多传感器融合的因果映射网故障诊断方法

Fault Diagnosis Method Based on Causality Mapping Net for Multi-sensor Fusion

针对异构、时变、耦合的复杂系统故障诊断问题,提出一种基于多传感器融合的因果映射网故障诊断方法。以时态数据挖掘为手段,实现包含原因层和结果层的快速映射网络。通过因果度定义节点连接强度,通过依赖度和影响度进行故障位置定位和影响预测。应用实例表明方法具有机理简单、算法快捷、结果直观优点。

For the fault diagnosis problem of complex system which has multiform architecture, time-varying character and coupling components, a causality mapping net methods based on multi-sensor fusion was proposed. By means of timing data mining, a fast mapping net with cause layer and effect layer was realized firstly. Then the node connection intension was defined by causality-degree, Aflerwards, the fault position was located by rely-degree and the fault influence was predicted by effect-degree. An application example demonstrates the method has advantages such as simple mechanism, fast arithmetic and visual result.