Measurement system analysis

from Wikipedia, the free encyclopedia

As a measurement system analysis and measurement equipment capability analysis or test equipment capability analysis , short-MSA (English: Measurement System Analysis ), is defined as the analysis of the capability of measuring equipment and complete measurement systems in quality management and Six Sigma .

Measurement errors

In general, one differentiates between five different categories of measurement errors : accuracy, repeatability , comparative precision , linearity and stability.

Accuracy, correctness, systematic measurement error (English accuracy, trueness, bias )

The accuracy describes the degree of approximation of a measurement value to a true value of a measured variable and does not constitute a quantitative indication. The measurement accuracy is determined by repeated measurement of the same test piece as a deviation of the mean value from a reference value. The difference between the mean value and the true value is known as the systematic measurement error ( bias ). This difference is then used to make a statement about the accuracy (correctness) of the measuring device.

Repeatability, reproducibility (Engl. Repeatability )

To determine the repeatability, the same test object is measured several times in quick succession by the same operator and with the same measuring equipment at the same location. However, the test item is always put back between the individual measurements. The standard deviation of the measured values ​​is then a measure of the repeatability.

Comparative precision, traceability (English reproducibility )

To determine the comparative precision, measurements are made on the same test objects in accordance with a specified measuring method

  • by different operators,
  • in different places or
  • with multiple devices of the same type

carried out. As a rule, there are two or three operators who measure the same parts several times in a row, or the same operator repeats the same measuring process at different locations or with different devices. However, only one of these three variables (operator, location, device) is changed within an examination. The measure of the comparative precision is then the differences between the mean values ​​observed by each operator (or at each location or with each device).

Stability (Engl. Stability )

In order to examine the stability, several measurements of one and the same test object are carried out according to a specified measuring method with the same equipment at the same location and by the same operator at specified time intervals. The mean value of the measured values ​​is calculated after each measurement series. The differences between the mean values ​​observed at different times are then used as a measure of the stability of the measuring device.

Linearity (Engl. Linearity )

To examine the linearity , measurements are carried out on several test objects by the same operator and with the same measuring equipment at the same location and according to a specified procedure, the feature values ​​of which cover the entire range of values ​​that can be expected in practice. Each test item is measured several times. The mean value of the measured values ​​observed is then calculated for each test item. Then the difference between the true value and the observed mean value (cf. accuracy above) is calculated for each test item . If these differences are of different size and if these differences are so large that they cannot simply be explained as random scatter, the behavior of the measuring device is not linear.

Measurement system analysis methods

Every measurement system analysis is preceded by an analysis of the resolution of the measurement equipment used. Apart from this, there are mainly two examination methods used in practice: Procedure 1 ( type-1 study ) and Procedure 2 ( type-2 study ). The data collected in this way is often analyzed using statistical software packages such as JMP (software) , Minitab , qs-STAT-solara or QS-1-2-3-4.

resolution

This procedure examines whether the measured value display (analog or digital) is displayed with sufficient resolution in relation to the tolerance. In general, 5% of the tolerance is the limit. (Example: resolution 0.001 with a tolerance of 0.02)

Procedure 1 (English type-1 study )

This procedure examines the accuracy and repeatability of a measuring system. A standard with a known feature value is used for the investigation . The normal is measured 50 times (at least 25 times). The normal is put back after each measurement. Based on the standard deviation of the measured values ​​and the systematic measurement deviation, the indices Cg and Cgk are then calculated. The calculation steps for this are analogous to those of the process capability study ; A company-specific percentage of the feature tolerance or feature spread is used as the “tolerance range”.

Method 2 (English type-2 study , Gauge R&R study )

This procedure examines the repeatability and reproducibility of a measuring device (English repeatability and reproducibility , hence R&R , also Gauge R&R or Gage R&R , after English ga [u] ge , measuring device) and is only used when the measuring device is in accordance with the procedure 1 has been classified as capable. Here, ten parts, which should cover the entire scatter range of the measured feature as far as possible, are measured two or three times by three different operators (or at three different locations or with three different devices of the same type). None of the operators may see the results of the other operators. The parts should also be measured on each run in a randomized order so that the operator can never remember the measurement result from the previous run.

When the measurements are complete, an overall mean and a mean span value (based on the differences between the largest and smallest measurement the operator determined for each part) are calculated for each operator. The difference between the largest and the smallest operator mean value allows a statement to be made about the comparative precision; the overall mean value of the average range values ​​calculated for the individual operators is used to make a statement about the repeatability. Based on the repeatability and comparative precision, the total scatter of the measuring device is then calculated and related to the characteristic scatter or tolerance.

Method 3 ( type-3 study , R&R study )

Method 3 is a special case of method 2, which assumes that the operator cannot influence the measuring device or that the influence is negligible. Typical purposes are automated measuring systems.

literature

  • Edgar Dietrich, Alfred Schulze: Proof of suitability of test processes . Hanser Fachbuchverlag, Munich 2007, ISBN 978-3-446-22320-2 .

Individual evidence

  1. a b c Burghart Brinkmann, "International Dictionary of Metrology - Basic and General Terms and Associated Terms ( VIM )", German-English version, ISO / IEC Guide 99: 2007, corrected version 2012, 4th edition 2012, Beuth- Publisher, ISBN 978-3-410-22472-3
  2. a b c d e f g Dietrich, Schulze: Statistical procedures for machine and process qualification . Hanser Verlag, 1995, ISBN 3-446-17984-4 , pp. 282-309.

Web links