Professional Documents
Culture Documents
Rocess Ontrol Tatistical: Drrswalia
Rocess Ontrol Tatistical: Drrswalia
Rocess Ontrol Tatistical: Drrswalia
PROCESS CONTROL
Dr R S Walia
Modern Importance of Quality
2
Three Issues
1. Productivity
2. Cost
3. Quality
3
Significant Issue
Off These three determinants of the profitability the most
significant factor in determining the long range success or failure
of an organization is Quality.
4
Cost of Quality
6
Four Cost Categories Related to Quality
• External failure cost: Cost that relates to all errors not detected
and corrected before delivery to the customer.
7
Normal Cost of Quality Distribution when
Quality System is NOT in Place
Appraisal Cost
Prevention
Cost
Internal
Failure Cost
Cost of
Quality
External
Failure
Cost
8
Optimum Cost of Quality Distribution
when Quality System is in Place
Appraisal Cost
Prevention
Cost of Cost
Quality Internal Failure Cost
9
Variance
(x − ) 2
n
–Mean is represented by and n is the
number of items.
Standard Deviation
Standard Deviation shows the variation in
data. If the data is close together, the standard
deviation will be small. If the data is spread
out, the standard deviation will be large.
– (12) = 144
2
(4) = 16
2
(−8) = 64
2
(−24) = 576
2
Find the variance and
standard deviation
– The math test scores of five students are:
92,88,80,68 and 52.
– 4) Find the sum of the squares of the
deviation from the mean:
– 256+144+16+64+576= 1056
– 5) Divide by the number of data
items to find the variance:
– 1056/5 = 211.2
Find the variance and
standard deviation
– The math test scores of five students are:
92,88,80,68 and 52.
–Answer Now
–The math test scores of five students are:
92,92,92,52 and 52.
–1) Find the mean: (92+92+92+52+52)/5 = 76
–2) Find the deviation from the mean:
– 92-76=16 92-76=16 92-76=16
– 52-76= -24 52-76= -24
– 3) Square the deviation from the mean:
– (16)2 = 256 (16)2 = 256 (16)2 = 256
(−) = (−) =
–4) Find the sum of the squares:
– 256+256+256+576+576= 1920
–The math test scores of five students
are: 92,92,92,52 and 52.
–Answer Now
Analyzing the data:
Class A: 92,88,80,68,52
Class B: 92,92,92,52,52
Class C: 77,76,76,76,75
Estimate the standard deviation for Class C.
a) Standard deviation will be less than 14.53.
b) Standard deviation will be greater than 19.6.
c) Standard deviation will be between 14.53
and 19.6.
d) Can not make an estimate of the standard
deviation.
–Answer Now
Analyzing the data:
Class A: 92,88,80,68,52
Class B: 92,92,92,52,52
Class C: 77,76,76,76,75
Estimate the standard deviation for Class C.
a) Standard deviation will be less than 14.53.
b) Standard deviation will be greater than 19.6.
c) Standard deviation will be between 14.53 and 19.6
d) Can not make an estimate if the standard
deviation.
– Answer: A
– The scores in class C have the same mean
of 76 as the other two classes. However, the
scores in Class C are all much closer to the
mean than the other classes so the standard
deviation will be smaller than for the other
classes.
Have you ever…
• Shot a rifle?
• Played darts?
• Played basketball?
Shot a rifle?
Played darts?
Played basketball? Ram
Arjun
29
Variability
Ram
Observations Deviations
10 10 - 8.4 = 1.6
9 9 – 8.4 = 0.6
8 8 – 8.4 = -0.4
8 8 – 8.4 = -0.4
7 7 – 8.4 = -1.4
averages 8.4 0.0
Arjun
30
Variability
Observations Deviations
7
7 7 – 6.6 = 0.4 6
7 7 – 6.6 = 0.4 7
7 7 – 6.6 = 0.4 7
6 Arjun
6 6 – 6.6 = -0.6
6 6 – 6.6 = -0.6
averages 6.6 0.0
31
Variability
8
7
10
• Variance = average distance between 8
observations and the mean squared 9
Ram
Variance Standard
Deviation
Ram 1.0 1.0
Arjun 0.24 0.4898979 Arjun
35
Variability
0.2
Probability
0.15 1 die
0.1 2 dice
0.05 3 dice
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Sum of dots
36
Data Distribution
• Data can be “distributed” (spread out) in
different ways
A Bell Curve
–The bell curve which represents a
normal distribution of data shows what
standard deviation represents.
