and describe

STEP 3 Failure analysis the Cause-Mode-Effect relations.
Write down available knowledge and

STEP 4 Action analysis evaluate the current situation.
Define and implement improvement

STEP 5 Optimization
actions. Evaluate the new situation.
Step 1: Structural Analysis
Objectives
• Overview of the analysed product/process
• Allow reuse of modules From: Apis IQ-RM
The system consists of individual system elements (SE) that are arranged
hierarchically for the description of the structural connections, for example the
hardware concept, or the process steps.
Depth of analysis:
• The detailing ends if failures are sufficient secured by actions
• With known and proven designs/processes a lesser detailing is necessary
• The lowest level of analysis are the characteristics of the components (the
characteristics level) or the process influence factors
Step 1: Structural Analysis: Product FMEA
Design FMEA
Component Design
Component
group
Component Design
Function
group Component
group
Product
Component
group
Function
group
Component
group
1 2 3
(effect) (failure mode) (cause)
Step 2: Functional Analysis
Purpose
• Recording of all quality requirements in a broader sense
• Verification against the customer requirements
• Which can be characteristics, component functions, properties, activities,
customer requirements
• The quality requirements are assigned to
structure elements
• Each structure element has at least one
function
• Description of functional relationships
• Basis for the failure analysis
From: Apis IQ-RM

Step 3: Failure Analysis
Identify potential failures

• The next step of a FMEA determines the possible failures for each function
(effect, failure modes and causes).
• Each function has at least one failure (no function, limited function,
exceeding function, intermittent function).
• A failure is linked to a function or characteristic.
• Do not describe a cause-effect relationship in a failure.
• The description must be clear to assign easily preventive
and detective actions.
• Remain realistic
Example
Step 4: Action Analysis
Action Analysis - Objectives

• Assigning existing and/or already established actions (controls) to the failures.
• Risk evaluation
Preventive Actions (controls)

• Facilitate in the product development phase the optimal system layout and/or the
characteristic layout so that the likelihood of the occurrence of the failure mode / cause is
very low.
• Facilitate in the process development phase the optimal process planning so that the
likelihood of the occurrence potential is very low.
Detective Actions (controls)
• By using detective actions, the potential failures are detected and/or the effectiveness of
the preventive controls are confirmed.
Risk Priority Number

Risk - Evaluations S x O x D = RPN
Severity – rating
Evaluates the severity of an effect from the customer point of view
Occurrence – rating
Evaluates the probability of a cause occurrence considering the preventive
actions (How strong are the preventive actions?)
Detection – rating
Evaluates the detection probability of a cause or failure mode (How strong are
the detection actions?)
Evaluation: Severity Rating
Severity is the score associated

with the effect of the failure mode.
Severity is a relative score

within the specific FMEA.
A reduction in the severity

score is only possible by
changing the design.
Severity can be estimated

using a Severity Rating table.
From: Apis IQ-RM
Evaluation: Occurrence Rating
The Occurrence score is a relative

score within the specific FMEA and
may not display the actual
probability of the presence of
the cause.
If the score cannot be

determined, the team cannot
agree with each other or there
are no preventive actions
(controls), then the score is 10!
From: Apis IQ-RM
Evaluation: Detection Rating
The detection score is a relative

score within the specific FMEA
and may not display the actual
probability of the presence of
the cause.
If the score cannot be

