针对目前测试性预计方法缺乏对不确定性全面考虑的现象,从测试和故障两个方面分析了测试性预计过程中的不确定性问题,基于贝叶斯网络测试性模型对不确定性测试问题进行了建模与分析,基于混合诊断模型对故障不确定性问题进行了建模与分析,并最终将贝叶斯网络测试性模型与混合诊断模型相融合,给出了一种兼容IEEEStd1522的基于混合诊断贝叶斯网络模型的测试性指标预计方法,测试性指标预计结果的可信度明显提高。

In view of ignorance of uncertain problem existed in test and diagnosis, current testability prediction method tends to be inaccurate. The hybrid diagnostic Bayesian networks can model the uncertain relationship between tests and failure sources, failure modes and functions. In the paper, a method of hybrid diagnostic Bayesian networks was given based on uncertain testability modeling and prediction, which is compatible with IEEE Std 1522. The model of hybrid diagnostic Bayesian Networks is a kind of enhanced diagnostic inference model which can provide testability prediction solution with high accuracy.