STATISTICAL QUALITY

CONTROL

CHAPTER THREE:
Statistical Process Control

Samuel.B
 STATISTICAL PROCESS CONTROL

Process variation
 Variation exist in all processes.
 The variation concept is a law of nature in that no two
natural items in any category are the same.
The variation may be quite large and easily noticeable.
The variation may be very small. It may appear that
items are identical; however, precision instruments
will show difference.
The ability to measure variation is necessary before it
can be controlled.
 Variation
Sources of Variation in production
processes:Operators Methods Measurement
Materials Instruments

INPUTS PROCESS OUTPUTS

Tools Human
Machines Environment Inspection
Performance
 Method
Environment

Man Material

Sources of
variability Machine
 Process variation

 Variation could be created due to:

 Equipment: tool wear, machine vibration, work
holding-device positioning, electrical fluctuation etc….
 Materials: tensile strength, thickness, porosity etc…
of the input material.
 Environment: temperature, light radiation, pressure,
humidity etc…
 Operator: physical and emotional wellbeing of the
operator, skill and knowledge etc…
 There is also a reported variation which is due to the
inspection activity.
Causes of variation
Common cause: take place as per chance or in
a random fashion as a result of the cumulative
effect of several minor causes which cannot be
identified. These causes are inherent in every
type of production.

Assignable cause: refers to those problems in

the quality of the products which can be assigned
or attributed to any particular causes like
defective materials, defective labour, etc.
 Characteristics of causes of variation

 Common cause
 Is naturally occurring variation inherent in all processes.
 Is unavoidable and inevitably occurs in a process.
 Its effect is minute and is tolerable
 They are problems inherent in the manufacturing system, such as
high temperature or excessive humidity.
 Assignable cause
 Typically caused by some problem or extraordinary occurrence in
the system.
 Are local, sporadic problems such as the failure of a particular
machine or a mistakenly recorded measurement.
 Causing factors can be clearly identified and possibly managed
(e.g.. Untrained worker, faulty machine, out of calibrated
instrument, etc.)
 Statistical Process Control (SPC)

 SPC is the application of statistical techniques to control a

process; measuring and analyzing the variation in processes
so that performance remains within specification limits.
 A process is considered in control or statistical control when
any variation in the specifications of the output product is
minute and is due to common cause of variation.
 However, it is said that a process is out of statistical control if
it implies that certain assignable causes are present affecting
the variability of quality.
 The objective of statistical process control is to detect such
assignable causes as soon as they occur in the production
process.
 Cont’d
The technique of process control implies:-
I. Determination of the way in which variations in
quality would be distributed when the process is
under control, and
II. Checking on a continuous basis whether the
variability in quality confirms to this distribution
or not.
 In other words, using statistical process control we want
to determine the amount of variation that is common or
normal and then we monitor the production process to
make sure production stays within this normal range.
 Cont’d
 Whether the process is in control or out of
control can be determined by the use of
control charts.
 A control chart provides a basis for deciding

whether the variation in the output is due to

random causes or due to assignable causes.
 It will assist us in making decisions whether

to adjust the process or not.

 Control charts
 A control chart is a graph that shows whether
a sample of data falls within the common or
normal range of variation.
 A control chart has upper and lower control

limits that separate common from assignable

causes of variation. The common range of
variation is defined by the use of these control
chart limits.
 We say that a process is out of control when a

plot of data reveals that one or more samples

 Cont’d

Fig: Control chart example

 Cont’d
 The upper and lower control limits on a
control chart are usually set at  3 standard
deviations from the mean. If we assume that
the data exhibit a normal distribution, these
control limits will capture 99.73 percent of
the normal variation.
 Control limits can be set at  2 standard

deviations from the mean. In that case,

control limits would capture 95.44 percent of
the values.
 Cont’d
 On the basis of probability distribution, if a
variable X is normally distributed, the
probability that a random observation on the
variable will lie between   3 is 99.73%,
which is extremely high.
 This means that the probability of a random

observation going beyond these limits is

nearly 0.26%.
 Cont’d
 Thus, the variable quality characteristic is
assumed to normally distribute and the
probability of a sample point going outside
3 limits when the process is in control is
very small.
 If a sample point goes beyond this limit, it

is highly likely that the normality

assumption of the process is not applicable.
 Types of Control charts
 The different characteristics that can be
measured by control charts can be divided into
two groups: variables and attributes.
 A control chart for variables is used to
monitor characteristics that can be measured
and have a continuum of values, such as height,
weight, or volume.
 A control chart for attributes, on the other
hand, is used to monitor characteristics that
have discrete values and can be counted. Often
TYPES OF CONTROL CHARTS
Types of Control Charts

Control Charts for Control Charts for

Variables Attributes

p- np C- U-
 R- σ-
Chart Chart Chart
Cha - Cha Cha
rt Cha rt rt
rt
 CONTROL CHARTS FOR
VARIABLES
 Control charts for variables monitor
characteristics that can be measured and
have a continuous scale, such as height,
weight, volume, or width.
 For example, if we were producing candles,

height might be an important variable.