determined, the team cannot
agree with each other or there
are no detective action
(controls), then the score is 10!
From: Apis IQ-RM
Step 5: Optimisation
Objectives
• Identify actions necessary for improvement
• Assessment of the risk (evaluate new situation)
• Check the effectiveness of the implemented actions
• Document the implemented actions
Working Order
• Concept modification to eliminate the failure cause and/or to obtain a
failure mode with a low severity
• Improve reliability of components to minimise the occurrence potential of the
failure cause
• Ensure effective detection of the failure causes / failure modes
• Severities can only be improved by changing the system and go beyond the
scope of your FMEA.
AIAG-VDA FMEA Alignment (since 2019)
• Automotive suppliers to both North American and German OEMs were required
to assess their failure modes and effects differently
• Differences between the Severity, Occurrence, and Detection rating tables in
the AIAG and VDA FMEA Manuals
• Caused confusion and added complexity to product development and process
improvement activities
• Alignment was needed in order to create a common set of requirements so
suppliers can have a single FMEA business process meeting needs and
expectations of any of their automotive customers
AIAG-VDA FMEA: New Unified, Aligned Approach & Handbook
[AIAG/VDA FMEA
© ISCN Ges.m.b.H
Handbook] 50
AIAG-VDA FMEA: 7 Step Approach
Applies to DFMEA, and PFMEA
[AIAG/VDA FMEA Handbook]
AIAG-VDA FMEA: 7 Step Approach
[AIAG/VDA FMEA
© ISCN Ges.m.b.H
Handbook] 52
AIAG-VDA FMEA: Step 1: Planning & Preparation
[QSG 2019]
AIAG-VDA FMEA: Step 2: Structure Analysis
DFMEA PFMEA
• Identification of design interfaces, • Identification of process steps and sub-
interactions, close clearance steps
• Tools: • Tools:
• Structure tree • Structure tree
• Block diagram • Process flow diagram
• Boundary diagram

AIAG-VDA FMEA: Step 3: Function Analysis
DFMEA PFMEA
• Association of requirements to • Association of characteristics to
functions functions
• • Tools: • Tools:
• Function analysis tree • Function analysis tree
• Parameter diagram • Parameter diagram

AIAG-VDA FMEA: Step 4: Failure Analysis
DFMEA PFMEA
• Potential Failure Effects, Failure Modes, • Potential Failure Effects, Failure Modes,
Failure Causes for each product Failure Causes for each process
function function
• Tools: • Tools:
• Parameter diagram • Fishbone diagram (4M)

AIAG-VDA FMEA: Step 5 – Risk Analysis
DFMEA PFMEA
• Assignment of Prevention Controls to • Assignment of Prevention Controls to
Risk Causes Risk Causes
• Risk Ratings (Sev, Occ, Det) • Risk Ratings (Sev, Occ, Det)
• Evaluation of Action Priority • Evaluation of Action Priority

AIAG-VDA FMEA: Step 6: Optimization
DFMEA PFMEA
• Assignment of responsibilities and due • Assignment of responsibilities and due
dates dates
• Implementation of actions • Implementation of actions

AIAG-VDA FMEA: Step 7: Results Documentation
• Communicate results and conclusions of the analysis

