Rocess Ontrol Tatistical: Drrswalia

S TATISTICAL
PROCESS CONTROL
Dr R S Walia
Modern Importance of Quality
“The first job we have, is to turn out quality

products that consumers will buy and keep on
buying. If we produce it efficiently and
economically, we will earn a profit, in which you
will share.”
- William Cooper Procter
Grandson of the founder of Procter and Gamble
(October 1887)
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.
Good quality of product and service can provide:

1. Competitive edge
2. Reduce cost due to returns, rework, and scrap

3. Productivity and Profits
4. Generates Satisfied customers:
Continued Patronage and word of mouth advertisement
4
Cost of Quality
Running a company by profit alone is

like driving a car by looking in the
rearview mirror.
It tells you where you’ve been, not where
you are going!
Dr. E. Deming
What is the Cost of Quality?
• Cost of Quality: the cost of ensuring that the
job is done right + the cost of not doing the job
right.
Cost of Conformance + Cost of Non-Conformance

(Prevention & Appraisal) (Internal/External Defects)
6
Four Cost Categories Related to Quality
• Prevention cost: Cost of planning and executing a project so it is

error-free or within an acceptable error range.
• Appraisal cost: Cost of evaluating processes and their outputs to

ensure quality.
• Internal failure cost: Cost incurred to correct an identified defect
before the customer receives the product.
• 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
External Failure Cost
9
Variance
Variance is the average squared

deviation from the mean of a set
of data. It is used to find the
standard deviation.
Variance
1. Find the mean of the data.

Hint – mean is the average so add up the values
and divide by the number of items.
2. Subtract the mean from each value – the
– result is called the deviation from the mean.
–3. Square each deviation of the mean.
–4. Find the sum of the squares.
–5. Divide the total by the number of items.
Variance Formula
–The variance formula includes the Sigma

Notation,  , which represents the sum of all
the items to the right of Sigma.
 (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.
– Standard Deviation is often denoted

by the lowercase Greek letter sigma, .
Find the variance and
standard deviation
– The math test scores of five students are:
92,88,80,68 and 52.
–1) Find the mean: (92+88+80+68+52)/5 = 76.
–2) Find the deviation from the mean:
– 92-76=16
– 88-76=12
– 80-76=4
– 68-76= -8
– 52-76= -24
standard deviation
The math test scores of five students are:
92,88,80,68 and 52.
– 3) Square the deviation from the
mean:
– (16) = 256
2
– (12) = 144
2
(4) = 16
2
(−8) = 64
2
(−24) = 576
2
standard deviation
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
standard deviation
92,88,80,68 and 52.
–6) Find the square root of the

variance: 211.2 = 14.53
–
–Thus the standard deviation of the

test scores is 14.53.
Standard Deviation
A different math class took the same

test with these five test scores:
92,92,92,52,52.
Find the standard deviation for this

class.
–Hint:
1. Find the mean of the data.
2. Subtract the mean from each value –
called the deviation from the mean.
3. Square each deviation of the mean.
4. Find the sum of the squares.
5. Divide the total by the number of items –
result is the variance.
6. Take the square root of the variance –
result is the standard deviation.
Solve:
A different math class took the same test with
these five test scores: 92,92,92,52,52.
Find the standard deviation for this class.
–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.
– 5) Divide the sum of the squares

by the number of items :
– 1920/5 = 384 variance
–6) Find the square root of the variance:
–
384 = 19.6
–Thus the standard deviation of the
second set of test scores is 19.6.
–Analyzing the data:
Consider both sets of scores. Both classes

have the same mean, 76. However, each class
does not have the same scores. Thus we use
the standard deviation to show the variation
in the scores. With a standard variation of
14.53 for the first class and 19.6 for the
second class, what does this tell us?
–Answer Now
Analyzing the data:
Class A: 92,88,80,68,52
Class B: 92,92,92,52,52
With a standard variation of 14.53 for the

first class and 19.6 for the second class, the
scores from the second class would be more
spread out than the scores in the second
class.
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?
What is System of point of

these sports?
27
Have you ever…
Shot a rifle?
Played darts?
Played basketball? Ram
Arjun
Who is the better shot? 28

Discussion
• What do you measure in your process?

