Transcript Capability & Continual Improvement

Capability and Improvement
- from Cpk to Six Sigma
© 2002 Systex Services
Capability & Improvement
Continuous
Improvement
Processes
© 2002 Systex Services
Business Processes
Parts
Statistical
Process
Control
Six Sigma
People
Manufacturing Processes
• Vital for credibility and results
Statistical Process Control
- theory and practice
© 2002 Systex Services
Use of Statistical Process Control
• Attributes and variables
Attribute
OK or Not OK
Variable
Measurable
13.456
© 2002 Systex Services
SPC Application
• Initially applied to mechanical components
– in mass production
– as a control mechanism
• Subsequently applied to any measurable
© 2002 Systex Services
Principles of SPC for Variable Data
• Rules apply when distribution is normal
Kgs
25.0
25.1
25.2
25.3
25.4
25.5
25.6
25.7
25.8
25.9
26.0
26.1
26.2
26.3
26.4
26.5
26.6
26.7
26.8
26.9
27.0
Qty
0
1
1
2
2
4
5
6
8
13
17
13
8
6
5
4
2
2
1
0
0
Plot
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Dimension: 26 Kgs ± 0.5
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1
1
Bell Curve
Total
© 2002 Systex Services
100
Principles of SPC for Variable Data
• Rules do not apply when distribution is abnormal
Skewed
Truncated
Multiple
Random Selection
© 2002 Systex Services
Principles of SPC for Variable Data
• Normal distribution has consistent variation
• Variation unit is ‘Standard Deviation’ - s (Sigma)
• Standard deviation is calculated using:
s = (fx2/n) - x2 (Root Mean Square method)
© 2002 Systex Services
Principles of SPC for Variable Data
• Standard deviation :s = (fx2/n) - x2 (Root Mean Square method)
Weight
X
25.1
25.2
25.3
25.4
25.5
25.6
25.7
25.8
25.9
26.0
26.1
26.2
26.3
26.4
26.5
26.6
26.7
26.8
Totals
Frequency
f
1
1
2
2
4
5
6
8
13
17
13
8
6
5
4
2
2
1
100
X = fX/n
= 2599.1/100
= 25.991
Total
Weight
fX
25.1
25.2
50.6
50.8
102.0
128.0
154.2
206.4
336.7
442.0
339.3
209.6
157.8
132.0
106.0
53.2
53.4
26.8
2599.1
s=
=
=
=
=
Squared
Weight
X2
630.01
635.04
640.09
645.16
650.25
655.36
660.49
665.64
670.81
676.00
681.21
686.44
691.69
696.96
702.25
707.56
712.89
718.24
Sum of
squares
fX2
630.0
635.0
1280.2
1290.3
2601.0
3276.8
3962.9
5325.1
8720.5
11492.0
8855.7
5491.5
4150.1
3484.8
2809.0
1415.1
1425.8
718.2
67564.3
[(fX2/n) - X2]
[(67564.270/100) - (25.991)2]
[675.643- 675.532]
74875.9
0.111 Kgs
© 2002 Systex Services
Principles of SPC for Variable Data
• Standard Deviation enables calculation of
probability of defects
1s
1s
1s
1s
1s
1s
68.26%
95.44%
99.73%
© 2002 Systex Services
Defects with spec. limits at:
1 sigma = 31.74% = 317,400 dpm
2 sigma = 4.56% = 45,560 dpm
3 sigma = 0.27% = 2,700 dpm
Principles of SPC for Variable Data
• Normal distribution relative to limits
– forecasts scrap
– defines process capability
– enables process control
‘Normal’ distribution
Lower control limit
Upper control limit
Lower spec. limit
Upper spec. limit
© 2002 Systex Services
Capability Studies
• Process capability is relative to:
– defined limits
– location of process mean
– spread of process
Moderate spread
Moderate placement
Narrow spread
Poor placement
Broad spread
Good placement
LSL
NOM
© 2002 Systex Services
USL
Capability Studies
• Purpose of capability studies
–
–
–
–
to define process capability
to help identify limiting causes
to demonstrate capability to customers
to improve process capability
• reduce defects, waste, cost, customer returns
• undertake higher spec. work
– to employ statistical process controls
© 2002 Systex Services
Capability Studies
• Two basic measures of capability
Cpk
Cp
– spec. range  6s
– lower value of
– no account of placement
2.40
2.430
2.385
2.475
s = 0.015
– (USL - X) / 3s or (X - LSL) /
2.503s
Cp = (USL-LSL)/6 s
= (2.50 - 2.40) / 6 x 0.015
= 1.111
Cpk = (X-LSL)/3 s
= (2.43 - 2.40) / 3 x 0.015
= 0.667
3s
LSL
3s
X
USL
© 2002 Systex Services
Capability Studies
• What is a good Cpk ratio?
