基于模糊积分的潜隐性故障状态转移过程分析

摘要

With the discussion of latent failure concept and the analysis of multi-state system Markov state transition process, a fuzzy integral fusion fault diagnosis method is proposed using HMM models based on weighted average method to determine the fuzzy density through the fuzzy integral value to achieve the recognition of comprehensive diagnosis of latent fault condition. A control module of radar fire control equipment unit as an example shows its effectiveness. This method overcomes difficulty in intelligent fault diagnosis of the training samples and improve the diagnostic accuracy of complex electronic equipment fault with good diagnostic capabilities.%探讨了潜隐性故障的概念,分析了多态系统状态转移的马尔可夫过程,提出了基于模糊积分信息融合技术的故障诊断方法,在利用HMM模型进行初级诊断的基础上,采用加权平均方法确定模糊密度,通过科学计算模糊积分值实现潜隐性故障状态综合诊断识别.最后以某雷达火控设备单元控制模块为例验证了其有效性.该方法克服了智能故障诊断训练样本获取难度大的问题,提高了诊断准确率,对复杂电子装备故障有较好的诊断能力.