초록 close

Degradation tests can be a useful tool for assessing the reliability when few or even no failures are expected in a life tests. In this paper, we use a simple random coefficient degradation model to compare statistical efficiency of degradation tests and traditional life tests.If failure is defined in terms of a critical level of degradation, a degradation model defines a time-to-failure distribution, i. e., when a distribution of degradation rate follows a Reciprocal-Weibull(or Lognormal) distribution, a time-to-failure distribution follows an Weibull(or Lognormal) distribution under a simple random coefficient degradation model.Compared with maximum likelihoods estimates of parameters of two tests in terms of the bias and mean squared error for a range of practical testing situations, degradation tests are in general more precision than life tests.