基于信息融合技术的电机故障诊断

为了从多方面反映电机系统状态,实现对电机故障模式的自动识别和准确诊断,将数据融合技术与神经网络相结合,建立电机故障诊断系统。首先在数据融合级上对故障特征量进行分类处理,然后采用多层神经网络进行故障特征级融合和电机故障的局部诊断,获得彼此独立的证据,再运用DS(DempserShafer)证据理论融合算法对各证据进行融合,最终实现对电机故障的准确诊断。诊断测试试验证明,该诊断系统提高了电机故障诊断的精度,能够满足诊断的实时性要求。

Motor Fault Diagnosis Based on Information Fusion Technology

The motor fusion diagnosis system was built for reflecting the motor system state in multi-aspect, realizing automatically identifying motor fault patterns and accurately diagnosing the faults by using neural network and evidence theory. After fault feature data were classed and processed in data fusion level, multi-neural networks were used to carry on fusion in motor feature level and carry on local motor fault diagnosis, in order to acquire independent evidence each other.Then D-S evidence theory fusion algorithm was used to fuse every evidence.Accurate motor fault diagnosis was fulfilled in the end.The diagnostic tests prove that the diagnosis system can improve the diagnostic precision and satisfy the requirement for real time.