Process
Capability
Analysis
1
• Process performance vs process
specifications
• Process capability studies
Course Outline
• Process capability (Cp, Cpk) and process
performance (Pp, Ppk) indices
• Short-term and long-term capability
Assumption
There are 5 assumptions to be kept for 4. Data must comply independently
study. normal distribution. Both of graphical
1. Parameter must be variable parameter. method and inferential method are
2. Process must be operated with normal applied to check this assumption
condition Such as.. • Histogram
• Box plot
• consuming only within spec material
• NOPP
• well-maintained machine
• Normality test
• working by well-trained operator
• avoid using data from laboratory. 5. Data must be measured value that
3. Data must be within statistical control. reflect on process variation.
Control chart is used to check this
assumption.
• 𝑋−𝑀𝑅
• 𝑋 ̅−𝑅
• 𝑋 ̅−𝑆
3
Natural Process Limits
• Natural Process Limits (NPL) refer to the statistical boundaries within which a process
typically operates when it is in control and only subject to inherent variability.
• These limits are calculated based on the process's historical data and are used to
differentiate between common cause variation (natural, expected variability) and special
cause variation (unusual, unexpected variability).
• Natural process limits, are determined from historical data of how the process will run if
undisturbed. The control limited are at the historical mean or target +/- 3 x the historical
standard deviation, representing approximately 99.73% of the data if the process is
normally distributed.
4
Natural Process Limits
6σ
LNTL UNTL
μ is the process mean.
σ is the standard deviation of the process.
5
Process Specifications
Process specifications, also known as specification limits or tolerance limits, are the pre-
determined criteria or requirements that define the acceptable range of variation for a
process output. These limits are typically set based on customer requirements, industry
standards, or design requirements.
Key Components:
• Upper Specification Limit (USL): The maximum acceptable value for a product
characteristic or process output.
• Lower Specification Limit (LSL): The minimum acceptable value for a product
characteristic or process output.
6
Process capability
• Process capability is a measure of a production process's ability to produce products
that meet standards and maintain high quality. Typically, process capability is measured
by comparing process variation with the voice of the customer or customer
specifications (customer requirements)."
Efficient Process – Less Process Inefficient Process – When the distribution of
Variation Distribution ensures that the process variables exceeds the specifications of
product meets all customer the Customer, some products may fail to meet
specifications customer requirements.
7
Why use a process capability study
• Process capability study helps in product and process design, vendor sourcing,
production planning, and manufacturing process control.
A few ways a process capability study helps:
• Assists designers in selecting or designing processes
• Establishing an interval between sampling
• Specifying requirements of new equipment
• Vendor selection based on process capability
• Reducing variability in the manufacturing process
8
Process Capability vs. Specification Limits
Natural Process Limits =
Process Specification =
9
Process Capability Indices – Cp , Cpk
• Cp is a measure of a process's potential capability to produce outputs within specified
limits, assuming the process is centered between the specification limits.
• Cpk measures the actual capability of a process by taking into account both the process
variation and how well the process is centered between the specification limits.
10
Interpretation of Cp & Cpk
• Cp and Cpk should always be evaluated and analyzed together.
• If the Cp and Cpk values are different, the overall data is not centrally located (compares
to specification)
• The larger the difference in the values, the most offset of the overall data.
• Cpk can never exceed Cp, so Cp can be seen as the potential Cpk if the process mean is
center at the target (mid of spec.).
11
Process Capability Indices – Pp , Ppk
• Pp is a measure of a process's overall performance in producing outputs within specified
limits over a period, considering all sources of variation, including both common and
special causes.
• Ppk measures the actual performance of a process over a specified period, considering
both the variation and how well the process is centered between the specification limits.
12
Short-term and long-term capability
• Short-term capability is measured over a very short time period. In other words, it
represents the variation at a certain time. For example, the process variation is relatively
small in one shift with one piece of equipment. At the same time.
• Long-term capability measures the variation over the duration of the entire study. If there
is a shift in the process, the overall variation could exceed the subgroup variation.
13
Within vs. Overall Standard Deviation
Within standard deviation (σ within)
• Taken over a relatively short time period.
• Takes in account only the variation within subgroup.
or Pooled Std. Deviation
Overall (or Total) standard deviation (σ overall)
• Taken over su icient time period.
• Calculated from many sample that represent the shift and drift that occurs in the
population due to all cause of variation
14
Process Capability Indices
• Process capability indices are statistical measures used to assess how well a process
can produce outputs within specified limits. These indices provide insights into the
capability of a process by comparing the inherent variability of the process to the
allowable variability defined by the specification limits. The most commonly used
process capability indices are Cp, Cpk, Pp, and Ppk.