 We could take samples of candles and

measure their heights.

 Cont’d
 Two of the most commonly used control charts
for variables monitor both the central tendency
of the data (the mean) and the variability of the
data (either the standard deviation or the range).
 Note that each chart monitors a different type of

information.
 When observed values go outside the control

limits, the process is assumed not to be in

control.
 Mean chart ( X -Chart)
 The mean chart is used for monitoring a shift in a
mean of the data. Thus, it is a measure of central
tendency of a process.

 x-bar chart displays the subgroup average of the

 Cont’d
 To construct a mean chart we first need to
construct the center line of the chart. To do
this we take multiple samples and compute
their means.
 Usually these samples are small, withX about
four or five observations. Each sample has
its own mean, .
 The center line of the chart is then computed

as the mean of all k sample means, where k

is the number of samples:
 Cont’d
 To construct the upper and lower control
limits of the chart, we use the following
formulas:
 Cont’d
 SinceX is an estimate of the population
mean , the control limits can also be
expressed as:
 
UC L =  + 3 and LCL =  - 3 .
n n
 Cont’d
 If the sample size is not large enough so that
the sample standard deviation cannot be
taken as an estimate of the population
standard deviation , then the value of  can
also be estimated by the following formula.
 R d
2

Where R-bar is the average of the ranges of all

 Cont’d
 Then the control limits can be redefined as:
3. R
 d2
UCL X  3 X 
and
n n
3. R
 d2
LCL X  3 X 
n n

 Let A2 = 3
.
d2 n

Then the control limits will be:

UCL = X + A2 R and, LCL = X - A2 R
 Cont’d
 Example: A new machine is installed to fill
aluminum cans with mixed dry fruits. The
machine is set to fill 40 ounces of dry fruit in
each can. The weight of the can is taken as 5
ounces, so that the total weight is 45 ounces. A
quality control inspector takes a sample of five
filled cans every hour and records the individual
weights as well as the average weight of each
sample as shown below. X Construct a control chart
for and check if the process is under control.
Cont’d

Table: Weight of
cans.
Cont’d
Solution
X of R
 First we calculate the values and .

X
 Cont’d

Then X = 180/4 = 45 Sample Mean Upper Lower

R= 12/4 = 3 Size, n Factor, A2 Range, D4 Range, D3
2 1.880 3.268 0
3 1.023 2.574 0
X R
UCL = + A2 4 0.729 2.282 0

X R 5 0.577 2.115 0
UCL = - A2 6 0.483 2.004 0
7 0.419 1.924 0.076
UCL = 45 + 0.577(3) 8 0.373 1.864 0.136

= 46.731 9 0.337 1.816 0.184

10 0.308 1.777 0.223
LCL = 45 - 0.577(3)
12 0.266 1.716 0.284
= 43.269 0.184

Table: Factors for Computing Control Chart Limits

The mean values of each sample can be
plotted as follows:

Figure: chart of weight in ounce.

 R- Charts

 Range charts monitor the dispersion or

variability of the process.
 It is complementary X to chart and usually
X
both chart and R chart are together
required for an adequate analysis of the
production process being considered.
 Mean and Range Charts
(a)
These (Sampling
sampling
distributions mean is shifting
result in the upward but
charts below range is
UCL
consistent)
(x-chart
x-chart detects shift in
LCL central
UCL tendency)
(R-chart does
R-chart not detect
LCL change in
mean)
 Mean and Range Charts
(b)
These (Sampling
sampling mean is
distributions constant but
result in the
charts below
dispersion is
UCL increasing)
(x-chart does
x-chart not detect the
LCL increase in
UCL dispersion)
(R-chart
R-chart detects
LCL increase in
dispersion)
 Cont’d
The center line and control limits are
given by K

R j1
j

CL  R 
K
 R   3d 3 
UCL R  3 R R  3d 3 
d  R 
 1  
 and ,
 2   d2 
 R   3d 3 
LCL R  3 R R  3d 3 
d  R 
1 d 

