3. Six Sigma
6 sigma is used by individual and
organizations to:
•Drive and sustain improvements
•Provide rigorous alignment of actions
with strategy
•Guide decision making with facts and
data
•Meet customer needs through improved
products and processes
•Deliver bottom-line results
8. The Six Sigma Evolutionary Timeline
1736: French
mathematician
Abraham de
Moivre publishes
an article
introducing the
normal curve.
1896: Italian sociologist Vilfredo
Alfredo Pareto introduces the 80/20
rule and the Pareto distribution in
Cours d’Economie Politique.
1924: Walter A. Shewhart introduces
the control chart and the distinction of
special vs. common cause variation as
contributors to process problems.
1941: Alex Osborn, head of
BBDO Advertising, fathers a
widely-adopted set of rules for
“brainstorming”.
1949: U. S. DOD issues Military
Procedure MIL-P-1629, Procedures
for Performing a Failure Mode Effects
and Criticality Analysis.
1960: Kaoru Ishikawa
introduces his now famous
cause-and-effect diagram.
1818: Gauss uses the normal curve
to explore the mathematics of error
analysis for measurement, probability
analysis, and hypothesis testing.
1970s: Imai develop Dr.
Deming concept called 14 keys
of Deming or called kaizen in
Japanese.
1986: Bill Smith, a senior engineer
and scientist introduces the
concept of Six Sigma at Motorola
1994: Larry Bossidy launches
Six Sigma at Allied Signal.
1995: Jack Welch
launches Six Sigma at
GE.
9. Customer
Competitive Price
High Quality Products
On-time Delivery, etc
Company
Profitability
Repeat Business
Growth/Expansion
$ CCaasshh ! !!!
VVaaluluee ! !!!
1
SSoommee P Prroofiftit 3
2
BBiiggggeerr PPrrooffiitt
1
3
2
PPrriiccee -- CCoosstt == PPrrooffiitt
Price to
Sell
Cost to
Produce
KEY BUSINESS CONCEPT OF SIX SIGMA
10. Where can Six Sigma be applied?
Service Design
Purchase
Six Sigma
Methods Production
HRM
Management
Administration
Quality
Depart.
M & S
IT
11. COPQ against sales revenue
COPQ (Cost of Poor Quality)
Sigma Level DPMO COPQ as sales percentile
1-sigma 691.462 (very low competitive) N/A
2-sigma 308.538 (Average Indonesia’s Industry) N/A
3-sigma 66.807 25-40% of sales
4-sigma 6.210 (Average USA’s Industry) 15-25% of sales
5-sigma 233 (Average Japan’s Industry) 5-15% of sales
6-sigma 3,4 (World Class Industry) < 1% of sales
13. Dupont Chronology
Periode Description Sigma
Before six sigma Dupont Total Cost of Poor Quality
= 20 -30 % of revenue
About 3 sigma
Implementing Cost of Implementing Six Sigma
= $ 20 million
-
1999 Q1 Pilot Project Six Sigma on Specialty Chemicals started
-Revenues $ 1.5 Billions
-Target $ 80 million savings (5% dari revenues)
1999 Q2 40 Black Belts done for training and start for project
1999 Q4 Total saving $ 35 million (initial target $ 25 million) 2.3% of Revenue
17.7% COPQ
4 Sigma
2000 Q4 Total saving $ 100 million (initial target $ 80 million) 6.7% of Revenue
13.3% COPQ
5 sigma
14. Difficult-to-Reach Fruit
Design for Six Sigma (DFSS)
Middle Fruit
Six Sigma tools
Lower Fruit
7 Basic Tools of QC
Ground Fruit
Logic and Intuition
66σσ BBaassiicc CCoonncceepptt
15. 66σσ BBaassiicc CCoonncceepptt
3 sigma level company 6 sigma level company
• <25~40% of sales is failure cost. • 5% of the sales is failure cost.
• Having 66,807 defects per million. • 3.4 defects per million.
• Depends on the detect to find
defect.
• Focusing on process not to produce
defects.
• Believes that high quality is expensive. • Realizes that high quality creates
low cost.
• Not available of systematic
approach.
• Uses know-how of measurement,
analysis, improvement & control.
• Benchmarking against competing
companies.
• Benchmarking to the best
in the world.
• Believes 99% is good enough.
• Define CTQ’s internally.
• Believes 99% unacceptable.
4 sigma Level? 1misspelled word per 30pages of newspaper.
5 sigma Level? 1misspelled word in a set of encyclopedias.
6 sigma Level? 1misseplled word in all of the books contained in a small library.
16. Understanding Basic Concept of Statistics
The concept of sigma
• It is important to understand the difference between accuracy and precision
• Sigma is a measure of vvaarriiaattiioonn (the data spread)
• It is a statistical measure unit displaying a process capability and the
measured sigma value is expressed by DPU(Defect Per Unit), PPM
• It is said that the process with higher sigma value is the process having smaller
defects
• The more increase the sigma value, the more decrease the quality cost and
Cycle Time
1σ
Inflection
Point
μ USL
T 3σ
: The size or a standard deviation
shows the distances between
the inflection point and the mean.
We could say the process has 3
sigma capability if 3 deviations
are fit table between the target
and the specification limit.
17. What does variation mean?
20
• Variation means that a
15
process does not produce
10
the same result (the “Y”)
5
every time.
0
• Some variation will exist in
-5
all processes.
-10
• Variation directly affects customer experiences.
CCuussttoommeerrss ddoo nnoott ffeeeell aavveerraaggeess!!
18. Measuring Process Performance
The pizza delivery example. . .
• Customers want their pizza
delivered fast!
• Guarantee = “30 minutes or less”
• What if we measured performance and found an
average delivery time of 23.5 minutes?
– On-time performance is great, right?
– Our customers must be happy with us, right?
19. The pizza delivery example. . .
How often are we delivering
on time?
Answer: Look at
the variation!
s
x
30 min. or less
0 10 20 30 40 50
• Managing by the average doesn’t tell the whole story.
The average and the variation together show what’s
happening.
20. Reduce Variation to Improve Performance
s
x
30 min. or less
How many standard
deviations can you
“fit” within
customer
expectations?
0 10 20 30 40 50
• Sigma level measures how often we meet (or fail to
meet) the requirement(s) of our customer(s).
21. SIX SIGMA Basic Concept
• All work occurs in a system of interconnected processes
• Variation exists in all processes
• Understanding and reducing variation are the keys to
improving customer satisfaction and reducing costs
22. Y = f(χ)
Question 1), Which one should we focus on the Y or X?
• Y
• Dependent Variable
• Output
• Effect
• Symptom
• Monitor
• X1 … Xn
• Independent Variable
• Input
• Cause
• Problem
• Control object
Question 2), Is needed to test and audit Y continually if the X is good?
6 Sigma activity is concerned about the problem happened(in the sector of
manufacturing and non manufacturing). They could be improved by focusing
the factor which causes the problem.
23. The Approach of 6 Sigma Step
Steps Activity
Define
Measurement
1. Clarifying improvement target object.
2. Forecasting improvement effect.
3. CTQ selection for products and process. Y
4. Understanding process capability for ‘Y’
5. Clarifying measurement method of ‘Y’
6. Specific description of Target object for improving
against ‘Y’
Focus
Y
Y
Y
Analysis 7. Clarifying Target for improving ‘Y’
8. Clarifying factors which affect ‘Y’
Y
X1 .... Xn
Improvement
9. Extract the vital few factors through screening
10. Understanding correlation of vital few factors
11. Process optimization and confirmation experiment
X1 .... Xn
Vital Few X1
Vital Few X1
Control
12. Confirm measurement system for ‘X’
13. Selection method how to control vital few factors
14. Build up process control system & audit for vital few
Vital Few X1
Vital Few X1
Vital Few X1
6 Sigma activity with 5 steps of D-M-A-I-C, will pass through the major 14 steps.
6 Sigma activity have D-M-A-I-C process breaking down the problem through the condition analysis, finding
the potential causal factor , and improving the vital few factors
After the condition identification, we have the first action about the part being improved at first, and then we
proceed continually the improvement activity at the next step.
25. SIX SIGMA CHALLENGES
• Six sigma less suitable for innovation.
• Six sigma emphasize process and cost,
while innovation constitutes something
new in which cost consuming.
• Six sigma only analyzing quantitative
data, qualitative data must be converted to
quantitative.
26. Six Sigma DMAIC Process
Develop Charter and
Business Case
Map Existing Process
Collect Voice of the
Customer
Specify CTQs / Requirements
Measure CTQs / Requirements
Determine Process Stability
Determine Process Capability
Calculate Baseline Sigma
Refine Problem Statement
Control
Identify Root Causes
Quantify Root Causes
Verify Root Causes
Institutionalize Improvement
Control Deployment
Quantify Financial Results
Present Final Project Results
and Lessons Learned
Close Project
Select Solution (Including
Trade Studies,
Cost/Benefit Analysis)
Design Solution
Pilot Solution
Implement Solution
Define
Measure
Analyze
Improv
e
DMAIC = Define, Measure, Analyze, Improve and Control
28. DDeeffiinnee
Defining the “Project Y”
Translate the external CTQ’s into internal product requirements or “Project Y”.
Example:
Voice of the CTQ Project Y
Customer
The range must heat
to the setting chosen
Call-takers must be
available to answer calls
The refrigerator must
stay dry
Answer rate
(% of incoming calls
answered within
20 seconds)
Call-takers must answer
95% of all incoming calls
within 20 seconds
(telephone promptness)
Calibration angle
of the thermostat
A thermostat setting of
350° must result in a
350° oven cavity
No sweat Foam density
29. PPrroocceessss MMaappppiinngg
Add the operating specification and process targets
For the controllable variable input
Start Finish
1 2
Process Targets and Specifications
Experimental
Input Parameters
Target Upper Spec. Lower Spec.
Y = f (X)
(SOP ) = Standard Operating
Parameters
( N ) = Noise Parameters
( X ) = Controllable Process
Parameters
30. Structure Tree
Example
ROTOR
Continue to ask “Why?” until you Reach the Root Cause…...
RPM
Courtesy of Daraius Patell
Losses
Inductance
OD
Core length
STATOR
ASSEMBLY
Electromagnetic
Mechanical
Area A
Area B
Lamination
Endrings
31. CCaauussee && EEffffeecctt DDiiaaggrraamm
•The final diagram will look like a fishbone with the backbone displaying every known
variable (Measurement, Method, Machine, People, Materials, Environment).
Measurement Method Machine
People Materials Environment
33. MMeeaassuurreemmeenntt
DDeetteerrmmiinnee PPrroocceessss CCaappaabbiilliittyy ffoorr
PPrroojjeecctt YY
Determining process capability for your "Project Y" allows you to do several important things.
– Establish a baseline for comparing the improvement of your product or process.
– Quantify the ability of your process to produce output that meets the performance
standard.
– Determine if there is a technology or control problem.
– Understand process capabilities for the design of future processes for DFSS (Design
for Six Sigma) projects.
– Compare your process with others (internally and externally) to judge relative
performance.
· Define the problem in
mathematical terms
· Predict probability of
producing defects
35. Structure Tree
Used to break down problem into manageable groups to identify root cause
or area of focus.
Process for creating a Structure Tree:
• List your problem statement on the left hand side of the page.
• Break the problem down into causes by asking ‘Why?’ and record on
tree branches. Typical categories of causes include:
Technical Transactional
Manpower People
Machine Price
Material Product
Method Promotion
Measurement Physical Distribution
• Assign a High, Medium or Low impact to each branch and select the
highest impact branch.
• Continue breaking down by asking ‘Why?’ until you reach the root cause.
36. More Frequently AAsskkeedd QQuueessttiioonnss
AAbboouutt MMeeaassuurreemmeenntt DDaattaa
What is a “Measurement System”?
- Everything associated with taking measurements:
the people, measurement tool, material, method
and environment is known as --
Observations
Measurements
Data
-- The “Measurement System”.
Parts
Inputs Outputs Inputs Outputs
Think of the “Measurement System” as a sub-process that
can add additional variation to measurement data. The goal
is to use a measurement process that has the smallest
amount of measurement error as possible.
37. SSoouurrcceess ooff MMeeaassuurreemmeenntt SSyysstteemm VVaarriiaattiioonn
Observed
Process
Variation
Actual Process
Variation
Measurement
Variation
Long Term
Process
Variation
slt
Short Term
Process
Variation
sst
Within
Sample
Variation
Variation due
to
Measurement
Equipment
Variation
due to
Operators
Accuracy Linearity Repeatability Stability Reproducibility
The Gage R&R methods we will study in this class will provide estimates of the total
measurement variation, the variation attributed to measurement equipment repeatability
and the variation attributed to the appraisers.