–+
–Theoretically, curve extends to infinity
–Mean, median, and mode are equal
7 - 46
–Normal Probability Distribution
–part of a “family” of curves
= 11 = 1
7 - 48
–The Standard
–Normal Probability Distribution
… is a normal distribution
with a mean of 0 and a standard deviation of 1.
–Also called the Z Distribution
–Formula z = (X− µ)
7 - 49
–The Standard
–Normal Probability Distribution
–Total Area under the curve is 100% or
1
$3300
z = 1.0
$3000
z=0
7 - 52
–The Standard
–Normal Probability Distribution
Formula z =(X− µ)
The bi-monthly
starting salaries of −$3,000
z= $2,550
recent MBA
graduates follows $300
the normal Z = -1.50
distribution with a
mean of $3,000 –A z-value of –1.50 indicates that
and a standard the value of $2,550 is
deviation of $300. one and a half (1.5)
–What is the standard deviations
z-value for a salary below the mean of $3,000.
of $2,550?
–Show curve
7 - 53
–The Standard
–Normal Probability Distribution
–A z-value of –1.50 indicates that the value of
$2,550 is one and a half standard deviations
below the mean of $3,000.
–$255
z 0= −1.0
–$300
0 z = 0
7 - 54
LET’S RECAP!
–The properties of a normal distribution:
• It is a bell-shaped curve.
• It is symmetrical about the mean, μ. (The mean, the mode and the median all have the same value).
• The total area under the curve is 1 (or 100%).
• 50% of the area is to the left of the mean, and 50% to the right.
–50 –50
% %
μ
–The properties of a normal distribution:
• It is a bell-shaped curve.
• It is symmetrical about the mean, μ. (The mean, the mode and the median all have the same value).
• The total area under the curve is 1 (or 100%).
• 50% of the area is to the left of the mean, and 50% to the right.
• Approximately 68% of the area is within 1 standard deviation, σ, of the mean.
–68
%
– –
μ -σ μ σ μ +
–
– –
σ σ
–The properties of a normal distribution:
• It is a bell-shaped curve.
• It is symmetrical about the mean, μ. (The mean, the mode and the median all have the same value).
• The total area under the curve is 1 (or 100%).
• 50% of the area is to the left of the mean, and 50% to the right.
• Approximately 68% of the area is within 1 standard deviation, σ, of the mean.
• Approximately 95% of the area is within 2 standard deviations of the mean.
–95
%
– – – –
μ -σ μ - σ μ σ–μ +σ –μ +
–
– –
2σ σ σ 2σ
–The properties of a normal distribution:
• It is a bell-shaped curve.
• It is symmetrical about the mean, μ. (The mean, the mode and the median all have the same value).
• The total area under the curve is 1 (or 100%).
• 50% of the area is to the left of the mean, and 50% to the right.
• Approximately 68% of the area is within 1 standard deviation, σ, of the mean.
• Approximately 95% of the area is within 2 standard deviations of the mean.
• Approximately 99% of the area is within 3 standard deviations of the mean.
–99
%
– – – – – –
μ -σ μ -σ μ - σ μ σμ +σ–μ +σ μ +
–
– – – – –
3σ 2σ σ σ 2σ 3σ
LET’S PRACTICE!
The Method
• Procedure, Set-up, temperature, etc.
The Material
• Moisture content, blending, contamination
The Machine
• Machine condition, inherent precision
Management
• Poor Process management, Poor System
67
Causes of Variations
Common Causes : Improvement in system
• Variation inherent in process
• Can be eliminated only through improvement in system
• No pattern
70
Control Charts:
Recognizing Sources of Variation
71
Control Charts: Types
Control Charts for Variables Data
_
X and R charts : for sample averages and ranges
_
X and s charts : for sample averages and standard deviations
73
Control Chart Selection: Variable Data
74
Control Chart Selection: Attribute Data
Defect or
Defective Data
Nonconformity Data
75
Charts may signal incorrectly!
