Measurement System Evaluation
The purpose of any measurement system analysis is to provide the best understanding about variation sources that can change the results generated by the system. This understanding allows you to quantify and inform the hedging of specific measurement systems.
As any process, the distribution that you can use to describe the variation of measurement system can be defined by:
1) Location:
| • | stability |
| • | Tendency |
| • | linearity |
2) Width or dispersion
| • | repeatability |
| • | reproducibility |
The primordial hypothesis of these techniques is that parts, or properties that has been measured, do not be changed or destructed by measurement.
Measurement System Variation Types
Normally, measurement is seen as exact. And, almost always, analysis and conclusions are based on this assumption. Variation in the system measurement is not always noticed; once this variation can affect individual measurement and, consequently, decisions based on this data.
Measurement system error can be classified by five types:
| • | tendencies |
| • | repeatability |
| • | reproducibility |
| • | stability |
| • | linearity |
One of the purposes of a measurement system study is to obtain information related to size and variation types of measurement associated with a measurement system when this one interacts with its environment. This is a valuable information as for the production common process is more practical to acknowledge calibration deviation and repeatability, and also to establish reasonable limits to them than to provide extremely exact measurement devices with high repeatability. Applications related to this study provide:
a) A standard to accept new measurement equipment.
b) A comparison between one measurement device and another.
c) A base to rank a measurement device that seems to be deficient.
d) A comparison for a measurement equipment before and after it has been repaired.
e) A necessary component to calculate process variation, and acceptance to a productive process.
f) Information needed to establish a measurement Device Performance Curve (CDD), to which
The follow definitions helps to report the error types or variations associated with a measurement system, so that each term be clearly understood during the next discussions. Every definition is supported by a picture in order to show you, graphically, the mean of each terms, being:
| • | Tendency |