 2   2 
 3d 3   3d 3 
To simplify the calculations, let 
 1  
  D 4 and 
 1  
  D3
 d2   d2 

Then UCL R= D4, and LCL

R =
D3
 Example
 An Electronics Company manufactures
batteries for use in flashlights. The company
claims that these batteries last on an average
of 100 hours in continuous use. The quality
control inspector takes a random sample of
four batteries every hour from the production
lot and puts these batteries to test. The
following table shows the number of hours
each battery in each sample lasted Xfor 10
consecutive samples taken on a given day.
Hours, the sample batteries last for
Sample Batteries Mean Range R
number 1 2 3 4 X

1 100 98 96 102 99.0 6

2 98 96 104 100 99.5 8
3 104 100 102 104 102.5 4
4 98 98 94 100 97.5 6
5 103 100 96 101 100.0 7
6 100 104 102 100 101.5 4
7 101 99 103 101 101.0 4
8 104 100 98 102 101.0 6
9 94 97 103 102 99.0 9
10 104 100 100 100 101.0 4
Cont’d
 For the X- bar chart
(for n=4, the value of A2 is 0.729)
X
CL = X = 1002.0/10
R = 100.2
UCL = + A
= 100.2 + (0.729*5.8)
= 104.43
X R
LCL = -A
= 100.2 - (0.729*5.8)
= 95.97
Cont’d
 For the R-chart
(for n=4; value of D4=2.282 and D3=0)
CL =R
= 58/10 = 5.8
UCL =DR4

=2.282*5.8 = 13.24
LCL =DR3

=0
 Analysis of Patterns on Control Charts

 The process is out of control if any one or more of the

following criteria is met.
I. Extreme Point Conditions: One or more points outside of
the control limits.
This pattern may indicate:
a) A special cause of variance from a material, equipment,
method, or measurement system change.
b) Wrong measurements of a part or parts.
c) Miscalculated or miss-plotted data points.
d) Miscalculated or miss-plotted control limits.
 Analysis of Patterns on……(Cont’d)

UCL

CL

LCL

Time

Figure : Process out of Statistical Control Points outside

Control limits
 Analysis of Patterns on……(Cont’d)

II. Control Chart Zones:

 Control charts can be broken into three zones, A, B, and C on
each side of the process center line.
 A series of rules exist that are used to detect conditions in
which the process is behaving abnormally to the extent that
an out of control condition is declared.
Analysis of Patterns on……(Cont’d)

3s
Zone A
UCL 2s
Zone B
1s
Zone C
CL
Zone C
1s
Zone B
LCL 2s
Zone A 3s

Sample number

Figure: The Three Zones in Control limits

 Analysis of Patterns on……(Cont’d)

a) Two of Three Consecutive Points in Zone A or Outside Zone A

X
Zone A
UCL
Zone B Z

Zone C
CL
Zone C

Zone B Y
LCL
Zone A

Time
 Analysis of Patterns on……(Cont’d)

b) Four of five consecutive points in zone B or beyond

X
UCL
Zone B
Z
Zone C
CL
Zone C
Y
Zone B
LCL
Zone A

Time
 Analysis of Patterns on……(Cont’d)

c) Runs above or below the Center Line

(Eight successive points on the same side of the centerline)
Zone A
UCL
Zone B

Zone C X
CL
Zone C

Zone B
LCL
Zone A

Time
 Analysis of Patterns on……(Cont’d)
d) Linear Trend
(Six successive points that increase or decrease)

Zone A
UCL
Zone B
X Y
Zone C
CL
Zone C

Zone B
LCL
Zone A

Time
 Analysis of Patterns on……(Cont’d)

e) Oscillatory trend: having 14 or more consecutive points

oscillating back and forth.

Zone A
UCL
X
Zone B

Zone C
CL
Zone C

Zone B
LCL
Zone A

Time
 Analysis of Patterns on……(Cont’d)

f) Avoidance of Zone C: The probability of having 8 or more consecutive

points occurring on either side of the center line and do not enter Zone C.

Zone A
UCL
Zone B

Zone C X
CL
Zone C

Zone B
LCL
Zone A

Time
 Analysis of Patterns on……(Cont’d)

g) Run in Zone C: The probability of having 15 or more consecutive

points occurring the Zone C.