77
Two types of error
Type II error: - results from inferring that a process is in
control when it is actually out of control.
(no of observations goes out of control limit)
78
Process Capability
–Nominal
–value
–Process distribution
–Lower –Upper
–specification –specification
–Process is capable
Process Capability
–Nominal
–value
–Process distribution
–Lower –Upper
–specification –specification
= =
x – Lower specification
Cpk = Minimum of , Upper specification - x
3
3
–We take the minimum of the two ratios because it gives the worst-
case situation.
Intensive Care Lab
Example
–The intensive care unit lab process has an average turnaround time
of 26.2 minutes and a standard deviation of 1.35 minutes.
–The nominal value for this service is 25 minutes with an upper
specification limit of 30 minutes and a lower specification limit of 20
minutes.
–The administrator of the lab wants to have three-sigma
performance for her lab. Is the lab process capable of this level of
performance?
Natural Pattern:
Np are indicative of a process that is in control; i.e. they
demonstrate the presence of a stable system of common cause.
Sudden Shift:
Sudden change or jump occurs because of changes – in process
setting as temp, pr., depth of cut, change in customer behavior no.
of tellers changed, new eq, process etc.
87
General Rules for Out of Control Situations
Gradual Shift :
→ new operator
R → decrease in worker skill
→ gradual improvement in vendors supply
88
General Rules for Out of Control Situations
Trending Pattern:
R → Operator skill
89
General Rules for Out of Control Situations
Cyclic Pattern:
90
Control Chart Construction
• Select the process to be charted;
• Decide what to measure and count; (Pareto sizing)
• Determine sampling method and plan;
– How large a sample needs to be selected? Balance the time and cost to collect a
sample with the amount of information you will gather.
– As much as possible, obtain the samples under the same technical conditions: the
same machine, operator, lot, and so on. (Rational Sub-grouping)
– Frequency of sampling will depend on whether you are able to discern patterns in
the data. Consider hourly, daily, shifts, monthly, annually, lots, and so on. Once
the process is “in control”, you might consider reducing the frequency with which
you sample.
– Generally, collect 20-25 groups of samples before calculating the statistics and
control limits.
91
Control Chart Construction
92
Set of Control Chart Interpretation
rules to determine "Out of Control"
–UCL
–LCL
–Sample observations
–LCL
–consistently below the
–center line
–Sample observations
–consistently above the
–center line
Control Chart Patterns (cont.)
–UCL
–UCL
–LCL
–Sample observations
–consistently increasing –LCL
–Sample observations
–consistently decreasing
Zones for Pattern Tests
UCL –=
–3 sigma = x + A2R
–Zone A
–=
–2 sigma = x + 2 (A2R)
3
–Zone B
–= 1
–1 sigma = x + (A2R)
3
–Zone C
Process =
–x
average
–Zone C
–=
–1 sigma = x - 1 (A2R)
3
–Zone B
–=
–2 sigma = x - 2 (A2R)
3
–Zone A
–=
LCL –3 sigma = x - A2R
–| –| –| –| –| –| –| –| –| –| –| –| –|
1 2 3 –4 –5 –6 –7 –8 –9 –10 –11 –12 –13
–Sample number
Control Chart Patterns
+ 3 Process
Average
- 3
LCL
TIME
–Levine, Prentic-Hall
Control Chart Representing Limits, Special
Causes, Common Causes
Centre Line
Lower
control limits
and R charts
101
Control Charts for Variable Type of Data
Centre line, upper, & lower control limit for R charts are calculated.
The formulae used are as following:
Ri = X max ( i )
− X min( i )
R
i =1
i
R = mean of ‘g’ samples
R =
g
Where D3 and D4 are
UCLR = D4 R (Upper control limit for R chart) constants and their
values can be obtained
from standard tables.
LCL R = D3 R (Lower control limit for R chart) These values depend
upon the size of sample.
102
Control
Chart
Constants
103
Control Charts for Variable Type of Data
Centre line, upper, & lower control limit for x bar chart are calculated.