• Why do those measures matter?
• Are those measures consistently the same?
• Why not?
29
Variability
• Deviation = distance between 8

7
observations and the mean (or 10
8
average) 9
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
• Deviation = distance between

observations and the mean (or
average) Ram
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
Observations Deviations Squared Deviations

10 10 - 8.4 = 1.6 2.56
9 9 – 8.4 = 0.6 0.36
8 8 – 8.4 = -0.4 0.16
Arjun
8 8 – 8.4 = -0.4 0.16
7 7 – 8.4 = -1.4 1.96
averages 8.4 0.0 1.0 Variance
32
Variability
• Variance = average distance between

observations and the mean squared
Ram
Observations Deviations Squared Deviations

7
7 7 - 6.6 = 0.4 0.16
6
7 7 - 6.6 = 0.4 0.16 7
7 7 - 6.6 = 0.4 0.16 7
6 6 – 6.6 = -0.6 0.36 6 Arjun
6 6 – 6.6 = -0.6 0.36
averages 6.6 0.0 0.24 Variance
33
Variability
• Standard deviation = square root

of variance
Ram
Variance Standard
Deviation
Ram 1.0 1.0
Arjun 0.24 0.4898979 Arjun
But what good is a standard deviation

34
Variability
The world tends to

be bell-shaped
Even very rare Fewer Most Fewer Even very rare

outcomes are in the outcomes in the outcomes are
possible “tails” occur in the “tails” possible
(probability > 0) (lower) middle (upper) (probability > 0)
35
Variability
Here is why: Even outcomes that are equally

likely (like dice), when you add
them up, become bell shaped
Add up the dots on the dice
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.
–One standard deviation away from the mean (  ) in either

direction on the horizontal axis accounts for around 68 percent
of the data. Two standard deviations away from the mean
accounts for roughly 95 percent of the data with three standard
deviations representing about 99 percent of the data.
What are some examples of things that
follow a Normal Distribution?
• Heights of people
• Size of things produced by machines
• Errors in measurements
• Blood Pressure
• Test Scores
Normal Distribution Curve
• mean=median=mode
• Symmetry about the center
• 50% of the values less than the mean and 50%
greater than the mean
–68% of
values are
within 1
The Standard standard
deviation of
Deviation : the mean
–95% of
–is a measure values are
within 2
of how standard
deviations
spread out of the mean
numbers are. –99.7% of
values are
within 3
standard
deviations
of the mean
Why do we need to know Standard
Deviation?
• Any value is
– likely to be within 1 standard deviation of the
mean
– very likely to be within 2 standard deviations
– almost certainly within 3 standard deviations
–Normal Probability Distribution
The normal The arithmetic

curve is mean, median,
bell-shaped and mode
of the
–and has a distribution
single peak are equal
at the exact and located at
centre of the the peak.
distribution.
–…thus half the area under the curve is above
the mean and half is below it.
7 - 44
The normal probability

distribution is
The normal asymptotic,
probability
distribution
is
–i.e. the curve gets
symmetrical closer and closer to
about its the x-axis
mean. but never actually
touches it.
7 - 45
–Summary
–Normal
curve is
symmetrical
–+
–Theoretically, curve extends to infinity
–Mean, median, and mode are equal
7 - 46
–part of a “family” of curves
–How does the Standard Deviation values

affect the shape of f(x)?
= , 3, 
 = 2
 =3
 =4
7 - 47
–part of a “family” of curves
–How does the following Expected values

affect the location of f(x)?
 = 10, 11, 1
 = 10
 = 11  = 1
7 - 48
–The Standard
… is a normal distribution
with a mean of 0 and a standard deviation of 1.
–Also called the Z Distribution
A z-value is the distance between a selected value (designated X)

and the population mean,
divided by the population Standard Deviation, .
A z-value is, therefore, a location on the standard normal curve.
–Formula z = (X− µ)
 7 - 49
–The Standard
–Total Area under the curve is 100% or
1
–Area = +0.5 –Area = + 0.5
–…since the mean is located at the peak of the curve,