Zone A
UCL
Zone B

Zone C X
CL
Zone C

Zone B
LCL
Zone A

Time
 - temperature and humidity changes, operator fatigue, rotation of
operators and electrical fluctuations
50 - operators fatigue, shift (day or night) and worn tools or dies

-mixtures can occur with over control

-difference in materials and measuring equipment

-incorrect calculation of the control limits

-change to a new type of material, new operator, new inspector,

new machines and so on
-change in motivation of the operators, new operators and new
equipment are few of the many

-caused due to gradual deterioration of equipment, worker

fatigue and accumulation of waste products
-improvement or deterioration of operator’s skill work fatigue
and gradual change in homogeneity of incoming quality are
come of the causes
 Analysis of Patterns on……(Cont’d)

A process is said to be “IN CONTROL” If the

following are true:

1) There are no sample points outside the control limits.

2) Most points are near the process average with out too many
close to the control limits.
3) There is approximately equal number of sample points above
and below the center line.
4) The points appear to be randomly distributed around the
center line.
CONTROL CHART FOR
ATTRIBUTES
Attribute Control charts include:
1. P- Chart
2. np- Chart
3. C- Chart
4. U- Chart
 P-chart: monitors the proportion (fraction) of
defective or unacceptable items in samples.
 np- chart: monitors the number of defectives

(unacceptable items) in samples; instead of

proportions.
 C- chart: monitors the number of defects or

nonconformities per inspection unit (ex: per day,

per square meter, etc.).
 U- chart is used when it is not possible to have

an inspection unit of a fixed size.

 P- Chart
 A chart for the proportion of nonconforming items
(p-chart) is based on a binomial distribution.
 For a given sample, the proportion nonconforming
is defined as
P (i ) 
x (i )
n
 where x is the number of nonconforming items in the sample and n
represents the sample size.
 With the P-chart, a sample is taken periodically
from the process and the proportion of defective
items in the sample is determined to see if it falls
with in the control limits in the chart.
 Steps in constructing P- chart
 Determine the rate of nonconformities in each subgroup
by using: x (i )
P (i ) 
n

Where:
P (i ) 

x(i ) 
the rate of nonconformities in subgroup i
n
the number of nonconformities in subgroup i

the size of subgroups

 Determine the center line:-
 1
CL = P 
k
 P (i )

Where:-
x (i )
P (i ) 
n
 The control limits are determined as:-
P 1  P 
UCL  P  3
n

P (1  P )
LCL P  3
n

 Plot the CL, LCL and UCL, and the process

measurements.
 Interpret the data to determine if the process is in
control
 Example
Quantity
Fifteen samples with Samp
le
different sample sizes are numb
chosen from a plastic- er
Inspected Rejected
injection molding machine 1 50 4
producing small containers. 2
3
50 7
50 5
The number of 4 125 12
nonconforming containers 5
6
100 8
75 4
for each sample is shown 7 50 13
in table below. Construct 8 50 2
9 125 16
the appropriate control 10 125 12
chart and comment 11 100 14
12 100 9
Plastic-injection molding example
 np Charts
 np charts are statistical
tools used to evaluate
the count of defectives,
or count of items
nonconforming, produced
by a process.
 np charts can be applied
to any attribute where
the appropriate
performance measure is
a unit count and the
subgroup size is held
constant.
 Example
 The data in Table below were obtained by opening
randomly selected crates from each shipment and
counting the number of bruised peaches. There
are 250 peaches per crate (constant n is required
for np charts). Using the above data the center
line and control limits are found as follows:
TABLE : Raw Data for np Chart

Ship Bruised Bruised

m. peache Shipm peache Using the data the center line and
no. s . no. s control limits are found as follows:
1 20 16 23
2 28 17 27
3 24 18 28
4 21 19 31
5 32 20 27
6 33 21 30
7 31 22 23
8 29 23 23
9 30 24 27
10 34 25 35
11 32 26 29
12 24 27 23
13 29 28 23
14 27 29 30
15 37 30 28
FIGURE: Completed np control
chart.
C - chart
 C-chart measures the number of
nonconformities (defects) per "unit"
which is commonly referred to as an
inspection unit and may be "per day" or
"per square foot" of some other
predetermined sensible rate.
There are k inspection units and c(i) is the
number of nonconformities in the ith
sample.
Hence:-  1
c   c (i )
k
CL =  
UCL  c  3 c
 
LCL  c  3 c
 Exercise
• The following table shows the number of
imperfections contained in a sample of fifteen
pieces of cloth taken from different rolls. Draw a
C-chart using these data and state whether the
process is in a state of statistical control.
Sample number 1 2 3 4 5 6 7 8 9 1 1 1 1 1 1
0 1 2 3 4 5

Number of 3 5 8 4 3 5 7 5 2 4 6 3 5 7 4
defects