Statistical Quality Control

By: Nadeem Irfan Bukhari / Nasir Abbas

 Basic Concepts
• Statistics – Planning, collecting, analyzing data for
reporting certain conclusion regarding data
• Descriptive Statistics – a presentation of facts, a
survey or summary of population with data
• Inferential Statistics – drawing conclusion about
whole population form a randomly selected small
sample.
• Random samples from a population truly represents
the population because every unit in a population
has an equal chance of being selected
 Basic Concepts
• Variable - a single quality characteristic that
can be measured on a numerical scale.
• Variables are characterized with Mean value
of characteristic and Variability (standard
deviation, range, etc) of characteristic.
 Basic Concepts
• Four parameters of descriptive statistics are
important:
– Mean or average (μ), measures central tendency
and is estimated by sample mean or xbar.
– Standard deviation (σ) measures spread of data
– Range, difference between largest/smallest
observations in a set of data
– Distribution of Data shape
• Normal or bell shaped – normally distributed
with , and  or Skewed

 x 
n n

 xi
2
i X
x i 1
σ i 1 Ri = xmax - xmin
n n 1
 Basic Concepts
• Distribution or shape of data is important to know for
analytical purposes. It must be bell shaped or normal
distribution.
Normal distribution Skewed distribution
 Basic Concepts
• Variation exists in all processes and may be:
– Common or Random variation: where causes
cannot be identified and are unavoidable, e.g.
differences in process variables, i.e., tablet
diameter, weight, etc.
– Common variation are also called as variation built
into process
– SD and Accuracy
 Basic Concepts
– Cause-Assignable variation: where causes of
variation can be identified and eliminated, e.g.
poor training, worn tools, machine needing repair
– Cause-assignable variation are also called as
special or non-routine events poorly
trained defective
raw
workers materials

poorly
calibrate worn or
d loose
machine parts
s

excessive tired or
machine indifferent
vibration workers
 Basic Concept
• Two aspects of quality
– Features: More attributes meet desirability = high
quality
– Freedom from defect = fewer defects = higher
quality
 Statistical Quality control
• SQC is set of statistical tools used in assessing quality
characteristics or parameters of a process, thus also
called as Statistical Process Control (SPQ).
• SQC or SPQ tool is a Control chart used in monitoring
variation in the characteristics of a product/process.
• Control chart is a graph, plotted in time order used to
study how a process changes (variates) over time.
 When to use control charts
• When controlling ongoing processes by finding and correcting
problems as they occur
• When predicting the expected range of outcomes from a
process
• When determining whether a process is stable (in statistical
control)
• When analyzing patterns of process variation from special
causes (non-routine events) or common causes (built into the
process)
• When determining whether your quality improvement project
should aim to prevent specific problems or to make
fundamental changes to the process
 Control Charts
• Control Chart always shows sample data on a graph
with central control limit (CL), upper control limit
(UCL), and lower control limit (LCL)
• Limits as lines on graph are determined from
historical data.
X
80
60 UCL
40 CL
20 LCL
0
1 2 3 4 5 6 7 8 9 10 11 12
Time
 Control Charts
• Upper and lower control limits represent 3 standard
deviations above and 3 standard deviations below
mean line, respectfully.

UCL = Process Average + 3 SD

Process Average + 3

- 3
LCL = Process Average – 3 SD

• Why 3 sigma is taken? [Next Slide]

 Control Charts
• Reasons for taking 3 
– Probability that a random value of measurement
falls in-between 3- limits is 0.9973, which is very
high.
– On other hand, probability that a random value of
sample falls outside of 3- limits is 0.0027, which
is very low.
 Control Charts
• Reasons for taking 3 
– Thus, when values fall in-between 3- limits,
variations are attributed due to chance variation,
then process is considered to be statistically
controlled.
– But, when one or many values fall out of 3-
limits, variations are attributed to assignable
variation and process is said to be not under
statistical control.
 Control Charts
• By comparing current data to these lines,
Conclusions can be drawn about whether process
variation is consistent (in control) or is unpredictable
(out of control, affected by special causes of
variation).
What do you understand ?
 Control Charts - Types
Control Charts