• Within organization
• With customers and/or suppliers (as appropriate)
• Document actions taken including effectiveness
AIAG-VDA FMEA: DFMEA Severity Table
[AIAG/VDA FMEA
Handbook]
AIAG-VDA FMEA: DFMEA Occurrence Table (1/2)
[AIAG/VDA FMEA
Handbook]
AIAG-VDA FMEA: DFMEA Occurrence Table (2/2)
[AIAG/VDA FMEA
© ISCN Ges.m.b.H Handbook] 62
AIAG-VDA FMEA: DFMEA Detection Table
[AIAG/VDA FMEA
Handbook]
AIAG-VDA FMEA: PFMEA Severity Table (1/2)
[AIAG/VDA FMEA
Handbook]
AIAG-VDA FMEA: PFMEA Severity Table (2/2)
[AIAG/VDA FMEA
AIAG-VDA FMEA: PFMEA Occurence Table
[AIAG/VDA FMEA
AIAG-VDA FMEA: PFMEA Detection Table (1/2)
[AIAG/VDA FMEA
Handbook]
AIAG-VDA FMEA: PFMEA Detection Table (2/2)
[AIAG/VDA FMEA
© ISCN Ges.m.b.H
Handbook] 68
AIAG-VDA FMEA: FMEA Action Priority
• High (H):
• Required to identify appropriate action to improve Prevention and/or
Detection Controls; OR justify and document why current controls are
adequate
• Priority Medium (M):
• Should identify appropriate actions to improve prevention and/or detection
controls; OR, at the discretion of management, justify and document why
current controls are adequate
• Priority Low (L):
• Could identify actions to improve prevention or detection controls
AIAG-VDA FMEA: FMEA Action Priority Table (1/2)
[AIAG/VDA FMEA
Handbook]
AIAG-VDA FMEA: FMEA Action Priority Table (2/2)
[AIAG/VDA FMEA
Handbook]
AIAG-VDA FMEA: Key Differences: Structure Analysis
AIAG 4th Ed FMEA New AIAG-VDA 1st Ed FMEA
[QSG 2019]
AIAG-VDA FMEA: Key Differences: DFMEA Function Analysis
[QSG 2019]
AIAG-VDA FMEA: Key Differences: DFMEA Failure Analysis
[QSG 2019]
AIAG-VDA FMEA: Key Differences: DFMEA Risk Analysis
[QSG 2019]
Removed – special characteristic identification not required in DFMEA; can use Filter Code column (optional)
AIAG-VDA FMEA: Key Differences: PFMEA Structure Analysis
[QSG 2019]
AIAG-VDA FMEA: Key Differences: PFMEA Function Analysis
[QSG 2019]
AIAG-VDA FMEA: Key Differences: PFMEA Failure Analysis
[QSG 2019]
AIAG-VDA FMEA: Key Differences: PFMEA Risk Analysis
[QSG 2019]
AIAG-VDA FMEA: Key Differences: DFMEA Optimization
[QSG 2019]
Summary
• The FMEA is the key tool for carrying out hazard and risk analysis in Six Sigma.
• The FMEA is very widely established in industry, and used extensively both for
design and manufacturing.
• The FMEA is carried out in five (new from 2019: seven) main steps in
interdisciplinary experts teams supported by specialized IT tools.
• In 2019, an alignment of the VDA-type FMEA with the AIAG-type FMEA has been
performed, leading to the new, harmonized AIAG-VDA FMEA.
• The AIAG-VDA FMEA is more detailed, specialized (DFMEA, PFMEA, FMEA-MSR),
and the RPN has been replaced by an AP (action priority) number.
Unit 2: Product Development
Element 2: Requirements
82
Customer Identification
Who and What is driving Six Sigma Projects?
External customers: VOC : Voice Of Customer
Identify the customer (stakeholder).

Customer demands, wishes & satisfaction
Quality improvement / defect (ppm) reduction
Internal customers: VOB : Voice Of Business
Organizational goals
Identify internal stakeholders, who are they?
Prevention / decrease of COPQ
Social relevance / Corporate Social Responsibility (CSR)
Customer Requirements
Voice of Customer
Customer satisfaction is important at all levels
Research has demonstrated that:

• the loss of relatively low amounts of money will make customers refrain from buying the same
product again;
• when supplying a bad product, the name of the supplier was mentioned twice as often as
when a good product was supplied.
Sales
Marketing Surveys
Audits
Interviews
Focus groups
VOC
Management Complaints
systems
Critical Requirements
Critical to Satisfaction
Project results are often linked to a particular area therefore the following ‘Critical to Satisfaction’
terms are often used:
CTQ – Critical to Quality Focus on Quality  Six Sigma
CTD – Critical to Delivery Focus on Delivery  Lean
CTC – Critical to Cost Focus on Costs
Sometimes we also see the following terms:

Quality Cost Delivery
CTP – Critical to Process
CTS – Critical to Safety
Safety Moral
CTQ Flowdown
Voice of Customer and External CTQ
Customer requirements are often expressed in ‘spoken language’ (VOC)

It is our task to interpret this into a specification that can be measured. This is called the CTQext
VOC CTQext
VOC : Customer requirements CTQext : Customer specification
• From standpoint of customer • Product characteristic

• Vague / Soft • Measurable / Hard
• Ambiguous • Specific / Unambiguous
CTQ Flowdown
Voice of Customer and External CTQ
Examples:
VOC CTQext
VOC : Customer requirements CTQext : Customer specification
• On Time Delivery • 14 days between order and delivery