38. EEvvaalluuaattiioonn CCrriitteerriiaa ffoorr
%%GGRR&&RR aanndd %%SSttuuddyy VVaarriiaattiioonn
Acceptable if less than 20%
Conditional if between 20% to 30%
Unacceptable if greater than 30%
Beware of the risk associated with using
Beware of the risk associated with using
data acquired from an unacceptable
data acquired from an unacceptable
measurement process.
measurement process.
s2
gage = s2
repeatability + s2
reproducibility
Repeatability: The variation in measurements taken by a single person or
instrument on the same or replicate item and under the same conditions.
Reproducibility: the variation induced when different operators, instruments, or
laboratories measure the same or replicate specimen.
40. Ho Ha
Correct
Decision
Correct
Decision
Type 1
Error
α
Type 2
Error
β
*Ho(Null Hypothesis) is assumed to be true.
This is like the defendant being assumed
to be innocent.
Ho
Ha
True
Ha(Alternative Hypothesis is alternatives
the Null Hypothesis.
Ha is the one that must be proved.
Accept
The ratio which is
being “Ha” even if it’s false.
Where “β” is usually
set up at 10%.
The ratio which is
being rejected Ho even
though certain thing is true
where “ α” is α error.
(usually 5%)
41. Data Types
Variable Discrete
◎ t-Test
(Compares means less than 2 population)
◎ ANOVA
(Compares variances more than 2 population)
◎ F-Test
(Compares variances of two population)
◎ Chi Square
(Compares counts and
frequencies.)
Before t-Test/ANOVA, confirm the
homogeneity of variance conducting
F-Test
the gap delta(δ)
T
The larger Means and expected gap is getting,
the more different two variances of average in population.
42. • The tool depends on the data type. We use ANOVA when
we have categorical input(s) and a continuous response.
Continuous Categorical
Categorical Continuous
Dependent Variable (Y)
Independent Variable (X)
Regression
ANOVA
Logistic
Regression
Chi-Square (c2)
Test
43. Variance Homogeneity Testing Means Testing
1.One population variance
testing
2.Two population variances
testing
3.Testing of population
variances for more than two
(Normal distribution)
4. Testing of population
variances for more than two
(Non-Normal distribution)
Chi Square
F
Homogeneity of
Variance
▶Bartlett’s Test
Homogeneity of
Variance
▶Levene’s Test
1.One population Mean
testing
1) When we know σ of the
population
2) When we know σ of the
population
2.Two population Mean
testing
1) When they know
σ1 and σ2
2) When they don’t know
σ1 and σ2
① σ1 = σ2
② σ1 ≠ σ2
Normal distribution
( 1 - Sample Z )
T distribution
( 1 - Sample t )
Normal distribution
T distribution
( 2 - Sample t )
The type of Hypothesis
44. IImmpprroovveemmeenntt
What is DOE ?
• DOE is mmoorree tthhaann jjuusstt aa ssttaattiissttiiccaall
tteecchhnniiqquuee..
• IItt iiss tthhee ccoommbbiinnaattiioonn ooff eeffffeeccttiivvee ppllaannnniinngg,,
ddiisscciipplliinnee,, ssuubbjjeecctt mmaatttteerr kknnoowwlleeddggee aanndd
ssttaattiissttiiccaall mmeetthhooddss tthhaatt mmaakkee tthhee
eexxppeerriimmeenntt aa ssuucccceessss..
46. BBaassiicc CCoonncceeppttss iinn DDOOEE
PPrreessssuurree
SSppeeeedd PPrroocceessss
QQuuaalliittyy CChhaarraacctteerriissttiicc ((YY))
Run Pr Spd Pr
Spd x Spd
Pr Y
1 5 70 - + 10
+ + - -
2 5 90 + - 4
3 10 70 - - 6
4 10 90 + + 12
Ave - 8 5
79
Ave 8 11
+
Y = 8
2 0 6
y^ == 88 ++ PPrr ++ 33 xx (( PPrr SSppdd))
PPrroocceessss
DDOOEE
PPrroocceessss
MMooddeell
47. FFMMEEAA
Failure Modes and Effects Analysis is a systematic method for identifying,
analyzing, prioritizing, and documenting potential failure modes and their
effects on a system, product, or process.
48. The Logic of SPC(Statistical Process Control)?
Upper Control Limit
Lower Control Limit
SAMPLES
Process
Capability
α/2
Controller α/2
Desired
Output
PROCESS OUTPUT
A B C D E
· Controllable factors
- Assignable causes
- Adjustable
- Special
L M N O P
· Uncontrollable factors
- Common causes
- Noise
- Inherent causes
INPUT
• SPC has been traditionally used to monitor and control the output of processes.
In this application, we are measuring the dimensions of finished parts or
characteristics of finished assemblies.
X
0.5
0.0
-0.5
S am p le Mean
Subgroup 0 50 100
• Six Sigma Quality focuses on moving control upstream to the leverage input characteristic for Y. If we can
measure and control the vital few X’s, control of Y should be assured.
UCL=0.4384
Mean=0.001188
LCL=-0.4360
1.0
0.5
0.0
S amp le R an g e
1 1 1
UCL=0.7596
R=0.2325
LCL=0
Xbar/R Chart for Sealing Angle Line #2
CCoonnttrrooll
49. Types of Control Charts Types of Control Charts
Variables Charts for
monitoring continuous X’s
• Average & Range
X bar & R
n < 10
typically 3~5
• Averages & Std Deviation
X bar & σ
n ≥ 10
• Median & Range
X & R
n < 10
typically 3~5
• Individual &
Moving Range
XmR
n = 1
Attributes Charts for
monitoring discrete X’s
• Fraction Defective
P Chart
typically n ≥ 50
tracks DPU/DPU
• Number Defective
np Chart
n ≥ 50 (constant)
tracks # def
• Number of Defects
c Chart
c > 5
• Number of Def/Unit
U Chart
n variable
• In order to select the appropriate control chart for monitoring your process, first
determine if your key process variables (X’s) are continuous or discrete. There are
control charts for both continuous data and discrete data.
50. Control Chart
12
10
8
·
·
·
·
·
·
·
· · · ·
·
·
·
·
·
·
·
·
·
Week
Upper Control Limit = Ave + 3 x Std Dev
14
13
7
6
Lower Control Limit = Ave - 3 x Std Dev
Central
Line =
Average
Note: Control limits should be established using subgroup standard deviation
52. Project title
Date …………………
Department / Team
Prepared : …………..
G ManagerDirector Pres. Dir.
Mgr.
6s Champion Review
Final Report
Contents
1. Define Step
2. Measure Step
3. Analysis Step
4. Improvement Step
5. Control Step
-Attachment
54. PJT Name
Ka Part Ka Group Project Registration
Period
E ngineer
Approva
Team
l
Div./Dept: Name
Breakthrough
KPI Current Wo rld B est Target
Main Improv ement Object
Team Fo rmation (Related Department Involved)
Name De p t. Level Role
Why ? How to do ?
(* Selection Background)
Quantitative
Qualitative
Expected
Results
New Idea for Target Achievement
Neck Point
55. M A I C
D
Background Expected result
4500
5000
11%
CCuurrrreenntt TTaarrggeett Unit:
(Nm3/day)
How to do:
Target Saving cost:
Others
Electric
O2
N2
LNG
0.1650 0.1309 0.1290 0.0605 0.0070
33.5 26.6 26.2 12.3 1.4
33.5 60.1 86.3 98.6 100.0
0.5
0.4
0.3
0.2
0.1
0.0
100
80
60
40
20
0
Defect
Count
Percent
Cum %
Percent
Count
Energy Usage Price
General Background
56. Add the operating specification and process targets
For the controllable variable input
Start Finish
1 2
Process Targets and Specifications
Experimental
Input Parameters
Target Upper Spec. Lower Spec.
Y = f (X)
(SOP ) = Standard Operating
Parameters
( N ) = Noise Parameters
( X ) = Controllable Process
Parameters
Process Mapping
M A I C
D
57. SIPOC – Suppliers, Inputs, Process, Outputs, Customers
You obtain inputs from suppliers, add value through your process, and
provide an output that meets or exceeds your customer's requirements.
Process Understanding
M A I C
D
60. Sampling
Training
Gage R&R
Result
Date:………
Collected 12 ea Sampling
Conduct a training……
How to see what kind ……..
Defect that happened in process.
Purpose : find out the operator ability.
Conduct Gage R&R for 4 men to know the
judgment capability (in different times & do
not know the inspection result of each other)
--> Repeat 2 times for each persons
% Gage R&R : 0 %
acceptable
M
TToo ddeetteerrmmiinnee iiff tthhee mmeeaassuurreemmeenntt eerrrroorr iiss
ssmmaallll aanndd aacccceeppttaabbllee rreellaattiivvee ttoo tthhee
Observed Process
Variation
Measurement Process Variation
Sample M a n
M a n
Machine
Method
Material
pprroocceessss vvaarriiaattiioonn,, wwee ccaann
uussee GGaaggee RR&&RR ssttuuddyy..
Gage R & R
D A I C
61. Process Capability
Current Condition
mm Units of Measure
Center of the bar
Smooth curve
interconnecting the
center of each bar
M
D A I C
Four Block Diagram
Current Target
1 2 3 4 5 6
2.5
2.0
1.5
1.0
0.5
Poor
Z Shift
Process Control
Good
Poor Good
Technology
Block A
Block C
Block B
Block D
Z Shift
A : Poor control, inadequate technology
B : Must control the process better, technology is fine
C : Process control is good, inadequate technology
D : World class
62. M
Factor Detail Analysis Plan Schedule
Mar 2nd Week Mar 3rd Week Mar 4th Week
X1.1 Inspect correlation between “Y” and inspector
X1.3
for each group
Analysis khole / dimension
Height, diameter, angle etc,
X1.2
X1.4
Analysis Plan
D A I C
64. A
D M I C
Use regression is to express and analyze a mathematical equation of describing a relationship.
That is, it is to fit a mathematical equation of describing a relationship between the “YY” and “XX’’ss”.
Regression Analysis: Angle Value versus Rotate Gear
The regression equation is
Angle Value = - 0,380 + 3,74 Rotate Gear
Predictor Coef SE Coef T P
Constant -0,3796 0,1693 -2,24 0,042
Rotate G 3,744 1,153 3,25 0,006
Regression Analysis: Angle Value versus Rotate Gear
The regression equation is
Angle Value = - 0,380 + 3,74 Rotate Gear
Predictor Coef SE Coef T P
Constant -0,3796 0,1693 -2,24 0,042
Rotate G 3,744 1,153 3,25 0,006
S = 0,1510 R-Sq = 43,0% R-Sq(adj) = 38,9%
S = 0,1510 R-Sq = 43,0% R-Sq(adj) = 38,9%
Analysis of Variance
Source DF SS MS F P
Regression 1 0,24032 0,24032 10,55 0,006
Residual Error 14 0,31905 0,02279
Total 15 0,55937
Unusual Observations
Obs Rotate G Angle Va Fit SE Fit Residual St Resid
2 0,230 0,4000 0,4815 0,1070 -0,0815 -0,77 X
7 0,130 0,5000 0,1071 0,0407 0,3929 2,70R
Analysis of Variance
Source DF SS MS F P
Regression 1 0,24032 0,24032 10,55 0,006
Residual Error 14 0,31905 0,02279
Total 15 0,55937
Unusual Observations
Obs Rotate G Angle Va Fit SE Fit Residual St Resid
2 0,230 0,4000 0,4815 0,1070 -0,0815 -0,77 X
7 0,130 0,5000 0,1071 0,0407 0,3929 2,70R
Normal Probability Plot of the Residuals
(response is Angle Va)
-0,2 -0,1 0,0 0,1 0,2 0,3 0,4
2
1
0
-1
Normal Score
Residual
GGeeaarr RRoottaattiioonn
Motor
65. D M I C
We represent characterized variation ““YY”” by the total sum of square, then this method is to find
what the factor’s level which influence enormously is, comparing both of them.
One-way ANOVA: Gr.A - 3, Gr.A - 4, Gr.B - 3, Gr.B - 4, Gr.C - 3, Gr.C - 4
Analysis of Variance
Source DF SS MS F P
Factor 5 0.8761 0.1752 2.59 0.030
Error 102 6.8897 0.0675
Total 107 7.7659
Since p-value < 0.05;
Ho (reject), Ha (accept).
