摘要
通过对电机发生故障情况下的振动信号进行小波包分解及重构,计算出信号在分解的各个频段内的能量谱,进行归一化后作为神经网络的训练样本,以此来实现电机故障类型的诊断。
The vibration signals in the cases of motor fault are decomposed and reconstructed based on wavelet packet, and the energy spectrum distribution of the signals is calculated. The normalized results can be taken as the training samples for the neural network algorithm for the diagnosis of motor fault types.
出处
《长春工业大学学报》
CAS
2013年第4期387-390,共4页
Journal of Changchun University of Technology
基金
吉林省重大科技攻关项目(10ZDGG002)
关键词
小波包
故障诊断
能量谱
故障特征
wavelet packet
fault diagnosis
energy spectrum
fault feature.