• Accurate • Maximum tolerance ±2%
• No damage • No scratches / Ra=0.01
• Good process • Yield >98%
CTQ Flowdown
Translate CTQext into Internal CTQ (CTQint) measures

CTQint
Identify Internal Measures from the process perspective
CTQint
It is critical to achieve external CTQ’s CTQext
CTQint
VOC
It can be measured
CTQint
A specification can be set to tell if CTQ has been achieved
CTQ Flowdown
Having warm coffee  VOC

and friendly service
Quality of
Time Quality of service
CTQ product CTD CTQ
Strength of Warm Waiting Friendly General  CTQext

coffee coffee time service satisfaction
Strength Friendly Satisfaction

Temperature of Waiting time Waiting after
judgment
(1-6)
coffee taking order placing order
judgment judgment
(1-6)
 CTQint
(1-6)
2.0 to 3.0 > 70 °C <2 min. < 2 min. > 4.0 > 4.0
Summary
• The VOC is of key importance for any Six Sigma project as it helps identify the
customer’s expectations.
• The VOC has to be coordinated with the VOB.
• The CTQ flowdown is a tools helping to make customer expectations clear and
quantifiable.
• The CTQ flowdown translates customer requirements into internal requirements
which can be measured.
Element 3: Design
91
Design of Experiments (DOE)
• Design of Experiments (DOE) is a systematic and highly efficient way to conduct

experiments, to examine the influence of factors and interactions on responses.
• The DOE approach is based on the pioneering work of Sir Ronald Fisher, who
applied DOE in agriculture in the early 1920’s. He improved the productivity of
the British farm by using the Full Factorial method. For this great contribution he
was knighted. Some 70 years later, Fisher's method, now known as ‘Design Of
Experiments’, has become an important tool for engineers and researchers.
DOE vs. ‘Normal’ Tests
What makes a DOE so different from a ‘normal’ test?
Some ‘Classic Problem Solver’ experts will tell you to change only one factor at a time . . . (OFAT)
C C C
- + -
A A A
+
B B - B
. . . Well Don’t !!
DOE vs. ‘Normal’ Tests
OFAT – "One Factor at a Time"
• OFAT is a way to conduct experiments but it is not as efficient as a well- designed and well-
planned DOE
The Yield Contours

are unknown 75
Trial Temp Pressure Yield
1 125 30 74
2 125 31 80
80 3 125 32 85
4 125 33 92
Temperature (°C)
135 85 5 125 34 86
6
130
6 130 33 85
1 3
90 7 120 33 90
2 4 5
125
95
120 7
Optimum found using
OFAT
30 31 32 33 34 35
True Optimum that can
Pressure (mbar) be found using DOE
2k Full Factorials
DOE reviews the design and evaluation of Full Factorial and Fractional Factorial
experiments.
More Factors
Factorial designs allow for the simultaneous study of the effects that several factors may have on a
process.
More Efficient
Varying the levels of the factors simultaneously rather than one at a time is more efficient in terms of
time and cost.
Interactions
It also allows for the study of interactions between the factors. Without DOE, interactions would
remain undetected.
2k Full Factorials
Basic DOE principles
Assume you have two input variables (Factors A & B) that have an influence on the output
(Response Y).
To test these, you can vary them one-by-one, or test all possible combinations which is even better.
-1, +1 +1, +1
There are four possible combinations
They form the corners of a ‘Square Plot’

B
The two factors levels are displayed with
the code values -1 en +1 (low and high)
-1, -1 A +1, -1
2k Full Factorials
Definitions: Responses, Factors and Levels

Example: let’s examine Temperature and Pressure as factors of influence
on the Response ‘Reaction Time’
Low level High level