That is we can claim there’s
a difference between the level
Of adhesive
Individual 95% CIs For Mean
Based on Pooled StDev
Level N Mean StDev -------+---------+---------+---------
Gr.A - 3 18 0.0572 0.3029 (-------*-------)
Gr.A - 4 18 -0.1217 0.2427 (-------*-------)
Gr.B - 3 18 0.0839 0.2231 (--------*-------)
Normal Probability Plot
Gr.B - 4 18 -0.1183 0.2403 (-------*-------)
Gr.C - 3 18 0.0156 0.2796 (-------*-------)
Gr.C - 4 18 0.0967 0.2626 (-------*--------)
-------+---------+---------+---------
Pooled StDev = 0.2599 -0.15 0.00 0.15
Anderson-Darling Normality Test
A-Squared: 0.452
P-Value: 0.242
Average: 0.0572222
StDev: 0.302865
N: 18
-0.5 0.0 0.5
.999
.99
.95
.80
.50
.20
.05
.01
.001
Probability
Gr.A - 3
See from sealing angle specifications there’s no problem, cause all operator adjustment
Still in range (-0.5o ~ 0.5o). But there’s a significant effect both of them seeing by characte
ristic variation result, each operator have a different mean adjustment.
G r.C - 4
Boxplots of Gr.A - 3 - Gr.C - 4
(means are indicated by solid circles) UCL
G r.C - 3
G r.B - 4
Target Line
G r.B - 3
G r.A - 4
G r.A - 3
0.5
0.0
LCL
-0.5
Operator Adjustment
Screen
Manual Adjustment
A
66. A
D M I C
Factor Analysis Purpose Detail Analysis Content Result Conclusion
Selected as
Vital “X” P = 0.000
Selected as
Vital “X”
X1.1 Find most effected
to “ Y “
X1.2
Find most effected
to “ Y “
P = 0.000
Selected as
Vital “X”
Bottle Neck
Regression, to compare Dew Point and
Purge Flow rate
X1.3
X1.4 Regression, to compare Dew Point
and Out Air Temperature
Find most effected
to “ Y “
Find most effected
to “ Y “
ANOVA, to compare Dew Point and
Heating Time
ANOVA, to compare Dew Point and
Drying Time
P = 0.000
P = 0.003
Analysis Result
73. I
Six Sigma Quality focuses on moving control upstream to the leverage input characteristic
for Y. If we can measure and control the vital few X’s, control of Y should be assured.
Desired Process Capability
Output
Controller
Input Process
Group Member
Controllable factors:
- Miss adjust causes
- Adjustable check
- Pad control
- Education
Upper Control Limit
X
●
Lower Control Limit
●
0.5
0.0
-0.5
Sample Mean
Output
Subgroup 0 50 100
UCL=0.4384
Mean=0.001188
LCL=-0.4360
1.0
0.5
0.0
Sample Range
1 1 1
UCL=0.7596
R=0.2325
LCL=0
Xbar/R Chart for Sealing Angle Line #2
Process Standard Change
C
D M A
74. I
C
D M A
P Chart enables us to control our process using statistical method's to signal when
process adjustments are needed.
0 10 20 30 40 50
0.004
0.003
0.002
0.001
0.000
Sample Number
Proportion
P Chart for Stem Crack
UCL=0.002466
P=0.001198
LCL=0
75. X-bar/R Chart use to control daily average for CTQ.
10.2
10.1
10.0
9.9
9.8
9.7
Sample Mean
1
Subgroup 0 5 10 15
UCL=10.13
Mean=9.943
LCL=9.755
1.0
0.5
0.0
Sample Range
UCL=1.016
R=0.5594
LCL=0.1030
Xbar/R Chart for Cullet Speed
CTQ’s daily control data
77. Optimalkan pemakaian Zinc
Pada Proses Galvanize
Date 20 MAY 2009
Team : Galvanize
Prepared by : Imam Mudawam
Prod Mgr CEO
.
Work Mgr
6s Champion Review
Final Report
Contents
1. Define Step
2. Measure Step
3. Analysis Step
4. Improvement Step
5. Control Step
NG DC ISK
78. PJTName
Period
M A I C
Work Mgr Prod Mgr
Tea
m Nam
e
Div./Dept:
CONST./GALV. Breakthrough
KPI Current Wo rld B est Target
Main Improvement Object
CEO
Approval
Team Fo rmation (Related Department Involved)
Name De p t. Level Role
Why ? How to do ?
(* Selection Background)
LSL Target USL
Process Data
Observ ed Performance
Exp. Within Performance
Quantitative
Qualitative
Expected
Results
New Idea for Target Achievement
Fab
Galv
Qc
Galv
Neck Point
Optimalkan Pemakaian Zinc
Z
Shift:
-
1.86.
Imam M
Lukman
Banbang
Spv
Supv
Form
• Ketebalan Galvanize
sesuai standard.
Leader
Leader
Member
Amin Form Member
Exp. Ov erall Performance
Rp / kg
Rp / kg
Slamet Galv
Member
D
Project Registration
D
M
A
I
C
-Making Theme Reg. Schedule
- Analyzing Aging Root-Caused
-Determine Potential X List
- Find Current situation
- Find Vital X by analyzing
Potential X
-Find Improvement idea
-Confirmation run
-Process Control by
Monitoring Aging
Amount
May – W5
Junl – W5
Jul –W5
Optimalkan Pemakaian Zinc
Pada Proses Galvanize
Potential (Within) Capability
Z.Bench -1.86
Lower CL -2.59
Z.LSL 4.62
Z.USL -1.86
Cpk -0.62
Lower CL -0.80
Upper CL -0.44
Ov erall Capability
Z.Bench -1.07
Lower CL -2.11
Saving Zinc
Rp. 295 Juta /Tahun
Galvanize
Reduce
cost
Form
Hans Ga Supv
Member
Sept –W3
Sept –W5
Galvanize
183.6 micr.
Galvanize
130 micron
120 160 200 240 280
LSL 100
Target 130
USL 150
Sample Mean 183.649
Sample N 35
StDev (Within) 18.0955
StDev (Ov erall) 31.8811
Z.LSL 2.62
Z.USL -1.06
Ppk -0.35
Lower CL -0.49
Upper CL -0.21
Cpm 0.11
Lower CL 0.09
PPM < LSL 0.00
PPM > USL 885714.29
PPM Total 885714.29
PPM < LSL 1.89
PPM > USL 968521.55
PPM Total 968523.44
PPM < LSL 4348.13
PPM > USL 854388.04
PPM Total 858736.18
Within
Ov erall
Pr ocess Capability of t 6.5mm +
(using 95.0% confidence)
Worksheet: Worksheet 3; 11/13/2009
ISK DC NG
79. D
General Background M A I C
• Hasil proses Galvanize ketebalannya melebihi standar yang
ditentukan dalam ASTM- A123 / A123M.
• Tujuan Proyek ini untuk dapat mengoptimalkan ketebalan lapisan zinc
pada hasil proses galvanize.
Ketebalan material
400
300
200
100
0
Material t.6.5s/d ...mm
Material t. 3.5s/d6.0mm
Material t.1.5s/d3.0mm
sample mean micron 183.6 84.2 83.6
Percent 52.3 24.0 23.8
Cum % 52.3 76.2 100.0
100
80
60
40
20
0
Produk tebal 6.0s/d….mm
sample mean micr on
Per cent
Pareto Char t of Ketebalan mater ial
Worksheet: Worksheet 3; 10/14/2009
80. Brainstorming Potential X’s List A I C
Big Y X1 X2 X3
M
D
Material
Degrising
Pickling 1 & 2
Ketebalan
Komposisi kimia
Konsentrasi Basa
Base Metal
Caustic Soda
Konsentrasi Keasaman
Optimalkan pemakaian Zinc
Pada Proses Galvanize
Temperatur
Waktu Pencelupan
1
Hcl
81. M
Brainstorming Potential X’s List A I C
Big Y X1 X2 X3
Optimalkan Pemakaian Zinc
Pada proses Galvanize
D
Fluxing Konsentrasi Keasaman
1
Temperatur
Zinc Amunium Chloride
Dipping Aluminium Alloy
Zinc Ingot
Temperatur
Komposisi Campuran
Waktu pencelupan
82. Measure Step - GAGE R&R A I C
Gage R&R take
from 2 Inspector
who check this
sample of
Galvanize coating
thickness in 10
chek point
In doing gage R&R
we take 2 times
repeat in check
Decide
operator
who take Gage
R&R
Do Gage
R&R
Change
Method,measurement, etc
Analyze
Result
Gage
R&R
Next Step
NG ;
Total Gage
R&R > 20%
OK ;
Total Gage
R&R < 20%
M
D
83. Gage R&R
Gage R&R < 20%
Acceptable
A I C
M
D
Sample pengukuran ketebalan galvanize sebanyak 2 sample dengan 30 titik pengukuran , tiap sampel diambil 15 titik
pengukuran
Diukur secara berurutan dan secara acak oleh dua orang operator .
Study Var %Study Var
Source StdDev (SD) (6 * SD)
(%SV)
Total Gage R&R 2.0064 12.0384
14.67
Repeatability 2.0064 12.0384
14.67
Reproducibility 0.0000 0.0000
0.00
oprtr 0.0000 0.0000
0.00
Part-To-Part 13.5243 81.1460
98.92
Total Variation 13.6724 82.0341
100.00
Gage R&R Repeat Reprod Part-to-Part
100
50
0
Per cent
% Cont ribut ion
% Study Var
8
4
0
Sam ple Range
UCL= 8.604
_
R=2.633
LCL= 0
A B
110
100
90
Sam ple Mean
UCL= 101.62
__
X=96.67
LCL= 91.71
A B
1 2 3
110
100
90
part no
A B
110
100
90
oprt r
1 2 3
110
100
90
part no
Average
The result of Gage R&R total is 14.67 % the acceptance percentage is
bellow 20% (< 20 % ) ,
meanwhile the result of measurement between < 20 %, it accepted.
oprt r
A
B
Gage name:
Date of study :
Reported by :
Tolerance:
Misc:
Components of Variation
R Chart by oprtr
Xbar Chart by oprtr
measure by part no
measure by oprtr
oprtr * part no I nteraction
OPERATOR PENGUKURAN KETEBALAN GALVANIZE
Worksheet: Worksheet 3; 9/9/2009
84. A I C
M
Current Condition D
Rata rata ketebalan Galvanize pada produk dengan ketebalan > 6.0 mm saat ini mencapai 183.6 micron , berdasarkan data
dari tgl.20 Juni 09 Sampai 27 Juni 09 ,jumlah sample 35 , alat ukur menggunakan COATING THICKNESS DIGITAL merek
TIME - TYPE TT 220 .
Untuk mendapatkan ketebalan sesuai target yang diinginkan grafik harus bergeser kekiri , dan harus mengurangi keteba-
Pelapisan Galvanize yang sesuai standard ASTM A 123/A 123M antara ( 100 s/d 150 ) micron .
Probability Plot Ketebalan Galvanize mater ial 6.5mm s/ d .......dst
100 150 200 250 300
99
95
90
80
70
60
50
40
30
20
10
5
1
t 6.5mm +
Per cent
Mean 183.6
StDev 31.88
N 35
AD 1.706
P-Value <0.005
Normal - 95% CI
Worksheet: Worksheet 3; 7/1/2009
LSL Target USL
120 160 200 240 280
Exp. Within Performance
Exp. Ov erall Performance
KETEBALAN GALVANIZE
Process Data
LSL 100
Target 130
USL 150
Sample Mean 183.649
Sample N 35
StDev (Within) 18.0955
StDev (Ov erall) 31.8811
Potential (Within) Capability
Z.Bench -1.86
Lower CL -2.59
Z.LSL 4.62
Z.USL -1.86
Cpk -0.62
Lower CL -0.80
Upper CL -0.44
Ov erall Capability
Z.Bench -1.07
Lower CL -2.11
Z.LSL 2.62
Z.USL -1.06
Ppk -0.35
Lower CL -0.49
Upper CL -0.21
Cpm 0.11
Lower CL 0.09
Observ ed Performance
PPM < LSL 0.00
PPM > USL 885714.29
PPM Total 885714.29
PPM < LSL 1.89
PPM > USL 968521.55
PPM Total 968523.44
PPM < LSL 4348.13
PPM > USL 854388.04
PPM Total 858736.18
Within
Ov erall
Process Capability of t 6.5mm +
(using 95.0% confidence)
Worksheet: Worksheet 3; 11/13/2009
85. M
Four Block Diagram D
A I C
Nilai sigma level saat ini adalah – 1.86 s, target yang ingin
dicapai 4.5 s
A B
C
2.5
2.0
1.5
1.0
0.5
1 2 3 4 5 6
Z shift
Poor
Process Control
Good
Poor Z st Good
Technology
EXPLANATION
Position of Current Condition was
column C , it was mean :
PROCESS CONTROL IS GOOD,
BUT TECHNOLOGY (METHOD)
IS BAD
D
SIGMA TARGET
SIGMA TARGET
Sigma current – 1.86
86. I C
A
D M
F(x) x
WAKTU PENCELUPAN
Optimalkan pemakaian Zinc
Pada proses Galvanize
Analysis
TEMPERATUR ZINC
KETEBALAN GALVANIZE
87. Analyze – Type of factor & Tools Using I C
Y Factor (x) Type Tools
Optimalkan
pemakaian
zinc pada
proses
Galvanize
(Continuous)
Waktu Pencelupan
A
D M
CONTINUE REGRESION
88. Analysis Waktu Fluxing
( Regresion )
Regression Analysis: Ketbln.Galva versus Wktu fluxing, Wktu.deeping, ...
