基于验前信息的测试性验证试验方案确定方法

针对目前测试性验证试验方案样本量过大、工程上实现困难的问题,提出了基于验前信息的复杂设备的Bayes测试性验证试验方案.首先,利用Beta分布对测试性验前信息的不确定性进行描述,运用不同来源的验前信息确定验前分布超参数;然后,定义了验前分布不确定性测度和支持度作为验前信息加权因子,设计了相应的融合算法;接着,利用融合后的验前信息建立成败型装备测试性验证试验方案的Bayes决策模型;最后,通过实例分析表明,与经典验证试验方案相比,新方案减少试验样本量40%左右,又克服了传统Bayes验证试验方案的冒进.

Method of confirming testability verification test scheme based on prior information

Existing testability verification test schemes need a large number of fault samples and the engineering implementation is difficult. To solve this problem, the Bayes testability verification test scheme was proposed based on the prior information for complex equipment. Firstly, the uncertainty of testability prior information was described using Beta distribution and the prior distribution hyper-parameter was determined by the prior information from different sources. Then, the uncertainty measure and supporting degree were proposed as the weight coefficient of prior information, and the corresponded fusion algorithm was designed. Finally, the Bayes decision model was estabilished for the testability qualification test scheme of the binomial equipments using the fused prior information. Compared with the classical counterpart, the new test scheme reduces the fault sample size by up to 40% or so, and avoids the aggressive of the traditional Bayes testability verification test scheme.