• Minimum normally 1.33 Cpk
– based on 4 sigma spread
– extra sigma compensates for
• larger spread over time & larger population
• particularly mean shift
– equivalent to 63 DPM on centred process
• Many companies now looking for 2.0 Cpk
– consistent with 6 sigma concept
– equivalent to 0 DPM
• based on centred process
• allowing up to 2 sigma shift
© 2002 Systex Services
Capability Studies
• Capability studies also indicate:
– trends
– cycles
– other influences
65
60
55
50
12
:0
0
11
:0
0
11
:3
0
10
:0
0
10
:3
0
09
:3
0
08
:3
0
09
:0
0
08
:0
0
45
Trend - Weight
Av. Viscosity
26.5
Upper Spec. Limit
26.0
91.5
91.0
90.5
90.0
89.5
25.5
25.0
Fri
Wed
Mon
24.0
23.5
23.0
22.5
22.0
Lower Spec. Limit
© 2002 Systex Services
49
47
45
43
41
39
37
35
33
31
29
27
25
23
21
19
17
15
13
9
11
7
5
3
21.5
1
Thu
Tue
Fri
Wed
Mon
Thu
Tue
Fri
Wed
Mon
Grams
24.5
X-Bar & Range Charts
• X-bar charts plot sample mean values
X-bar
24.70
24.60
UCL
24.50
24.40
24.30
24.20
24.10
24.00
LCL
23.90
23.80
1
2
3
4
5
6
7
8
9
Process Mean
24.24
24.24
24.24
24.24
24.24
24.24
24.24
24.24
24.24
Subgroup Mean
24.36
24.04
24.50
24.04
24.50
24.00
24.28
24.10
24.50
UCL (Mean+A2*Av.R)
24.56
24.56
24.56
24.56
24.56
24.56
24.56
24.56
24.56
LCL (Mean-A2*Av.R)
23.92
23.92
23.92
23.92
23.92
23.92
23.92
23.92
23.92
© 2002 Systex Services
X-Bar & Range Charts
• Range charts plot sample range values
Range Chart
1.40
UCL
1.20
1.00
Grams
0.80
0.60
0.40
0.20
0.00
1
2
3
4
5
6
7
8
9
10
Range
0.70
0.30
0.80
0.40
0.70
0.20
0.70
0.20
0.70
0.90
UCL (D4*Av.R)
1.18
1.18
1.18
1.18
1.18
1.18
1.18
1.18
1.18
1.18
© 2002 Systex Services
Capability Reports
STUDY:
RESULTS:
Customer:
Part Number:
Type:
Dimension:
Cavity Number:
Conducted by:
Date:
Excellence plc
344 834 890
Preliminary
24.00
1
John Ashcroft
36678
Mean: 24.24
Std Dev: 0.31
U-Ppk: 1.92
L-Ppk: 2.45
USL: 26.00
LSL: 22.00
+ 3 Sigma: 25.16
- 3 Sigma: 23.33
Histogram
12
10
COMMENTS:
Histogram Here
6
4
2
26.6
26.3
26.0
25.7
25.4
25.1
24.8
24.5
24.2
23.9
23.6
0
23.3
24.2
24.4
24.8
24.9
24.2
23.8
24.0
23.9
24.2
24.7
23.0
24.0
24.1
24.1
24.7
24.5
24.2
24.0
24.0
24.2
24.1
22.7
24.2
24.1
24.8
24.7
24.7
23.9
24.0
24.0
24.1
24.0
22.1
24.0
24.6
24.5
24.8
24.6
24.1
23.8
23.9
24.2
24.2
21.8
24.0
24.1
24.4
24.6
24.7
24.2
24.0
24.0
23.9
24.1
8
22.4
DATA:
Trend
25
25
24
Trend Chart Here
24
23
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49
© 2002 Systex Services
Capability
study
results
required
by many
major
customers
Six Sigma
- achieving quantum leaps in
competitiveness
© 2002 Systex Services
Six Sigma Application
• Applies Statistical Process Control to ALL
business process - not just manufacturing
• Combined with classical Continuous
Improvement Techniques
© 2002 Systex Services
Six Sigma Objective
Critical customer
requirement
e.g. 3 day delivery
Reduced
variation
Defects:
> 3 days
Service output
© 2002 Systex Services
© 2002 Systex Services
37.80
37.70
37.60
37.50
37.40
37.30
37.20
37.10
37.00
36.90
2
36.80
Mean 35.25
36.70
4
36.60
6
36.50
8
1350 DPM - stable process
22750 DPM - 1 Sigma mean shift
10
36.40
CCR 36.0
36.30
12
36.20
36.10
36.00
35.90
3s
35.80
35.70
35.60
35.50
1 Sigma = 0.25
35.40
35.30
35.20
35.10
35.00
34.90
34.80
34.70
34.60
34.50
34.40
34.30
34.20
34.10
Six Sigma Example
36.75
6s
0
6 Sigma & Quality Loss Function
Quality Loss Function
Normal Distribution
Taguchi: Quality Loss Function = k(x - T)2
Where:
k = constant for scrap value
x = value of quality characteristic
T = target
© 2002 Systex Services
What is Six Sigma?
