Reliability evaluation method of torpedo loading based on zero-failure data

Abstract

Reliability evaluation is significant in torpedo reliability engineering. But it is difficult to evaluate the reliability of torpedo loading due to the high cost and long period of loading test. In this study, a Bayesian evaluation method based on information fusion is proposed by considering that the torpedo loading test data are small samples zero-failure ones. Specifically, the torpedo lifetime distribution obeys the exponential distribution and the prior distribution of the failure rate is the Gamma distribution. In this case, the information fusion model for determining the hyperparameter values of the Gamma distribution is firstly derived on the basis of the maximum entropy method and expert information. Then, the Bayesian estimate of the failure rate is obtained from the prior distribution. Finally, the comprehensive Bayesian evaluation method of failure rate is deduced by introducing failure information. The research results show that this method avoids the radical problem when using zero-failure data of small samples for reliability evaluation, and the evaluation results are also robust. This method not only provides an effective means for torpedo loading reliability evaluation, but also applies to the reliability evaluation of other test data systems with small samples zero-failure data.