half the area under the curve is
above the mean and –half is below the mean
7 - 50
–The Standard
Formula z =(X− µ)
The bi-monthly
starting salaries of

recent MBA z = $3,300−–$3,000
graduates $300
follows the normal z = 1.00
distribution with a
mean of $3,000 –A z-value of 1 indicates that
and a standard the value of $3,300 is
deviation of $300. one standard deviation
What is the above the mean of $3,000.
z-value for a
salary of $3,300?
–Show curve
7 - 51
–A z-value of 1 indicates that
the value of $3,300 is
one standard deviation
above the mean of $3,000.
$3300
 z = 1.0
$3000
z=0
7 - 52
–The Standard
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
–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
• Approximately 68% of the area is within 1 standard deviation, σ, of the mean.
–68
%
– –
μ -σ μ σ μ +
–
– –
σ σ
• 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
–99
%
– – – – – –
μ -σ μ -σ μ - σ μ σμ +σ–μ +σ μ +
–
– – – – –
3σ 2σ σ σ 2σ 3σ
LET’S PRACTICE!
You can use a white board or a scrap

piece of paper!
WE DO: 95 % of students at school are
between 1.1 m and 1.7 m tall.
• Assuming this data is normally distributed can
you calculate the mean and standard deviation?
YOU DO: 68% of American’s own a
dog between 4 and 6 years old.
–Assuming this data is normally distributed can
you calculate the mean?
WE DO: The reaction times for a hand-eye coordination test
administered to 1800 teenagers are normally distributed with
a mean of .35 seconds and a standard deviation of .05
seconds.
• Represent this information on a bell curve:
• About how many teens had reaction times

between .25 and .45 seconds?
• What is the probability that a teenager selected at
random had a reaction greater than .4 seconds?
YOU DO: The waiting times for an elevator are normally
distributed with a mean of 1.5 minutes and a standard deviation of
20 seconds.
• Represent this information on a bell curve:
• Find the probability that a person waits longer

than 2 minutes 10 seconds for the elevator.
“Normal” bell shaped curve
Add up about 30 or most things

and you start to be “normal”
Normal distributions are divided up

into 3 standard deviations on
each side of the mean
Once you know that, you

know a lot about
what is going on
And that is what a standard deviation

is good for 66
Potential Reasons of Variations
The Operator
• Training, supervision, technique
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
Assignable Causes : Control of Process

• Variation due to identifiable factors
• Can be modified through operator or management action
• May exhibit a pattern
• Examples of special causes include: wrong tool, wrong production
method, improper raw material, operator’s skill, wrong die etc.
SPC uses samples to identify that Assignable causes have occurred 68

Control Charts:
Recognizing Sources of Variation
• Why Use a Control Chart?
– To monitor, control, and improve process performance over time by
studying variation and its source.
What Does a Control Chart Do?

• Focuses attention on detecting and monitoring process variation over
time;
• Distinguishes special from common causes of variation, as a guide to
local or management action;
• Serves as a tool for ongoing control of a process;
• Helps improve a process to perform consistently for higher quality, lower
cost, and higher effective capacity;
• Provides a common language for discussing process performance. 69
Control Charts:
How Do I Use Control Charts?
▪ There are many types of control charts. The control
charts that you or your team decides to use should be
determined by the type of data that you have.
▪ Data are of two types

– Variables
– Attributes
70
Control Charts:
Variables “Things we measure”
Length Blood Pressure Diameter

Weight Volume Tensile strength
Time Temperature
Attributes “Things we count”
Number or percent defective items

Number of defects per item
Types of defects etc.
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
Md and R charts : for sample medians and ranges
X and Rm charts : for individual measures and moving ranges
Control Charts for Attributes Data
p charts : proportions of units nonconforming
np charts : number of units nonconforming
c charts : number of defects/nonconformities
u charts : number of defects/nonconformities per unit 72

Control Charts: Selection
• How Do I Select Control Charts?
– Use the following tree diagram to determine which

chart will best fit your situation.
– Only the most common types of charts are addressed.
73
Control Chart Selection: Variable Data
n=1 2<n<9 n is ‘small’ n is ‘large’

median 3<n<5 n > 10
X & Rm X&R X&R X&S
74
Control Chart Selection: Attribute Data
Defect or
Defective Data
Nonconformity Data
Constant Variable Constant Variable

sample size sample size n > 50 n > 50
c chart u chart p or np chart p chart
75
Charts may signal incorrectly!
Charts repeatedly apply hypothesis testing!
– Type I error with charts:

