Statistical Quality Control

Dr. Samir Parikh

Quality Improvement and Statistics
 Definitions of Quality

 Quality means fitness for use

 - quality of design
 - quality of conformance
 Quality is inversely proportional to variability.

 Quality Improvement
 Quality improvement is the reduction of variability
in processes and products.
 Alternatively, quality improvement is also seen as
“waste reduction”.
Basic Statistics
Descriptive Statistics
A straightforward presentation of facts. A
survey or summary of a population in which all
data are known.
Statistics used to describe quality characteristics and
relationships.
Inferential Statistics
Drawing conclusions about a population from a
random sample
Inferential Statistics
Inferential statistics is a valuable tool because it
allows us to look at a small sample size and make
statements on the whole population.
Samples must be pulled RANDOMLY from a
population so that the sample truly represents the
population. Every unit in a population must have a
equal chance of being selected for the sample to be
truly random.
The distribution or shape of the data is important to
know for analytical purposes.
The most common distribution is the bell shaped or
normal distribution.
Inferential Statistics
Parameters can be estimated from sample statistics.
Two of the most common parameters are the mean
and standard deviation.
The mean (or average, denoted by μ) measures central
tendency
This is estimated by the sample mean or x-bar.
The standard deviation (σ ) measures the spread of
the data and is estimated by the sample standard
deviation
 σ =
 Three SQC Categories
Statistical quality control (SQC) is the term used to
describe the set of statistical tools used by quality
professionals to evaluate organizational quality.
SQC encompasses three broad categories of;
Descriptive statistics
 e.g. the mean, standard deviation, and range

Statistical process control (SPC)

 Involves inspecting the output from a process
 Quality characteristics are measured and charted

 Helpful in identifying in-process variation

Acceptance sampling used to randomly inspect a

batch of goods to determine acceptance/rejection
 Does not help to catch in-process problems
Kinds of data
All of these statistical techniques may be applied to two kinds of
data.
1. Attribute Data: when the quality characteristic being
investigated is noted by either its presence or absence and then
classified as Defective or Non-Defective.
Example: Conforming or non-conforming
 Pass or fail
 Good or bad

2. Variable Data: The characteristics are actually measured and can

take on a value along a continuous scale.
Example: Length, Weight
Thus, attribute data does not have information of “how much
good or how much bad?” which the variable data would have,
because it would record the exact measurements of each shaft.
Dr. Walter A. Shewhart
Dr. Walter A. Shewhart,
called the father of
quality control
analysis, developed the
concepts of statistical
quality control
The foundation for
Statistical Process Control
was laid by Dr. Walter
Shewart
What is Statistical Process Control?
Statistical Process Control is an analytical decision
making tool which allows you to see when a process is
working correctly and when it is not.
Variation is present in any process, deciding when the
variation is natural and when it needs correction is the
key to quality control.
The foundation for Statistical Process Control was laid by
Dr. Walter Shewart working in the Bell Telephone
Laboratories in the 1920s conducting research on methods
to improve quality and lower costs. He developed the
concept of control with regard to variation, and came up
with Statistical Process Control Charts which provide a
simple way to determine if the process is in control or not.
Statistical process control
Statistical process control is a collection of tools
that when used together can result in process
stability and variance reduction
TOYATA’S ADVANTAGE OF USING SQC Based on
the corporate philosophy of customer first and
quality first since its founding, Toyota Motor Co.,
Ltd. won the Deming Application Prize in 1965and
the Japan Quality Control Award in 1970, following
the introduction of statistical quality control
(SQC) in 1949.
OPERATION PROCESS
The goal of every operation or production system is to generate
a useful product.
The product may be a service, information, or physical object.
Each production cycle begins with inputs that are transformed
by a process into a more desired state or into the product.
 NEED OF QUALITY CONTROL:
In each process, excessive variations and errors can cause
nonconformities, which leads to three undesirable
consequences:
(a) scrapped or wasted resources;
(b) degraded process throughput;
(c) “contamination” from undetected nonconformities, reducing
the value of the product to the customer.
GOAL OF QUALITY CONTROL
 The goal of quality control in every production
system is to
(a) eliminate nonconformities and their
consequences,
(b) eliminate rework and wasted resources, and
(c) achieve these goals at the lowest possible cost
The eventual goal of SPC is the elimination of
variability in the process.
Sources of Variation
Variation exists in all processes.
Variation can be categorized as either;
Common or Random causes of variation, or
 Random causes that we cannot identify
 Unavoidable