Potential
Performance
15
Process Capability Indices
• Cp results from the variation with process • Short term capability is generally studied
design. Process design means all of designed when process starting, to monitor existing
parameter ; such as Material, Machine, Man, process, or acknowledge the current
Method. situation of process before improvement.
• Cpk results from the variation with combination • Long term capability index and Short term
of process design and setting. Variation with capability index are different by process
setting will be demonstrated by a shift of mean control. Thus, capability index structure can
of data from mean of specification limit. be decomposed to..
• Pp results from the variation with combination Decomposing summary
of process design and process control. Variation
of Process control is the spread due to how well Cp = Design
you can control the process by consider result,
not how to control. Cpk = Design + Setting
• Ppk results from combination of all process Pp = Design + Control
variation. That is Process design, Process
setting, and Process Control. Ppk = Design + Setting + Control
16
Interpretation of Capability Indices
• Cpk and Ppk less than 1: The process has poor capability and needs to be improved
urgently.
• Cpk and Ppk between 1 – 1.33: The process capability is moderate and needs to be
improved.
• Cpk and Ppk between 1.33 – 2: The process capability is at a good to very good level.
• Cpk and Ppk greater than or equal to 2: The process capability is at a World Class level
(Six Sigma Level).
• Ppk is always less than or equal to Cpk.
• If Cpk and Ppk are close to each other, the process is not destabilized by the movement
of the average value, or the process is stable in the long run. Conversely, if Ppk is much
less than Cpk, the process is destabilized.
17
Indies value and Proportion of defective
Proportion of Defective (ppm) Recommended value to be used for
Indies value
1-side spec 2-side spec determine capability of process is above
2.00 0.001248 0.002496 below table, depend on condition of
1.83 0.02124 0.04248
process.
1.67 0.3002 0.6
Process 1-side spec 2-side spec
1.50 3.45 7
1.33 31.79 64 Existing process 1.13 1.33
1.17 232.70 465
New process 1.20 1.50
1.00 1,350 2,700
0.83 6,210 12,420
0.67 22,750 45,500 However, this is just recommended value. If
0.50 66,810 113,620
customer has determined required value,
you will compare to the value.
0.33 158,700 317,400
18
Cpk vs. Sigma Level
Mean
Ex. Point of interest = USL
ZUSL = ?
Therefore: Sigma Level (Z) = Cpk x 3
19
Cpk vs. Sigma Level
20
Example 1 : Short term study
• An engineer at a metals company wants Use data set
to determine whether the diameter of a CableWires
cable wire that the company
manufactures meets specifications. Function
• Each hour, operators take a subgroup of Stat > Quality Tools > Capability Analysis >
five consecutive cable wires from the Normal…
production line and measure the
diameters.
• A cable wire must be 0.55 ± 0.05 cm in
diameter to meet engineering
specifications.
21
Example 1 : Short term study
• Call function “Capability Analysis –
Normal” then complete dialog box as
below
• Select ‘C1’ in Single column
• Enter ‘C2’ in Subgroup size
• Enter lower specification limit ‘0.50’ in
Lower spec
• Enter upper specification limit ‘0.60’ in
Upper spec
22
Example 1 : Short term study
• In Estimate, • In Option,
• Uncheck Use unbiasing constants. • Uncheck Overall analysis
Note : (It’s due to number of subgroup is 1. That is Note : According to short term analysis consumes
less than 20 and cannot estimate by unbiasing one group of data then there will be not overall
constant) analysis. Thus, long term analysis is not necessary
now.
23
Interpreting the result
2 3 2. Descriptive information portion shows concerned statistic.
• Result with controlled factors (Mean) is 0.545615.
• Result with uncontrolled factors (Standard deviation) is
0.01823.
3. In capability indices portion, Pp and Ppk are needed to be
1 considered.
• Cp = 0.91. That’s not passed recommended target at 1.33.
• Cpk = 0.83. That’s not passed recommended target at 1.33
4 4. Performance of process, unit ppm.
• PPM total of observed = 20833.33. So, there are 20,834
defective of million of product out-of-spec from collected
samples. (2 pcs from 96 pcs.)
• PPM total of expected = 7627.23. So, there will be 7628
defective of a million of product when produce with this
1. Start at considering shape of Histogram whether process.
follow Normal distribution (Bell shape) or not.
