基于模糊神经网络的数据通信子系统全局故障诊断

Global Fault Diagnosis to Data Communication Subsystem based on Fuzzy Neural Network

基于模糊神经网络理论,提出一种基于模型构建的数据通信子系统(DCS)全局故障诊断方法。全局故障诊断模型的输入空间由故障征兆集组成,诊断过程由全局故障诊断规则实现,输出空间由故障类别集组成。基于对DCS系统结构的分析,选取了一些关键设备信息作为故障征兆信息。将故障征兆信息中的物理向量分析转化为算术数值判断,创建决策矩阵,构建全局故障诊断规则,实现了故障类别综合判定,从而完成全局故障诊断模型构建。以工程实例中的DCS典型故障类别为验证对象,对全局故障诊断模型进行了试验验证。该方法丰富了DCS故障诊断方法,总体精度可达到91.29%。

Based on fuzzy neural network, a global fault diagnosis method to Data Communication Subsystem (DCS) based on model building is introduced. Input space of global fault diagnosis model is mainly comprised of fault symptom set, diagnostic procedure is realized by global fault diagnosis rule, and output space of global fault diagnosis model was mainly comprised of fault type set. Besed on the analysis of DCS framework, some key equipment information is selected as fault symptom information. Through transforming physical vector analysis of fault symptom information to arithmetic numerical value judgement, creating decision matrix and constructing global fault diagnosis rule, global fault diagnosis model buildup is completed and synthetic judgement of fault type is realized. Regarding the case studies of typical fault types of DCS as validation objectives, experiment validation is carried out by global fault diagnosis model. The experiment result indicated that global fault diagnosis model has better to analyze cause and effect between fault symtom and fault types of DCS, enriched fault diagnosis methods of DCS, and collectivity precision is 91.29%.