Shewhart Process
Control Charts Acceptance
Control Charts

Variables Attributes

x R p c
Chart Chart Chart Chart
 Control Charts - Shewhart
• Control chart for variables are used to monitor a
measurable continuous characteristics, e.g. weight,
diameter, etc,
• Control charts for variable are:
– X-bar Chart
– R-Chart
• Both, x-bar chart and R-chart are used as pair
This type of chart graphs the means (or averages) of a set of samples, plotted in
order to monitor the mean of a variable, for example the length of steel rods, the
weight of bags of compound, the intensity of laser beams, etc.. In constructing this
chart, samples of process outputs are taken at regular intervals, the means of each
set of samples are calculated and graphed onto the X bar control chart. This chart
can then be utilized to determine the actual process mean, versus a nominal process
mean and will demonstrate if the mean output of the process is changing over
time.Mean of the process changing over time
This type of chart demonstrates the variability within a process. It is suited to processes
where the sample sizes are relatively small, for example <10. Sets of sample data are
recorded from a process for the particular quality characteristic being monitored. For
each set of date the difference between the smallest and largest readings are recorded.
This is the range “R” of the set of data. The ranges are now recorded onto a control chart.
The center line is the averages of all the ranges.
 Control Charts - Types
• Control charts for attributes are used to monitor
feature that have discrete values and can be counted,
e.g. % defective, number of flaws in a syrups, number
of chipped tablets
• Control charts for attributes are:
– p-Chart and
– c-Chart
 Control Charts for Variables
• x-bar and Range charts are SQC tools used to
monitor variables data when samples are
collected at regular intervals from a process.
• Assume process is normally distributed, and
fraction nonconforming can be found by: P(x <
LCL) + P(x > UCL), i.e., beyond limits.
 Control Charts for Variables
• x-bar chart detects changes in mean value of a quality
characteristics across all units or a process
• x-bar monitor central tendency.
• R-chart detects dispersion or variability, changes in range
(standard deviation of quality characteristics across all
units of a process
• R-charts monitor the dispersion
• In general, xbar chart is insensitive to small departures
from normality.
• R chart is more sensitive to non-normality than the xbar
chart
 Control Charts for Variables
• Process can show acceptable central tendencies on x-
bar chart but unacceptable variability on R-chart OR
acceptable variability on R-chart but unacceptable
central tendencies on x-bar
– Thus, R-chart is interpreted first:
– If R-chart is in control, interpret the X-bar chart (i) if
in control, process is in control; (ii) if out of control:
process average is out of control
– If R-chart is out of control, process variation is out of
control then investigate cause; no need to interpret
the X-bar chart
 Control Charts for Variables
• X-bar & R-bar Charts
reveal different
problems

• Shift in mean in x-bar

chart but not be R-chart

• Shift in dispersion in R-
chart but not be x-bar
chart
 Control Charts for Attributes
• P-Charts are used for quality characteristics that are
discrete and involve yes/no or good/bad decisions
– Number of leaking tubes in a lot of 48
• C-Charts are used for discrete defects when there
can be more than one defect per unit
– Number of flaws on a tablets
 Control Charts for Attributes
• p-charts – a statistical control chart that plots
proportion for defective sample (p) over time
– Involves taking of sample and each item is
inspected for determining sample proportion
defective by dividing the number of defective
items by sample size
P charts are utilized where there is a pass / fail determination on a unit inspected. The p
chart will show if the proportion defective within a process changes over the sampling
period (the p indicates the portion of successes). In the p chart the sample size can vary
over time. A similar chart to the p chart is the np chart. However, with the np chart the
sample size needs to stay constant over the sampling period. An advantage of the np
chart is that the number non-conforming is recorded onto the control rather than the
fraction non conforming. Some process operators are more comfortable plotting the
number rather than the fraction of non-conformances.
 Control Charts for Attributes
– Plot the sample proportion defective on the
control chart and compare with UCL and LCL to
determine if process is out of control
– Mean = p, Sd = Sigma p = SQR (p(1-p)/n) where p =
proportion defective item and n is the sample size

p 1  p  p 1  p 
CL= p UCL= p  3 n
LCL= p  3 n
 Control Charts for Attributes
• c-chart for number (count) of defects per unit plots
movement in the number of defects per unit
• Randomly selected one item and number of defects
are counted in that item
• Number of defects are plotted on control chart which
are compared with UCL and LCL to determine if
process is out of control
• Sd=SQR of mean
• UCL= C + 3.SQR of mean
• LCL=C – 3. SQR of mean
The c chart is used where there can be a number
of defects per sample unit and the number of
samples per sampling period remains constant.

In the u chart, again similar to the c chart, the

number of defects per sample unit can be
recorded, however, with the u chart, the
number of samples per sampling period may
vary.
 Control Charts – Process Acceptance
Control Charts
• Charts that give information about capability of
process relative to its specification limits (not mean).
• These charts estimates fraction of nonconforming
items deviating specification limit for any process
• In these charts, CL is Specification limit (SL), LCL is LSL
and USL
• Specification for assay of Active ingredient is 95-105
% and an ideal target specification is 100%
• In the real world, results vary ...90% , 98.5%, 104% ,
110 % etc and mean may be different from the target
specification, i.e., 108% vs 100%.
 Control Charts – Process Acceptance
Control Charts

USL

SL

LSL
Control Charts – Process Acceptance Control
Charts vs Shewart Chart

USL

SL

LSL Mean based CL

 Difference between 2 types of Charts
Shewhart Control Charts Acceptance Control charts
Central Line is practical average/mean Central line is the target specification
deduced from data , 101% 100 %

Indicates how process is behaving from Indicates how process is behaving from
practical calculated mean ? theoretical set target specification?

