齿轮故障振动信号在强噪声背景下常常是非线性、非平稳的,具有不容易提取的特征.为解决故障特征难以提取的问题,提出了基于小波滤波与三次样条极值点插值拟合的故障特征提取方法.该方法首先对信号进行滤波降噪,之后提取数据的极值点,再运用三次样条对极值点进行插值拟合,进行频谱分析,提取故障特征频率.对齿轮局部断齿故障振动信号进行试验处理分析,证明,该方法提取齿轮局部断齿故障特征有效,比只用小波滤波降噪效果好.%Gear vibration signal is usually nonlinear and non-stationary in strong noise background, and it is not easy to extract fault feature from it. In order to solve the problem, the paper proposed a fault feature extraction method based on wavelet filtering and cubic spline interpolation fitting of extremum points. Firstly, the signal was filtered to de-noise, and then the extreme points of the data were extracted. After that, interpolation fitting of the extreme points by using cubic spline was conducted, and the frequency spectrum was analyzed to extract the fault feature frequency. The test for the vibration signal of partial gear tooth frature was carried out, and the results showed the method was effective in extracting the fault feature of partial gear tooth fracture and superior to the method with only wavelet filtering and de-noising being applied.
展开▼