 e.g.
slight differences in process variables like
diameter, weight, service time, temperature
Assignable causes of variation
 Causes can be identified and eliminated
 e.g.
poor employee training, worn tool, machine
needing repair
Traditional Statistical Tools
Descriptive Statistics include
The Mean- measure of n
central tendency x i
x  i 1

The Range- difference n

between largest/smallest
observations in a set of data

 x 
Standard Deviation n 2
measures the amount of data i  X
dispersion around mean σ  i1

Distribution of Data n 1
shape
 Normal or bell shaped or
 Skewed
Distribution of Data
Normal distributions Skewed distribution
Introduction to Control Charts
Basic Principles
 A process that is operating with only chance
causes of variation present is said to be in
statistical control.
 A process that is operating in the presence of
assignable causes is said to be out of control.
 The eventual goal of SPC is the elimination of
variability in the process.
Introduction to Control Charts
Process control charts are fairly simple-looking
connected-point charts. The points are plotted on an
x/y axis with the x-axis usually representing time.
The plotted points are usually averages of subgroups
or ranges of variation between subgroups, and they
can also be individual measurements.
Some additional horizontal lines representing the
average measurement and control limits are drawn
across the chart. Notes about the data points and any
limit violations can also be displayed on the chart.
Control charts have four key features:
1) Data points are either averages of subgroup measurements
or individual measurements plotted on the x/y axis and joined
by a line. Time is always on the x-axis.
2) The Average or Center Line is the average or mean of the
data points and is drawn across the middle section of the
graph, usually as a heavy or solid line.
3) The Upper Control Limit (UCL) is drawn above the
centerline and often annotated as "UCL". This is often called
the “+ 3 sigma” line.
4) The Lower Control Limit (LCL) is drawn below the
centerline and often annotated as "LCL". This is called the “- 3
sigma” line.
The x and y axes should be labeled and a title specified for the
chart
Introduction to Control Charts
A typical control chart has control limits set at values such
that if the process is in control, nearly all points will lie
within the upper control limit (UCL) and the lower control
limit (LCL)

Control chart
Statistical Basis of the Control Chart
Out-of-Control Situations
If at least one point plots beyond the control
limits, the process is out of control
If the points behave in a systematic or nonrandom
manner, then the process could be out of control.
Statistical Basis of the Control Chart
Relationship between hypothesis testing and
control charts
Control limits can be set at 3 standard deviations from the
mean.
This results in “3-Sigma Control Limits”
UCL = 74 + 3(0.0045) = 74.0135
CL= 74
LCL = 74 - 3(0.0045) = 73.9865
Choosing the control limits is equivalent to setting up the
critical region for testing hypothesis
H0:  = 75
H1:   75
Introduction to Control Charts

Process
Improvement
using the
control chart.
Steps Involved In Using Statistical
Process Control
Proper Statistical Process Control
starts with planning and data
collection. Statistical analysis on the
wrong or incorrect data is rubbish, the
analysis must be appropriate for the
data collected.
Be sure to PLAN, then
constantly re-evaluate your
situation to make sure the plan
is correct.
The key to any process
improvement program is the
PDSA cycle described by
Walter Shewart.
PDSA
 Plan: Identify the problem and the possible causes. The QC
tools can help organizations identify problems and possible
causes, and to prioritize
 corrective actions.
 Do: Make changes designed to correct or improve the situation.
 Study: Study the effect of these changes on the situation. This is
where control charts are used – they show the effects of changes
on a process over time. Evaluate the results and then replicate
the change or abandon it and try something different.
 Act: If the result is successful, standardize the changes and then
work on further improvements or the next prioritized problem.
If the outcome is not yet successful, look for other ways to
change the process or identify different causes for the problem.
Basic Principles

