19. Process Capability Flashcards
Process Capability P.365
Inherent variability of the characteristics of a process. A process is capable when the output always conforms to the process specifications (6σ).
Process Capability Index P.365
Used to represent the ability to meet customer specification limits..
Cp Index
Describes process capability in relation to the specified tolerance of a characteristic divided by the natural process variation
Cp = USL - LSL / 6σ 1 = capable 2= represent 6σ performance
Cpk Index
Cpk = min (Cpkl, Cpku)
Cpkl = x̅- LSL / 3σ Cpku = USL - x̅ / 3σ
- 33 = goals
- 67 = mission critical / safety items
Cpm Index
Process capability index of the mean. An index that accounts for the location of the process average relative to a target value.
Process performance P.370
Statistical measure of the outcome of a characteristic from a process that may not have been demonstrated to be in a state of a statistical control.
Process Performance Index P.371
A dimensionless number that is used to represent the ability to meet specification limits on a characteristic of interest.
Tend to be a bit larger than their corresponding process capability indices since they are often based on smaller sample sizes.
Pp Index
Pp = USL - LSL / 6s
Ppk Index
Ppk = min. (Ppkl, Ppku)
Ppm Index
Analogoous to Cpm
Conducting a Process Capability Study P.372
- Select a quality characteristic of a specific process for study.
- Confirm the measurement system.
- Gather the data.
- Verify process stability and ensure that the process is in control.
- Verify that the individual data are normally distributed.
- Obtain the process specifications.
- Determine process capability indices and interpret them.
- Update the process control plan.
Lower confidence Limit on the Cpk P.373
LCL Cpk = Cpk - Zα√(1/9n+(Cpk^2)/(2n-2))
Process Capability for Attributes Data P.374
Equivalent Ppk% = Equivalent Ppk based on percent
Equivalent Pp% = Equivalant Pp based on percent
Process Capability for Non-Normal Data P.377
- Find another known distribution that fits the data.
- Use nonlinear regression to fit a curve to the data.
- Transform the date to produce a second variable that is normally distributed.
- Box-Cox
- Johnson transformation
- Weibull
Natural Process Limit P.381
+- 3σ limits (6σ spread) around a process averages. (VoP)