Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
MIN-310
Taguchi’s Methodology
Six Sigma Methodology
Benchmarking
Six Sigma
Cost of poor quality Fig. 1 Cost of poor quality versus Sigma level
as % of earnings
30%
20%
10%
0%
3 4 5 6 7
Sigma Level
Mean Value
*Distribution shifted +/- 1.5
4
Sigma = = Deviation n 1
( Square root of variance )
1
2
3
5
6
7
0
4
between + / - 1 68.27 % 317300 ppm outside (deviation)
Contd….
Green Belts
Project Champions
Master Black
Black Belts
Belts
Fundamental Themes of
Six Sigma
• Six Sigma is based around a number of central
themes
– A genuine focus on the customer
– Data and fact driven Management
– Team based problem solving projects
– Process focus, management & improvement
– Proactive management
– Drive for perfection, tolerance for failure
Contd….
Fundamental Themes of
Six Sigma
• A genuine focus on the customer
– Six Sigma measures start with customer satisfaction
– Emphasis on understanding customer expectations and
requirements
Contd….
Fundamental Themes of
Six Sigma
• Process focus, management & improvement
– Understanding the process is the key aspect
– Control the inputs, and improve process
• Proactive management
– Understanding of Six Sigma principles
– Defining the root cause of problems
– Setting and maintaining aggressive improvement targets
– Challenging ‘why’ things are done this way
Contd….
Fundamental Themes of
Six Sigma
• Team based collaboration
– Six Sigma is teamwork focused
– Project lead by process expert
– Facilitated by Six Sigma Black Belt
– Multi-skill experts within project team
Six Sigma
Supporting Evidence
• General Electric started to use Six Sigma in 1995
• Total savings estimated to be $750 million by 1998
• Over 10,000 GE managers and employees have been trained in Six
Sigma methods
Six Sigma
Supporting Evidence
• Motorola established the first Six Sigma
improvement programme in 1987
• Used to track and compare customer requirements
• Over the ten years 1987-97
– Five-fold increase in sales
– Profits increasing at nearly 20% per year,
– Share price equal annual increase of over 21% per year
– Cumulative savings based on Six Sigma at over $17 billion
(1986-2004)
Design of
Experiments Balanced
Scorecards
Process Design/
Six Sigma
Redesign Voice of the
Customer
Continuous
Improvement Creative
Process Thinking
Management
Six Sigma
Improvement Strategy
Existing Product
• Phase 0: Process Definition
• Phase I: Process Measurement
• Phase II: Process Analysis
• Phase III: Process Improvement
• Phase IV: Process Control
23
DMAIC
Six Sigma Improvement Model
24
Six Sigma
Improvement Strategy
New Product
• Phase 0: Define
• Phase I: Measure
• Phase II: Analyse
• Phase III: Design
• Phase IV: Verify
25
Six Sigma
The four fundamentals
• Process … Everything Is Process
• Correlation … Y=f(x1, x2, x3)
• Discipline … DMAIC
• Data … All Learning Resides In It
26
27
DMAIC
28
DMAIC
29
DMAIC
30
DMAIC
31
DMAIC
32
• Many of the objectives of TQM are echoed within the Six Sigma
approach
• Focus on customers
• Team based activity
• Dependent on people
• Produces an improved business position
Differences
• Focused on hard ‘process based’ deliverables
• Precise definition and simple message
• Cross functionality of approach
• Well defined project time-scales
• Well defined project structure
• Well defined training objectives
34
Summary
• Six Sigma
- It is different
- It works if you follow the process
- Produces significant bottom line results
• Change is required by all persons who get
involved with the initiative
• Six Sigma will last as long as it produces bottom
line results.
