论文部分内容阅读
Ｂａｙｅｓ方法是在小子样产品可靠性评定中应用较多的一种方法，但产品验前分布确定得是否合理，将直接影响评定结果。在采用Ｂａｙｅｓ方法对鱼雷可靠性进行评定时，常常是将成败型分系统试验的验前分布取为β（０，０）。这种对鱼雷的成败型分系统验前分布的取法是不够科学的。本文从理论上和工程实践上说明了鱼雷产品的成败型分系统验前分布应取为β（１／２，１／２），从而给出了由分系统试验数据折合成系统试验数据的Ｂａｙｅｓ方法修正公式。
The Bayes method is one of more methods used in the reliability assessment of small sample products. However, the pre-test distribution of products is reasonable and will directly affect the assessment results. When using Bayes method to assess the reliability of torpedo, the pre-test distribution of the success or failure subsystem test is often taken as β (0,0). This torpedo success or failure subsystems pretrial distribution method is not scientific enough. This paper shows theoretically and engineering practice that the pre-test distribution of the success or failure subsystems of torpedo products should be taken as β (1/2, 1/2), so that the Bayesian data deducted from the subsystems test data into system test data Method to correct the formula.