Six Sigma Business Improvement...
Markets
Suppliers
Inputs
Processes
Critical
customer
requirements
Process
outputs
Defects
Root cause analysis of
defects leads to permanent
defect reduction
Variations in
process output
cause defects
… a data driven approach to root cause analysis
© 2002 Systex Services
Six Sigma Success Factors
Committed Leadership
Integration with top
level strategy
Business process
framework
Customer & market
intelligence network
Projects produce real
savings or revenue
Full time six sigma
team leaders
Incentives for all
© 2002 Systex Services
6 Sigma & Business Strategy
Business Strategy
Development
Key Performance
Measures
Core business
Process
Process Output
Measures
Marketplace
Critical Customer
Requirements
Process
Sigma
© 2002 Systex Services
6 Sigma Process
Project by project
MEASUREMENT:
- selection
- measurability
- acceptability
ANALYSIS:
- process capability
- experimentation
- root cause
CONTROL:
- selection
- maintenance
- reaction
IMPROVEMENT:
- actions
- process trials
- proving
© 2002 Systex Services
Implementing 6 Sigma
Organisational
Assessment
Exec. Planning
Workshop
Pilot Business
Unit Workshops
6 Sigma Leader
Training
• Appoint core team
• Process mapping
• Current measures
• Process owners
• Customer knowledge
• Customer surveys
• Current capabilities
• Competitive data
• Accountabilities
• Vision/ goals/ 6 sigma
• Basis for improvement
• Tools & methods
• 5 year plan:
• net earnings
• growth
• improvements
• Opportunities
• Select pilot units
• Communication plan
• Leadership criteria
• Resource planning
• Commitment
• Strategy outline
• 6 sigma methods
• Integration process
• Status assessment
• Identify projects
• Benefit targets
• Force field analysis
• Select leaders
• Training schedules
• Project milestone
• Set regular reviews
• High profile launch
• Interactive training
• Project definition
• Mapping
• Measurement
• Analysis
• Analytical tools
• Design of experiment
• Process sigma
• Apply
• Facilitate teams
• Measurable benefits
Over 4 weeks
2 days
2 days
Typical time scales
© 2002 Systex Services
4 - 15 weeks
6 Sigma Roll Out
Organisational
Assessment
Executive
Workshop
Pilot Unit
Workshop
Team Leader
Training
Projects
Executive
Review
Unit Review
Unit Workshops
Team Leader
Training
Projects
© 2002 Systex Services
WEEK 1
WEEK 3
6 sigma & planning overview
Process mapping
Quality function deployment
Failure mode effects analysis
Organizational effectiveness
Basic statistics
Process capability
Measurements systems analysis
Design of experiments
- factorial
- fractional factorials
- balanced block design
- evolutionary operation EVOP
- response surface designs
ANOVA (Analysis of Variance)
Regression (multiple)
Facilitation tools
WEEK 2
WEEK 4
Review of key week 1 topics
Statistical thinking
Hypothesis testing
Correlation
Passive multi-vari analysis &
regression (simple)
Team assessment
Control plans
Statistical process control
(advanced)
Mistake proofing
Team development
Wrap-up of tools
© 2002 Systex Services
Black Belt
Green Belt
6 Sigma Black Belt Training
6 Sigma Comments
Benefits
Risks
• Large cost reductions:
– AlliedSignal $800M (95/7)
– GE $600M (3Q97 gain)
• Performance bonus link
• Capability quantified
• Investors & stakeholders
understand financial gain
• Customer needs measured
• Year one payback - ROI
potential 20% + thereafter
• Large initial investment
off putting
• Poor follow through
• Short term thinking
• Changed priorities
• New leadership
• ‘Tried that’ (no longer use
it!)
• Fear of statistics
© 2002 Systex Services