Reliability Demonstration: Testing and Acceptance Procedures
Reliability demonstration is a procedure for testing whether the reliability of a given device (system) at a certain age is sufficiently high. More precisely, a time point t0 and a desired reliability R0 are specified, and we wish to test whether the reliability of the device at age t0, R(t0), satisfies the requirement that R(t0) ≥ R0. If the life distribution of the device is completely known, including all parameters, there is no problem of reliability demonstration — one computes R(t0) exactly and determines whether R(t0) ≥ R0. If, as is generally the case, either the life distribution or its parameters are unknown, then the problem of reliability demonstration is that of obtaining suitable data and using them to test the statistical hypothesis that R(t0) ≥ R0 versus the alternative that R(t0) < R0. Thus, the theory of testing statistical hypotheses provides the tools for reliability demonstration. In the present section we review some of the basic notions of hypothesis testing as they pertain to reliability demonstration.
Unable to display preview. Download preview PDF.