Process Capability Analysis

Agenda
• Introduction to Capability
– What is capability
– Histograms
– Normal Distribution

• Capability Indices
– Cp
– Cpk
– Pp
– Ppk

• Calculating Sigma
– Relating capability to percent nonconforming

• Capability with attribute data

– DPO, DPMO
– DPU
 What is Capability?

• Process capability so sánh đầu ra của một in-control

process với specification limits.

• Việc so sánh được thực hiện bằng cách hình thành

tỷ lệ chênh lệch giữa process specifications (the
specification "width") với sự phân tán của process
values, được đo bằng process standard deviation.

• Capable process là process mà hầu hết các giá

trị đo được đều nằm trong specification limits.
 Graphical Representation

LSL USL

Mean = 75 6S on each side of

SD = 0.3 the mean to the
USL = 73 specification limit.
USL = 77
6S = 1.8

73.2 73.8 74.4 75.0 75.6 76.2 76.8

 Histogram

73.66 74.11 74.41 74.71 75.31 75.61 75.91 76.36

±1S (68.3%)
40

±2S (95.4%)

30
±3S (99.7%)
Frequency

20 ±4.5S (99.9993%)

10

0
74.0 74.4 74.8 75.2 75.6 76.0 76.4
 Examples of Capability

• Some examples of where capability

analysis can be used:

• Process that is not centered

• Process with large variability
• One-sided specifications
• Setting/confirming customer specifications
 Percent Out of Specifications

• Dựa trên normal distribution, có thể tính được phần trăm số sản
phẩm out of spec.

• This is best explained using an example.

• Giả sử chúng ta có một quá trình với mean = 50, standard deviation
= 4, USL = 58 and LSL = 46.

• Chúng ta chia vấn đề này thành hai phần. First the percent out of
specification on the high end (greater than the USL) and then the
percent out on the low end (less than the LSL).
 The Normal Distribution

• The normal distribution is:

USL  X X  LSL
Z ;
S S
• Z is là số standard deviations từ specification đến mean.

• Normal probability tables give you the percent of the distribution that
would exceed the specification limit for a given z value

• Remember that 68.3% of the data is within ±1S (therefore 31.7% is

outside of ±1S).
 Out of Specification Calculations
USL  X X  LSL
Z ;
S S

58  50 50  46
Z ;
4 4

Z  2 for the upper specifcation and 1for the lower specification

• A z = 2 is 2.28% out of specification; z=1 is 15.9%.

• The total percent expected to be out of specification would be

18.18%

• We will discuss summary statistics for this later.

 Estimating Sigma

• There are several methods for estimating

sigma (S) used in capability analysis.

• Control charts
– Rbar
– Sbar
– Moving Range
– MSSD
• Pooled standard deviation
• Total standard deviation
Sigma within (subgroup size > 1)
Sigma within (subgroup size > 1)
Sigma within (subgroup size > 1)
Sigma within (subgroup size = 1)
Sigma within (subgroup size = 1)
Sigma within (subgroup size = 1)
Sigma overall
 Assumptions

• There are two critical assumptions to consider when performing

process capability analyses with continuous data, namely:

• The process is in statistical control.

• The distribution of the process considered is Normal.

• If these assumptions are not met, the resulting statistics may be

highly unreliable.

• In a later modules we will discuss capability analysis for non-normal

data.
 Capability Indices

• There are several statistics that can be used to measure the

capability of a process: Cp, Cpk, Pp and Ppk.

• The statistics assume that the population of data values is normally

distributed.

• Variability can be stated as either within or overall.

• Cp and Cpk are based on sigma within

• Pp and Ppk are based on sigma overall

 Cp

• Approximately 99.7% of the data from a normal distribution is

contained between ±3.

• If the process is in control and the distribution is well within the

specification limits then the difference between the Upper
specification (U) and Lower specification (L) should be larger
than 6.

• If the specifications are larger than 6, the ratio will be less than 1.