where
k = distance of the control limit from the
center line
w = mean of some sample statistic, W.
w = standard deviation of some statistic, W.
Basic Principles
Important uses of the control chart
1. Most processes do not operate in a state of statistical
control.
2. Consequently, the routine and attentive use of control
charts will identify assignable causes. If these causes can
be eliminated from the process, variability will be
reduced and the process will be improved.
3. The control chart only detects assignable causes.
Management, operator, and engineering action will be
necessary to eliminate the assignable causes.
4. Out-of-control action plans (OCAPs) are an important
aspect of successful control chart usage.
Basic Principles
 Types of control charts
 Variables Control Charts
 These charts are applied to data that follow a
continuous distribution.
 They are used to monitor characteristics that can be
measured, e.g. length, weight, diameter, time
 Attributes Control Charts
 These charts are applied to data that follow a
discrete distribution.
 They are used to monitor characteristics that have
discrete values and can be counted, e.g. % defective,
number of flaws in a shirt, number of broken eggs in
a box
Basic Principles
 Control Charts show sample data plotted on a graph with CL,
UCL, and LCL
Popularity of control charts
 1) Control charts are a proven technique for
improving productivity.
 2) Control charts are effective in defect prevention.
 3) Control charts prevent unnecessary process
adjustment.
 4) Control charts provide diagnostic information.
 5) Control charts provide information about process
capability.
Choice of Control Limits
General model of a control chart

UCL   W  L W
Center Line   W
LCL   W  L W
where L = distance of the control limit from the
center line
 W = mean of the sample statistic, w.
W = standard deviation of the statistic, w.

Introduction to Statistical Quality Control, 4th

Edition
Choice of Control Limits
“99.7% of the Data”
 If approximately 99.7% of the data lies within 3 of
the mean (i.e., 99.7% of the data should lie within the
control limits), then 1 - 0.997 = 0.003 or 0.3% of the data
can fall outside 3 (or 0.3% of the data lies outside the
control limits). (Actually, we should use the more exact
value 0.0027)
 0.0027 is the probability of a Type I error or a false alarm
in this situation.
Choice of Control Limits
Three-Sigma Limits
 The use of 3-sigma limits generally gives good
results in practice.
 If the distribution of the quality characteristic
is reasonably well approximated by the
normal distribution, then the use of 3-sigma
limits is applicable.
 These limits are often referred to as action
limits.
Design of a Control Chart
 Suppose we have a process that we assume the
true process mean is  = 74 and the process
standard deviation is  = 0.01. Samples of size 5
are taken giving a standard deviation of the
sample average, is

 0.01
x    0.0045
n 5
Design of a Control Chart
 Control limits can be set at 3 standard deviations
from the mean in both directions.
 “3-Sigma Control Limits”
 UCL =  + 3 xbar
 = 74 + 3(0.0045) = 74.0135
 CL = 74
 LCL =  - 3 xbar
 = 74 - 3(0.0045) = 73.9865
Design of a Control Chart

X-bar control chart for piston ring diameter.

Design of a Control Chart
 Choosing the control limits is equivalent to setting up the
critical region for hypothesis testing
 H0:  = 74
 H1:   74
An important factor in preparing for SPC charting is
determining if you will measure every product of the
process, such as measuring every part, or if you will use
subgroups. Subgroups are a sample of data from the total
possible data. Subgroups are used when it is impractical or
too expensive to collect data on every single product or
service in the process. Decisions to use subgroups or not
needs to be carefully thought out to ensure they accurately
represent the data.
Rational Sub grouping
Subgroups or samples should be selected so that if assignable
causes are present, the chance for differences between
subgroups will be maximized, while the chance for differences
due to these assignable causes within a subgroup will be
minimized.
Subgroups need to be homogenous within themselves so
that special causes can be recognized, so problem areas
stand out from the normal variation in the subgroup.
For example, if you are in charge of analyzing processes
in a number of facilities, a separate group should
represent each facility, since each facility has different
processes for doing the same tasks. Each facility
subgroup should probably be broken down even further,
for example by work shifts.
Design of a Control Chart
 Constructing Rational Subgroups
 Select consecutive units of production.
 Provides a “snapshot” of the process.
 Good at detecting process shifts.
 Select a random sample over the entire sampling
interval.
 Good at detecting if a mean has shifted
 out-of-control and then back in-control.
Analysis of Patterns on Control Charts
Nonrandom patterns can indicate out-of-control
conditions
 Patterns such as cycles, trends, are often of
considerable diagnostic value
 Look for “runs” - this is a sequence of observations of
the same type (all above the center line, or all below
the center line)
 Runs of say 8 observations or more could indicate an
out-of-control situation.
 Run up: a series of observations are increasing
 Run down: a series of observations are decreasing
Analysis of Patterns on Control Charts
An X-bar chart with a cyclic pattern.
Analysis of Patterns on Control Charts
(a) Variability with the cyclic pattern.
(b) Variability with the cyclic pattern eliminated.
Control Charts Performance
Process mean shift of 2
Distribution of Data
Skewed
Normal distribution
distributions