A: Temperature 400°C 800°C
B: Pressure 1bar 2bars
400, 2 800, 2
The Square Plot shows all four

combinations
B
Each combination (cell / run) has it’s own
response yi
© ISCN Ges.m.b.H 400, 1 A 800, 1 97
2k Full Factorials
Definition: Main Effects
 The response for a combination

is yi y3 y4
Main Effect of A :
The effect of a factor is called a
 y2  y4 y1  y3 1
Main Effect.   ( y1  y2  y3  y4 )
B 2 2 2
The Main Effect is the average Main Effect of B :

response at level +1 minus the y3  y4 y1  y2 1
A   ( y1  y2  y3  y4 )
average response at level -1 y1 y2
2 2 2
The Main Effect is calculated as:
2k Full Factorials
Exercise
Draw Cube Plot
Create Factorial Design Matrix
List Responses in Factorial Design Matrix
Calculate the Main Effect for each factor
A: Temperature, B: Pressure
level -1 +1
Factor A 400 800
Factor B 1 2
Responses:
y1= (-1,-1) = 2
y2= (+1,-1) = 4
y3= (-1,+1) = 1
y4= (+1,+1) = 6
2k Full Factorials
Answer:
y3 = 1 Y4 = 6
Run Temp Pressure A B Time

1 400 1 -1 -1 2
2 800 1 1 -1 4
B
3 400 2 -1 1 1
4 800 2 1 1 6
y1 = 2 A y2 = 4
Main Effect of A : Main Effect of B :

y2  y4 y1  y3 1 y3  y4 y1  y2 1
  ( y1  y2  y3  y4 )   ( y1  y2  y3  y4 )
2 2 2 2 2 2
4  6 2 1 1 1 6 2  4 1
  (2  4  1  6)  3.5   (2  4  1  6)  0.5
2 2 2 2 2 2
2k Full Factorials
Definition: Interactions
The combined effect of A and B on the Result Y is called the AB interaction
The Interaction Effect is calculated as follows:
Run Temp Pressure A B AB Time

1 400 1 -1 -1 1 2
2 800 1 1 -1 -1 4
y3 y4
3 400 2 -1 1 -1 1
4 800 2 1 1 1 6
Half the difference in effect of A on both B levels:

B
1
2
  1
ABhigh  ABlow  (y 4  y3 )  (y2  y1 )
2
Interaction Effect of AB:
y1 A y2
1
(y 4  y3 )  (y2  y1 )  1 (y1  y2  y3  y 4 )
© ISCN Ges.m.b.H 2 2 101
2k Full Factorials
Definition: Interactions
Some Interaction No Interaction Full Reversal
High High High
B- B-
B-
Y B+
B+ Y B+ Y
B+
Low Low Low
- A + - A + - A +
Strong Interaction Moderate Reversal
High High
B- B-
Y Y
B+
B+ B+
Low Low
- A + - A +

2k Full Factorials
Definition: Level Designs
2-Level Design : 2k
Experiment with all factors set at two levels
Low (-1) and High (+1)
3-Level Design : 3k
Experiment with all factors set at three levels

Low (-1), Medium (0) and High (+1)
Using a 3-level design allows you to evaluate the quadratic effects of the model.
3-level designs require more experiments than 2-level designs.
k = number of Factors
The number of factors in your model is the number of "effects" you wish to include in your
model. (Interactions not included)
Example: Your model consists of three factors: temperature, pressure, and humidity.

2k Full Factorials
Definition: Transfer Function
The outcome is a mathematical model of your experiment
Y = a + b×A + c×B + d×C + e×AB + f×AC + g×BC + h×ABC + Error
First order Second order Third order

effect effect effect
Main Effects Interaction

Fractional Factorials
Why Fractional Designs?

• A Fractional Factorial DOE is ‘smaller’ than a Full Factorial, meaning it require
less runs.
• Number of runs for a Full Factorial (without Centre Points): 2k.
• You can imagine that if the number of factors is increasing, the number of
runs will become unpractical and expensive.
#Factors #Runs
2 4
3 8
4 16
5 32
6 64
7 128
Resolution
• Resolution indicates how much the design is confounded.
• When you run a fractional factorial design, one or more of the effects are confounded. The
effects cannot be estimated independently
• The lower the resolution, the more confounding
• Resolution III, IV, and V designs are particularly important.