The formulae used are as following:
n
X i
Xi = i =1
n
X i = mean of the ith sample
n = sample size,
Xi = ith data
g
X
X = i =1
i
= CL X = mean of ‘g’ samples
g x
Where A2 is
constants and its
value can be
UCLx = X + A2 R (Upper control limit for X bar chart) obtained from
standard table. Its
LCLx = X − A2 R (Lower control limit for X bar chart) value depend
upon the size of
sample. 104
Control Charts for Variable Type of Data
Example
Mean values and ranges of data from 20 samples (sample size = 4) are shown
in the table below:
1 10 4 5 9 5 9 10 4 13 12 4 17 12 4
2 15 4 6 11 6 10 11 6 14 12 3 18 15 3
3 12 5 7 11 4 11 12 5 15 11 3 19 11 3
4 11 4 8 9 4 12 13 4 16 15 4 20 10 4
20
R
Average of Ranges of 20 samples = i =1
20
= 4.15 (Center Line of R Chart)
Upper Control Limit of R chart = D4*4.15 (D4= 2.282 for sample size 4)
= 9.47 9.5
Lower Control Limit of R chart = D3*4.15 (D3 = 0 for sample size 4) = 0 105
Control Charts for Variable Type of Data
Example
Mean values and ranges of data from 20 samples (sample size = 4) are shown
in the table below:
1 10 4 5 9 5 9 10 4 13 12 4 17 12 4
2 15 4 6 11 6 10 11 6 14 12 3 18 15 3
3 12 5 7 11 4 11 12 5 15 11 3 19 11 3
4 11 4 8 9 4 12 13 4 16 15 4 20 10 4
20
X
i =1
Average of mean values of 20 samples =
20
= 11.6 (Center Line of x bar Chart)
Upper Control Limit of x bar chart = 11.6 + A2*4.15 = 14.63 (A2 = 0.729 for sample size 4)
Lower Control Limit of x bar chart = 11.6 - A2*4.15 = 08.57 (A2 = 0.729 for sample size 4)
106
Control Charts for Variable Type of Data
10
8
All the data are within
the LCL and UCL in
6
R Chart.
4
Hence variability of
2 the process data is not
0
an issue to worry.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
R Chart
Sample data at S.N 2,
16
16, and 18 are slightly
14
above the UCL.
12
Efforts must be made
10
to find the special
8
causes and revised
6
limits are advised to
4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 calculate after
X-Bar Chart
deleting these data. 107
Control Charts for Variable Type of Data
Sample Observation (ohms) X R Comments
1 20,22,21,23,22 21.60 3
2 19,18,22,20,20 19.80 4
New vendor
3 25,18,20,17,22 20.40 8
4 20,21,22,21,21 21.00 2
5 19,24,23,22,20 21.60 5
6 22,20,18,18,19 19.40 4
7 18,20,19,18,20 19.00 2
8 20,18,23,20,21 20.40 5
9 21,20,24,23,22 22.00 4
10 21,19,20,20,20 20.00 2
11 20,20,23,22,23 21.00 3
12 22,21,20,22,23 21.60 3
13 19,22,19,18,19 19.40 4
14 20,21,22,21,22 21.20 2
15 20,24,24,23,23 22.80 4
16 21,20,24,20,21 21.20 4
17 20,18,18,20,20 19.20 2
18 20,24,22,23,23 22.40 4
19 20,19,23,20,19 20.20 4
20 22,21,21,24,22 22.00 3
21 23,22,22,20,22 21.80 3
22 21,18,18,17,19 18.60 4 High Temperature
Wrong Die
23 21,24,24,23,23 23.00 3
24 20,22,21,21,20 20.80 2
25 19,20,21,21,22 20.60 3
R i
87
R= i =1
= = 3.48
g 25
For a sample of size 5, D4 = 2.114 and D3 = 0. The trial control limits for the
R-chart are calculated as follows:
109
Control Charts for Variable Type of Data
g
X i
521.00
X= i =1
= = 20.840
g 25
110
Control Charts for Variable Type of Data
111
Control Charts for Variable Type of Data
The revised control limits on the R-chart are
UCLR = D4 R = (2.114)(3.273) = 6.919
LCLR = D3 R = (0)(3.273) = 0
459
X= = 20.864
22
Note that sample 15 falls slightly above the upper control limit on the chart.
