Process Capability

Analysis Overview
 Statistical Process Control (S.P.C.)
• This is a control system which uses statistical techniques
for knowing, all the time, changes in the process.
• It is an effective method in preventing defects and helps
continuous quality improvement.

What does S.P.C. mean?

• Statistical:
• Statistics are tools used to make predictions on
performance.
• There are a number of simple methods for analysing
data and, if applied correctly, can lead to
predictions with a high degree of accuracy.
 What does S.P.C. mean?
• Process:
• The process involves people, machines, materials,
methods, management and environment working
together to produce an output, such as an end product.

What does S.P.C. mean?

• Control:
• Controlling a process is guiding it and comparing
actual performance against a target.
• Then identifying when and what corrective action is
necessary to achieve the target.
 S.P.C.
• Statistics aid in making decisions about a process based
on sample data and the results predict the process as a
whole.
• S.P.C. is statistical analysis of the predictability and
capacity of a process to give a uniform product.
The Benefits of S.P.C.
• Assesses the design intent.
• Achieves a lower cost by providing an early warning system.
• Monitors performance, preventing defects.
• Provides a common language for discussing process
performance.
 Process Variations
• No industrial process or machine is able to produce
consecutive items which are identical in appearance,
length, weight, thickness etc.
• The differences may be large or very small, but they are
always there.
• The differences are known as ‘variation’. This is the
reason why ‘tolerances’ are used.
 Statistical Process Control
• Statistical process control (SPC) is the application
of statistical methods to identify and control the
special cause of variation in a process.
− Common cause variation – random fluctuation
of response caused by unknown factors
− Special cause variation – non-random variation
caused by a specific factor
Upper Specification Limit
Upper Control Limit
3σ
Target
Lower Control Limit
Lower Specification Limit
Special cause variation?
Process Capability vs. Process Control
• Evaluating Process Performance
− Ability of process to produce parts that conform to
engineering specifications (CONFORMANCE)
− Ability of process to maintain a state of statistical
control; i.e., be within control limits (CONTROL)
• Process must be in statistical control before
assessing process capability. Why?
• Statistical aspects
− Process Control…use summary statistics from a
sample (subgroup); dealing with sampling
distributions, e.g.,
Xand R
(LCL, UCL)
− Process Capability…dealing with individual
measurements, e.g., X
(LSL, USL)
Numerical Measures

• Measurement of central tendency

− Mean
− Median
− mode
• Measurement of variations
− Range
− Variance
− Standard deviation
Mean

• The mean is the average value

Σx i
Sample mean x=
n
Σx i
Population mean µ=
N
Median

• The middle value when the numbers are

arranged in ascending or descending
order.
1
2
Ex: 3 median
4
5
Mode

• The data value that occurs with greatest

frequency

1
Ex: 1 mode
1
3
5
Central Tendency Example
• Test scores:

98 83 Σx i (98 + 96 + 94 + ...+ 64 + 44)

µ= = = 82
96 83 N 16
94 83
93 80 (86 + 83)
Median = = 84.5
88 75 2
87 72
86 64 Mode = 83
86 44
Range

• Difference between the largest and

smallest value in the dataset

1
2
Ex: 3 Range = 5-1 =4
4
5
Standard Deviation

• The positive square root of the variance

2
Sample standard deviation s= s

2
Population standard deviation σ= σ
Normal Distribution

34.13% 34.13%

13.06% 13.06%
2.14% 2.14%
0.13% 0.13%

-3σ −2σ −1σ µ 1σ 2σ 3σ

68.26%
95.46%
99.73%

68.26% of the population is within +/- 1?? of the ?

Process Variation

Mean
Standard Deviation

−1σ +1σ
Rejects

LL UL
−3σ +3σ
3σ Capability

A frequency plot shows how a

process varies from the mean
Decrease Variation

Decrease Standard Deviation

Reduce rejects

LL UL
−3σ +3σ
3σ Capability
Mean Shifts Happen

Mean Shift

Mean variation is time-dependent

due to tool wear, temperature, etc.
 Measures of Process Capability

LL UL
Process Capability Index:
±3 σ
UL − LL
CP =
6σ
σ

k
Cp adjusted for k: LL UL

Cpk= Cp (1-k)

± 6σ
 Data Interpretation

• As a set of numbers it is difficult to see any pattern.

• Within the table, numbers 30 and 37 were outside the
tolerance – but were they easy to spot?
• A way of obtaining a pattern is to group the
measurements according to size.
Data Interpretation – Tally Chart
36.39

36.38
• The tally chart groups the
measurements together by
36.37
size as shown.
36.36

36.35
• The two parts that were
out of tolerance are now
36.34
easier to detect (36.38mm).
36.33

36.32

36.31

36.30

36.29

36.28
 Tally Chart - Frequency
36.39

36.38
• The tally chart shows
2 patterns and we can obtain
36.37
6 the RANGE - 36.32mm to
36.36
7 36.38mm.
36.35
16
36.34
12
36.33 • The most FREQUENTLY
5 OCCURRING size is
36.32
2 36.35mm.
36.31

