Control Charts for

Attributes
Prof. Dr. Mahmoud Moustafa El_Sherbiny
 Introduction

Data that can be classified into one of several

categories or classifications is known as
attribute data.
Classifications such as conforming and
nonconforming are commonly used in quality
control.
Another example of attributes data is the count
of defects.
 Control Charts for Fraction
Nonconforming

Fraction nonconforming is the ratio of the

number of nonconforming items in a
population to the total number of items in
that population.
Control charts for fraction nonconforming
are based on the binomial distribution.
 Control Charts for Fraction
Nonconforming

Recall: A quality characteristic follows a binomial

distribution if:
1. All trials are independent.
2. Each outcome is either a success or failure.
3. The probability of success on any trial is given as
p. The probability of a failure is 1-p.
4. The probability of a success is constant.
 Control Charts for Fraction
Nonconforming

The binomial distribution with parameters n0

and 0 < p < 1, is given by
n x nx

p( x ) p (1 p)
x
The mean and variance of the binomial
distribution are
np np (1 p)
2
 Control Charts for Fraction
Nonconforming

Development of the Fraction Nonconforming

Control Chart
Assume
n = number of units of product selected at random.
D = number of nonconforming units from the sample
p= probability of selecting a nonconforming unit from
the sample.
Then:
n x nx
P(D x ) p (1 p)
x
 Control Charts for Fraction
Nonconforming

Development of the Fraction Nonconforming

Control Chart
The sample fraction nonconforming is given as
D
p
n
where p is a random variable with mean and
variance
p(1 p)
p
2

n
 Control Charts for Fraction
Nonconforming

Standard Given
If a standard value of p is given, then the
control limits for the fraction nonconforming
are
p(1 p)
UCL p 3
n
CL p
p(1 p)
LCL p 3
n
 Control Charts for Fraction
Nonconforming
No Standard Given
If no standard value of p is given, then the control
limits for the fraction nonconforming are

p (1 p )
UCL p 3
n
CL p
p (1 p )
LCL p 3
n m m
where Di p i
p i 1
i 1
mn m
 Control Charts for Fraction
Nonconforming

Trial Control Limits

Control limits that are based on a preliminary set
of data can often be referred to as trial control
limits.
The quality characteristic is plotted against the
trial limits, if any points plot out of control,
assignable causes should be investigated and
points removed.
With removal of the points, the limits are then
recalculated.
 Control Charts for Fraction
Nonconforming

Example
A process that produces bearing housings is
investigated. Ten samples of size 100 are
selected.
Sample # 1 2 3 4 5 6 7 8 9 10
# Nonconf. 5 2 3 8 4 1 2 6 3 4

Is this process operating in statistical control?

 Control Charts for Fraction
Nonconforming

Example
n = 100, m = 10
Sample # 1 2 3 4 5 6 7 8 9 10
# Nonconf. 5 2 3 8 4 1 2 6 3 4
Fraction
0.05 0.02 0.03 0.08 0.04 0.01 0.02 0.06 0.03 0.04
Nonconf.

m
p i
p i 1
0.038
m
 Control Charts for Fraction
Nonconforming
p (1 p )
Example UCL p 3
n
Control Limits are: CL p
p (1 p )
LCL p 3
n
0.038(1 0.038)
UCL 0.038 3 0.095
100
CL 0.038
0.038(1 0.038)
LCL 0.038 3 0.02 0
100
 Control Charts for Fraction
Nonconforming
Example
P Ch art for C1

0.10
3.0SL= 0.09536
Proporti on

0.05

P= 0.03800

0.00 - 3.0SL= 0.000

0 1 2 3 4 5 6 7 8 9 10

S a m pl e N u m b er
 Control Charts for Fraction
Nonconforming

Design of the Fraction Nonconforming Control

Chart
The sample size can be determined so that a shift of
some specified amount, can be detected with a stated
level of probability (50% chance of detection). If is
the magnitude of a process shift, then n must satisfy:
p(1 p)
L
n
Therefore,
2
L
n p(1 p)