On further investigation, no special causes could be identified for this sample.
So, the revised limits will be used for future observations until a subsequent
revision takes place. 112
Data
• From Gillette example 15 hours become 15
subgroups representing the total population of
razor blades and each subgroup will have 5
data values n = 5.
Table of Data from Gillette
Example–5 measurements for
each subgroup
–Subgroups
Calculations for the Data
• Find the mean (X) and the range for each
subgroup
1. X = ( Σ xi ) / n
2. R = Max (xi) – Min (xi)
• Find the mean of the X and Range
1. X = ( Σ X i ) / (#of sub groups )
2. R = ( Σ Ri ) / (# of sub groups )
Data with Calculated figures
Create Control Charts
Limits to our Control charts
• Upper and lower control limits represent 3
standard deviations above and 3 standard
deviations below the mean line respectfully.
Centre line, upper, & lower control limit for s charts are calculated.
The formulae used are as following:
g
−
s
1
S= i (Center Line for ‘s’ chart) Where B3 and B4
g i =1 are constants and
their values can
UCLR = B4 S (Upper control limit for ‘s’ chart) be obtained from
standard tables.
These values
LCLR = B3 S (Lower control limit for ‘s’ chart) depend upon the
131
size of sample.
Control Charts for Variable Type of Data
Factors for Calculating Limits for and S Charts
n A3 B3 B4
2 1.880 0 3.267
3 1.023 0 2.575
4 0.729 0 2.282
5 0.577 0 2.115
6 0.483 0 2.004
7 0.419 0.076 1.924
8 0.373 0.136 1.864
9 0.337 0.184 1.816
10 0.308 0.223 1.777
132
Control Charts for Variable Type of Data
Centre line, upper, & lower control limit for x bar charts are calculated.
The formulae used are as following:
n
X i
Xi = i =1
n
th
X i = mean of the i sample
n = sample size,
Xi = ith data
g
g
−
s
1
Average of Ranges of 30 samples = S= i = 7.41 (Center Line of s Chart)
g i =1
Upper Control Limit of s chart = B4*7.41 (B4 = 2.115 for sample size 5)
= 15.48
Lower Control Limit of R chart = B 3*7.41 (B3 = 0 for sample size 5) = 0
134
Control Charts for Variable Type of Data
Example
Mean values and SD of data from 30 samples (sample size = 5) are shown in
the table below:
X
i =1
Average of mean values of 30 samples = = 49.63 (Center Line of x bar Chart)
30
Upper Control Limit of x bar chart = 49.63 + A3*7.41 = 60.21 (A3 = 1.427 for sample size 5)
Lower Control Limit of x bar chart = 49.63 - A3*7.41 = 39.05 (A3 = 1.427 for sample size 5)
135
Control Charts for Variable Type of Data
60
Sample data at S.N 2, is
slightly above the UCL.
Sample Mean
55
Efforts must be made to find
50 the special causes and
revised limits are advised to
45 calculate after deleting these
data.
40
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
X Bar chart
14
10
LCL and UCL in R Chart.
6 Hence variability of the
process data is not an issue to
2
worry.
-2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
S chart
136
Control Charts for Attribute Type of Data
137
Control Charts for Attribute Type of Data
– Attribute information indicates whether a certain quality
characteristic is within specified limits. It does not state
the degree to which specifications are met or not met.
Thus Variable control chart provide more information on the
of performance of the process
– Attribute charts require larger sample sizes
➢N >> large
➢If p = 2.5%, n=25,
nonconforming item per
sample = 0.625
➢ p = 2.5%, n =100,
nonconforming item per
sample = 2.5
138
Control Charts for Attribute Type of Data
n!
P( X = x) = p x (1 − p) n − x
(n − x)! x!
p = prob.of nonconforming
n = sample size
E(p̂) = p
p(1 - p)
var(p̂) =
n 139
Control Charts for Attribute Type of Data
140
Control Charts for Attribute Type of Data
Centre line, upper, & lower control limit for p charts are calculated.
