The case mentions that the cans are being checked by using an inspection procedure known as s sampling. Let's assume that 1% of the cans coming from the process are out of specification (but our inspector doesn't know this). Let's assume a very simple sampling plan is being used where the inspector takes a random sample of five units and rejects the batch if one or more cans are filled beyond the maximum acceptable pressure of 24 psi. A step toward statistcal control Now let's assume that as a first step to improving the filling machine, pressures were taken from a sample of 40 units coming from the machine.
|
19
|
20
|
19
|
18
|
22
|
17
|
20
|
19
|
17
|
18
|
|
18
|
22
|
21
|
22
|
19
|
18
|
21
|
16
|
18
|
16
|
|
21
|
24
|
22
|
21
|
21
|
20
|
18
|
17
|
19
|
19
|
|
21
|
20
|
24
|
17
|
22
|
20
|
16
|
20
|
21
|
22
|
If the specification is such that no can should have a pressure greater than 24, what is the Cpk for this process? Assuming the pressures are distributed normally, what percentage of the output is expected to be out of tolerance? What if there is an upper and lower specification, where the lower pressure limit is 14, what are the Cpk and the Cp for the process? What percentage of the output is expected to be out of tolerance?