 Control Charts for Fraction
Nonconforming

Positive Lower Control Limit

The sample size n, can be chosen so that the lower
control limit would be nonzero:

p(1 p)
LCL p L 0
n
and
(1 p) 2
n L
p
 Control Charts for Fraction
Nonconforming

Interpretation of Points on the Control Chart for

Fraction Nonconforming

Care must be exercised in interpreting points

that plot below the lower control limit.
They often do not indicate a real improvement in
process quality.
They are frequently caused by errors in the inspection
process or improperly calibrated test and inspection
equipment.
 Control Charts for Fraction
Nonconforming

The np control chart

The actual number of nonconforming can also be
charted. Let n = sample size, p = proportion of
nonconforming. The control limits are:
UCL np 3 np (1 p)
CL np
LCL np 3 np (1 p)
(if a standard, p, is not given, use p )
 Variable Sample Size

In some applications of the control chart for

the fraction nonconforming, the sample is a
100% inspection of the process output over
some period of time.
Since different numbers of units could be
produced in each period, the control chart
would then have a variable sample size.
 Variable Sample Size

Three Approaches for Control Charts with

Variable Sample Size

1. Variable Width Control Limits

2. Control Limits Based on Average Sample Size
3. Standardized Control Chart
 Variable Sample Size

Variable Width Control Limits

Determine control limits for each individual
sample that are based on the specific sample
size.
The upper and lower control limits are

p(1 p)
p3
ni
 Variable Sample Size

Control Limits Based on an Average Sample Size

Control charts based on the average sample size
results in an approximate set of control limits.
The average sample size is given by
m
ni
n i 1
m

The upper and lower control limits are

p(1 p)
p3
n
 Variable Sample Size

The Standardized Control Chart

The points plotted are in terms of standard
deviation units. The standardized control chart
has the follow properties:
Centerline at 0
UCL = 3 LCL = -3
The points plotted are given by:
p i p
zi
p(1 p)
ni
 The Operating-Characteristic Function and
Average Run Length Calculations

The OC Function
The number of nonconforming units, D, follows
a binomial distribution. Let p be a standard
value for the fraction nonconforming. The
probability of committing a Type II error is
P(p UCL | p) P(p LCL | p)
P(D nUCL | p) P(D nLCL | p)
 The Operating-Characteristic Function and
Average Run Length Calculations

Example
Consider a fraction nonconforming process
where samples of size 50 have been collected
and the upper and lower control limits are
0.3697 and 0.0303, respectively.It is important
to detect a shift in the true fraction
nonconforming to 0.30. What is the probability
of committing a Type II error, if the shift has
occurred?
The Operating-Characteristic Function and
Average Run Length Calculations

Example
For this example, n = 50, p = 0.30, UCL =
0.3697, and LCL = 0.0303. Therefore, from the
binomial distribution,
P(D nUCL | p) P(D nLCL | p)
P(D 50(0.3697) | 0.30) P(D 50(0.0303 | 0.30)
P(D 18.48 | 0.30) P(D 1.515 | 0.30)
P(D 18 | 0.30) P(D 1 | 0.30)
0.8594 0
0.8594
 The Operating-Characteristic Function and
Average Run Length Calculations

OC curve for the fraction nonconforming control chart

with p = 0.20, LCL = 0.0303 and UCL = 0.3697.

1.0

0.5
B

0.0

0.0 0.1 0.2 0.3 0.4 0.5 0.6

p
The Operating-Characteristic Function and
Average Run Length Calculations

ARL
The average run lengths for fraction
nonconforming control charts can be found as
before:
1
The in-control ARL is ARL 0

1
The out-of-control ARL is ARL 1
1
Control Charts for Nonconformities (Defects)

There are many instances where an item will

contain nonconformities but the item itself is
not classified as nonconforming.
It is often important to construct control charts
for the total number of nonconformities or the
average number of nonconformities for a given
area of opportunity. The inspection unit
must be the same for each unit.
Control Charts for Nonconformities (Defects)

Poisson Distribution
The number of nonconformities in a given area can be
modeled by the Poisson distribution. Let c be the
parameter for a Poisson distribution, then the mean and
variance of the Poisson distribution are equal to the
value c.
The probability of obtaining x nonconformities on a
single inspection unit, when the average number of
nonconformities is some constant, c, is found using:
ecc x
p( x )
x!
 Procedures with Constant Sample Size

c-chart
UCL c 3 c
Standard Given:
CL c
LCL c 3 c

UCL c 3 c
No Standard Given: CL c
LCL c 3 c
 Procedures with Constant Sample Size

