多传感器数据融合技术在故障诊断中的应用

Application of multi-sensors data fusion in fault diagnosis

利用多传感器数据融合的方法进行故障诊断,建立融合故障诊断系统。将故障诊断系统按数据融合的方法分为数据级融合模块、特征级融合模块和决策级融合模块。数据级融合模块主要对多传感器的测量信号进行处理,提取出故障诊断的特征信息。特征级融合模块采用3个结构相同的并行神经网络,一是进行局部诊断;二是获得决策级D-S证据理论的基本概率赋值。决策级采用D-S证据理论的方法对特征级局部诊断的结果加以融合,得到最终的诊断结果。利用此系统在汽轮机转子试验台架上进行了故障诊断,得到了令人满意的结果。

The multi-sensors data fusion method is applied to make a fault diagnosis, and build the fusion fault diagnosis system. The fault diagnosis system is classed into data fusion level module, feature level module and decision fusion level module according data fusion method. The data fusion level module mainly handles the metrical data of multi-sensors and extracts the feature of faults. The feature level module adopts three collateral neural networks which structures are same, its function is to do local diagnosis and to get basic probability assignment (BPA) of D-S evidence theory. The decision fusion level module uses D-S evidence theory to fuse the local diagnostic results of feature fusion level, then get the final diagnostic results. The fault diagnosis system is tested on the gas turbine rotor, and the satisfying results are gained.