Concluding that a process is not in control
when it actually is
– Type II error with charts:
Concluding that a process is in control when
it is not
76
Two types of error
Type I error: - results from inferring that a process is out of control
when it is actually in control.
Prob. of type I error is a
Type I error = a/ + a/ = a

= sum of two tails area outside
control limits
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
• Process capability is the ability of the

process to meet the design specifications
for a service or product.
• Nominal value is a target for design
specifications.
• Tolerance is an allowance above or
below the nominal value.
Process Capability
–Nominal
–value
–Process distribution
–Lower –Upper
–specification –specification
–20 –25 –30
–Process is capable
Process Capability
–Nominal
–value
–Process distribution
–Lower –Upper
–specification –specification
–20 –25 –30

–Process is not capable
Process Capability Ratio, Cp
Process capability ratio, Cp, is the tolerance width divided by

6 standard deviations (process variability).
Upper specification - Lower specification

Cp =
6
Process Capability Index, Cpk
–Process Capability Index, Cpk, is an index that measures the

potential for a process to generate defective outputs relative to
either upper or lower specifications.
= =
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?
–Upper specification = 30 minutes

–Lower specification = 20 minutes
–Average service = 26.2 minutes
 = 1.35 minutes
Intensive Care Lab
Assessing Process Capability
–Example –Upper specification = 30 minutes
–Lower specification = 20 minutes
 = 1.35 minutes
= =
x –Lower specification
Cpk = Minimum of , Upper specification – x
3
3
Cpk = Minimum of 26.2 – 20.0 30.0 – 26.2
3(1.35) 3(1.35)
–Process
Cpk = Minimum of 1.53, 0.94 = 0.94 Capability
Index
Intensive Care Lab
Assessing Process Capability
–Example
Upper specification - Lower specification
Cp =
6
30 - 20 = 1.23 –Process Capability Ratio
Cp =
6(1.35)
–Does not meet 3 (1.00 Cpk) target due to a shift in mean
(Note variability is ok since Cp is over 1.0)
–Before Process Modification

–Upper specification = 30.0 minutes Lower specification = 20.0
minutes
 = 1.35 minutes Cpk = 0.94 Cp = 1.23
General Rules for Out of Control Situations
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
Gradual Shift :
- Change in the Q. of coming material overtime 

X  
 - Change in the maintenance program 
→ new operator 
 
R → decrease in worker skill 
→ gradual improvement in vendors supply 
 
88
Trending Pattern:
X → Tool wear, die wear, gradual deterioration of Eq.
R → Operator skill 
89
Cyclic Pattern:
→ repetitive periodic behavior

X : Periodic change in temp rotation of operator, seasonal variation
R : Operators fatigue periodic maintenance
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
• Initiate data collection:

– Run the process untouched, and gather sampled data.
– Record data on an appropriate Control Chart sheet or
other graph paper. Include any unusual events that occur.
• Calculate the appropriate statistics and control limits:
– Use the appropriate formulas.
• Construct the control chart(s) and plot the data.
92
Set of Control Chart Interpretation
rules to determine "Out of Control"
General rules for detecting out of control or non-random situations

Any Point Above +3 Sigma
------------------------------------------------------------------------ +3 LIMIT
2 Out of the Last 3 Points Above +2 Sigma
---------------------------------------------------------------------- +2 LIMIT
4 Out of the Last 5 Points Above +1 Sigma
---------------------------------------------------------------------- +1 LIMIT
8 Consecutive Points on This Side of Control Line
=========================================== CENTER LINE
8 Consecutive Points on This Side of Control Line
------------------------------------------------------------------------ -1 LIMIT
4 Out of the Last 5 Points Below - 1 Sigma
-------------------------------------------------------------------------- -2 LIMIT
2 Out of the Last 3 Points Below -2 Sigma
------------------------------------------------------------------------- -3 LIMIT
Any Point Below -3 Sigma
TREND RULE :6 in a row trending up or down.