* Wktu.deeping is (essentially) constant
* Wktu.deeping has been removed from the equation.
* NOTE * All values in column are identical.
* Temp.deeping is (essentially) constant
* Temp.deeping has been removed from the equation.
The regression equation is
Ketbln.Galvanize (micron ) = 110 - 1.17 Wktu fluxing
Predictor Coef SE Coef T P VIF
Constant 109.667 3.658 29.98 0.000
Wktu fluxing -1.1668 0.4555 -2.56 0.034 1.000
S = 4.13685 R-Sq = 45.1% R-Sq(adj) = 38.2%
Analysis of Variance
Source DF SS MS F P
Regression 1 112.32 112.32 6.56 0.034
Residual Error 8 136.91 17.11
Total 9 249.22
Unusual Observations
Wktu Ketbln.Galvanize
Obs fluxing (micron ) Fit SE Fit Residual St Resid
1 3.0 114.00 106.17 2.43 7.83 2.34R
R denotes an observation with a large standardized residual.
Durbin-Watson statistic = 1.69422
I C
A
D M
Ketebalan Galvanize VS
Waktu Fluxing
Nilai p-value < 0.05, maka
Ho ditolak, yang
menandakan variabel
waktu Fluxing
berpengaruh terhadap
ketebalan galvanize.
89. Analysis Temperatur
Dipping ( Regresion )
Regression Analysis: Ktbl . Galva versus Tmprt deepin,
Wkt.deeping, ...
* Wkt.deeping is (essentially) constant
* Wkt.deeping has been removed from the equation.
* NOTE * All values in column are identical.
* Wkt.fluxing is (essentially) constant
* Wkt.fluxing has been removed from the equation.
The regression equation is
Ktbl . Galvanize (micron ) = - 98 + 0.443 Tmprt deeping
Predictor Coef SE Coef T P VIF
Constant -98.4 157.1 -0.63 0.548
Tmprt deeping 0.4428 0.3562 1.24 0.249 1.000
S = 6.47038 R-Sq = 16.2% R-Sq(adj) = 5.7%
Analysis of Variance
Source DF SS MS F P
Regression 1 64.70 64.70 1.55 0.249
Residual Error 8 334.93 41.87
Total 9 399.63
Durbin-Watson statistic = 1.36705
I C
A
D M
Ketebalan Galvanize VS Temperatur Zinc
Nilai p-value > 0.05, maka
Ho diterima, yang
menandakan variabel
temperatur tidak
berpengaruh terhadap
ketebalan galvanize.
90. Analysis Waktu Dipping
( Regresion )
Regression Analysis: Ktbalan Galvanize micron versus Waktu
deeping
The regression equation is
Ktbalan Galvanize micron = 43.7 + 19.9 Waktu deeping
Predictor Coef SE Coef T P VIF
Constant 43.674 5.279 8.27 0.000
Waktu deeping 19.9082 0.6573 30.29 0.000 1.000
S = 5.96979 R-Sq = 99.1% R-Sq(adj) = 99.0%
Analysis of Variance
Source DF SS MS F P
Regression 1 32698 32698 917.49 0.000
Residual Error 8 285 36
Total 9 32983
Durbin-Watson statistic = 2.41646
I C
A
D M
Waktu Dipping VS Ketebalan Galvanize
Nilai p-value < 0.05, maka
Ho ditolak, yang
menandakan variabel
waktu dipping
berpengaruh terhadap
ketebalan galvanize.
91. I C
A
D M
Analysis Result ( Regression )
Item Content Result Remarks
Selected not Vital View
Selected as Vital View
Galvanizing
P = 0.034
P = 0.024
P = 0.000
Waktu Fluxing
Temperatur Dipping
Waktu Dipping
Selected not Vital View
Berdasarkan data diatas variabel waktu Dipping paling berpengaruh terhadap hasil ketebalan pada proses galvanize (99%)
92. D M A I
Improvement C
• Berdasarkan hasil analisa, untuk
produk dengan ketebalan 6 mm
didapatkan hubungan linier antara
lamanya waktu pencelupan dengan
hasil ketebalan galvanize.
• Grafik di bawah ini dapat dijadikan
acuan untuk menentukan tebal
galvanize yang diinginkan.
HUBUNGAN WAKTU PENCELUPAN DAN KETEBALAN GALVANIZE
PADA PRODUK DENGAN KETEBALAN 6 mm
260
250
240
230
220
210
200
190
180
170
160
150
140
130
120
110
100
90
80
70
60
50
40
30
20
10
0
0 1 2 3 4 5 6 7 8 9 10 11
Waktu Pencelupan (menit)
Ketebalan Galvanis (micron)
Standar 6mm
Y = 43.7 + 19.9 x
93. 94
D M A I
Improvement C
Catatan : Proses improvement masih sedang berjalan dan akan dilaporkan
kemudian hari.
94.
95. DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
Measurement Analysis Improvement Control
Session Start : 23.01.02
□○ Theme Register
□○ Team Organization
□○ Process Map
□○ Cpk Analysis(Current)
□○ Problem Description
□○ CTQ Selection
□○ Measurable Y Value
Selection
□○ 4 Block Diagram
Session Start : Session Start : Session Start :
□○ Brainstorming
□○ Logic Tree Analysis
□○ Analysis by Minitab
□○ Process Benchmarking
□○ CTQ Selection
□○ Process Map
□○ ANOVA
□○ Regression Analysis
□○ Factor Level Decision
□○ DOE
□○ Statistical Interpretation
□○ Data Gathering & Analysis
□○ Main Factor Analysis
□○ Hypothesis Test
□○ ANOVA
□○ Control Chart
□○ Rational Tolerance
selection
□○ Document Control plan
□○ Training Process Controller
□○ CTQ Process Monitoring
System set up
Session Finished :
□○ Y Identification
□○ Gauge R&R
□○ 4 Block Focus(Zst & Zshift)
□○ Problem analysis
reaffirmation
□○ Statistical skill of Y
□○ Graphical skill of Y
□○ Gap Analysis
□○ The 1st improvement of
X(Factor)
□○ Conclusion(the fixed X
factor)
□○ Test plan
□○ Control Plan
Implementation
□○ CTQ Process Monitoring
System Build-Up
□○ Double check of all the
problems
Main Schedule
Session Finished : Session Finished : Session Finished :
96. Definition Of Defect
A period of time taken off by an employee which is neither
planned or authorised
DEFINITION
Sickness
Unauthorised absence
Lateness
DEFECT TYPES
Payment for overtime to cover absence = £800,000:00 pa
Loss in revenue due to lost production = £700,000:00 pa (est)
Adverse morale issues
Deterioration in productivity and quality
VALUE
DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
97. Absence Logic Tree
Absence
Management
Style
DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
Accidents
Morale issues
Cleanliness
Repetitive Work
Sex
Time With Company
=CTQ’s
Age
Use Of Policy
Section
Team Leader
Guidelines
Employee
Target Setting
Environment
Method
Consistent Approach
Shift Pattern
98. Sigma Level Calculation
DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
Graph shows reduction in absence, since the beginning of the project (DEC) spotlight effect has
reduced absence. One problem we do have is the calculation of absence data there is a difference
between the HR data collection method and the tube plant data collection method.
8
6
4
2
0
-2
Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun
Tube 4.27 4.66 5.23 4.7 5.01 5.34 3.86 3.13 2.75
HR 4.31 5.22 5.86 4.56 5.09 5.02 3.97 3.45 2.72
Variance -0.04 -0.56 -0.63 0.14 -0.08 0.32 -0.11 -0.32 0.03
Using The Percentage Defective (4.605%) We Calculated The DPMO Since Merger
DPMO = 46050
Using statistical tables the SIGMA LEVEL = 3.21
99. VOC - Harp Questionnaire
219 People Questioned Across The Tube Plant
Answered by all employees.
Name Marital Status
Clock Number Age
Department Children
Shift Home
Time with Company Job position
Time in current job
Have you taken any unauthorised time off since 5th July 2001? YES / NO
Do you currently have any warnings that relate to sickness or absence? YES / NO
Do you understand LGPDW's absence and sickness policy? YES / NO
Do you understand the affect of absence on our business? (e.g. cost, pressure on colleagues)
Only to be answered by employees with no absence history
How do you think high absence levels affects your ability to do your job?
What do you think of LG. PHILIPS Displays as an employer?
Are you happy working for LGPDW?
Ask employee all questions below, and mark the score they give ie strongly agree
= 10, agree = 5 and strongly disagree = 0.
Do you think absence is due to Management style?
Do you think absence is due to accidents in work?
Do you think absence is due to the current shift pattern?
Do you think absence is due to your working environment?
If I was late for work, I would take the rest of the shift off because it still counts as an absence
Are there any other reasons that you think contribute to absence that are not listed above?
DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
100. Catchment Area Of People Questioned
Overall
catchment area
Main catchment
area
DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
101. Others
Job rotation
P ay rise
Music
Mgt at ti tude
Att Bonus
137 51 33 14 5 12
54.4 20.2 13.1 5.6 2.0 4.8
54.4 74.6 87.7 93.3 95.2 100.0
250
200
150
100
50
0
100
80
60
40
20
0
Defect
Count
Percent
Cum %
Percent
Count
How would y ou tackle high absence?
Operators
Pareto’s Of Absence Data
DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
Teambuilding
gather info on persistant absentees
E xtra manning
1 - 1Communication
A tendance bonus
4 2 2 1 1
40.0 20.0 20.0 10.0 10.0
40.0 60.0 80.0 90.0 100.0
10
5
0
100
80
60
40
20
0
Defect
Count
Percent
Cum %
Percent
Count
How would y ou tackle absence?
Section Leaders
Conclusions
Operators would like to be paid
more to come to work
Operators see a problem in the
way they are treated by Mgt
Music might help??
