基于改进相关性模型的子系统级故障诊断研究

Researches on Fault Diagnosis of Sub-system Based on Elevated Relative Model

对三种典型的相关性模型的基本理论、相关性矩阵生成方法进行分析和研究的基础上,结合某电路子系统级故障诊断的实际需求,从两个层面对相关性模型的故障诊断方法进行改进;为提高故障诊断过程中相关性矩阵的生成效率,首先提出了"近亲传递循环"算法;其次针对已有的诊断策略优化算法难以准确快速搜索到最优解的问题,提出新的测试点优选办法;实例已验证,对于多个测试节点的优选过程中,采用改进的相关性模型,依据故障检测信息量最大的原则选取的测试节点达到了快速、全面诊断的要求。

Based on the researches of three typical relative model theories and relative matrix generation methods, a elevated fault diagnosis method based on relative model is proposed which combined with the real acquirements of a sub-system. Firstly, in order to improve the generation efficiency, the ＂close relative transfer circulation＂ arithmetic is presented; secondly a new test points selective method is proposed too. All the methods are verified by realistic diagnosis experiments.