• If Cp is greater than 1 then the process has the potential to meet

specifications as long as the mean is centered.
 Cp
LSL USL

Mean = 0.045
SD = 0.005
LSL = 0.042
USL = 0.048
Cp = 0.19

UpperSpec LowerSpec
Cp 
0.028 0.032 0.036 0.040 0.044 0.048 0.052 0.056

6 S LSL

Mean = 0.045
USL

SD = 0.005
LSL = 0.03
USL = 0.06
Cp = 0.97

0.030 0.036 0.042 0.048 0.054 0.060

 Cpk

• Cpk is an process capability index that assesses how close

the process mean is from the specification limit.

• If the process is in control and the distribution is well within

the specification limits then the difference between the
Upper specification (U) and the mean or the difference
between the Lower specification (L) and the mean should
be larger than 3.

• If Cpk is greater than 1 then the process mean is sufficiently

far from the specification limit.
 Cpk
LSL USL

UpperSpec  X
C pU  Mean = 0.045
SD = 0.005
3 S
LSL = 0.042
USL = 0.048
CpL = 0.17
CpU = 0.22
Cpk = 0.17

X  LowerSpec 0.028 0.032 0.036 0.040 0.044 0.048 0.052 0.056

C pL 
3 S LSL USL

Mean = 0.045
SD = 0.005
LSL = 0.03
USL = 0.06
CpL = 0.95
CpU = 0.99
C pk min(CpL ,C pU ) Cpk = 0.95

0.030 0.036 0.042 0.048 0.054 0.060

 Cpk

– Cpk greater than 1 shows the process is probably centered

and usually able to meet specifications

– Cpk less than 1 indicates either the mean is not centered

between the specifications or there is problem with variability

– Cpk is meant to be used with processes that are in control –

gives us a measure of whether the in-control process is
capable of meeting specifications

– Cpk is not an appropriate measure if there are trends, runs,

out-of-control observations or if the process is too variable
 Pp

• Pp is an overall capability similar to Cp.

• Sigma overall is used in the denominator instead of the

within.

• If the process is stable and in control the estimate of Pp is

similar to the estimate of Cp.

• If Pp is greater than 1 then the process is meeting the

specifications as long as the mean is centered.
 Pp
LSL USL

Mean = 0.045
SD = 0.0054
LSL = 0.042
USL = 0.048
Pp = 0.19

UpperSpec  LowerSpec
Pp 
0.028 0.032 0.036 0.040 0.044 0.048 0.052 0.056

6S LSL USL

Mean = 0.045
SD = 0.0054
LSL = 0.03
USL = 0.06
Pp = 0.93

0.030 0.036 0.042 0.048 0.054 0.060

 Ppk

• Ppk is an process capability index that assesses how close the

process mean is from the specification limit.

• Sigma overall is used in the denominator instead of the short term.

• If the process is in control and the distribution is well within the

specification limits then the difference between the Upper
specification (U) and the mean or the difference between the Lower
specification (L) and the mean should be larger than 3.

• If Ppk is greater than 1 then the process mean is sufficiently far from
the specification limit.
 Ppk
LSL USL

Mean = 0.045
SD = 0.0054
X LowerSpec
PpL  LSL = 0.042
USL = 0.048
3S PpL = 0.17
PpU = 0.21
Ppk = 0.17

UpperSpec  X
PpU 
0.028 0.032 0.036 0.040 0.044 0.048 0.052 0.056

3S
LSL USL

Mean = 0.045
SD = 0.005
LSL = 0.03
USL = 0.06
Ppk  min(PpL, PpU ) PpL = 0.91
PpU = 0.95
Ppk = 0.91

0.030 0.036 0.042 0.048 0.054 0.060

 Example

• An engineer at a metals company wants to determine

whether the diameter of a cable wire that the company
manufactures meets specifications. Each hour, operators
take a subgroup of five consecutive cable wires from the
production line and measure the diameters. A cable wire
must be 0.55 ± 0.05 cm in diameter to meet engineering
specifications.

• High Cpk and a Funny-Looking Histogram: Is My Process

Really that Amazing?
 Attribute Data

• When examination of an item or event results in a PASS

or FAIL rather than a measurement, the capability
analysis must be based on a discrete distribution.