Conclusions
Section Leaders see the benefit
in some form of attendance
bonus, but more emphasis on
communication and data
collection
102. Pareto’s Of Absence Data
Others
deaths
accidents insi de work
car problems
i njury out side work
low morale
f lexi ble fl oatdays
si ck ness
f lu
domest ic
4 4 4 3 2 2 1 1 1 1
17.4 17.4 17.4 13.0 8.7 8.7 4.3 4.3 4.3 4.3
17.4 34.8 52.2 65.2 73.9 82.6 87.0 91.3 95.7 100.0
20
10
0
100
80
60
40
20
0
Defect
Count
Percent
Cum %
Percent
Count
Main Reasons For Absence On Your Shif t
DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
Main issues af f ecting shif ts ability to meet output targets
Others
experience
absence
poor maint enance
suppl y
low manning
11 5 3 2 1 1
47.8 21.7 13.0 8.7 4.3 4.3
47.8 69.6 82.6 91.3 95.7 100.0
20
10
0
100
80
60
40
20
0
Defect
Count
Percent
Cum %
Percent
Count
Conclusions
No significant patterns have
emerged
Conclusions
Absence is not a major issue
103. Analysis Using M System
DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
Age
Sex
Time With Company
Employee
NEXT STEP IS TO CARRY OUT ANALYSIS BY
EMPLOYEE TO ESTABLISH IF THERE ARE
ANY STATISTICAL DIFFERENCES
DATABASE WAS CREATED LOOKING AT THE
NINE MONTH PERIOD (5th July-4th April) PRIOR
TO AND DURING PROJECT
Employment Details Time with Company Personal Details Unauthorised Absence
empn
o Emp Name dept
cd Dept Name Grade shift no Distance
Details
date_hir
ed Year/Month gender Age Age
Group
Count
Absence Abs Occ
12345 A N OTHER1 81000 Tube Manufacturing Senior Engineer DT CF1 1AB 17/03/00 02/01 Male 34 D 0 0
12346 A N OTHER2 81000 Tube Manufacturing Senior Engineer DT CF1 1AB 06/10/99 02/06 Male 34 D 0 0
12347 A N OTHER3 81000 Tube Manufacturing Team Leader DT CF1 1AB 11/08/97 04/08 Male 33 D 3 1
12348 A N OTHER4 81100 Screen Production Section Leader T4 CF1 1AB 18/05/98 03/11 Male 51 H 0 0
12349 A N OTHER5 81100 Screen Production Section Leader T2 CF1 1AB 26/04/99 02/12 Male 42 F 1 1
12350 A N OTHER6 81100 Screen Production Section Leader T1 CF1 1AB 10/05/99 02/11 Male 41 F 0 0
12351 A N OTHER7 81100 Screen Production Section Leader DT CF1 1AB 20/04/98 03/12 Male 41 F 0 0
12352 A N OTHER8 81100 Screen Production Section Leader T2 CF1 1AB 06/10/97 04/06 Male 39 E 1 1
12353 A N OTHER9 81100 Screen Production Team Leader DT CF1 1AB 12/01/98 04/03 Male 39 E 0 0
12354 A N OTHER10 81100 Screen Production Technician DT CF1 1AB 06/10/97 04/06 Female 35 D 0 0
104. Analysis By Age
DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
Created Age Groupings
Age Group Age Range
A 16-20
B 21-25
C 26-30
D 31-35
E 36-40
F 41+
Analysed Percent Absent By Age Group
Conclusions
There is a significance showing that
D = 31-35 and F = 41+ are more likely to
not to take unauthorised absence and that
B = 21-25 are more likely to take
unauthorised absence
Chi-Square Test: A, B, C, D, E, F
Expected counts are printed below observed counts
A B C D E F Total
1 188 175 183 132 91 62 831
183.80 187.26 182.08 125.12 94.06 58.68
2 25 42 28 13 18 6 132
29.20 29.74 28.92 19.88 14.94 9.32
Total 213 217 211 145 109 68 963
Chi-Sq = 0.096 + 0.802 + 0.005 + 0.378 + 0.099 + 0.188 +
0.603 + 5.050 + 0.029 + 2.378 + 0.626 + 1.183 = 11.438
DF = 5,
P-Value = 0.043
19.35%
13.27%
8.97%
16.51%
8.82%
13.71%
11.74%
A B C D E F Grand
Total
105. Analysis By Gender
DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
Split By Gender Analysed Percent Absent By Gender
Conclusions
There is significance showing that a
male person would more likely take
unauthorised absence
Female Male Grand Total
22.28% 30.88% 29.08%
157 526 683
45 235 280
Chi-Square Test: F, M
Expected counts are printed below
observed counts
F M Total
1 188 643 831
174.31 656.69
2 14 118 132
27.69 104.31
Total 202 761 963
Chi-Sq = 1.075 + 0.285 +
6.767 + 1.796 = 9.924
DF = 1,
P-Value = 0.002
15.51%
13.71%
6.93%
Female Male Grand Total
106. Analysis By Time Served
Group Range
A <1Year
B 1-2 Years
C 2-3 Years
D 3-4 Years
E >4 Years
DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
Analysed Percent Absent By Time Served
Conclusions
There is a significance in that
E = 4 years +
is more likely to take unauthorized
absence
Created Time Served Groupings
Chi-Square Test: A, B, C, D, E
Expected counts are printed below observed counts
A B C D E Total
1 117 129 151 324 110 831
113.04 126.85 154.46 314.97 121.67
2 14 18 28 41 31 132
17.96 20.15 24.54 50.03 19.33
Total 131 147 179 365 141 963
Chi-Sq = 0.138 + 0.036 + 0.078 + 0.259 + 1.120 +
0.872 + 0.229 + 0.489 + 1.630 + 7.050 =
11.902
DF = 4,
P-Value = 0.018
12.24%
15.64%
11.23%
21.99%
13.71%
10.69%
A B C D E Grand
Total
107. Analysis By Department
DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
Created Department Sub-Groups Percent Absent By Sub Group
Group dept_cd Dept Name Total Abs Never
A 83210 Tube Material Warehouse 46.15% 6 7
B 81400 Design/Process Engineering 35.71% 5 9
C 81320 Sputter section 32.50% 13 27
D 81310 Spin section 21.21% 7 26
E 84120 Tube OQC section 16.13% 5 26
81140 Maintenance 1 section
81510 Maintenance Section(2/3)
81230 Gunseal and Exhaust section
81330 I.T.C. section
81340 Outgoing section
81450 CS-Reinspection section
81210 Assembly Section
81220 Inner Dag section
I 84110 Tube IQC section 7.69% 1 12
J 81240 1st Inspection section 5.63% 4 67
81110 Screen Coating section
81120 Chemical section
L 81130 Shadow Mask Section 2.50% 2 78
Average 12.72% 109 748
Conclusions
This shows that there is a significance in departments
A = Tube Material Warehouse, B = Design/Process Engineering,
C = Sputter section.
In which all are more likely to take unauthorised absence
F
G
H
14.55%
13.68%
10.29%
47
284
61
8
45
7
K 5.45% 6 104
35.71%
32.50%
21.21%
16.13%14.55%13.68%
10.29%
7.69%
5.63% 5.45%
2.50%
46.15%
A B C D E F G H I J K L
108. Analysis By Shift
DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
109. Improvement Suggestions
Improvement Actions/Suggestions by CTQ.
DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
CTQ Proved By Improvement Recommendation
Age Proved By Chi Square
Test Use best fit when recruiting.
Gender Proved By Chi Square
Test Use best fit when recruiting.
Length Of Service Proved By Chi Square
Test Use best fit when recruiting.
Shift Pattern Proved By Chi Square
Test Build In More flexibility for day shift workers.
Department Proved By Chi Square
Test Compare Management Styles
Morale Proved By HARP
Survey New Incentive Scheme (Ongoing)
Accidents Proved By HARP
Survey New Health & Safety Structure In Place
Envirionment Proved By HARP
Survey Music & Improved Rest Room Facilities
Management Style Proved By HARP
Survey
Training Courses For Manager On Interpersonal Skills.
Management Attitude Improvement Plan Next Slide
Aggressive Target Setting Proved By HARP
Survey Unable to improve due to the nature of our business.
110. DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
Improvement Suggestions - Management Attitude
Morale Absence
More
Information
Team Building
Improved
Interpersonal
Skills
Treat
Operators As
Equal
More 1 To 1
Communication
Follow Correct
Procedures
Improved
Grading
System
HIGH MORALE = LOW ABSENCE
111. DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
Improvement Suggestions - Attendance Bonus
£100 Per
Year
Deductions?
All Authorised
Non- Sickness
Absence
Paid Quarterly
Cash/Vouchers/Sa
vings
Decided By
Incentive Scheme
Working Party
All Absence
Resulting In
Warnings
Verbal -25%
Written - 50%
F Written - 100%
Total
Savings
£1,086,000
All
Deductions
For 1 Year
From Date Of
Warning
112. Output
DDeeffiinnee MMeeaassuurree AAnnaallyyssee IImmpprroovvee CCoonnttrrooll
Initial Current
s level 3.21 3.43
PPM (Month) 46021 27300
Loss (£) £117000 £69405
Est. Monthly Saving of: £47595
(Based on hours lost)
113. Process Sterilization
Capability Up
Contents:
1. Define Step
2. Measure Step
3. Analysis Step
4. Improvement
5. Control
6s Champion Review
Final Report
114. Background M A I C D
Develop working efficiency and found 66 SSiiggmmaa control for free
Salmonella in sterilization process.
Process Capability Current Condition
Process Capability Sixpack for Sealing Angle Line #2
0 100 200 300 400 500
1.0
0.5
0.0
-0.5
Individual and MR Chart
Obser.
Ind ivid ua l V a lue
UCL=0.7674
Mean=-0.01707
LCL=-0.8016
0.9
0.6
0.3
0.0
Mo v.Ra ng e
UCL=0.9638
R=0.2950
LCL=0
Last 25 Observations
480 490 500
0.3
0.0
-0.3
-0.6
Observation Number
V a lue s
Capability Histogram
-0.5 0.0 0.5
Normal Prob Plot
-0.5 0.0 0.5
Capability Plot
Process Tolerance
Within
I I I
Overall
I I I
Specifications
I I
-0.5 0.5
Within
StDev:
Cp:
Cpk:
0.261507
0.64
0.62
Overall
StDev:
Pp:
Ppk:
0.288617
0.58
0.56
2
0.65
1.5
0.51
Target Current
Cp
Cpk
115. D M A I C
Potential X’ List
Executing analysis with Logic Tree for sterile product.
Big Y X1 X2 X3
Sterile Product Machine Sterile Time Fixed
Rotate
Temp Warm
Hot > 97
Steam Supply Spec (6.5 ~ 9) bar
Pineapples Bracket Stand
Material
Juice pH
Methods Automatic
Manual
116. Gage R&R D A I C M
Sealing angle Line 2
Mar 3rd,2006
novi m
0.05
0.01
Gage name:
Date of study:
Reported by:
Tolerance:
Misc:
Gage R&R (ANOVA) for Measurement
0
0.5
0.0
-0.5
R Chart by Operator
Xbar Chart by Operator
A B C
Sam ple Mean
UCL=0.01198
LUMCeCLaLn==0=0.0.00.0233370201767
0
0.010
0.005
0.000
A B C
Sample Range
R=0.003667
LCL=0
1 2 3 4 5 6 7 8 9 10
1 2 3 4 5 6 7 8 9 10
0.4
0.3
0.2
0.1
0.0
-0.1
-0.2
-0.3
-0.4
Part
Operator
Operator*Part Interaction
Average
A
B
C
A B C
0.4
0.3
0.2
0.1
0.0
-0.1
-0.2
-0.3
-0.4
Operator
By Operator
0.4
0.3
0.2
0.1
0.0
-0.1
-0.2
-0.3
-0.4
Part
By Part
%Contribution
%Study Var
Gage R&R Repeat Reprod Part-to-Part
100
50
0
Components of Variation
Perc ent
Gage R&R
%Contribution
Source VarComp (of VarComp)
Total Gage R&R 0.000020 0.03
Repeatability 0.000020 0.03
Reproducibility 0.000000 0.00
Operator 0.000000 0.00
Part-To-Part 0.077747 99.97
Total Variation 0.077767 100.00
StdDev Study Var %Study
Var
Source (SD) (5.15*SD) (%SV)
Total Gage R&R 0.004437 0.02285 1.59
Repeatability 0.004425 0.02279 1.59
Reproducibility 0.000323 0.00166 0.12
Operator 0.000323 0.00166 0.12
Two-Way ANOVA Table With Interaction
Source DF SS MS F P
Part 9 4.19852 0.466502 21530.8 00..0000000000
Operator 2 0.00004 0.000022 1.0 0.38742
Operator*Part 18 0.00039 0.000022 1.2 0.33365
Repeatability 30 0.00055 0.000018
Total 59 4.19950
If significant, P-value < 0.05 indicates
that a part is having a variation for
Some measuring
Ok for “product acceptance”
considering a products
tolerance.
117. M
D A I C
Measurement
Through analysis of process capability , getting sigma level 11..8855 s
Four Block Diagram
Target
1 2 3 4 5 6
2.5
2.0
1.5
1.0
0.5
Poor
Z Shift
Process Control
Good
Poor Good
Technology
Block A
Block C
Block B
Block D
1.85 s
4.5 s
Z Shift
Process capability for Sterile Product
Process Capability Analysis for Sealing
Angle Line #2
LSL Target USL
-1.0 -0.5 0.0 0.5 1.0
Potential (Within) Capability
Overall Capability Observed Performance Exp. "Within" Performance Exp. "Overall" Performance
PPM < LSL
PPM > USL
PPM Total
PPM < LSL
PPM > USL
PPM Total
PPM < LSL
PPM > USL
PPM Total
USL
Target
LSL
Mean
Sample N
StDev (Within)
StDev (Overall)
Z.Bench
Z.USL
Cpk Z.LSL
Cpm
Z.Bench
Z.USL
Z.LSL
Ppk
47140.05
36602.01
83742.06
32395.08
24004.66
56399.74
4008.02
0.00
4008.02
0.500000
0.000000
-0.500000
-0.017074
499
0.261507
0.288617
1.59
1.98
1.85
0.62
0.58
1.38
1.79
1.67
0.56
Process Data
Within
Overall
A : Poor control, inadequate technology
B : Must control the process better, technology is fine
C : Process control is good, inadequate technology
D : World class
118. A
Analysis - Regression D M I C
Use regression is to express and analyze a mathematical equation of describing a relationship. That is, it is
to fit a mathematical equation of describing a relationship between the “YY” and “XX’’ss”.
