滚动轴承的故障特征提取

Picking up Fault Character of Rolling Bearings

滚动轴承故障特征提取采用在C++ Builder中嵌入Matlab对载波信号消噪后,将振动信号用小波包进行3层分解,提取各节点重构信号的方差,作为故障特征参量,归一化后输入神经网络进行故障诊断。其交互步骤是用C++ implib命令将外部def文件生成含有Matlab函数的lib文件,并将其加到新建工程中,编写目标函数。Matlab完成数值分析处理、绘制数据图形。

After noise of carrier signal was erased with C++ Builder embed Matlab, vibration signal was decomposed into three layers with wavelet packet, and variance of reconstructed signal of each node picked up used as characteristic parameter of fault. After normalization processing, characteristic parameter of fault was sent in neural network, thereby, picking up fault character and fault diagnoses of rolling bearings was realized. Interactive step realized by software: external def files was created into lib files which including Matlab functions with C++ implib command, lib files were added in new project, and object function was compiled, numerical value analysis and protracting data figure was finished by Matlab.