• It’s shape is similar to Bell shape. So, data that
be collected, is normal distribution
24
Example 2 : Short & Long term study
• An engine manufacturer uses a forging • Use data set
process to make piston rings. The quality PistonRingDiameter
engineers want to assess the process
capability. They collect 25 subgroups of
five piston rings and measure the • Function
diameters. The specification limits for Stat > Quality Tools > Capability Analysis >
piston ring diameter are 74.0 mm ± 0.05 Normal…
mm.
25
Example 2 : Short & Long term study
• Call function “Capability Analysis –
Normal” then complete dialog box as
below
• In Single column, enter Diameter.
• In Subgroup size, enter 5.
• In Lower spec, enter 73.95.
• In Upper spec, enter 74.05.
26
Interpreting the result
2 3 • Descriptive data
• Sample mean = 74.0012
• SD (short term) = 0.01005
• SD (long term) = 0.01019
• Capability index
1
• Cp = 1.66
• Cpk = 1.62
• Pp = 1.63
4 • Ppk = 1.60
• Ppk and Cpk are not significantly different.
• Cpk and Cp are not significantly different.
• Performance by PPM
• Observed PPM = 0
1. Consider shape of histogram, it is bell shape. • Short term expected PPM = 0.77
So, Data is normal distribution. • Long term expected PPM = 1.11
27
Example 2 : Short & Long term study
• An engine manufacturer uses a forging process to • Use data set
make piston rings. The quality engineers want to
assess the process capability. They collect 25 PistonRingDiameter
subgroups of five piston rings and measure the
diameters. The specification limits for piston ring
• Function
diameter are 74.0 mm ± 0.05 mm. Stat > Quality Tools > Capability Sixpack >
Normal…
• Note : This is the same monitoring to example 2
but using different method because last method
didn’t determine other concerned details
28
Example 2 : Short & Long term study
Call function for capability analysis
Stat > Quality Tools > Capability Sixpack > Normal…
In Single column, enter Diameter.
• In Subgroup size, enter 5.
• In Lower spec, enter 73.95.
• In Upper spec, enter 74.05.
29
Interpreting the result
3
2
1 3
1 4
30
Interpreting the result
1. R-chart and Individual plot are consider for process 4. Process capability index
consistency. • Cp = 1.66
• No data out off specification • Cpk = 1.62
2. Xbar chart is used to consider how well of process • Pp = 1.63
control. • Ppk = 1.60
• All data within control • Final conclusion is similar to previous interpreting even if
capability indices value are not different.
3. Histogram shows a single peak as bell shape. So, you
can conclude that data are normal distribution. 2 3
• Not only histogram, but also NOPP and P-value can be
used for consideration of normal distribution of data.
• When you see NOPP, you can conclude that data are
normal distribution. 1 3
• P-Value of normality test = 0.892. So you have not enough
evidence to reject H0. That mean data be normal
distribution.
1 4
31
Cpm in Minitab and interpreting the result
• To modify function, just click ‘Options...’ button and
enter desired value in Target box. Minitab will add Cpm
in calculation and return you.
• This option can be used both of Capability analysis and Cpm = 1.63
Capability sixpack. That means process proceeding good capability when
• Note : Above picture and result come from Capability using desirability concept.
analysis.
32
Cpm in Minitab and interpreting the result
Acceptability
• Both of Cpk and Ppk are based on the concept of acceptability, process proceeds a
product within specification limit. So, this concept is called “Goal post”
• Disadvantage of this acceptability concept is when you decide quality of process, you
interest only whether it is out of spec or not but customer does not think so.
Desirability
• In fact, customer need not only produce be within specification limit but they need the
produce be on the desired value also.
• That means Cpk and Ppk are not the best one to consider process capability. You
should interest desirability concept, process proceeds a process and try it be on the
target. So, Cpm was developed to determine the process capability base on the
concept.
33
Exercise 1
An engineer wants to assess the capability of Use data set
a process that coats large rolls of paper with FilmThickness
a thin film. The engineer collects three
measurements of coating thickness from 25 Function
consecutive rolls. Because the machine is Stat > Quality Tools > Capability
reset for each new roll, the engineer needs to Analysis > Normal…
consider variation between rolls in addition to
variation within a roll. The film thickness must
be 50 ± 3 microns to meet engineering
specifications.
34
Exercise 2
A manufacturer uses machines to produce Use data set
connector pins. A quality engineer wants to PinLength
assess the process capability of machine.
The engineer randomly samples five pins per Function
box from machine. The specification range Stat > Quality Tools > Capability
for the length of the pin is 13 mm to 25 mm. Sixpack > Normal…
The engineer performs normal capability
analysis to evaluate how well the lengths of
the connector pins meet customer
requirements for the two machines.
35
Salakjitt Buddhachakara
Thank you salakjitt@[Link]
093-2939978