Choice of Sample Size: The u Chart

If we find c total nonconformities in a sample of n
inspection units, then the average number of
nonconformities per inspection unit is u = c/n.
The control limits for the average number of
nonconformities is

u
UCL u 3
n
CL u
u
LCL u 3
n
 Procedures with Variable Sample Size

Three Approaches for Control Charts with

Variable Sample Size

1. Variable Width Control Limits

2. Control Limits Based on Average Sample Size
3. Standardized Control Chart
 Procedures with Variable Sample Size

Variable Width Control Limits

Determine control limits for each individual
sample that are based on the specific sample
size.
The upper and lower control limits are

u
u3
ni
 Procedures with Variable Sample Size

Control Limits Based on an Average Sample Size

Control charts based on the average sample size
results in an approximate set of control limits.
The average sample size is given by
m
ni
n i 1
m
The upper and lower control limits are

u
u3
n
 Procedures with Variable Sample Size

The Standardized Control Chart

The points plotted are in terms of standard
deviation units. The standardized control chart
has the follow properties:
Centerline at 0
UCL = 3 LCL = -3
The points plotted are given by:
ui u
zi
u
ni
 Demerit Systems

When several less severe or minor defects

can occur, we may need some system for
classifying nonconformities or defects
according to severity; or to weigh various
types of defects in some reasonable
manner.
 Demerit Systems

Demerit Schemes

1. Class A Defects - very serious

2. Class B Defects - serious
3. Class C Defects - Moderately serious
4. Class D Defects - Minor

Let ciA, ciB, ciC, and ciD represent the number

of units in each of the four classes.
 Demerit Systems

Demerit Schemes
The following weights are fairly popular in
practice:
Class A-100, Class B - 50, Class C 10, Class D - 1

di = 100ciA + 50ciB + 10ciC + ciD

di - the number of demerits in an inspection unit

 Demerit Systems

Control Chart Development

Number of demerits per unit:
D
ui
n
where n = number of inspection units
n
D = di
i 1
 Demerit Systems

Control Chart Development

UCL u 3 u
CL u
LCL u 3 u
where u 100u A 50u B 10u C u D
and
100 u A 50 u B 10 u C u D
2 2 2 1/ 2

u
n
The Operating-Characteristic Function

The OC curve (and thus the P(Type II

Error)) can be obtained for the c- and u-
chart using the Poisson distribution.
For the c-chart:
P( x UCL| c ) P( X LCL| c )
where x follows a Poisson distribution
with parameter c (where c is the true mean
number of defects).
 The Operating-Characteristic Function

For the u-chart:

P( x UCL | u ) P( x LCL | u )
P(c nUCL | u ) P(c nLCL | u )
 Dealing with Low-Defect Levels

When defect levels or count rates in a process

become very low, say under 1000 occurrences
per million, then there are long periods of time
between the occurrence of a nonconforming
unit.
Zero defects occur
Control charts (u and c) with statistic
consistently plotting at zero are uninformative.
 Dealing with Low-Defect Levels

Alternative
Chart the time between successive occurrences
of the counts or time between events control
charts.
If defects or counts occur according to a
Poisson distribution, then the time between
counts occur according to an exponential
distribution.
 Dealing with Low-Defect Levels

Consideration
Exponential distribution is skewed.
Corresponding control chart very asymmetric.
One possible solution is to transform the exponential
random variable to a Weibull random variable using
x = y1/3.6 (where y is an exponential random variable)
this Weibull distribution is well-approximated by a
normal.
Construct a control chart on x assuming that x
follows a normal distribution.
 Choice Between Attributes
and Variables Control Charts

Each has its own advantages and disadvantages

Attributes data is easy to collect and several characteristics
may be collected per unit.
Variables data can be more informative since specific
information about the process mean and variance is
obtained directly.
Variables control charts provide an indication of
impending trouble (corrective action may be taken before
any defectives are produced).
Attributes control charts will not react unless the process
has already changed (more nonconforming items may be
produced.
 Guidelines for Implementing
Control Charts

1. Determine which process characteristics to

control.
2. Determine where the charts should be
implemented in the process.
3. Choose the proper type of control chart.
4. Take action to improve processes as the result
of SPC/control chart analysis.
5. Select data-collection systems and computer
software.