Process Capability - Tool
Process Capability - Tool
CE
Process capability
Process A combination of equipment, materials, people, methods and environment that act together to produce an output. The quality of the process is judged by the quality ( characteristic ) of the output.
FICCI
CE
Process capability
Process capability The inherent variability of a quality characteristic that the process is capable of maintaining, when in a state of statistical control under a given set of conditions. Assuming normal distribution of the quality characteristic Process Capability = +/- 3 SD = Total spread of 6 SD.
FICCI
CE
Process capability
Process capability is judged by comparing process performance with process requirements. Since meeting specification limits is one of the most basic requirements of a process capability study, it is extremely important to accurately verify and define the specification limits.
FICCI
CE
FICCI
CE
1. Condition of machine/ equipment. 2. Type of operation and operational conditions. 3. Raw materials. 4. Skill of operators. 5. Measurement method / instruments. 6. Inspectors skill.
FICCI
CE
Visual from run chart. Frequency distribution and histogram. Control charts. Analysis of variances. Probability papers.
Note : Reduction of variability can be achieved through use of SPC tools and Design of Experiments.
FICCI
CE
FICCI
CE
LSL
USL
5 - 3 SD
10
11 12 13 + 3 SD
14
15
16
FICCI
CE
A major reason for quantifying process capability is to compute the ability of a process to hold product tolerances. A measure of this relationship is process capability ratio or Cp. Process capability is also known as potential capability. The Cp index is given by:Cp = Tolerance 6 SD
Where Tolerance = USL - LSL SD = Standard Deviation
FICCI
CE
Interpretations of Cp Cp > 1 : The process is quite capable Cp = 1 : The process is just capable Cp < 1 : The process is incapable The recommended value of Cp is 1.33 ( minimum) In order to achieve Six Sigma quality in the organization, we must reduce the variation in the process so as to achieve the value of Cp=2.
FICCI
CE
z 2 3 4 5 6
Refer normal distribution table for finding defective parts per million ( PPM ) for corresponding z values.
FICCI
CE
For individual parts, the ideal design is Cp = 2; in other words, the design specification is twice as wide as the true capability of the process. This is where the phrase Six Sigma Quality originated. Since the process capability is +/- 3SD, a design specification twice as wide would be +/- 6 SD.
FICCI
CE
However, it is learnt from the industry experience across the world, that the processes rarely stay centered on their targeted nominals, shifts of 1.5 SD to either side of the mean are common, even in well controlled processes. Therefore in practice, an ultimate z-value of 4.5 is considered Six Sigma Quality. Consulting the z-table, we find that the probability of producing non conformities associated with a zvalue of 4.5 is 0.0000034. If this value is obtained, then only 3.4 out of every million parts manufactured will be defective.
FICCI
CE
Calculating defective PPM with permissible shift of 1.5 SD at various levels of sigma. ( practical situation )
Refer normal distribution table for finding defective PPM for corresponding z values.
FICCI
CE
Drawbacks of Cp
Cp however is not a very reliable measure as it does not tell us all. Consider the following four processes producing the same output X with specification 20+/- 4. Each of these processes have the Standard deviation of 1.
FICCI
CE
Process 1
X=20 SD=1
LSL
USL
Cp=1.33
FICCI
CE
Process 2
X=22 SD=1
LSL
USL
Cp=1.33
FICCI
CE
Process 3
X=15 SD=1
LSL
USL
Cp=1.33
FICCI
CE
Process 4
X=25 SD=1
LSL USL
Cp=1.33
FICCI
CE
Cpk is a measure of process performance capability The process performance index Cpk is given by:Cpk = Min [ USL - x , x - LSL ] 3SD 3SD Example : Specification : 20 +/- 4, SD = 1 Cp = Tolerance/6SD = 8/6 = 1.33
FICCI
CE
FICCI
CE
In the previous slide we observe that, although the Cp value = 1.33 in all the four cases, but because of the shift in the process setting level we are getting Cpk values as 0.67 in 2nd case and hence the non conformities. Similar observations are noticed in 3rd and 4th case where we get the Cpk as -0.33.
FICCI
CE
Thus Cpk = Cp means the process is centered. Cpk < 1 means non- conformances are being produced. Cpk < 0 indicates that the process has been set beyond either of the two specification limits.
Note : Cpk is always less than or equal to Cp.
FICCI
CE
Therefore, the first step is to bring Cpk=Cp by proper centering of the process. The second step should be to improve the Cp value by decreasing the variation.
FICCI
CE
One-sided Tolerances - The Cp and Cpk indexes are used for characteristics with two-sided tolerances - that is, processes with both upper and lower specification limits. Since many characteristics have only one-sided specifications, it is also convenient to have one-sided capability indexes.
FICCI
CE
Correspondingly, when only an Upper specification exists, we define an upper capability index by:Cpu = USL - x 3 SD