Regression Analysis: Sterile product versus time and temp
The regression equation is
Sterile =0.000303 + 0.00113 time + 0.000060 temp
Regression Analysis: Sterile product versus time and temp
The regression equation is
Sterile =0.000303 + 0.00113 time + 0.000060 temp
Predictor Coef SE Coef T P
Constant 0.0003033 0.0002999 1.01 0.345
TTiimmee 00..0000111122885599 00..0000000033993399 2288..6655 00..000000
Temp 0.00006005 0.00005694 1.05 0.327
Predictor Coef SE Coef T P
Constant 0.0003033 0.0002999 1.01 0.345
TTiimmee 00..0000111122885599 00..0000000033993399 2288..6655 00..000000
Temp 0.00006005 0.00005694 1.05 0.327
S = 0.0003891 R-Sq = 100.0% R-Sq(adj) = 100.0%
S = 0.0003891 R-Sq = 100.0% R-Sq(adj) = 100.0%
Normal Probability Plot of the Residuals
(response is Angle)
Analysis of Variance
Source DF SS MS F P
Regression 2 0.82500 0.41250 2.725E+06 0.000
Residual Error 7 0.00000 0.00000
Total 9 0.82500 -0.0005 0.0000 0.0005
Analysis of Variance
Source DF SS MS F P
Regression 2 0.82500 0.41250 2.725E+06 0.000
Residual Error 7 0.00000 0.00000
Total 9 0.82500
1
0
-1
N o rm a l S c o re
Residual
p-value < 0.05 :
Significant factor
R2 and R2-adj are over 90% :
which indicates a potentially good fit
119. Analysis – Regression D M I C
Comparing of Sterile product and steam supply to find what the factor’ level’s which influence
enormously by represent characterized variation ““YY”” by the total sum of square.
Regression Analysis: Sterile product versus Steam supply
The regression equation is
Sterile product = - 0,380 + 3,74 Steam supply
Predictor Coef SE Coef T P
Constant -0,3796 0,1693 -2,24 0,042
SStteeaamm SSuuppppllyy 33,,774444 11,,115533 33,,2255 00,,000066
Regression Analysis: Sterile product versus Steam supply
The regression equation is
Sterile product = - 0,380 + 3,74 Steam supply
Predictor Coef SE Coef T P
Constant -0,3796 0,1693 -2,24 0,042
Steam Supply 3,744 1,153 33,,2255 00,,000066
S = 0,1510 R-Sq = 43,0% R-Sq(adj) = 38,9%
S = 0,1510 R-Sq = 43,0% R-Sq(adj) = 38,9%
Analysis of Variance
Source DF SS MS F P
Regression 1 0,24032 0,24032 10,55 0,006
Residual Error 14 0,31905 0,02279
Total 15 0,55937
Analysis of Variance
Source DF SS MS F P
Regression 1 0,24032 0,24032 10,55 0,006
Residual Error 14 0,31905 0,02279
Total 15 0,55937
The P-value < 0.05
Reject Ho ; Accept ha
A
-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3
4
3
2
1
0
Residual
Frequency
Histogram of the Residuals
(response is Cullet S)
120. A
Analysis – Chi-square D M I C
This Chi-Square is used to Test hypotheses about the frequency of occurrence of some event
happening with equal probability.
Conclusion:
◆ At least no one region is different, because a
dependence exists. (P > 0.05)
◆ It no appears that the dependence may exist with
Region 1 due to the large difference between the
observed and the expected values.(must
subtract the expected and observed values)
Since P-Value >> 0.05; there’s no significant
Effect between product sterile and factor.
Chi-Square Test: matang, Stngh matang,
juice
Expected counts are printed below observed counts
Chi-Square contributions are printed below
expected counts
Stngh
matang matang juice Total
OK 1000 995 1013 3008
1000.01 993.03 1014.96
0.000 0.004 0.004
NG 3 1 5 9
2.99 2.97 3.04
0.000 1.308 1.269
Total 1003 996 1018 3017
Chi-Sq = 2.585, DF = 2, P-Value = 0.275
3 cells with expected counts less than 5.
121. Analysis –
A
two sample T-test D M I C
Hypothesis tests help to determine if a difference is real, or if it could be due to
chance
Two-Sample T-Test and CI: Automatic, Chart
Two-sample T for Automatic vs Chart
N Mean StDev SE Mean
Automatic 12 14.70 1.47 0.42
Manual 12 14.13 2.19 0.63
Difference = mu (Automatic) - mu (Chart)
Estimate for difference: 0.562500
95% CI for difference: (-1.030754,
2.155754)
T-Test of difference = 0 (vs not =): T-Value
= 0.74 P-Value = 0.469 DF = 19
D a t a
Automatic Chart
19
18
17
16
15
14
13
12
11
Boxplot of Automatic, Chart
There is no statistically significant difference
if the confidence interval for m1 - m2 does
include 0.0.
122. Improvement – D M A C
Response Surface Experiment
MMaaiinn EEffffeeccttss PPlloottss ffoorr TTiimmee,, TTeemmpp && SStteeaamm ssuuppppllyy AAvveerraaggee aanndd SSttaannddaarrdd DDeevviiaattiioonn ooff RReessiidduuee..
9
Cube Plot (data means) for Salmonella
6.5
97
90
10
7
0
7.5 10
Steam
Temp
Time
3
4 11
15 12
I
Main Ef fects Plot (data means) for Salmonel la
From the Main Effects Plot for the average of residue we
conclude:
• Temp has the greatest effect on average residue
• Time has a lesser effect on average residue
• Steam supply shows little or no effect (within the test range)
on the average residue
Best
Condition:
• Salmonella : 0
• Temp max : 97 oC
• Time : 7.5 min
• Steam : 6.5 bar
Mean of Salmonella
7.5 10.0
10.0
7.5
5.0
90 97
6.5 9.0
10.0
7.5
5.0
Time Temp
St eam
123. Improvement –
Response Surface Experiment
D M A I C
Contour Plot
Lines of target
response for
“0” Salmonella
1. Get to know the condition giving lower salmonella.
2. To get the regular response, we realize what variables
is important to control (temp & time)
3. Determine the level of independent variance needed for
getting salmonella 00 (When temp is approximately 97oC)
97
96
95
94
93
92
91
90
Contour Plot of Salmonella vs Temp, Time
7.5 8.0 8.5 9.0 9.5 10.0
Time
Te m p
Salmonella
< 2.0
2.0 - 4.5
4.5 - 7.0
7.0 - 9.5
9.5 - 12.0
12.0 - 14.5
> 14.5
Hold Values
Steam 6.5
Interpretation: to make sterile product (no salmonella) move towards the center corner of the
Contour Plot (samonella = 00). Read off potential “Time” and “Temp” values that will provide Salmonella
< 2.
124. D M A I C
Improvement –
Response Surface Experiment
Generally, main effect is more important than interaction. If interaction is regarded as a important thing,
then interaction can be used as a factor of interaction and another interaction might be confounded.
96
Salmonella
10
5
0 94
T emp
15
8 92 9 10 90 T ime
Hold Values
Steam 6.5
Surface Plot of Salmonella vs Temp, Time
125. Improvement –
Result
D M A I C
Four Block Diagram
1 2 3 4 5 6
2.5
2.0
1.5
1.0
0.5
Poor
Z Shift
Process Control
Good
Poor Good
Technology
Block A
Block C
Block B
Block D
4.25 s
Z Shift
1.85 s
Improvement Result:
Saving Cost estimated: 2.7K U$/Year
Process Capability Analysis for Sealing
Angle Line #2
LSL Target USL
0.500000
0.000000
-0.500000
0.014762
84
0.121123
0.124193
Potential (Within) Capability
3.94
4.01
1.34 4.25
-0.50 -0.25 0.00 0.25 0.50
Overall Capability Observed Performance Exp. "Within" Performance Exp. "Overall" Performance
PPM < LSL
PPM > USL
PPM Total
PPM < LSL
PPM > USL
PPM Total
PPM < LSL
PPM > USL
PPM Total
USL
Target
LSL
Mean
Sample N
StDev (Within)
StDev (Overall)
Z.Bench
Z.USL
Cpk Z.LSL
Cpm
Z.Bench
Z.USL
Z.LSL
Ppk
17.00
46.70
63.70
10.69
30.86
41.55
0.00
0.00
0.00
1.34
3.83
3.91
4.14
1.30
Process Data
Within
Overall
126. Control
D M A I C
Six Sigma Quality focuses on moving control upstream to the leverage input characteristic for Y.
If we can measure and control the vital few X’s, control of Y should be assured.
Desired Process Capability
Output
Input Controller
Process
Group Member
Controllable factors:
- Miss adjust causes
- Adjustable check
- Pad control
- Education
Upper Control Limit
X
●
Lower Control Limit
●
0.5
0.0
-0.5
Sample Mean
Output
Subgroup 0 50 100
UCL=0.4384
Mean=0.001188
LCL=-0.4360
1.0
0.5
0.0
Sample Range
1 1 1
UCL=0.7596
R=0.2325
LCL=0
Xbar/R Chart for Sealing Angle Line #2
Process Standard Change
127. Optimizing Material DIO
Contents
6s Champion Review
DMAIC-Step Report
1. DEFINE
2. MEASURE
3. ANALYSIS
4. IMPROVEMENT
5. CONTROL
128. ‘06 6s Project Registration
PJT Name
Period
Team
Name
Optimizing Material D I O
Main Improvement Object Breakthrough (KPI) Current World Best Target New Idea for Target Achievement
Just In Time Purchasing
Vendor Managed Inventory
Team Formation, Related Department Involved
Name Dept. Position Main Role
NECK POINT
- Warehouse Inventory Amount D I O
2.6 days 2.3 days - Working In Process Inventory Amount
How to do ? Why ?
(* Selection Background)
Expected
Results
Quantitative
1. JIT delivery system for press part
2. Door to door delivery for glass
3. Hub delivery system from Korea
4. Raw material issue control to process
5. Weekly stock taking
6. Minimize NG and rework stock in process
Qualitative
at end of the month
7. PO issued based on the latest production
plan
1. D I O is one of key performance indicators in
inventory management.
2. Good level of inventory will support production
line in effective and efficient way.
3. Fluctuated material D I O
Current Condition
USL : 2.9
LSL : -
Means : 2.58091
Sample N : 14
Z Bench : 1.03
( Days Inventory
Outstanding )
Process Capability Analysis for C2
1.5 2.0 2.5 3.0 3.5
Potential (Within) Capability
Overall Capability Observed Performance Exp. "Within" Performance Exp. "Overall" Performance
PPM < LSL
PPM > USL
USL
Target
LSL
Mean
Sample N
StDev (Within)
StDev (Overall)
Z.Bench
Z.USL
Z.LSL
Cpk
Cpm
Z.Bench
Z.USL
Z.LSL
2.90000
*
*
2.58091
14
0.310413
0.380447
1.03
1.03
*
0.34
*
0.84
0.84
*
1,117 K $ US
*
214285.71
Current
PPM < LSL
PPM > USL
*
151989.40
1,050
- Reduced warehouse and WIP inventory amount 1NG and rework stock in the end of month
- All material used efficiently in production line
PPM < LSL
PPM > USL
*
200815.40
Target - Continuous material supplies to production line
2. Fluctuate production schedule
K 67
UUSSLL
PPM Total
PPM Total
PPM Total
Ppk
200815.40
151989.40
214285.71
0.28
Process Data
Within
Overall
129. Background
D
M A I C
One of the material inventory management control is DIO (Days Inventory Outstanding) that
has the formula:
Inventory Amount
Material DIO = ------------------------- X total days of current month
Sales Amount
The elements of material DIO are:
1. Warehouse Inventory (Raw Material)
2. Working In Process inventory (Semi Finished Goods )
3. Material In Transit Inventory
Current Material Inventory Condition :
Average Amount : $ K 1,117
D I O : 2.6 days
This large amount and high DIO have some effects :
Risk in obsolescence, expired, lost, and
defect
High inventory carrying cost
Optimizing inventory amount and D I O will bring material inventory management in a more
efficient cost
130. D
M A I C
Y X1 X2 X3 X4 X5
M. RATIO
Net BOM
Material Price
Src.
Performance
BOM Quantity
M. YIELD
CPT Price
Experience
Negotiation
Education
M. DIO
Material Loss
Net Req. Material
Material Cost Amt
Market Situation
Key Part
Chemical
Others
Mat. Inv. Amt
Production Qty
Sales Amt
Beginning Stock
Sales Target
Forecast Skill
Calculation Skill
Purchasing Qty
Ending Stock
Receive Amt
Total Days
In Transit Inv.
