Scribd
Upload
110 views21 pages

2.1 Capability Analysis

The document discusses different approaches to assessing process capability including parametric and non-parametric methods. Parametric methods assume the data is normally distributed while non-parametric methods do not make distributional assumptions. The document provides examples of assessing capability using normal and non-normal distributed data as well as when the data does not fit a known distribution.

Uploaded by

melimaul07
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
110 views21 pages

2.1 Capability Analysis

The document discusses different approaches to assessing process capability including parametric and non-parametric methods. Parametric methods assume the data is normally distributed while non-parametric methods do not make distributional assumptions. The document provides examples of assessing capability using normal and non-normal distributed data as well as when the data does not fit a known distribution.

Uploaded by

melimaul07
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 21

Capability Analysis for

Continuous Data
Basic Concepts

Quantifiable comparison of Voice of Customer


(spec limits) to Voice of the Process (control limits)

Allowed variation (spec .)


Cp =
Normal variation of the process

USL - LSL
Cp =
6
Stable Process

Total
Variation

Slide 3
UNStable Process

Total
Variation

Slide 4
Capable Process

CAPABLE

Spec
Limits

NOT
CAPABLE

Slide 5
Four Indices are Typically Used in Industry
to Measure Process Performance

In-Control Process NOT In-Control Process

Process is Process Capability Process Performance


Centered (Cp) (Pp)
Process Capability Process Performance
Process is
Adjusted for Process Shift Adjusted for Process
Not
Shift
Centered
(Cpk) (Ppk)
How Does Statistical Control Relate to Cpk/Ppk?

❑ Cp/Cpk provides an estimate of potential process


performance or capability
▪ How the process would perform in the absence of
special cause variation.

❑ Pp/Ppk provides the actual process performance given


all of the variance that is currently present in the
process
▪ Pp/Ppk is generally <= Cp/Cpk
Assessing Capability Analysis Assumptions

Capability Estimates are meaningful if both of the


following are true:
1. The process is in-control
2. We have information about the distribution of the data

Use Capability Six Pack


Assessing Capability Analysis

Case 1
Conduct Capability Analysis
Normal BASED ON Normal Data

Continuous
Transformation
Data

Not Fit with


Appropriate
Conduct Capability
Analysis BASED ON
Normal Distribution THAT Distribution
Case 2 Non-
Parametric
Approach
Case 3
Case 1

Capability
Normal Data
Analysis
Example 1: Capability Analysis With Normal Data

The production operator measures the thickness of the


sugar-coated multivitamin tablets. The thickness is critical
property of the tablet since it influences the blistering
process. Based on the specification listed in the batch
record, the lower specification limit is 3.9 mm and the
upper specification limit is 4.3mm
Data Collection
Operators randomly select five tablet samples at regular intervals for
measuring the sugar-coated tablet thickness. These samples
adequately represent the inherent variation of the process at that time.
The operators record the thickness each tablet.

Operators evaluate the capability of the thickness to ensure that it can


meet customer requirements.

Dataset: CoatingCapability.MPJ
Tasks:
1. Assess capability analysis assumptions using Capability SixPack
2. Subset worksheet, not to include process with adjustment
3. Conduct Capability Analysis
Exercise

An inspector must asses the capability of tablet weight to


meet customer requirements.

Data Collection

Inspector record the tablet weights every 2 hours using


rational subgroups of size 5. The tablet weights have an
allowable tolerance of 24  3 grams

Dataset: CapTablet.MPJ
Exercise: Instruction

1. Asses the assumptions of process capability for the data


in Weight by using Capability Sixpack > Normal
2. After a root cause investigation, you discover an
incorrect testing set up for subgroup 4/29 10 am and
incorrect entries that are due to server problem for
subgroup 4/30 4 pm. That data have been removed and
the remaining data are placed in the Revised Data
column. Using the revised data, reassess the capability
of the process
3. Compare the results in step 1 and 2, including Cp versus
Pp and Cpk versus Ppk
Case 2

Capability Non-Normal
Analysis Data
Example 1: Capability Analysis With NON Normal Data

A validation specialist wants to evaluate the


diameter of animal health tablet. The manufacturer
requires that the diameter be less than 1.567 cm
Data Collection
Operator collect samples of the tablets in subgroups of size
10 every 3 hours for 12 days.

Dataset: TabDiameter.MPJ
Tasks:
1. Assess capability analysis assumptions using
Capability SixPack
2. Individual Distribution Identification
3. Capability Analysis > Nonnormal
4. Transformation
Case 3

Capability Non-Normal
Analysis Data

No suitable transformations

No distributions fit to the data


NON-Parametric Capability Analysis

• There are other nonparametric capability analysis methods


available.
• Only the empirical percentile method is shown here.
• It was founded on 2000
D. W. McCormack Jr., Ian R. Harris, Arnon M. Hurwitz, and
Patrick D. Spagon, 2000. "Capability Indices for Non-Normal
Data," Quality Engineering, 12 (4), 489-495.
NON-Parametric Capability Analysis

Requirement:
• At least 100 numeric observations
• Nonparametric Capability MACRO (ECAPA.Mac)

Running the Macro


Suppose your data is in C1 and you have a lower spec limit (LSL)
of 12 and an upper spec limit (USL) of 16.
To run the macro, choose Edit > Command Line Editor and type:
%ECAPA C2;
LSL 12;
USL 16.
Example

The data are measured individual weight data (in mg)


of a pharmaceutical company from a certain batch.
The specification of the weight are LSL = 620 and
USL=680.

Dataset: Bobot Individu.xlsx


Instruction:
1. Import the data into Minitab
2. Assess the capability of the data with Normal/Non-Normal
Distribution Approach
3. Assess the capability of the data with Non-Parametric
Approach

You might also like