Professional Documents
Culture Documents
TOPIC 9 & 10
TOOLS TO IMPROVE
QUALITY AND QUALITY
DIAGNOSIS PROCEDURE :
STATISTICAL QUALITY
CONTROL
KAORU ISHIKAWA
STATISTICAL
FUNDAMENTALS
1. What Is Statistical Thinking
Is a decision-making skill demonstrated by
the ability to draw conclusions based on data.
Statistical thinking is based on three
concepts:
1. Random variation
2. Non-random
variation
Random Variation
Is uncontrollable
In centered around a mean and occurs
with a somewhat consistent amount of
dispersion.
The amount of random variation in a
process may be either large or small
Non-Random Variation
10
Process Stability
11
12
14
QUALITY CONTROL
APPROACHES
Statistical process control (SPC)
Monitors the production process to prevent
poor quality
Acceptance sampling
16
STATISTICAL PROCESS
CONTROL
Take periodic samples from a process
Plot the sample points on a control chart
Determine if the process is within limits
17
Hospitals
timeliness, responsiveness, accuracy
Grocery Stores
Check-out time, stocking, cleanliness
Airlines
luggage handling, waiting times, courtesy
Fast food restaurants
waiting times, food quality, cleanliness
PROCESS CONTROL
CHART
Upper
control
limit
Process
average
Lower
control
limit
2
Sample number
10
19
CONSTRUCTING A CONTROL
CHART
Decide what to measure or count
Collect the sample data
Plot the samples on a control chart
Calculate and plot the control limits on the control chart
Determine if the data is in-control
If non-random variation is present, discard the data (fix the
problem) and recalculate the control limits
A Process Is In Control If
20
THE NORMAL
DISTRIBUTION
95 %
99.74 %
-3
-2
-1
= 0 1 2 3
Ch 4 - 14
CONTROL CHART Z
VALUES
Smaller Z values make more
sensitive charts
Z = 3.00 is standard
Compromise between sensitivity
and errors
+3
Mean
-3
23
LCL
TYPES OF DATA
Attribute data (p-charts, c-charts)
Product characteristics evaluated with a
discrete choice (Good/bad, yes/no, count)
24
CONTROL CHARTS
FOR ATTRIBUTES
p Charts
Calculate percent defectives in a sample;
an item is either good or bad
c Charts
25
P - CHARTS
Based on the binomial distribution
p=
LCLp =
P-CHART CALCULATIONS
Proportion
Sample Defect Defective
1
2
3
6
0
4
.06
.00
.04
.
20
18
.18
200
1.00
100 jeans in each sample
total defectives
p =
total sample observations
200 =
20 (100) 0.10
UCL = p + 3 p(1-p) /n
0.2
0.18
0.14
0.12
0.1
0.08
0.06
0.04
0.02
20
18
16
14
12
..
10
Sample number
28
0
0
Proportion defective
0.16
C - CHART
CALCULATIONS
Count # of defects per roll in 15 rolls of denim fabric
c = 190/15 = 12.67
UCL = c + z c
.
= 12.67 + 3 12.67
= 23.35
LCL = c - z c
= 12.67 - 3 12.67
= 1.99
30
Sample Defects
1
12
2
8
3
16
.
.
.
15
15
190
EXAMPLE C - CHART
24
18
15
12
9
6
3
14
12
10
Sample number
31
Number of defects
21
CONTROL CHARTS
FOR VARIABLES
Mean chart (X-Bar Chart)
Measures central tendency of a sample
Range chart (R-Chart)
Measures amount of dispersion in a sample
32
5.02
5.01
4.94
4.99
4.96
4.98
5.01
5.03
5.07
4.95
4.96
5.00
4.99
5.00
4.93
4.92
4.99
4.97
5.03
4.91
5.01
4.98
4.89
4.96
4.95
4.92
5.03
5.05
5.01
4.99
4.97
5.06
5.06
4.96
5.03
5.01
5.05
5.01
5.10
4.96
4.99
5.02
5.09
5.10
5.00
4.99
5.08
5.05
5.14
5.10
4.99
5.08
5.09
5.08
5.01
4.98
5.08
5.07
4.99
50.09
1.15
5.03
0.10
34
EXAMPLE X-BAR
CHART
5.10
UCL
5.08
5.06
e
g 5.04
a
r
e
v 5.02
a
e
l 5.00
p
m
a 4.98
S
4.96
4.94
4.92
1
Sample number
10
LCL
CONSTRUCTING AN
RANGE CHART
UCLR = D4 R = (2.11) (.115) = 2.43
LCLR = D3 R = (0) (.115) = 0
where R =
R / k = 1.15 / 10 = .115
k = number of samples = 10
36
3 CONTROL CHART
FACTORS
D4
3.27
2.57
2.28
2.11
2.00
1.92
1.86
X-chart
R-chart
A2
D3
1.88
1.02
0.73
0.58
0.48
0.42
0.08
0.37
0.14
37
Sample size
EXAMPLE R-CHART
UCL
0.3
0.25
0.2
e
g
n 0.15
a
R
0.1
0.05
0
1
Sample number
10
LCL
CONTROL CHART
PATTERNS
UCL
UCL
LCL
LCL
Sample observations
consistently below the
center line
Sample observations
consistently above the
center line
Ch 4 - 34
CONTROL CHART
PATTERNS
UCL
UCL
LCL
LCL
Sample observations
consistently increasing
Sample observations
consistently decreasing
Ch 4 - 35
CONTROL CHART
PATTERNS
UCL
UCL
LCL
LCL
Sample observations
consistently below the
center line
1998 by Prentice-Hall Inc
Russell/Taylor Oper Mgt 2/e
Sample observations
consistently above the
center line
Ch 4 - 36
CONTROL CHART
PATTERNS
1.
2.
3.
4.
5.
LCL
5.08
Zone A
x + 3 sigma
Zone B
x + 2 sigma
Zone C
x + 1 sigma
Zone C
x - 1 sigma
Zone B
x - 2 sigma
Zone A
x - 3 sigma
5.05
5.03
5.01
4.98
4.965
4.94
43
UCL
SAMPLE SIZE
DETERMINATION
Attribute control charts
50 to 100 parts in a sample
Variable control charts
2 to 10 parts in a sample
Ch 4 - 41