35
Case Study
Bombay Tiffenwallas
36
28
37
Case Study
Bombay Tiffenwallas
(Source: Forbes Magazine)
Bombay Tiffanwallas (Raghunath Medge – president)
- They make
one Error on every 16 millions transactions
5000 Tiffanwallas deliver 2,00,000 lunches
everyday and take the empty tiffin back
- They make
one mistake in 2 month
2,00,000x2(transaction per day) x30 (days)=12,000,000
12,000,000x2 months= 24,000,000 (24 millions)
38
Case Study
Bombay Tiffenwallas
(Source: Forbes Magazine)
The charge for this service = Rs. 500/- pm
Tiffenwallah income= Rs 10,000/- pm
Takeaways are passion, commitment, consistency,
100% execution, accuracy, dedication, time
management and customer satisfaction
Superb!!!
Service and Excellence -----------
“Can anyone ask for more????”
39
THANKS!!!
40
41
GENICHI TAGUCHI
• Electrical Engineer
Taguchi’s Method
43
Quality
“Conformance to requirements”
“Customer satisfaction”
Taguchi on Quality
46
Total Cost
it includes the cost of problems in manufacturing
and the cost of problems in the field.
47
Concept of Quality
Loss
• Traditional concept:
– Good in the limit and bad
outside the limit Pass Fail
– Step function m m m
– loss off the limit
52
L(y) = (y-m)2
Ao = k (o)2
A
k 0
2
0
B A
L( y ) 0
{ y m}
2
2
0
m- m+
53
Example
Calculate the average loss per unit to a manufacturing company if
deviation is 0.00, 0.01, 0.02 when specifications on a part dimensions are
0.50 ± 0.02 cm. The cost for return/replacement is Rs. 50 for parts with
dimensions of 0.48 or 0.52 (i.e., 0.50 ± 0.02).
L(y) = (y-m)2, here tolerance is 0.02, i.e., (y-m) = (0.02)
L(y) = Rs. 50 or = 125,000
Loss function for a single part is L(y) = 125,000 (y-m)2
Example
• A company received an average of 10 complaints per month
last year. In November they received 15 complaints (y).
Management sets an acceptable level at 2 (tolerance).
Example Continued…
= ₹ 150/22 = ₹ 37.50
L(y) = * (y-m)2 m- m m+ y
R
• The larger, the better:
j1
Ao
Example: Strength of structure
y
L(y) = / y2
S/N 10log 1 1
R
R y
HB 2
j1
j
57
System Design: The focus of the system design phase is on determining the
suitable working levels of design parameters. It includes designing and testing a
system based on engineer's judgment of selected materials, parts and nominal
product/process parameters based on current technology. Thus it helps to
identify the working levels of the design parameters
Parameter design: It seeks to determine the parameter levels that produce the
best performance of the product/process under study. The optimum condition is
selected so that the influence of uncontrollable factors causes minimum
variation to system performance. The orthogonal arrays, variance and signal to
noise analysis are the essential tools of parameter design.
Tolerance design: It is a step to fine tune the results of parameter design by
tightening the tolerance of parameters with significant influence on the product.
60
Steps
• Identify the problem / response characteristic
• Conduct Brainstorming Sessions to identify the various factors
• Develop a Cause and Effect Diagram
• Classify the factors as control, noise, or signal
• Conduct Pareto Analysis for vital parameters (down sizing)
• Select the experimental range of factors
• Decide the Orthogonal Array for Control factors as well as for
noise factors
• Conduct experiments
• Analyze the results
• Find optimum setting (Robust Setting of Parameters)
• Confirm the results
61
Classification of Factors
Control Factors
Design factors that are to be set at optimal levels to improve quality and
reduce sensitivity to noise
Noise Factors
Factors that represent the noise that is expected in production or in use
Adjustment Factors
Affects the mean but not the variance of a response
Signal Factors
Set by the designer to communicate desires of the user
62
Number of Factors
LN (2k)
Total Number of Runs Number of Levels Per Factors
63
64
Analysis
65
Analysis (Cont’d)
Taguchi uses signal to noise (S/N) ratios as response variables.