Warehouse Inv.
W I P Inv.
Sales Qty.
Current Month
Supply Condition
Glass
Mask
Press Part
Sub Mount
Process Part
Assy Part
Logic Tree
131. Logic Tree D M A I C
Y X1 X2 X3 X4 X5
Mat Inv Amt
In Transit Inv Delivery Sched.
Material
DIO
Sales Amt
B
Graphite G 72 B
Warehouse Inv Experience
Sales Qty
Price
A
Education
Total Days
Simulation Skill
Glass
Stock
Bulb
D Y
Part Stock
Assy Stock
Current Month
Sales Target
Production Capa
CMA
M/Assy
Marketing Nego
Market Situation
W I P Inventory
YS
TCL Prod.
Comp. Supply
Sub Month
132. D M A I C
This project with potential X-List will be focused to control Warehouse Inventory and working In
Process Inventory
X Level
X Level
2
2
X Level
3
DD Y Y
BBigig Y Y
MMaateteriraial lD DIOIO
PP A A D D
PPaartr tS Stotockck
AAssssy yS Stotockck
X Level
4
FFlalat tM Maasksk
GGlalassss
SSuubb M Moouunnt t
PPhhoosspphhoor r
GG 7 722 B B
X Level
X Level
1
1
MMaateteriraial lI nInvv. .A Ammtt
WW/h/hoouussee I nInv.v.
WWIPIP I nInv.v.
X Level
3
X Level
4
ggoooodd
AAssssyy
CCMMAA
Brainstorming Potential X
133. D M A I C
Gage R & R
To test validation of measurement , 3 (three) persons, twice inspection and 14 months calculation for
material D I O has carried out, and the result was acceptable.
Gage R&R
StdDev Study Var %Study
Var
Source (SD) (5.15*SD) (%SV)
Total Gage R&R 0.109148 0.56211 16.58
Repeatability 0.109046 0.56159 16.56
Reproducibility 0.004723 0.02432 0.72
Operator 0.004723 0.02432 0.72
Part-To-Part 0.649363 3.34422 98.62
Total Variation 0.658472 3.39113 100.00
Number of Distinct Categories = 8
Gage name:
Date of study:
Reported by:
Tolerance:
Misc:
Gage R&R (ANOVA) for Measure
0
4
3
2
R Chart by Operator
Xbar Chart by Operator
1 2 3
S a m p le M e a n
LUMCeCLaLn==2=2.2.67.3624749
0
1.0
0.5
0.0
1 2 3
S a m p le R a n g e
LRUC=C0LL.==0002.40075857
1 2 3 4 5 6 7 8 9 10 11 12 13 14
1 2 3 4 5 6 7 8 9 10 11 12 13 14
4
3
2
Part
Operator
Operator*Part Interaction
A v e r a g e
1
2
3
1 2 3
4
3
2
Operator
By Operator
4
3
2
Part
By Part
%Contribution
%Study Var
%Tolerance
Gage R&R Repeat Reprod Part-to-Part
350
300
250
200
150
100
50
0
Components of Variation
P e r c e n t
Gage R&R
<20%
Acceptable
The result of Gage R&R total is 16.58, the acceptance percentage is below 20 (<20), meanwhile the
result of measurement between <20, it accepted.
134. Current Condition DM A I C
Process Capability Analysis for C2
UUSSLL
1.5 2.0 2.5 3.0 3.5
Potential (Within) Capability
Overall Capability Observed Performance Exp. "Within" Performance Exp. "Overall" Performance
PPM < LSL
PPM > USL
PPM Total
PPM < LSL
PPM > USL
PPM Total
PPM < LSL
PPM > USL
PPM Total
USL
Target
LSL
Mean
Sample N
StDev (Within)
StDev (Overall)
Z.Bench
Z.USL
Z.LSL
Cpk
Cpm
Z.Bench
Z.USL
Z.LSL
Ppk
*
200815.40
200815.40
*
151989.40
151989.40
*
214285.71
214285.71
2.90000
*
*
2.58091
14
0.310413
0.380447
1.03
1.03
*
0.34
*
0.84
0.84
*
0.28
Process Data
Within
Overall
Z-Bench :
1.03
135. Block Diagram D M
A I C
Z shift has been identified , the Z shift will be 0.19, regarding this issue the target
of the project is 4.5 sigma
A B
C
2.5
2.0
1.5
1.0
0.5
1 2 3 4 5 6
Z shift
Poor
Process Control
Good
Poor Z st Good
Technology
EXPLANATION
Position of DIO was column C ,
it was mean :
PROCESS CONTROL IS GOOD,
BUT TECHNOLOGY (METHOD)
IS BAD
Z st : 1.03
Z shift : Z st – Z lt
Z st : 1.03
Z shift : Z st – Z lt
: 1.03 – 0.84
: 0.19
: 1.03 – 0.84
: 0.19
D
136. D M A I C
WIP Assy Stock Analysis
The Pareto analysis has been done, the result shows that tube stock, Furnace, CMA, and Assy
have the highest contribution to WIP Assy Amount
Bare Tub e
165826 48983 37174 33504 7725 14962
53.8 15.9 12.1 10.9 2.5 4.9
53.8 69.7 81.8 92.6 95.1 100.0
Others
CP T D o ngbang
Mo unt Assy
C MA
B ulb
300000
200000
100000
0
100
80
60
40
20
0
Defect
Count
Percent
Cum %
P e rc e n t
C o u n t
Avg WIP Stock 1Q '04
Analysis
BBBBaaaarrrreeee TTTTuuuubbbbeeee
45 % conveyor stock ($ 78 K)
55 % ( $ 97 K) consists
-O Nf G: & rework,
- Pending Lot
- Remained Prod. Stock
BBBBuuuullllbbbb
F’ce Stock
C/V Stock
CCCCMMMMAAAA
Stock to keep production
MMMMoooouuuunnnntttt AAAAssssssssyyyy
Safety Stock to secure
supply
2 shift Mount Assy Process
Pareto Chart for WIP Assy Inv.
137. WIP Part Stock Analysis D M A I C
The Pareto analysis result for WIP part stock shows that Mask Stock, Glass, Sub Mount and
Phosphor have the highest contribution to WIP part Amount Analysis
Others
Pho spor
S ub M ount
Ma sk
la ss
G 179511 50418 32501 10535 7375
64.0 18.0 11.6 3.8 2.6
64.0 82.0 93.6 97.4 100.0
200000
100000
0
100
80
60
40
20
0
Defect
Count
Percent
Cum %
P e rc e n t
C o u n t
Pareto Chart for Desc.
FFFFllllaaaatttt MMMMaaaasssskkkk
86 % stock at PT YSI($ 156 K)
- Flat Mask (RM)
- Annealing
- Forming & Blackening
14 % stock at SM ( $ 23 K)
GGGGllllaaaassssssss
Stock
loading
SSSSuuuubbbb MMMMoooouuuunnnntttt
Mount Assy Process stock
PPPPhhhhoooosssspppphhhhoooorrrr
Safety Stock for aging time
High price
Pareto Chart for WIP Part Inv.
138. Warehouse Stock Analysis D M A I C
The Pareto analysis result for warehouse stock shows that Stock, Graphite, Phosphor and PAD
have the highest contribution to Warehouse Stock amount Analysis
DDDD YYYY
Inner Supply
- Revised Production Plan
So frequently
- Quality Problem
- Comp. Request to achieve
sales target
GGGGrrrraaaapppphhhhiiiitttteeee
Hardly PO revision
PO issued three month before
PPPPhhhhoooosssspppphhhhoooorrrr
PPPP AAAA DDDD
Decreasing consumptions
G la ss
P AD
P ho spo r
Graphi te
D Y
47520 41443 39654 17572 16007
29.3 25.6 24.4 10.8 9.9
29.3 54.8 79.3 90.1 100.0
150000
100000
50000
0
100
80
60
40
20
0
Defect
Count
Percent
Cum %
P e rc e n t
C o u n t
Pareto Chart for C1
Blue Stock
Pareto Chart for W/House Inv.
Minimum order 2304 kg
139. D M A I C
Item High Amount Stock Remarks
Selected as Vital View
Selected as Vital View
Not Selected as Vital View
Assy
Remained production stock & Pending lot
conveyor stock
W I P Selected as Vital View
Selected as Vital View
Part
NG and rework Stock
Mask Stock outsourcing process
Warehouse
Analysis Result
Glass Stock
Sub Mount
Selected as Vital View
Not Selected as Vital View
Y D Y
G 72 B & G 355
P A D
Selected as Vital View
Selected as Vital View
Phosphor
Selected as Vital View
140. D M A I C
Bottle Neck
Analysis Result
1. High stock of NG, rework, pending lot in the end of month due to quality
problem, sourcing team can not fully controlled this situation. Actually, it’s depend
on process and quality performance.
2. Outsourcing Mask Annealing process at Shin require more raw material
stock to keep production and secure supply.
3. Frequently change production plan for CIT and Y DY quality problem cause influence
high stock Y DY.
141. Improvement
I
D M A
C Item Improvement Remarks
Stock
Mask Stock
- Identify and checking 3 days before closing
- Communicate Stock to related dept and
push for action
- Working closely with PCT Team to input remained stock
- Partial raw material delivery
Glass Stock
- Daily vendor managed inventory
In the end of month
- Pending Lot
- Remained Prod.
from March ‘04
Annealing
Outsourced
- Optimized in out stock control
- Daily input glass to production line
- Communicate and push process to minimize stock
- Check stock condition & make any necessary action
Y D Y
- Confirm I production plan
- Best effort to match PCT Production Plan & actual
- Just In Time purchasing
Phosphor
Graphite
- Improve Import delivery simulation skill
Weekly control
from March ‘04
Weekly control
from March ‘04
- Tightly control on ETD & ETA
- Maintain actual delivery performance on SRS
from March ‘04
from March ‘04
142. D M A
I
Improvement Result C
Result analysis after improvement actions : Z bench 2.60
UUSSLL
Process Capability Analysis for DIO
1.5 1.7 1.9 2.1 2.3 2.5 2.7 2.9
Potential (Within) Capability
Overall Capability Observed Performance Exp. "Within" Performance Exp. "Overall" Performance
PPM < LSL
PPM > USL
PPM Total
PPM < LSL
PPM > USL
PPM Total
PPM < LSL
PPM > USL
PPM Total
USL
Target
LSL
Mean
Sample N
StDev (Within)
StDev (Overall)
Z.Bench
Z.USL
Z.LSL
Cpk
Cpm
Z.Bench
Z.USL
Z.LSL
Ppk
*
1307.83
1307.83
*
4651.78
4651.78
*
0.00
0.00
2.90000
*
*
2.25093
6
0.249574
0.215663
2.60
2.60
*
0.87
*
3.01
3.01
*
1.00
Process Data
Within
Overall
Z-Bench :
2.60
143. D M A
I
Sigma Value C
After improvement action, we can compare it with previous condition.
Improved Condition is better than Previous Condition.
PREVIOUS CONDITION
IMPROVED CONDITION
A B
C
2.5
2.0
1.5
1.0
0.5
1 2 3 4 5 6
Poor
Z shift
Process Control
Good
Poor Z st Good
Technology
2.5
2.0
1.5
1.0
Poor
Sigma = 1.03
D
Z st : 2.60
Z shift : Z st – Z lt
Z st : 2.60
Z shift : Z st – Z lt
: 2.60 – 3.01
: -0.41
: 2.60 – 3.01
: -0.41
A B
C
0.5
1 2 3 4 5 6
Z shift
Process Control
Good
Poor Z st Good
Technology
D
Z st : 1.03
Z shift : Z st – Z lt
Z st : 1.03
Z shift : Z st – Z lt
: 1.03 – 0.84
: 0.19
: 1.03 – 0.84
: 0.19
Sigma = 2.60
144. D M A
I
Saving Cost C
TARGET
1,117
K $ US
Current
1,050
Target
K 67
(6%)
RESULT
1,117
K $ US
Before
1,075
After
K 42
(4%)
Previous Average Material Inventory Amount : $ K 1,117
Current Average Material Inventory Amount : $ K 1,075
Saving Cost : $ K 42
149. Reduce LNG Usage
Date Nov 24th , 2004
Process Technique Group
Prepared : Novi Muharam
6s Champion Review
DMAIC Report
Contents
1. Define Step
2. Measure Step
3. Analysis Step
4. Improvement
5. Control
150.
151. Background
D M A I C
LNG (Liquid Natural Gas) is source of energy for combustion process in F’ce
Up to 3rd quarter LNG usage for exhaust furnace still higher, it’s about 5000 Nm3/day.
