Mathematically analysis of measurement instrumentation on the test benches at DLR/Lampoldshausen

Kurzfassung

To appraise the quality of measurements of a test bench, different statistical methods can be used. Two statistical methods are described in this paper, the classical Gaussian error propagation and the general uncertainty method (GUM). We will show how these statistical methods can be used to give a reliable indication of the quality of the measurements. We will also show the limits as well as the advantages and drawbacks of the two methods. The statistical results of the two methods indicate measurement problems at an early state and furthermore it is possible with these methods to narrow down the sources of an error. Hence these methods help to identify problems in measurement chains swiftly enabling repairs to be carried out before data loss occurs. Moreover these results can be taken into consideration and lead to criteria for measurement quality standards.