Signal to Noise Ratio
• A single response which makes a tradeoff between setting the
mean to a desirable level while keeping the variance low.
• Always try to MAXIMIZE a S/N Ratio
• There are three types:
– Smaller is Better
– Target is Best
– Larger is Better
66
SN s 10 log 10 y 2j / n y
j
•The larger, the better:
Ao
1 2
SN l 10 log10 2 / n y
j y j
67
CASE STUDY
68
Experimentation Methodology
Work piece
Grit size
Slurry concentration
Power rating
Tool material
Parameter
S. No. Factors Level 1 Level 2 Level 3
Designation
1. Work piece A ASTM Gr. 2 ASTM Gr. 5 *
2. Grit Size B 220 320 500
3. Slurry Concentration C 25% 30% 35%
4. Power Rating D 40% 60% 80%
5. Tool E HCS HSS WC
The most appropriate array for this study was L18 (21*37) OA with
17 DOF for experimentation.
Assignment of factors has been done using the linear graph and
triangular table. The unassigned columns are treated as error.
-8.0
2.2
Surface Roughness
-6.0
μm
1.8
-4.0
1.4
-2.0
0.0 1
HCS
HSS
ASTM Gr 2
ASTM Gr 5
CC
220
320
500
25
30
35
40
60
80
Response Curves Showing Effect of Process Parameters on Surface Roughness
S/N Ratio Surface Roughness
Surface Roughness
Second level of grit size (B2 ) 320
First level of slurry concentration (C1 ) 25%
First level of power Rating* (D1 ) 40%
Second level of tool* (E2 ) High speed steel
First Level of work piece* (A1) ASTM Gr. 2
Summary
• The effect of work piece, grit size, slurry concentration, power
rating and tool as process parameters on the quality of hole
drilled by USD process were investigated and optimized.
• Titanium was chosen as work material.
• Taguchi’s L18 orthogonal array was used to design the
experiments.
• The parameters that were kept fixed during the entire
experimentation were frequency, static load and flow rate of
slurry.
• Analysis of variance was performed on raw data as well as S/N
data.
• The optimal result obtained was validated by conducting
confirmation experiments.
Parameter
S. No. Factors Designation Level 1 Level 2 Level 3
82
Benchmarking
Benchmarking
Even if you're on the right track, you'll get run over if you just sit there. --- Will Rogers
US humorist & showman
Benchmarking
Benchmarking
Benchmarking
Benefits of Benchmarking
Prevents reinventing the wheel (Why invest the time and
costs when someone else may have done it already--and
often better, cheaper, and faster?).
Benefits of Benchmarking
Benefits of Benchmarking
• Improves Strategic Planning
Levels of Benchmarking
Levels of Benchmarking
Levels of Benchmarking
Levels of Benchmarking
Levels of Benchmarking
• Internal Benchmarking
– Comparisons within the organizations
• Generic Benchmarking
– Comparisons to unrelated organizations known for product innovation
Phase 1: Planning
Phase 2: Analysis
M7 M3
M6 M4
Current Performance
M5 (Baseline)
104
Phase 3: Integration
Phase 4: Action
Phase 5: Maturity
Off-the-shelf benchmarks are often historic in nature and may not reflect
technological trends. Benchmarks must consistently be updated and
contextualized within their areas.
Better to select one or several processes that form a part of the total system,
work with it initially and then move on to the next part of the system.
Choosing a benchmarking topic that is not aligned with the overall strategy
and goals of the business; or worse, cuts across some other initiative
the organization is already taking.
Going out to make benchmarking visits before you have analyzed your
own process thoroughly.
Benchmarking assumes that you already know your own process and its
level of performance thoroughly. After all, that information is what
you have to offer to your benchmarking partners in exchange for the
information you are seeking from them. Make sure your
benchmarking team is very clear about what it wants to learn before
you approach potential benchmarking partners.
Thanks!!!