Practical Applications of Bayesian Reliability

5Reliability Demonstration Testing

There are two types of reliability demonstration tests: substantiation tests and reliability tests. The purpose of a substantiation test is to demonstrate that a redesigned part or component has either eliminated or significantly improved a known failure mode. In this case, the distribution of the existing failure mode is assumed to be known (e.g. the Weibull scale parameter η and the shape parameter β are known). This is a classical statistical concept where we assume that the parameters of a distribution are fixed. A frequentist test can be developed with a specified number of test units or specified test duration to show that the new design is significantly better than the old design.

The purpose of a reliability test is to demonstrate that a certain reliability objective has been achieved. A reliability objective may be stated as a desired reliability goal (e.g. 95%, 99%, etc.) at a specified test duration (e.g. 5 days or 150 000 cycles, etc.) or mean time to failure (MTTF). Again, in this case a frequentist test can be developed with a fixed number of test units or fixed test duration to demonstrate that the specified reliability objective is met. In reliability or substantiation testing involving a fixed number of test units, how long each unit needs to be tested must be determined. A test plan specifies how many units to test and for how long. It also specifies a stopping rule which can be stated in terms of number of failures or the ...