摘要
针对旋转机械故障征兆与故障模式映射的复杂性,将反向传播(BP)网络、径向基(RBF)网络和概率网络(PNN)用于风机进行故障诊断,并比较了3种网络的诊断精度。以风机振动信号的7段频谱能量峰值作为故障特征,采用训练好的神经网络进行故障辨识,结果表明,RBF网识别精度高于PNN网络,BP网络表现较差。
Aiming the problems of mapping complex between fault symptoms and fault patterns of rotary machines, BP neural network, RBF neural network and PNN were utilized to diagnose a fan and their classification results were compared. Peak energy of seven frequency bands for vibration signals of a fan being the fault symptoms, neural networks are trained to diagnose a fan, the results shows that RBF network has the best classification accuracy and PNN has a better result.
出处
《价值工程》
2012年第18期24-25,共2页
Value Engineering