Example:
5 factor experiment with Resolution 3, requires 8 runs.
Notice that for a 5 factor experiment we have two possible Fractional Factorial Designs available
Summary
• DOE is a highly efficient method to plan experiments in a way that with a

minimum amount of statistical tests, a maximum amount of information about
the process can be collected.
• DOE assumes that the identified process can be described with a specific system
transfer function.
• The use of DOE is strongly supported by common statistical tools.

Element 4: Test and Integration
109
Testing
Purpose
• To optimize the settings of the design parameters to achieve the best possible responses or
characteristics of a product.
• To investigate the sensitivity for parameter variation (robustness).
• To demonstrate preliminary capabilities which meets the customer requirements
• To demonstrate that the product meets lifetime requirements
Tools
• Design of Experiments (DOE)
• Capability Analysis
• Lifetime testing

Design of Experiments
Optimize design parameters

• DOE is a tool to determine in a very efficient way optimal parameter settings
• DOE unveils also the interactions between the parameters
• DOE is much better than the “One factor at a time” approach
• DOE determines the Y=f(X) transfer function.
Tools
• Full 2-level factorials with center points
• Fractional 2-level factorials with center points
• Central Composite and Box-Behnken designs for non-linear relations

Design of Experiments
Sensitivity analysis empirical

• Build products and test them using DOE
• Use DOE to show that the variability's of components and processes are small
enough to assure good products.
• Use DOE to determine realistic tolerances of components and specifications
of process parameters.
Sensitivity analysis - simulation

• Use the transfer function to simulate the sensitivity’s
• The parameters are randomly generated from known or assumed probability
distributions

Capability Analysis
Preliminary Capability
• Determine the capability of the design and process before start of
production.
• Preliminary Process Capability Studies are short term studies conducted to
obtain early information on the performance of new or revised processes
relative to internal or customer requirements.
• In many cases, preliminary studies should be conducted at several points in
the evolution of new processes (e.g., at the equipment or tooling
subcontractor's plant, after installation at the supplier's plant)
• These studies should be based on as many measurements as possible.
• Preferable measurements of at least 20 subgroups of 2 – 5 parts

Lifetime Testing
Demonstrate lifetime
• A product is expected to perform its functions during, with the customer
agreed, lifetime.
• By performing lifetime (reliability) tests under conditions based on usage
profile, it is possible to demonstrate, with specified confidence, that an
agreed percentage of the products will survive the specified time.
• A test can last very long.
• Accelerated life testing (ALT) can reduce the needed testing time
• ALT is the process of testing a product by subjecting it to conditions (stress,
strain, temperatures etc.) in excess of its normal service parameters.
• ALT will uncover faults and potential modes of failure in a shorter amount of
time.

Lifetime Testing Strategy
Life Time Testing Strategies
• Test until all product have failed

• Test until specified number of products have failed
• Test until specified time T
• Allow zero failure in test
• Allow specified number of failures in test
• Testing when failure probability density function is known
• Testing when failure probability density function is unknown
• Test reduced or extended time
• Accelerated lifetime testing

Lifetime Testing
Reliability analysis is the tool to
• Identify the best fitting “Failure probability density function” and the related
Reliability function (e.g. Exponential, Weibull, Lognormal,..)
• Determine the sample size for the test
• Demonstrate Reliability at time T
• Determine ‘Mean time to failure (MTTF)”
• Quantify the confidence of the estimated values

Lifetime Testing
Reliability test
How many products have to…

Be tested for how long to…….
Prove that
With a confidence of C %
R % of the population……
Will keep functioning during time T
R95C90 at time T
At time T, 95% of the products will still function
The confidence of this statement is 90%

Lifetime Testing
Problem: we want to know the time till failure of parts
e.g.
- what is the mean time till failure?
- what is the probability that an item fails before a specified
time?
If a product lasts for many years, how do you quickly get an idea of the failure time?
accelerated life testing:

Compressed-time testing: product is tested under usual conditions
but more intensively than usual (e.g. a washing machine used
almost continuously)
Advanced-stress testing: product is tested under harsher conditions
than normal so that failure happens soon (e.g. refrigerator motor run
at a higher speed than if operating within a fridge). - requires some
assumptions

