基于Elman神经网络的电机故障诊断

通过Elman神经网络训练采集到的电机特征频率振动信号,建立电机故障的模型.并以转子故障为对象,在电机实验台上分别采集水平和垂直方向上的振动信号,通过放大、滤波和A/D转换后,进行频谱分析.为验证网络对故障判断的效果,对故障电机采集振动信号后,代入网络检验.验证表明该网络模型可有效识别电机常见故障.

Study on Motor Fault Detection Based on Elman Neural Network

The motor fault model was established by using the frequency vibration signal of motor feature which acquired by Elman neural network. And the rotor fault was taken as the object; the vibration signals of level and uprightness directions were gathered in motor test platform; after amplification, filtering, and A/D conversion, the spectrum was analyzed. For testing the effect of network fault estimation, after acquiring vibration signal of fault motor, and the results were taken into network. The test showed that this network model could recognize the general faults of motor.