A process consistency – may not be within Tells that process consistency with respect
regulatory spec, for example mean is 108 to the Regulatory Registered Specification.
% , SD 1.2, YES, the process is consistent,
but it is not meeting SPECIFICATION

Tells, once specification is set, how the

Used during development setting of process is behaving with respect to the
specifications set specification.

Parameters include UCL, CL, LCL Parameters include USL, SL and LSL and
the Process Capability ratio (Cp) and
Process Capability Index (CPk)
 Control Charts – Process
• 1. Setting limits and Zones
• 2. Process procedure
• 3. Pattern Analysis
• 4. Interpretation
 Control Charts – Setting limits
• Control limits are functions of natural variability of
process
• Natural tolerance limits represent natural variability
of process (usually set at 3-sigma from mean)
• Specification limits are determined by regulatory
bodies, e.g., FDA.
 Control Charts – Setting limits
• Specification width
and process
variability

Depends on
Nature of a
process or
narrowest control limits industry
norms

middle control limits

widest control limits

=
-3 z -2 z -1 z x +1 z +2 z +3 z
THE NUMBER OF NORMAL STANDARD DEVIATIONS ( Z )
Control Chart – Assigning different Zones

• For interpretation of pattern on charts, data is

represented over 6 standard deviations, 3 standard
deviations from mean line to UCL and 3 from mean
to the LCL.
• To help analyze charts, the chart area is divided into
6 sections A, B, and C representing the standard
deviations.
• On next slide there is an example of R-chart with
assigned zones
Control Chart – Assigning different Zones
UCL 3 sigma
Zone A
2 sigma

Zone B
1 sigma
Zone C
Process =
x
average
Zone C
1 sigma

Zone B
2 sigma
Zone A
LCL 3 sigma
| | | | | | | | | | | | |
1 2 3 4 5 6 7 8 9 10 11 12 13
Sample number
Control Chart – Process Procedure
• Random sampling of output of a process
• Inspecting output from a process for measuring the
quality characteristics
• Charting as graph with set CL, UCL and LCL
• Helpful in identifying in-process variation
• Acceptance sampling used to randomly inspect a
batch of products to determine acceptance/rejection
 Control Chart – Patterns of points
• Position of points on control chart makes patterns of
data distribution
• Pattern of point distribution with reference to UCL
and LCL and 6 zones on the graphs are observed
• An unusual or nonrandom pattern in the data:
– Number of points outside of control limits
– Cyclic Patterns
– Shift in process level (all above/below CL)
– Trend (increasing/decreasing)
– Stratification (3 up and 3 down)
 Control Chart – Patterns of points
• Patterns of points provide useful diagnostic
information on process which is used to make
process modifications that reduce variability.
• Pattern tell when there is a problem in process
• Process is out of control if one or a combination of
following is observed in pattern of points distribution
[Next Slides]:
 Control Chart – Interpretation
UCL
• Sample observation (points)
consistently below CL LCL

UCL
• Points consistently above CL
LCL

• Points consistently increasing UCL

LCL
• Points consistently decreasing
UCL
LCL
 Control Chart – Interpretation
• A cyclic trend.