Lifetime Testing
How do you deal with items which are still working at the end of the test programme?
An example of censored data.
– we do not know all the failure times at the end of the test
Exponential data (failure rate 𝜈 independent of time)
Test components up to a time 𝑡0
- assuming a rate, can calculate probability of no failures in 𝑡0 .
- calculate probability of getting any set of failure times (and non failures by 𝑡0 )
- find maximum-likelihood estimator for the failure rate in terms of failure times
For failure times 𝑡𝑖 , with 𝑡𝑖 = 𝑡0 for parts working at 𝑡0 , and 𝑛𝑓 failures

𝑛𝑓
⇒ Estimate of failure rate is 𝜈Ƹ =
σ𝑖 𝑡𝑖
Reliability Function and Failure Rate
For a probability distribution function (pdf) 𝑓(𝑥) for the time till failure, define:
Reliability function
Probability of surviving at least till age 𝑡. i.e. that failure time is later than 𝑡
∞
𝑅 𝑡 = 𝑃 𝑇 > 𝑡 = න 𝑓 𝑥 𝑑𝑥 = 1 − 𝐹(𝑡)
𝑡
𝑡
𝐹 𝑡 = ‫׬‬0 𝑓 𝑡 𝑑𝑡 is the cumulative distribution function.
Failure rate
𝑓 𝑡
𝜙 𝑡 = This is failure rate at time 𝑡 given that it survived until time 𝑡:
𝑅 𝑡

Reliability Function and Failure Rate
Example: Find the failure rate of the Exponential distribution

Answer:
∞
The reliability is 𝑅 𝑡 = ‫ 𝑒𝜈 𝑡׬‬−𝜈𝑥 𝑑𝑥 = 𝑒 −𝜈𝑡
𝑓 𝑡 𝜈𝑒 −𝜈𝑡
Failure rate, 𝜙 𝑡 = 𝑅 = =𝜈 Note: 𝜈 is a constant
𝑡 𝑒 −𝜈𝑡
The fact that the failure rate is constant is a special “lack of

ageing property” of the exponential distribution.
However, often failure rates actually increase with age.

The Weibull Distribution
A way to model failure rates that are not constant
Failure rate: 𝜙 𝑡 = 𝑚𝜈𝑡 𝑚−1
Parameters 𝑚 (shape parameter) and 𝜈 (scale parameter)
𝑚 = 1: failure rate constant, Weibull=Exponential
𝑚 > 1: failure rate increases with time
𝑚 < 1: failure rate decreases with time

The Weibull Distribution
Failure rate: 𝜙 𝑡 = 𝑚𝜈𝑡 𝑚−1

𝑡
⇒ ln 1 − 𝐹 𝑡 = − න 𝑚𝜈𝑥 𝑚−1 𝑑𝑥 = −𝜈𝑡 𝑚
0
𝑚
⇒ 𝐹 𝑡 = 1 − 𝑒 −𝜈𝑡
𝑚
Reliability: 𝑅 𝑡 = 𝑒 −𝜈𝑡
𝑑𝐹 𝑡 𝑚
Pdf: 𝑓 𝑡 = = 𝑚𝜈𝑡 𝑚−1 𝑒 −𝜈𝑡
𝑑𝑡

Summary
• Design of Experiments is an important tool for testing in Six Sigma.

• Determining the reliability of parts and products is an important objective of
lifetime testing.
• The Weibull distribution function is often used to model the failure rate over time.