The formulae used are as following:
p (1 − p )
LCL = p − 3 (Lower control limit for p chart)
n
Where n is the sample size. Sample size in p chart must be 50
Sometimes LCL in p chart becomes negative, in such cases LCL should be taken 141
as 0
Control Charts for Attribute Type of Data
Example
Data for defective CDs from 20 samples (sample size = 100) are shown in the
table below:
Sample No. No. of Proportion Sample No. No. of Proportion
Defective Defective = Defective Defective =
CDs = x x/sample size CDs = x x/sample size
1 4 .04 11 6 .06
2 3 .03 12 5 .05
3 3 .03 13 4 .04
4 5 .05 14 5 .05
5 6 .06 15 4 .04
6 5 .05 16 7 .07
7 2 .02 17 6 .06
8 3 .03 18 8 .08
9 5 .05 19 6 .06
10 6 .06 20 8 .08
p (1 − p ) .051(1 − .051)
UCL = p + 3 = 0.051 + 3 = .12
sum of defectives
n 100
101
CL = = = 0.051
sum of allsamples 2000
p (1 − p ) .051(1 − .051)
LCL = p − 3 = 0.051 − 3 = −.02 0
n 100 142
Control Charts for Attribute Type of Data
P CHART
0.14
0.12
0.10
MEANP
0.08 UCLP
0.06 LCLP
P
0.04
0.02
0.00
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
There is important observation that is clearly visible from the data points that
there is an increasing trend in the average proportion defectives beyond
sample number 15. Also, data show cyclic pattern. Process appears to be out
of control and also there is strong evidence that data are not from
independent source. 143
Control Charts for Attribute Type of Data
–(c-chart) Poisson Distribution
–for PD
1. Opportunity area should be large
2. Occurrence of non-conformity must be independent to each other
3. Each sample should have equal chance of the occurrence of
nonconformity
e−c c x
p ( x) =
x!
p = prob.of x nonconforming
x = number of nonconformity in the sample
c = mean
144
Control Charts for Attribute Type of Data
c-charts counts the number of defects in an item. c-charts are used only
when the number of occurrence per unit of measure can be counted such
as number of scratches, cracks etc.
145
Control Charts for Attribute Type of Data
Centre line, upper, & lower control limit for c charts are calculated.
The formulae used are as following:
146
Control Charts for Attribute Type of Data
Example
Data for defects on TV set from 20 samples (sample size = 10) are shown in
the table below:
14
12
10
Defects
8
6
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
None of the sample is out of the LCL and UCL. But the chart shows cyclic
trend.
148
Control Charts for Attribute Type of Data
u-chart counts the number of defect per sample. The u chart is used when
it is not possible to have a sample size of a fixed size.
149
Control Charts for Attribute Type of Data
Centre line, upper, & lower control limit for u charts are calculated.
The formulae used are as following:
k
u= =
i =1
k = centre line of u chart
total number of items in all samples n
i =1
i
u
UCL = u + 3 (Upper control limit for u chart)
ni
4.68
c
k
UCL1 = 4.68 + 3 = 9.27
−
u= i =1
i
=
192
= 4.68
2
k
n
41
4.68
LCL1 = 4.68 − 3 = 0.092
i
i =1
151
2
S No. Sample size No. of UCL LCL S No. Sample size No. of UCL LCL
ni Defects per ni Defects per
100 product 100 product
1 2.0 2.50 9.3 0.1 11 3.0 3.00 8.4 0.9
2 3.0 4.67 8.4 0.9 12 2.5 6.40 8.8 0.6
3 2.5 3.20 8.8 0.6 13 2.0 6.00 9.3 0.1
4 1.5 5.33 10.0 0.0 14 2.5 4.00 8.8 0.6
5 2.5 4.80 8.8 0.6 15 1.0 6.00 11.2 0.0
6 1.0 6.00 11.2 0.0 16 2.0 4.00 9.3 0.1
7 2.0 10.00 9.3 0.1 17 2.0 2.50 9.3 0.1
8 1.5 6.67 10.0 0.0 18 1.0 5.00 11.2 0.0
9 1.5 4.00 10.0 0.0 19 3.0 4.67 8.4 0.9
10 2.5 4.00 8.8 0.6 20 2.0 4.00 9.3 0.1
12
10
8
Defects
-2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
152