14 in a row alternating up and down
93
Using x- bar and R-Charts
Together
▪ Process average and process variability must be

in control.
▪ It is possible for samples to have very narrow
ranges, but their averages is beyond control
limits.
▪ It is possible for sample averages to be in
control, but ranges might be very large.
Control Chart Patterns
–UCL
–UCL
–LCL
–Sample observations
–LCL
–consistently below the
–center line
–consistently above the
–center line
Control Chart Patterns (cont.)
–UCL
–UCL
–LCL
–consistently increasing –LCL
–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
▪ 8 consecutive points on one side of the center line

▪ 8 consecutive points up or down across zones
▪ 14 points alternating up or down
▪ 2 out of 3 consecutive points in zone A but still
inside the control limits
▪ 4 out of 5 consecutive points in zone A or B
Control Charts
• UCL = Process Average + 3 Standard Deviations
• LCL = Process Average - 3 Standard Deviations
X
UCL
+ 3 Process
Average
- 3
LCL
TIME
–Levine, Prentic-Hall
Control Chart Representing Limits, Special
Causes, Common Causes
Special cause Common cause

Upper
control limits
Centre Line
Lower
control limits
Stable process Unstable process

100
Control Charts for Variable Type of Data
and R charts
In the x bar chart the sample means are plotted in order to

control the mean value of a variable.
In R chart, the sample ranges are plotted in order to control

the variability of a variable
101
Centre line, upper, & lower control limit for R charts are calculated.
The formulae used are as following:
Ri = X max ( i )
− X min( i )
Ri = range of ith sample

Xmax(i) = maximum value of the data in ith sample
Xmin (i) = minimum value of the data in ith sample
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
Centre line, upper, & lower control limit for x bar chart are calculated.
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
Example
Mean values and ranges of data from 20 samples (sample size = 4) are shown
in the table below:
Mean Mean Mean Mean Mean

S.N of Range S.N of Range S.N of Range S.N of Range S.N of Range
Sample Sample Sample Sample Sample
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
Example
Mean values and ranges of data from 20 samples (sample size = 4) are shown
in the table below:
Mean Mean Mean Mean Mean

S.N of Range S.N of Range S.N of Range S.N of Range S.N of Range
Sample Sample Sample Sample Sample
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
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
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
Sum=521.00 Sum=87 108

Consider a process by which coils are manufactured. Samples of size 5 are
randomly selected from the process, and the resistance values (in ohms) of the
coils are measured. The data values are given in Table
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:
UCLR = D4 R = (2.114)(3.48) = 7.357

LCLR = D3 R = (0)(3.48) = 0
109
g
X i
521.00
X= i =1
= = 20.840
g 25
UCL X = X + A 2 R = 20.84 + (0.577)(3.48) = 22.848
LCL X = X − A 2 R = 20.84 − (0.577)(3.48) = 18.832
110
111
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
The revised center line on the X BAR chart
459
X= = 20.864
22
The revised control limits on the
UCL X = X + A 2 R = 20.864 + (0.577)(3.273) = 22.753
LCL X = X − A 2 R = 20.864 − (0.577)(3.273) = 18.975
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.
• Instead of calculating the standard deviation

for all of the sample data we use these simple
equations
Control Limit Equations
• X (UCLx) = X + A2*R
• X (LCLx) = X - A2*R
Where A2 can be found on page 237 of your Intro to IE textbook
• R (UCLR) = D4*R
• R (LCLR) = D3*R
Where D4 and D3 can be found on page 237 of your Intro to IE
textbook.
Finished Control Charts
Control Chart Analysis
• To analyze the control charts it is important to
remember that the data is represented over six
standard deviations, there are three standard
deviations from the mean line to the upper
control limit and three from the mean to the
lower control limit. To help analyze the
charts, it is important to divide the chart area
into six sections A, B, and C representing the
standard deviations.
Purpose of Control Charts
• Control charts provide a graphical means for
testing hypotheses about the data being
monitored.
Example of Different Zones
Interpreting the Control Chart
• To interpret the data we first look at the R
chart. The R chart represents the variety in
the data and if the variety is to great than
there is no need to look at the X control
chart.
• The chart is out of control if one or a
combination of the following four examples
occur:
Process out of Control
1. If one point falls outside of the 3 sigma

control limits (beyond zone A)
2. If two out of any three successive points fall
in zone A of the same side
3. If four out of any five successive points fall in
or beyond zone B of the same side
4. If eight successive points fall in or beyond
zone C of the same side
How does the Co-ops Control Charts Look?
A Process Is in Control If
…
1. … no sample points outside limits