36.30

36.29

36.28
 Tally Chart - Information
• The tally chart gives us further information:
• The number of bores at each size;
• The number of bores at the most common size;
• The number of bores above and below the most common
size (36.35mm) -
• number above 36.35mm is 7+6+2=15
• number below 36.35mm is 12+5+2=19
 Histogram
We can redraw the frequency chart as a bar chart known
as a histogram:

16

14

12

10

0 36.31 36.32 36.33 36.34 36.35 36.36 36.37 36.38 36.39

 Smoothed Frequency
If we now draw a smooth curve through the top of each
box we get a bell-shaped pattern:

16

14

12

10

0 36.31 36.32 36.33 36.34 36.35 36.36 36.37 36.38 36.39

 Distribution
• The bell-shaped pattern is fairly typical of most
industrial processes.
• There is a central value at the highest point of the
curve and the pattern of results spread out equally on
both sides of the central value.
• The further we move from the central value, the fewer
values we will find.
• This bell-shaped pattern is known as the ‘NORMAL
DISTRIBUTION’.
• The bell-shaped pattern known as ‘Normal
Distribution’ is one that we would expect to see where
the process is running in a stable condition.
 Distribution Curves - shape
• Shape

B A

• ‘A’ is a Normal Distribution Curve.

• ‘B’ is a Skewed Distribution Curve.

Distribution Curves - spread
• Spread

• Spread of curve ‘B’ is greater than spread of curve

‘A’.
 Distribution Curves - location
• Location
X Y

X Y

• Line XX is location of curve ‘A’.

• Line YY is location of curve ‘B’.

 Distribution Curves - location
For a Normal Distribution Curve:

• The mean,
median and
mode all lie
at the same
point.

Mean
Median
Mode
 Process Capability
• Capability is a measure of the ability of a process to
meet specified requirements, usually expressed as an
index.
• Process capability measurement can be made using
the data that is obtained for the control chart.
• The capability assessment compares the process
spread with the specified requirements.
• The process chart does not make any reference to
the specified requirements.
 Process Capability

PROCESS SPREAD SPECIFIED TOLERANCES

` SPECIFIED TOLERANCES ` PROCESS SPREAD

• A Capable • An Incapable
Process Process
 Process Capability - Types
• There are two types of process capability study:
• Initial Process Capability Study;
• On-going Process Capability Study.

• An Initial Process Capability Study is conducted to

obtain early information on a new or revised process.
• An On-going Process Capability Study uses data from
a longer time period to include all common causes of
process variation.
• Both studies use control charts.
 Visualizing Process Capability
Lower Upper Lower Upper
Spec. Spec. Spec Spec.
Limit
0.4 Limit 0.4
Limit Limit
0.3 0.3

0.2 0.2

0.1 0.1

0.0 0.0
-4 -3 -2 -1 0 1 2 3 4 -8 -6 -4 -2 0 2 4 6 8

Process
Process Capability Capability

Cp=1 Cp=2
What if Process is Off-Target?

Lower Upper
Spec. Spec.
0.4
Limit Limit
0.3
Cp = 1.33
0.2

Cpk = 1.33 0.1

0.0
-5.33 -4.0 -2.67 -1.33 0 1.33 2.67 4.0 5.33

Lower Upper
Spec. Spec.
0.4
Limit Limit
0.3

Cp = 1.33 0.2

0.1
Cpk = 0.83
0.0
-5.33 -4.0 -2.67 -1.33 0 1.33 2.67 4.0 5.33
 Measures of Process Capability

LL UL
Process Capability Index:
±3 σ
UL − LL
CP =
6σ
σ

k
Cp adjusted for k: LL UL

Cpk= Cp (1-k)

± 6σ
 Capability Index
Capability : Ratio of the specification width to 6 times
the process or standard deviation.

LSL USL

6σ
tolerance

USL − LSL
CapabilityIndex(C p ) =
6σ
Capability Index, Cp

USL - 20
LSL - 10
Tolerance=USL - LSL=10
σ=1

tolerance 10
Cp = = =1.8
6σ 6
 Actual Capability Index, Cpk
Actual Capability : Ratio of the difference between the process mean
and the nearest specification limit to 3 times the standard deviation.

LSL USL
X

3σ
tolerance

CPL = X-LSL CPU = USL-X CPK = Min {CPL, CPU}

3σ 3σ
Actual Capability Index, Cpk

X −LSL 5
Cpk = = =1.8
3σ 3

X −USL 2.5
Cpk = = =0.83
3σ 3

4−LSL 4
Cpk = = =1.33
3σ 3
Interpreting Process Capability Indexes

CP CPK Interpretation
1.80 1.80 LOW VARIABILITIY, CENTERED

1.80 0.50 LOW VARIABILITY, NOT CENTERED

0.30 0.30 HIGH VARIABILITY, CENTERED

0.40 0.15 HIGH VARIABILITY, NOT CENTERED

When to undertake process capability studies
• Periodically as a measure of quality control
• Mandatory when new product started
• Mandatory when process altered
• Mandatory when specs changed
• When vendor rating is to be done
• Shortlisting the vendors