摘要： 通过构造模糊神经网络分类器,将基于模糊神经网络的数据融合方法引入到故障诊断中。通过在机组中设置温度传感器测量冷却介质进口温差、定子温度、冷却介质出口温差等模拟量,并求出各传感器对故障的隶属度,然后通过基于信息融合故障诊断算法确定故障 原因,并且通过单传感器与多传感器融合结果的对比表明,多传感器数据融合诊断比单个传感器具有更高的准确性。

关键词: 信息融合, 故障诊断, 模糊神经网络, 隶属函数

Abstract: Data fusing method is introduced in this paper via constructing blur neural network model. In order to measure stimulant data difference in input temperature of refrigerant, stator temperature and output temperature of refrigerant multi-sensors are set in the hydro-generator group and membership of every sensor about fault is computed and the cause of fault is determined via data fusing method .At last via contrasting fusing result of single sensor and fusing result of multi-sensor it is indicated that fault diagnosis of multi-sensor data fusing is more exact than those of single sensor.

中图分类号: 

• TP183