Capability Info

Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 14

Process Capability

Introduction

MEAN - HOW IS THE PROCESS CENTRED


RANGE - VARIATION IN PROCESS
CONTROL CHARTS - PICTURE OF PROCESS

PROCESS CAPABILITY INDICES - PERFORMANCE


MEASURE.
Process Capability
Cp

Capability of a process can be defined as its natural


tolerance, or the degree of variability it suffers from due to
common causes, compared to the customer specified
tolerance.

Capable process has a natural tolerance < customer


tolerance.

Incapable process has a natural tolerance > customer


tolerance.
COMPARES WHAT THE PROCESS CAN DO
WITH WHAT WE WANT IT TO DO.
Process Capability
Before examine process capability:
1. Process must be in STATISTICAL CONTROL -
no special causes of variation
2. The process distribution curve is NORMAL
(Bell-shaped)

Cp represents process capability. It is the


ratio of specification width to variation width.

Cp = CUSTOMER TOLERANCE = T
PROCESS TOLERANCE 6
Process Capability
Cp Exercise
Exercise
Show the natural and customer tolerances on the following
process distributions and comment on the capability of each.

Mean Mean

LCL UCL LCL UCL


Process Capability
Calculating Standard Deviation, 
Population standard deviation can be estimated:

 = R
D2
D2 is a constant and can be looked up
Remember: 6 = 99.73 % population data in this range

IF A PROCESS IS BARELY ACCEPTABLE Cp = 1


IF A PROCESS IS LESS THAN ACCEPTABLE Cp < 1
IF A PROCESS IS BETTER THAN ACCEPTABLE Cp > 1
Process Capability
Cp Exercise
Exercise
For the figures below indicate which process has Cp = 1, Cp < 1, Cp > 1.
Comment on the implications for consistently meeting requirements and what
happens to the value of Cp if the mean changes?

USL

Mean

LSL

What conclusions do you draw about the usefulness of Cp?


Process Capability

BUT, Cp only considers the specification (tolerance)


range.
Therefore it is only an indication of process potential
Need to consider both:
1. Variation in the process
2. Location of the distribution curve

WHERE IS X?
IS IT INSIDE OR OUTSIDE SPECIFICATION LIMITS?
IF INSIDE IS IT CLOSER TO USL OR LSL?
Process Capability
Cpk
Cpk
is a measure of actual process performance

Cpk considers location and variation and indicates


distance between X and nearest spec limit.

Cpk = USL - X OR Cpk = X - LSL


3 3

WHICHEVER IS THE LEAST

NOTE: WHEN PROCESS ON TARGET Cp = Cpk


Process Capability
Cpk Exercise
Exercise
For the process distributions below indicate whether the Cpk value is > 1, < 1
or = 1.
Why do 3 of the distributions have the same spread but different Cpk values?

USL

Mean

LSL
Process Capability
Cpk Exercise 1

Example
Going back to the chart example 1, supplier have
accepted an average torque reading in joint
assemblies as 50nm with accepted tolerance of plus
and minus 25nm.

How capable is the process ?


Process Capability
Cpk Exercise 2

Example
Going back to the last control chart example 2, MIS
customers have accepted an average response time of
6.25 seconds as reasonable and have requested a
tolerance of plus and minus 0.25 seconds.

How capable is the process ?


Process Capability
Cpk Exercise 3
For the following data calculate Cp and CpK
and comment on the process, assume subgroup
size of 5.

1. MEAN RANGE = 91 mg, PROCESS MEAN = 2500 mg


USL = 2650 mg, LSL = 2350 mg
2. MEAN RANGE = 91 mg, PROCESS MEAN = 2650 mg
USL = 2750 mg, LSL = 2250 mg
Process Capability
Conclusion
A Cpk requirement of greater than 1.0 is a statement that
"anywhere within tolerance" isn't acceptable, and is a
repudiation of the widely accepted "goalpost mentality" of
the past.

Design tolerances are not intended for the benefit of the


producers. To use as such and is likely to lead to unreliable
products and high warranty costs.

Maintaining Cpk higher than 1.0 implies that processes need


to be better maintained and reset more often (higher costs).
These can, be offset against lower warranty claims, better
customer satisfaction and enhanced reputation.

You might also like