基于两级信息融合的凝汽器故障诊断方法  

An Approach to Condenser Fault Diagnosis Based on Hierarchical Information Fusion

作　　者：侯炜[1] 田沛[1] 

机构地区：[1]华北电力大学,河北保定071003

出　　处：《汽轮机技术》2003年第6期392-394,共3页Turbine Technology

摘　　要：采用多级信息融合,提出了一种新的凝汽器故障诊断方法。首先,利用神经网络和距离测度的方法进行初级融合;然后,再利用D S证据理论进行决策级的融合。结果表明,该方法降低了神经网络误判的可能性,同时也得到比基于距离测度的方法更直观的判决结果。An approach to condenser fault diagnosis is analyzed by hierarchical information fusion. First of all, the omens of fault are fused by neural network and distance estimate; secondly, the transformed outputs of them are used as the inputs of D-S evidence theory. The result shows that the approach reduce the possibility of misanalysis of neural network, and it is more intuitionistic then the result of distance estimate.

关 键 词：信息融合 BP网络 距离测度 D—S证据理论 凝汽器 故障诊断 

分 类 号：TK264.11[动力工程及工程热物理—动力机械及工程]