摘要
实测信号往往受到多种因素的干扰,如高频噪声。提出了一一种小波降噪神经网络的故障诊断方法,利 用小波的多重分辨率分析,有效降低高频噪声干扰,从而简化了有效特征信号的提取。建立了基于小波变换和BP 神经网络的混合诊断模型,成功地对故障进行了智能诊断。最后实验验证了此种方法的有效性。
The combination usage of wavelet transform and artificial neural network in the diagnosis of rotation machine. The signal of rotation machine has analyzed based on wavelet theory, and a feasible means is put forward for the signal de - noising, so fault character can be simplified by applying wavelet transform upon the FFT. A mixed model based on WT and BP network is constructed, which provides a feasible technique support for fault diagnosis of rotation machine.
出处
《噪声与振动控制》
CSCD
北大核心
2005年第5期24-28,共5页
Noise and Vibration Control
基金
华北电力大学博士基金
项目编号:92104392