Statistical Quality Control

Analysis of Patterns on Control Charts
 Western Electric Handbook Rules
 A process is considered out of control if any of the
following occur:
 1) One point plots outside the 3-sigma control limits.
 2) Two out of three consecutive points plot beyond
the 2-sigma warning limits.
 3) Four out of five consecutive points plot at a
distance of 1-sigma or beyond from the center line.
 4) Eight consecutive points plot on one side of the
center line.
Analysis of Patterns on Control Charts
The Western Electric zone rules.
Analysis of Patterns on Control Charts
When do you have a problem with your process?
One or more points outside of the control limits
A run of at least seven points (up, down or above or
below center line)
Two or three consecutive points outside the 2-sigma
warning limits, but still inside the control limits
Four or five consecutive points beyond the 1-sigma
limits
An unusual or nonrandom pattern in the data
(From Douglas C. Montgomery “Introduction to Statistical Quality Control”)
 Setting Control Limits
Percentage of values Control limits balance
under normal curve risks like Type I error

Statistics tell us that in normal data dispersion, we can expect the following
percentages of data to be included within the sigma limits:
1 sigma – 68.3%
2 sigma – 95.5%
3 sigma – 99.7%
Hypothesis Tests
Results of hypothesis tests fall into one of four
scenarios:
H0 is true H0 is false

We reject H0 Type I Error OK

We do not OK Type II Error
reject H0

Type I Error- ART (Alpha, Reject Ho when true)

Type II Error- BAF (Beta, Accept Ho when false)
Specification Limits, Process Limits
and Revised Limits
Over a period of time, knowingly or unknowingly a
sort of understanding between the consumers and the
producers about the quality of the product.
For example if the stipulated length a screw is 1” and
if the variation of 0.05” in the length of the screw is
acceptable to consumers, consumer will accept the
screw of the length between 0.95” and 1.05”. These
limits are known as “Specification Limits”
Here 0.95” is the lower specification limit and
1.05” is the upper specification limit
Process Limits
If the expected level of quality is not known, the value
of the characteristic W are found out from samples
and their average is obtained. From the average value
of W and its standard deviation, the control limits are
derived. These limits are known as “Process Limits”.
Centre Line = Average value of W
Lower Control Limit = Average value of W – 3 S.D.
Upper Control Limit = Average value of W + 3 S.D.
Thus the process limits are unknown the control
charts are drawn using these Process Limits.
Revised Limits
If the process is not under statistical control one or more
points fall outside of the Process Control Limits.
If the quality characteristics of these samples are removed and
the control limits are found on the basis of remaining samples,
the control charts can be prepared with these new limits.
Even in this chart if one or more points fall outside the new
limits, the above process is repeated. In this way control limits
are revised until all points fall within the limits.
When all points fall within the limits, these limits are taken as
final limits and are used for the control of the future
production.
The limits obtained this way are known as Revised Control
Limits.
Uses of Control Charts
1. To define the standards of the process
2. To attain such standards
3. To maintain the standards achieved by the production
process, and use them for future production
4. To detect assignable causes of variation and to remove them
5. The efficiency of the workers improves as they know that
their work is being checked
6. Quality is maintained and goodwill is established
7. A permanent record is created, which will be useful
8. Machine efficiency can be determined , for advance
planning replacement
9. The cost of production and the cost of inspection can be
reduced
Control Charts for Variables
Use x-bar and R charts
together
Used to monitor different
variables
X-bar & R Charts reveal
different problems
In statistical control on
one chart, out of control
on the other chart? OK?
Control Charts for Variables
Use x-bar charts to monitor the changes in the mean of a
process (central tendencies)
Use R charts to monitor the dispersion or variability of the
process
System can show acceptable central tendencies but
unacceptable variability or
System can show acceptable variability but unacceptable
central tendencies
X and R-Charts (mean and range charts) together are
commonly used in dealing with variable data to monitor
the quality of a manufacturing process.
The reason that both the charts have to be used together is
that both the mean and the variation (spread) have to be
under control.
X-bar and R or S Control Charts
3-sigma control limits:

The grand mean (  ):

X-bar and R or S Control Charts
For R chart: Sample size < 10, and must be constant

The average range:

An unbiased estimator of :
X-bar and R or S Control Charts
x Control Chart (from R ):

R Chart:
Factors for Control Limits
n A2 D4 D3
2 1.880 3.268 0.0
n=
3 1.023 2.574 0.0 sample size
4 0.729 2.282 0.0
5 0.577 2.114 0.0
6 0.483 2.004 0.0
7 0.42 1.98 0.08
8 0.37 1.86 0.14
9 0.34 1.82 0.18
10 0.31 1.78 0.22
X-bar and R or S Control Charts
For S chart: Sample size >10, and need not be constant
3-sigma control limits for S:

An unbiased estimator of :
X-bar and R or S Control Charts
S Chart:

x Control Chart (from S ):

Construction of X-bar and R Charts
The central line and both the control limits are drawn
parallel to X-axis.
For X-bar chart the subgroup numbers are taken on X-
axis and the values of X-bar are taken on Y-axis. The
value of X-bar for each sample is plotted on the graph
paper. On joining these points in order X-bar chart is
obtained.
For R chart the values of R are taken on Y-axis. The
values of R obtained from each sample are plotted on
graph paper. On joining them in order R chart is
obtained.
Example -1
Example -1
Example -1

X and R control
charts for vane
opening.
Example -1
Example -1
Example -1
The S control chart for vane opening
 Example -1
*

The X-bar
and R
control charts
for vane
opening.
 Example -1: Revised Control Charts

The X-bar
and R
control charts
for vane
opening.
Example- 2
The St. Patrick's Hospital is starting a quality
improvement project on the time to admit a patient
using X and R Charts. Determine the limits for the X
and R charts and check to see if there are any out-of-
control points.
Assess the statistical stability of the Process
Example- 2 Assess the statistical stability of the Process
Subgroup Subgroup
Number X1 X2 X3 Xbar R Number X1 X2 X3 Xbar R
1 6.0 5.8 6.1 13 6.1 6.9 7.4
2 5.2 6.4 6.9 14 6.2 5.2 6.8
3 5.5 5.8 5.2 15 4.9 6.6 6.6
4 5.0 5.7 6.5 16 7.0 6.4 6.1
5 6.7 6.5 5.5 17 5.4 6.5 6.7
6 5.8 5.2 5.0 18 6.6 7.0 6.8
7 5.6 5.1 5.2 19 4.7 6.2 7.1
8 6.0 5.8 6.0 20 6.7 5.4 6.7
9 5.5 4.9 5.7 21 6.8 6.5 5.2
10 4.3 6.4 6.3 22 5.9 6.4 6.0
11 6.2 6.9 5.0 23 6.7 6.3 4.6
12 6.7 7.1 6.2 24 7.4 6.8 6.3
Control Charts for Individual
Measurements
 What if you could not get a sample size greater than 1 (n
=1)? Examples include
 Automated inspection and measurement technology is
used, and every unit manufactured is analyzed.
 The production rate is very slow, and it is inconvenient
to allow samples sizes of N > 1 to accumulate before
analysis
 Repeat measurements on the process differ only
because of laboratory or analysis error, as in many
chemical processes.
 The individual control charts are useful for samples of
sizes n = 1.
Moving Range Chart – MR Chart
This type of chart displays the moving range of
successive observations. A moving range chart can be
used when it is impossible or impractical to collect
more than a single data point for each subgroup.
This chart can be paired with an individual chart,
which is then called an Individual Moving Range (IR)
chart. An individual chart is used to highlight the
changes in a variable from a central value, the mean.
 The moving range chart displays variability among
measurements based on the difference between one
data point and the next.
Control Charts for Individual
Measurements
 The individual control charts are useful for
samples of sizes n = 1.
 The moving range (MR) is defined as the
absolute difference between two successive
observations:
MRi = |xi - xi-1|
 which will indicate possible shifts or changes in
the process from one observation to the next.
Individual And Moving Range
Charts – IR Charts
This pair of variable control charts is often offered
together for quality control analysis.
The Individual chart displays changes to the
process output over time in relation to the center
line which represents the mean.
The Moving Range chart analyzes the variation
between consecutive observations, which is a
measure of process variability.
Control Charts for Individual
Measurements
Interpretation of the Charts
 X Charts can be interpreted similar to X-bar charts. MR
charts cannot be interpreted the same as X-bar or R
charts.
 Since the MR chart plots data that are “correlated” with
one another, then looking for patterns on the chart does
not make sense.
 MR chart cannot really supply useful information about
process variability.
 More emphasis should be placed on interpretation of the
X chart.
Charts Used with Attribute Data
P-Chart, also known as the fraction or percent defective
chart, is commonly used in dealing with attribute data to
monitor the quality of a manufacturing process.
The mean proportion defective ( p bar ) is the central line.
The upper and lower control limits are constructed as
follows:
p bar= Total Number of Defectives / Total Number
Inspected
The standard deviation ( ) of p bar:

where n = sample size.

 Attribute Control Charts
p and np Chart (Control Chart for Proportions)
p Chart ( It follows Binomial Distribution)

If instead of p-charts np-chart is to be drawn, it is not necessary to find fraction

defective p for each sample. Number of defectives d = np is obtained from each
sample. Average number of defectives np-bar is found out. It follows Binomial
Distribution.
np-bar =
The control limits of np chart are
C.L. = np-bar
L.C.L. = np-bar -3 U.C.L. = np-bar +3
81
Attribute Control Charts Example -

82
Attribute Control Charts Example
P chart for a ceramic substrate

83
Attribute Control Charts Example

84
Example - A computer manufacturer collects data from the final test of its
product starting from the end of January and all through February. Each day a
sample of 2000 items are inspected and the number of items in the sample that
do not conform to specifications is recorded. The data is shown below:
Subgroup Number Number Fraction Subgroup Number Number Fraction
Number Inspected Defective Defective Number Inspected Defective Defective
(day) (day)
1 2000 55 13 2000 47
2 2000 18 14 2000 31
3 2000 50 15 2000 38
4 2000 42 16 2000 28
5 2000 39 17 2000 30
6 2000 52 18 2000 113
7 2000 47 19 2000 58
8 2000 34 20 2000 34
9 2000 29 21 2000 19
10 2000 53 22 2000 30
11 2000 45 23 2000 17
12 2000 26 24 2000 46
Attribute Control Charts
U Chart (Control Chart for Defects per Unit)
Sample size need not be constant

86
Attribute Control Charts
Example

87
Attribute Control Charts
Example

88
Attribute Control Charts
Example

U chart of defects per unit on printed circuit boards.

89
The c-Chart
(Control Chart for No. of Defects)
In many quality control problems the particular items being
subjected to inspection may have more than one defect.
We may wish to count no. of defects instead of merely
classifying at item as to whether or not it is defective.
If Ci denotes the no. of defects observed in the ith inspected
item, we can safely assume that Ci has a Poisson distribution.
Let this Poisson distribution have a mean of λ for a process in
control.
For us to be 99.74% sure, almost all the Ci’s should fall within
three standard deviations of the mean if the process is in
control.
The c-Chart
(Sample size must be constant)
Since E  Ci    and Var  Ci    , the control limits are :
  3 .

If k items are inspected, then the unbiased estimator of  is :

1 k
C   Ci .
k i 1

The estimated control limits then become :

C 3 C .