摘要
为了提取强背景噪声下滚动轴承故障特征信息,提出了一种多小波预处理的约束独立分量分析(MW-cICA)算法。该算法首先对输入信号进行多小波降噪处理,提高信号信噪比;然后应用约束独立分量分析方法提取故障特征。与传统的小波独立分量分析(W-ICA)方法相比,该方法具有如下优势:1)由于多小波具有单小波所不能同时具有的正交性、对称性、紧支性和高阶消失矩等特点,因而对信号的降噪效果更加明显;2)引入参考信号作为约束条件,使得算法直接收敛于期望信号,提高了运算效率;3)建立基于故障模型的参考信号能够更加接近于真实期望信号,提高算法性能。仿真结果表明,多小波比单小波具有更好的降噪效果,基于故障振动模型的约束独立分量分析比传统的FastICA算法运算效率更高。将该算法运用于滚动轴承内圈故障试验中,可成功提取出内圈故障特征信号。
A novel approach based Multi-Wavelet and Constrain Independent Component Analysis （MW-cICA） was proposed to extract fault feature of rolling element bearing which is usually buried in strong background noise. Firstly, the input signal was processed to improve the index of SNR, and then clCA method was selected to extract fault feature. Comparing with the traditional Wavelet Independent Component Analysis （W-ICA） ,the advantages of the proposed method were shown as follows：1 ）because of the good characteristics of orthogonality, concurrent symmetry, short support and high order vanishing moment, the effect of noise reduction by multi-wavelet is more effective than scalar wavelet ;2 ）the reference signal was introduced as the constrain function to guide the convergence direction to the desired signal which can reduce the running time;3 ）the fault model reference signal was introduced to improve the performance of the proposed method, because it is similar to the real desired signal. The result of simulations was indicated that multi-wavelet is more effective than scalar wavelet, and the operation efficiency of clCA algorithm based fault model reference signal is better than traditional FastlCA method. The fault rolling element bearing experiment shows that the proposed method can effectively extract the inner fault feature.
出处
《现代制造工程》
CSCD
北大核心
2016年第10期126-134,共9页
Modern Manufacturing Engineering
关键词
多小波预处理的约束独立分量分析
故障模型
参考信号
滚动轴承
故障诊断
Multi-Wavelet and Contrain Independent Component Analysis （MW-cICA）
fault model
reference signal
rolling ele-ment bearing
fault diagnosis
