基于遗传算法的匹配追踪小波分析及其在非平稳故障信号中的应用

闰美观分析理论在信号检测、故障诊断领域已得广泛应用，尤其是在处理非平稳故障信号方面，小波分析更为有效，匹配追踪小波分析是专门处理高度不平稳信号的一种算法，但该算法每次计算需要不断地调整小波库中的各个参数来寻找最佳的匹配小波，增加了算法实现的难度，从而限制了该算法的应用，本文提出基于遗传算法的匹配追踪小波分析，把遗传算法良好的多参量全局寻优特性和匹配追踪的基本原理有机地结合起来，并采用了处理非平稳信号的有力工具Wigner谱。利用该方法分别对仿真信号和试验信号进行分析，结果表明本方法十分有效，而且对实测故障信号的分析诊断结果与实际十分吻合。

WAVELET ANALYSIS OF MATCHING PURSUIT BASED ON GENETIC ALGORITHM AND ITS APPLICATION IN NON-STATIONARY FAULT SIGNALS

The application of wavelet analysis in fault diagnosis is growing rapidly. Especially in dealing with non stationary fault signals, wavelet transform behaves much preponderance over Fourier Transform. Of all the wavelet analysis methods, matching pursuit is a developing technique which is specially used in analyzing high non stationary signal in recent years. But the main problem in realizing this algorithm is how to select the best parameters in the wavelet dictionary. In this paper, matching pursuit wavelet analysis based on genetic algorithm is presented. The characteristic of genetic algorithm for searching the best value in the whole domain makes it easy to complete the matching pursuit algorithm. The simulated and tested signals are analyzed using this method respectively. The result shows that this method is effective and suitable to diagnose the non stationary faults of impact. It is shown that matching pursuit wavelet analysis has many advantages over other wavelet analysis method and has a broad application prospect in the domain of non stationary fault diagnosis.