References
International Organization for Standardization (ISO): ISO 13053. Quantitative methods in process
improvement — Six Sigma (2011)
H.C. Theisens, Climbing the mountain mindset - skill set and tool set: Lean Six Sigma Green Belt. LSSA B.V.,
2016.
[AIAG/VDA FMEA Handbook] AIAG & VDA FMEA Handbook, 2019
[QSG] Laura Halleck, Failure Modes and Effects Analysis (FMEA)New 7 Step Approach!,
QualitySupportGroup, 2019

Reference to Authors
This Training Material has been certified according to the rules of ECQA –
European Certification and Qualification Association and is taught by:.
VSB – Technical University of Ostrava, Faculty of Electrical Engineering and Computer Science, Czech Republic, www.fei.vsb.cz/en
Graz University of Technology, Institute for Technical Informatics, Austria, www.iti.tugraz.at/en
University of Applied Sciences JOANNEUM, Austria, fh-joanneum.at/en
Grenoble Alps University, Institute of Engineering (Grenoble INP), France, www.grenoble-inp.fr/en
Hochschule Düsseldorf, Germany, hs-duesseldorf.de/en
University of Maribor, Slovenia, www.um.si/en/
ISCN Ges.m.b.H, Austria, www.iscn.com
ECQA – European Certification and Qualification Association, Austria, www.ecqa.org
This project (2015 – 2017) has been co-funded by the Erasmus+ Programme of the European Union - 2015-1-CZ01-KA203-013986 -
2015 – 2017.
AQUA (2013 – 2014) – Knowledge Alliance for Training Quality and Excellence in Automotive has built up quality alliance – LLP project
EAC-2012-0635 - 2013 – 2014.
This publication reflects the views only of the authors, and the Commission cannot be held
responsible for any use which may be made of the information contained therein.

Certification
By ECQA European Certification and Qualification
Association
www.ecqa.org
ECQA Certification is based on standard procedures,

attendees need to register for the exam.
127

R4.aqua Sixsigma Idesigner - En.v3

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

R4.aqua Sixsigma Idesigner - En.v3

Uploaded by

Copyright:

Available Formats

Building a Community of SPI* Knowledge

https://www.eurospi.net https://soqrates.eurospi.net https://academy.eurospi.net

Conference and book series since 1994.

AQUA JRCo Job Role Committee

Subcomponent: Six Sigma

Six Sigma is a rigorous and systematic methodology that utilizes

What does Six Sigma Capability mean?

What does “out of specification” mean?

Listen carefully to what the customer wants

Voice of the Process

Defects Defects Defects Defects

LSL USL LSL USL

Inherent Process Variation

Short Term vs. Long Term Capability

Short term Capability

Long term Performance

A Process has a tendency to shift by approximately 1.5

Cpk  min CpL, CpU 

Ppk  min PpL, PpU 

Capability vs. Performance

Process Capability : The “Potential” Capability of the process

Six Sigma Process

A Ppk=1.5 is needed to get under 3.4 ppm

4.5σ 3.4 ppm

Why Six Sigma Level?

DfSS - Design for Six Sigma

DfSS is a systematic and rigorous methodology using tools,

DfSS - Design for Six Sigma

20-24 months 14-17 months 1-3 months 1st day of

It is applicable to organizations using manufacturing processes as well as service

Define Measure Analyze Improve Control

Lean Principles SIPOC Brainstorming 5S Housekeeping Statistical process control

Initial (Level 1): no description of any process in the

Managed (Level 2): reactive only on customer Level 5:

Optimized (Level 5): the processes can be optimized

• Six Sigma is a holistic methodology aiming at reducing the variability of processes

Subcomponent: Six Sigma

Six Sigma Organisation

There should be a broad support system throughout the whole organization

Black belt Black belt Black belt

Green Green Green Green Green Green

Mikel Harry & Schroeder - 2000

• Six Sigma qualification levels are indicated by colored belts.

Subcomponent: Six Sigma

Design for Six Sigma:

Develop design, test/optimize

What do we need to improve? Identify and implement

Subcomponent: Six Sigma

FMEA: Failure Mode & Effect Analysis

• Identifies and evaluates the potential failures of a product or process and

Concept of a Failures in the

Concept of a Failures in the

Customer Concept Design Product Process

Designer / Manufacturer Customer

Structuring Make a tree-structure of the

Write down functions and describe the

Write down failures and describe

Write down available knowledge and

Define and implement improvement

From: Apis IQ-RM

Identify potential failures

Action Analysis - Objectives

Preventive Actions (controls)