2. … most points near process average
3. … about equal number of points above
and below centerline
4. … points appear randomly distributed
and s charts
Centre line, upper, & lower control limit for s charts are calculated.
Suppose we have g preliminary samples at our disposition,

each of size n, and let si be the standard deviation of the ith sample.
Then the average of the g standard deviations is
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.
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
Centre line, upper, & lower control limit for x bar charts are calculated.
n
X i
Xi = i =1
n
th
X i = mean of the i sample
n = sample size,
Xi = ith data
g
X X = mean of ‘g’ samples

X = i =1
i
= CL x
g Where A3 is
constants and its
value can be
obtained from
UCLx = X + A3 S (Upper control limit for X bar chart)
standard table. Its
value depend
LCLx = X − A3 S (Lower control limit for X bar chart) upon the size of
133
sample.
Example
Mean values and SD of data from 30 samples (sample size = 5) are shown in
the table below:
SN Mean SD SN Mean SD SN Mean SD SN Mean SD SN Mean SD

1 55.6 9.63 7 46.8 6.72 13 44 14.35 19 50.2 7.6 25 44.6 8.96
2 61 8.63 8 44.2 8.53 14 51.6 5.18 20 44 8.46 26 46.8 6.5
3 45.2 7.4 9 50.8 11.95 15 53.2 5.36 21 50 5.15 27 49.2 3.19
4 46.2 4.09 10 48.4 6.19 16 52.4 9.48 22 47 5.15 28 45.6 7.96
5 46.8 7.22 11 51.2 6.83 17 50.6 3.44 23 50.6 5.55 29 57.6 14.38
6 49.8 8.76 12 49.4 5.46 18 56 7 24 48.8 6.5 30 51.4 6.8
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
Example
Mean values and SD of data from 30 samples (sample size = 5) are shown in
the table below:
SN Mean SD SN Mean SD SN Mean SD SN Mean SD SN Mean SD

1 55.6 9.63 7 46.8 6.72 13 44 14.35 19 50.2 7.6 25 44.6 8.96
2 61 8.63 8 44.2 8.53 14 51.6 5.18 20 44 8.46 26 46.8 6.5
3 45.2 7.4 9 50.8 11.95 15 53.2 5.36 21 50 5.15 27 49.2 3.19
4 46.2 4.09 10 48.4 6.19 16 52.4 9.48 22 47 5.15 28 45.6 7.96
5 46.8 7.22 11 51.2 6.83 17 50.6 3.44 23 50.6 5.55 29 57.6 14.38
6 49.8 8.76 12 49.4 5.46 18 56 7 24 48.8 6.5 30 51.4 6.8
30
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
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
All the data are within the

Sample SD
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
Control chart for attributes

Attribute: - Quality characteristic for which a numerical
value is not specified.
Taste, paint quality, SF, Quality of output,
–Nonconformity (defect)
–Nonconforming (defective)
Attribute - at all levels, cost effective
Variable - at lowest level
 rules are not used because distribution in attributes
charts are non-normal
137
– 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
–(p-chart) Binomial Distribution

–for BD:- validity the Pnc=constant from sample to
sample and the sample must be independent.
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
p chart: chart for proportion nonconforming
p-charts calculates the percent defective in sample. p-charts

are used when observations can be placed in two categories
such as yes or no, good or bad, pass or fail etc.
140
Centre line, upper, & lower control limit for p charts are calculated.
sum of defectives in all samples = centre line of p chart

p=
total number of items in all samples
p (1 − p) (Upper control limit for p chart)