This project have a purpose to decreasing LNG usage in furnace
LNG Usage
4500
5000
CCuurrrreenntt TTaarrggeett
11%
Unit:
(Nm3/day)
Energy Unit Price (U$)
How to do:
LNG & Air pressure system
Adjustment to find best ratio
Both of them.
Target Saving cost:
= (5000 –4500 )Nm3/day x 0.165 U$/Nm3 x 30 x 12
= 500 Nm3/day x 0.165 x 30 x 12
= 2299,,770000 UU$$//YYeeaarrss
Others
Electric
O2
N2
LNG
0.1650 0.1309 0.1290 0.0605 0.0070
33.5 26.6 26.2 12.3 1.4
33.5 60.1 86.3 98.6 100.0
0.5
0.4
0.3
0.2
0.1
0.0
100
80
60
40
20
0
Defect
Count
Percent
Cum %
Percent
Count
Energy Usage Price
152. G/S
Process Mapping
Load Exh
P-pipe
Exhaust Furnace
Zone #1 ~ #44
In
Out
Tip off B.B.D
-F/F C/V -Cart
Unload
Robot
Robot Control Panel
Keeping
Zone
Up Slope
Zone
Down Slope
D M A I C
CTQ Area : - LNG & Air Usage
153. Big Y X1 X2 X3
LNG Ratio
Usage
Material LNG Pressure Ratio
Air Pressure Ratio
Machine TIC Temperature
RC Fan RPM Motor
Dumper Valve
Exh Blower Pressure
D M A I C
Brainstormed Potential X List
154. Gage R&R for Exhaust F’ce Line #1 Measurement
GaR StdDev Study Var
%Study Var
Source (SD) (5,15*SD) (%SV)
Total Gage R&R 0,009704 0,04998 1,40
Repeatability 0,002582 0,01330 0,37
Reproducibility 0,009354 0,04817 1,35
Operator 0,002566 0,01321 0,37
Operator*Part 0,008995 0,04633 1,30
Part-To-Part 0,692590 3,56684 99,99
Total Variation 0,692658 3,56719 100,00
Gage name:
Date of study:
Reported by:
Tolerance:
Misc:
Exhaust F'Ce Measurement
Oct 19th, 2005
Novi M
Gage R&R (ANOVA) for Auto 14"
100
50
0,010 Eng'r Gijo Maker 1 maker 2 PQC
0,005
0,5 Eng'r Gijo Maker 1 maker 2 PQC
0
0,0
-0,5
-1,0
-1,5
Xbar Chart by Operator
S a m p le M e a n
UCL=0,004356
LUMCCeLLa=n=-=-00-,,055,0058032257
0
0,000
R Chart by Operator
S a m p le R a n g e
R=0,001333
LCL=0
a b c d e f
a b c d e f
0,5
0,0
-0,5
0,5
0,0
-0,5
0,5
0,0
-0,5
-1,0
Part
Operator
Operator*Part Interaction
A v e r a g e
Eng'r
Gijo_1
Gijo_2
Gijo_3
PQC
Eng'r Gijo Maker 1 maker 2 PQC
-1,0
Oper
Response By Operator
-1,0
Part
Response By Part
%Contribution
%Study Var
Gage R&R Repeat Reprod Part-to-Part
0
Components of Variation
P e r c e n t
Gage R&R
D M A I C
Gage R&R <20%
Acceptable
155. Capability Process
Exhaust F'ce 14" Capa'
LSL USL
0 2 4 6 8 10
Potential (Within) Capability
Overall Capability Observed Performance Exp. "Within" Performance Exp. "Overall" Performance
PPM < LSL
PPM > USL
PPM Total
PPM < LSL
PPM > USL
PPM Total
PPM < LSL
PPM > USL
PPM Total
USL
Target
LSL
Mean
Sample N
StDev (Within)
StDev (Overall)
7.50000
*
2.50000
5.27297
37
0.98276
1.66305
Z.USL Z.Bench
Z.LSL
Cpk
Cpm
Z.Bench
Z.USL
Z.LSL
Ppk
47716.94
90265.06
137981.99
2389.10
11722.77
14111.87
0.00
81081.08
81081.08
2.27 2.19
2.82
0.76
*
1.09
1.34
1.67
0.45
Process Data
Within
Overall
D M A I C
Four Block Diagram
1 2 3 4 5 6
2.5
2.0
1.5
1.0
0.5
Poor
Z Shift
Process Control
Good
Poor Good
Z st
Technology
Block A
Block C
Block B
Block
D
A : Poor control, inadequate technology
B : Must control the process better, technology is fine
C : Process control is good, inadequate technology
D : World class
156. Analysis
A
Analysis of Temp furnace with result has significant effect to Pressure LNG
Regression Analysis: Press LNG versus Temp
The regression equation is
Press LNG = 3.7 + 21.9 Temp
Regression Analysis: Press LNG versus Temp
The regression equation is
Press LNG = 3.7 + 21.9 Temp
Fail to reject Ho
Predictor Coef SE Coef T P
Constant 3.75 36.18 0.10 0.918
Temp 21.934 6.556 3.35 0.002
S = 64.96 R-Sq = 24.2% R-Sq(adj) = 22.1%
Predictor Coef SE Coef T P
Constant 3.75 36.18 0.10 0.918
Temp 21.934 6.556 3.35 0.002
S = 64.96 R-Sq = 24.2% R-Sq(adj) = 22.1%
Analysis of Variance
Analysis of Variance
Accept Ha
Source DF SS MS F PP
Source DF SS MS F PP
Regression 1 47242 47242 11.19 00..000022
Residual Error 35 147702 4220
Total 36 194945
Regression 1 47242 47242 11.19 00..000022
Residual Error 35 147702 4220
Total 36 194945
The p-value < 0.05, Fail to accept
Ho, which demonstrate statistical
significance (an equation with a
“good” fit)
LNG
Pressure
Manometer
Gauge
D M I C
Normal Probability Plot
Anderson-Darling Normality Test
A-Squared: 0.414
P-Value: 0.321
Average: 5.27297
StDev: 1.65154
N: 37
2.2 3.2 4.2 5.2 6.2 7.2 8.2 9.2
.999
.99
.95
.80
.50
.20
.05
.01
.001
Probability
Qty LNG
Hypothesis Analysis :
Ho : Temperature furnace has no significant effect to LNG Pressure
Ha : Temperature furnace has significant effect to LNG Pressure
157. Test for Equal Variances for LNG Ratio
95% Confidence Intervals for Sigmas
One-way ANOVA: Air Press, Press LNG,
Temp
0 100 200 300 400 500 600 700 800 900
Analysis of Variance
Source DF SS MS F PP
Factor 2 1769458 884729 118.10 00..000000
Error 108 809084 7492
Total 110 2578542
Bartlett's Test
Test Statistic: 3.282
P-Value : 0.858
Levene's Test
Test Statistic: 2.096
P-Value : 0.131
Factor Levels
75
75
75
75
75
75
75
75
75
75
75
75
75
75
75
75
75
75
75
75
75
75
75
75
75
75
0.1
0.2
0.5
0.6
1.7
2.0
3.0
4.0
5.0
6.0
6.8
7.0
7.3
8.0
9.0
9.1
9.3
9.8
11.0
11.6
11.8
12.0
12.9
13.5
14.2
14.6
One-way ANOVA: Air Press, Press LNG,
Temp
Analysis of Variance
Source DF SS MS F PP
Factor 2 1769458 884729 118.10 00..000000
Error 108 809084 7492
Total 110 2578542
Individual 95% CIs For Mean
Based on Pooled St Dev
Individual 95% CIs For Mean
Based on Pooled St Dev
Level N Mean StDev --+---------+---------+---------+----
Air Pres 37 424.73 73.53 (-*--)
Press LN 37 117.00 73.04 (--*--)
Temp 2 37 297.54 108.32 (--*--)
Level N Mean StDev --+---------+---------+---------+----
Air Pres 37 424.73 73.53 (-*--)
Press LN 37 117.00 73.04 (--*--)
Temp 2 37 297.54 108.32 (--*--)
--+---------+---------+---------+----
--+---------+---------+---------+----
Pooled StDev = 86.55 100 200 300 400
Pooled StDev = 86.55 100 200 300 400
A
Temp
Press
LNG
Air Pres
600
500
400
300
200
100
0
Boxplots of Air Press - Temp
(means are indicated by solid circles)
Analysis
D M I C
Analysis of Temp furnace with result has significant effect to LNG & Air press
Hypothesis Analysis :
Ho : Temperature furnace has no significant effect to LNG & Air Pressure
Ha : Temperature furnace has significant effect to LNG & Air Pressure
158. A
D M I C
Analysis of RPM motor with result has significant effect to Pressure LNG
Regression Analysis: LNG Press versus RPM
Motor
The regression equation is
RPM = 1448 + 0.0344 LNG Press
The regression equation is
RPM = 1448 + 0.0344 LNG Press
Fail to reject Ho
Predictor Coef SE Coef T P
Constant 1448.00 5.02 288.20 0.000
LNG press 0.03440 0.01568 2.19 0.035
Predictor Coef SE Coef T P
Constant 1448.00 5.02 288.20 0.000
LNG press 0.03440 0.01568 2.19 0.035
S = 8.711 R-Sq = 12.1% R-Sq(adj) = 9.6%
S = 8.711 R-Sq = 12.1% R-Sq(adj) = 9.6%
Analysis of Variance
Analysis of Variance
Source DF SS MS F PP
Source DF SS MS F PP
Regression 1 365.11 365.11 4.81 00..003355
Residual Error 35 2655.97 75.88
Total 36 3021.08
Regression 1 365.11 365.11 4.81 00..003355
Residual Error 35 2655.97 75.88
Total 36 3021.08
RC Fan
Rotation
Analysis
Accept Ha
Normal Probability Plot
Anderson-Darling Normality Test
A-Squared: 0.624
P-Value: 0.097
Average: 1462.22
StDev: 19.1748
N: 37
1420 1430 1440 1450 1460 1470 1480 1490
.999
.99
.95
.80
.50
.20
.05
.01
.001
Probability
RPM
The p-value < 0.05, Fail to accept Ho, which
demonstrate statistical significance (an
equation with a “good” fit)
Hypothesis Analysis :
Ho : RPM motor has no significant effect to LNG Pressure
Ha : RPM motor has significant effect to LNG Pressure
159. Analysis
A
D M I C
Analysis of Air Blower with result has no significant effect to Pressure LNG
Regression Analysis: LNG Press Versus Air Blower
The regression equation is
Air Blower = 475 - 0.169 LNG Press
Regression Analysis: LNG Press Versus Air Blower
The regression equation is
Air Blower = 475 - 0.169 LNG Press
Reject Ho
Predictor Coef SE Coef T P
Constant 474.97 35.14 13.52 0.000
Air Blower -0.1689 0.1111 -1.52 0.138
Predictor Coef SE Coef T P
Constant 474.97 35.14 13.52 0.000
Air Blower -0.1689 0.1111 -1.52 0.138
S = 72.23 R-Sq = 6.2% R-Sq(adj) = 3.5%
Analysis of Variance
Source DF SS MS F PP
S = 72.23 R-Sq = 6.2% R-Sq(adj) = 3.5%
Analysis of Variance
Source DF SS MS F PP
Regression 1 12044 12044 2.31 00..113388
Residual Error 35 182604 5217
Total 36 194647
Regression 1 12044 12044 2.31 00..113388
Residual Error 35 182604 5217
Total 36 194647
Exhaust F’ce
Blower
Fail to accept Ha
.999
.99
.95
.80
.50
.20
.05
.01
Average: 424.730
StDev: 73.5314
N: 37
Normal Probability Plot
300 400 500 600
Anderson-Darling Normality Test
A-Squared: 0.200
P-Value: 0.875
.001
Probability
Blower
Hypothesis Analysis :
Ho : RPM motor has significant effect to LNG Pressure
Ha : RPM motor has no significant effect to LNG Pressure
Hinweis der Redaktion
Six sigma digunakan baik secara individu maupun secara kelompok, untuk (1) pengerak dan penyokong improvement (proses peningkatan/pengembangan kualitas), (2) menyediakan arah yang tepat/akurat/teliti dari suatu kegiatan dengan sebuah strategi, (3) memandu membuat keputusan dengan fakta dan data, (4) mencapai kebutuhan pelanggan dengan meningkatkan product dan proses, dan (5) mengantarkan/mengantarkan hasil yang bottom-line (benefit, ukuran kinerja yg berhubungan dengan uang).
Definisi 6 sigma mempunyai 3 elemen, ke-3 elemen tersebut tujuannya sama yaitu untuk CI .