When it comes to analyzing the current parameters against customer requirements or designate metrics, than Process Capability comes into a play. It is typically defines as,
The inherent variability of characteristic of a product
It represents the performance of the process over a certain period of stable operations. If process always conforms to the process specification, that it is said to be capable. The purpose is to quantity defects or failures in terms of Process sigma (Where do we stand right now in terms of meeting customer requirement)
The power of the data is best understood, in the form of visualization. Cited by Mr. Shankar, raw data summarization tend to follow following form.
Following points can be concluded from the pic above,
- Raw Data must be broken down with respect to central tendency & spread measures, answering the question (where lies most of our data & typical attributes).
- Voice of the process (LCL/USL) must be broader than customer specification limit i.e.: CSL must reside in between VOP
- Graphical summary of raw data will produce all of the required values in a matter of seconds.
Let us have a look, cited by Tom Kubiak& Donald Benbow at the two major type of indices encountered in process capability.
To give a best illustration, let us consider an example of vehicle size & road. Think of your process as the vehicle you are driving a customer specification as a road on which you are driving.
- If size of vehicle is half of roads width i.e.: vehicle is going in center with a tolerance are on both sides than Cpk=2.
- If size of the vehicles same as the width of road, than Cpk=1.
- If size of chicle (process) is wider than customer specifications (road width), than Cpk=0.5
Thus, Cpk is the ration of difference in customer specification and normal process variation.
As Cpk tells us the size of vehicle, Cpk predicts our driving ability. Cpk is a measure of our process ability to meet specification.Thus,
- If Cpk=1, than it means our process average and specification are at the same point.
- If Cpk=1.33, than our vehicle have a maneuvering limit of 1 S.D. on both sides.
- If Cpk=2 than we have plenty of room to go around.
So, it determines the proximity of the process average to the nearest customer specification limit.
Sigma Level =3Cpk
Process Performance Indices
Process Performance is a statistical measure of the outcome of characteristic from a process that many not have demonstrated to be in a statistical control.
Pp & Cpk are also given as,
Cpm is a process capability index that monitors the process average with respect to target value.
- When the process average and the target value are equal, Cpk = Cpm.
- When process average drifts from the target value, Cpm < Cpk.
(Note: Ppm is analogues to the Cpm)
Short Terms Vs Long Term Capability
Some basic points regarding it are the following,
- AIAG suggests that capability indices are long term while performance indices are short term indicators.
- Also, it follows following formula,
Long term variability = Short term variability +- 1.5 Sigma
- Variable data process sigma is considered short term while attribute data sigma is considered long term.
In next blogs, light will be shed on the computation of Capability for Normal/ Binominal Poison data,with Minitab.