Measurement system analysis (MSA)
The accuracy of quality improvement analyses and methodologies performed within an organization depends solely on the data on which they are based. Thus, it becomes essential to ensure that the measurement system is stable and capable of measuring required data before embarking on the process improvement efforts.
Measurement System Analysis (MSA) is defined as a mathematical and experimental method to determine the degree of variation within the measurement process, which will contribute to overall process variability. Measurement System Analysis is done on the basis of five parameters – stability, linearity, bias, reproducibility, and repeatability.
The purpose of measurement system analysis is to ensure the measurement system provides:
- adequate resolution
- results which are not unduly biased
- little variability when compared to specified tolerances
The general rule of acceptability for measurement system is:
- Under 10 percent - which suggests that the error is acceptable
- 10 percent to 30 percent - which suggests that the system can be acceptable depending upon the vitality of the application, the cost of measurement device, cost of repair, and other such factors
- Above 30 percent - which suggests that the error is unacceptable, and that the measurement system should be improved
It is stated by AIAG that the number of categories arrived at by the measurement system by dividing a process should be equal to or greater than 5. In addition to the number of distinct categories and the percentage error, it is also essential to review the graphical analyses over time to decide whether a measurement system is acceptable or not.
Thus, measurement system analysis is utilized for assessing the statistical properties of process measurement systems. This analysis is essential for measuring the business process capability of an organization and for meeting the specific needs of customers.