• When you are counting defective items, the binomial is

used to estimate capability.

• When you are counting defects per unit, then the

capability analysis is based on the Poisson
distribution.
 DPU, DPO, DPMO

• Defects per unit (DPU) is the number of defects in a sample divided

by the number of units sampled.

• Defects per opportunity (DPO) is the number of defects in a sample

divided by the total number of defect opportunities.

• Defects per million opportunities (DPMO) is the number of defects in

a sample divided by the total number of defect opportunities
multiplied by 1 million. DPMO standardizes the number of defects at
the opportunity level and is useful because you can compare
processes with different complexities.
 DPU, DPO, DPMO
Your printing business prints custom stationary orders. Each order is
considered a unit. Fifty orders are randomly selected and inspected and
the following defects are found.
• Two orders are incomplete
• One order is both damaged and incorrect (2 defects)
• Three orders have typos
 Six of the orders have problems and there are a total of 7 defects out
of the 50 orders sampled; therefore DPU = 7/50 = 0.14. On average,
this is your quality level and each unit of product on average contains
this number of defects.
 Six of the orders have problems, and there are a total of 7 defects
out of the 200 opportunities (50 units * 4 opportunities / unit);
therefore DPO = 7/200 = 0.035.
 There are a total of 7 defects out of the 200 opportunities. Therefore,
DPO = 0.035 and DPMO = 0.035 * 1000000 = 35,000. If your
process remains at this defect rate over the time it takes to produce
1,000,000 orders, it will generate 35,000 defects.
 Binomial Data

• First plot your data on a p-chart. This allows you to assess if the
process is in statistical control. Remove any points outside the limits
and recalculate the average percent defective.

• Using the overall percent defective, calculate the upper and lower
95% confidence interval on percent defective.

• Multiply the percent defective by 1,000,000 to get the DPMO.

• Calculate Process Z by determining the Z-value that corresponds to

the percent defective. A higher Process Z is desirable.
 Example

• The following are the number of defectives

found during the inspection of 20 units in each
lot.
1 4 1 2 4
2 3 1 0 7
3 2 5 8 8
2 3 4 9 6
3 2 3 0 0
4 1 4 4 5
3 2 5 5 6
2 3 9 0 0
3 4 4 1 2
5 2 3 2 4
 P-Chart

P Chart
0.5
1 1

UCL=0.4156
0.4

0.3
Proportion

0.2
_
P=0.166

0.1

0.0 LCL=0

1 6 11 16 21 26 31 36 41 46
Sample
 Updated P-Chart

Updated P Chart
0.4
UCL=0.3786

0.3
Proportion

0.2

_
P=0.1435

0.1

0.0 LCL=0

1 6 11 16 21 26 31 36 41 46
Sample
Summary Statistics
 Poisson Data

• First plot your data on a u-chart. This

allows you to assess if the process is in
statistical control. Remove any points
outside the limits and recalculate the
average defects per unit (DPU).

• Using the overall defects per unit,

calculate the upper and lower 95%
confidence interval on percent defective.
 Example

• The following are the number of defectives

found per unit during the inspection of 50 units.

7 2 2 3 5
4 4 2 4 4
4 5 2 7 3
7 3 2 2 1
0 1 9 6 2
2 3 0 2 5
1 4 4 1 3
0 2 1 4 1
3 2 3 3 1
5 2 2 2 1
 U-Chart

U Chart
1
9

8 UCL=8.12

7
Sample Count Per Unit

4
_
3 U=2.96

0 LCL=0

1 6 11 16 21 26 31 36 41 46
Sample
 Updated U-Chart

Updated U Chart
8 UCL=7.890

6
Sample Count Per Unit

_
3 U=2.837

0 LCL=0

1 6 11 16 21 26 31 36 41 46
Sample
Summary Statistics
 Conclusions

• Process capability analysis can be predictive of

expected out of specification results.

• The process must be in statistical control prior to

doing a capability analysis.

• Data transformation can be used for non-normal

data.

• Attribute data can use confidence intervals to

predict process performance.
Questions?
THANK YOU!