UCL = p + 3
n
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
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
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
–(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
c charts: Chart for the number of nonconformities
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
Centre line, upper, & lower control limit for c charts are calculated.
sum of defects in all samples

c= = centre line of c chart
total number of items in all samples
UCL = c + 3 c (Upper control limit for c chart)
LCL = c − 3 c (Lower control limit for c chart)
146
Example
Data for defects on TV set from 20 samples (sample size = 10) are shown in
the table below:
Sample No. of Sample No. of Sample No. of Sample No. of

No. Defects No. Defects No. Defects No. Defects
1 5 6 4 11 6 16 5
2 4 7 5 12 5 17 4
3 5 8 6 13 4 18 6
4 6 9 8 14 7 19 6
5 4 10 7 15 6 20 6
sum of defects 109

CL = = = 5.45
number of samples 20
UCL = c + 3 c = 5.45 + 3 5.45 = 12.45
LCL = c − 3 c = 5.45 − 3 5.45 = −1.55 = 0

147
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
u charts: Chart for number of nonconformity per unit
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
Centre line, upper, & lower control limit for u charts are calculated.
k
sum of defects in all samples c i
u= =
i =1
k = centre line of u chart
total number of items in all samples n
i =1
i
ci =number of defects in ith sample

k = number of samples
ni = size of ith samples
u
UCL = u + 3 (Upper control limit for u chart)
ni
u (Lower control limit for uc chart)

LCL = u − 3
ni 150
Example
Data for defects in a product from 20 samples are shown in the table below.
The number of products varies from sample to sample. Construct a control
for number of defects per 100 products.
Sample No. of No. of Sample No. of No. of Sample No. of No. of

No. products Defects No. products Defects No. products Defects
inspected inspected inspected
1 200 5 8 150 10 15 100 6
2 300 14 9 150 6 16 200 8
3 250 8 10 250 10 17 200 5
4 150 8 11 300 9 18 100 5
5 250 12 12 250 16 19 300 14
6 100 6 13 200 12 20 200 8
7 200 20 14 250 10
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

Rocess Ontrol Tatistical: Drrswalia

Uploaded by

Copyright:

Available Formats

You might also like

Rocess Ontrol Tatistical: Drrswalia

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Rocess Ontrol Tatistical: Drrswalia

Uploaded by

Copyright:

Available Formats

S TATISTICAL

“The first job we have, is to turn out quality

Good quality of product and service can provide:

2. Reduce cost due to returns, rework, and scrap

Running a company by profit alone is

Cost of Conformance + Cost of Non-Conformance

• Prevention cost: Cost of planning and executing a project so it is

• Appraisal cost: Cost of evaluating processes and their outputs to

External Failure Cost

Variance is the average squared

1. Find the mean of the data.

–The variance formula includes the Sigma

– Standard Deviation is often denoted

–6) Find the square root of the

–Thus the standard deviation of the

A different math class took the same

Find the standard deviation for this

Find the standard deviation for this class.

– 5) Divide the sum of the squares

Consider both sets of scores. Both classes

With a standard variation of 14.53 for the

What is System of point of

Who is the better shot? 28

• What do you measure in your process?

• Deviation = distance between 8

• Deviation = distance between

Observations Deviations Squared Deviations

• Variance = average distance between

Observations Deviations Squared Deviations

• Standard deviation = square root

But what good is a standard deviation

The world tends to

Even very rare Fewer Most Fewer Even very rare

Here is why: Even outcomes that are equally

Add up the dots on the dice

–One standard deviation away from the mean (  ) in either

The normal The arithmetic

The normal probability

–How does the Standard Deviation values

–How does the following Expected values

A z-value is the distance between a selected value (designated X)

–Area = +0.5 –Area = + 0.5

–…since the mean is located at the peak of the curve,

You can use a white board or a scrap

• About how many teens had reaction times

• Find the probability that a person waits longer

Add up about 30 or most things

Normal distributions are divided up

Once you know that, you

And that is what a standard deviation

Assignable Causes : Control of Process

SPC uses samples to identify that Assignable causes have occurred 68

What Does a Control Chart Do?

▪ Data are of two types

Variables “Things we measure”

Length Blood Pressure Diameter

Attributes “Things we count”