Cost-reduced determination of the measurement uncertainty of complex measurement processes

Steckbrief

Eckdaten

Duration:
01.01.2018 to 31.12.2019
Organizational Unit:
Chair of Production Metrology and Quality Management, Model-based Systems
Funding:
German Research Foundation DFG
Status:
Closed

Contact

Telephone

work Phone
+49 241 80 28211

E-Mail

  Test stand Copyright: WZL

Each measurement of a quantity is subject to uncertainty. This uncertainty of measurement can lead to miscalculations when testing product conformity. In order to be able to make a reliable decision on product conformity, the uncertainty of the measurement or of the entire measurement process must therefore be known and lie below a tolerance-dependent limit value.

With the existing standardized metrological methods such as VDA5 or MSA, it is possible to determine the measurement uncertainty for a large part of the measurement processes. One challenge remains the cost-reduced determination of measurement uncertainty for measurement processes in which none of the methods mentioned can be applied and more complex methods such as the procedure according to the "Guide to the expression of uncertainty in measurement" (GUM) are too complex. These measurement processes are summarized under the term "complex measurement processes". These include, for example, measurements with a large number of influencing factors or an underlying non-linear relationship. For these cases there is a need for a procedure in industry and research.

The goal of the DFG research project "MUKOM" is the development of a generally valid method for the determination of the measurement uncertainty of complex measurement processes with reduced effort. The method is based on the GUM and supplements it with a cost-reduced procedure for modelling measurement processes. This is achieved through the use of machine learning algorithms, the initial provision and integration of process steps for the relevance assessment of influences as well as the systematic integration of empirical knowledge. The result is a procedure that reduces the effort for the determination of measurement uncertainty and in particular the modelling of the measurement process and its influencing factors.