UCL

CL

LCL
 Control Chart – Interpretation
• One point outside of the 3 sigma control limits
(beyond zone A)
 Control Chart – Interpretation
• Two out of any three successive points fall in zone A
of the same side
 Control Chart – Interpretation
• Four out of any five successive points fall in or
beyond zone B of the same side
 Control Chart – Interpretation
• Eight successive points fall in or beyond zone C
of the same side
“The Process ”behaviour
 Control Chart - Interpretation
• The process is considered in control for n samples
collected if all points are inside control limits and no
systematic behavior is identified
• Above indicates that trial control limits are suitable
for controlling current or future production.
 Control Chart - Interpretation
• If points plot out of control, then control limits must
be revised.
• Before revising, identify out of control points and
look for assignable causes.
– If assignable causes can be found, then recalculate
control limits.
– If no assignable causes can be found then consider
control limits as appropriate for current control.
 Advantages (Applications)
• Applied where a process has tendency to go out of
control which may be harmful and costly
• May be used to study past performance and or to
evaluate present conditions.
• Helps controlling ongoing processes by finding (and
correcting) problems as they occur.
• Data collected from control chart may form basis for
process improvement.
• Helps determine what process adjustments need to be
made
 Advantages (Applications)
• Help predicting expected range of outcomes in a
process.
• Help determining whether a process is stable or in
control.
• Help making defects visible.
• Help analyzing patterns of process variation from
special causes or common causes.
• Target quality improvement aim to prevent specific
problems or to make fundamental changes to
process.
 Advantages (Applications)
• Tool to:
– Improve process
– Improve productivity
– Make defects visible
– Zero defects
 Process Capability – an Application of
Process Acceptance Control Chart
• Acceptance charts help in determining
capability/feasibility of a process
• Two parameters of process capability are:
– Process Capability Ratio (Cp)
– Process Capability Index (Cpk)
 Process Capability – an Application of
Process Acceptance Control Chart
• Process-Capability Ratio (Cp) is used to express
process capability.
• A way to deal with quality of product by devising
quantifiable measurements of quality element
• Involves assessing process variability relative to
preset product specifications
• Cp assumes that the process is centered in the
specification range
USL  LSL
Cp 
6
 Process Capability – an Application of
Process Acceptance Control Chart
• Cp > 1 indicates that process exceeds minimal
specifications (a low number of nonconforming items
are being produced).
• Cp = 1 reflects process variability just meets the
specifications (about 0.27% nonconforming units are
being produced).
• Cp < 1 reveals that process is not capable of
producing within specifications (a large number of
nonconforming items are being produced.
 Process Capability – an Application of
Process Acceptance Control Chart
• Process Capability Index (Cpk) helps to address a
possible lack of centering of the process
 USL  μ μ  LSL 
Cpk  min  , 
 3σ 3σ 
– min is a function which outputs the minimum value of the
given value from above 2 resultant values.

• Cpk is less than 1, revealing that the process is not

capable
 Process Capability – an Application of
Process Acceptance Control Chart
• Cpk Values – Interpretation and Actions
Cpk Interpretation Actions

Less than 0.6 NOT CAPABALE STOP / COMPLETE RE

VALIDATION REQUIRED

 Validation of test method

Greater than 0.6 and less POORLY CAPABLE  Process mapping /Process
than 1 validation
 Warning limit review (early
warning of problem)
 Identify Critical control
points
 Scope for Regulatory
Implications
 Process Capability – an Application of
Process Acceptance Control Chart

Cpk Interpretation Actions

Greater than or equal to Marginally Capable Look into the
1 but less than 1.33 Likely to pass the CONTINIOUS
specifications routinely PROCESS
IMPROVEMENTS

Greater than 1.33 Statistically very capable. No action Required

Greater than 1.5 but less Process is much better Risk controlled – Reduced
than 3 than needed. testing.
 Process Capability – an Application of
Control Chart
Variation and Design
Specifications
Process capability
(a) Natural variation
exceeds design
specifications; process
is not capable of
meeting specifications
all the time.
Process
Design
Specifications

(b) Design specifications

and natural variation the
same; process is capable
of meeting specifications
most of the time.

Process
 Process Capability – an Application of
Control Chart
Variation and Design
Specifications
Process capability
(c) Design specifications
greater than natural
variation; process is
capable of always
conforming to
specifications.
Process
Design
Specifications

(d) Specifications greater

than natural variation, but
process off center;
capable but some output
will not meet upper
specification.
Process
 Process Capability – an Application of
Control Chart
• Process capability with
minimum variability improves Nominal value
process capability and Six sigma
minimum defective products
• This has led a concept of 6 Four sigma
sigma
Two sigma
• 6 sigma means 3.4 defects per
million and Lower Upper
specification specification
• Cpk = 2

Mean
 Process Capability – an Application of
Control Chart
• ±6 Sigma versus ± 3
Sigma PPM Defective for ±3σ VS ±6σ quality
• Motorola coined “six-
sigma” to describe their
higher quality efforts
back in 1980’s
• Six-sigma quality
standard is now a
benchmark in many
industries including
Pharma Industry
 Process Capability – an Application of
Control Chart
• Six sigma refers to variation that exists within  six
standard deviations of the process outputs
• Six Sigma is a comprehensive philosophy and set of
methods used to achieve, sustain and maximize
profit by eliminating defective products and
processes and variability in process
 Process Capability – an Application of
Control Chart
• It relies heavily on the principles and tools of total
quality management (TQM), quality by design (QBD),
and process analytical technique (PAT).
• Seeks to reduce variation in the processes that lead
to product defects
• In 6 sigma, more stringent specifications are
followed, i.e., the tolerance limit is very narrow
Process Capability – an Application of
Control Chart
 Reference
• Douglas C. Montgomery “Introduction to
Statistical Quality Control”