基于小波包-GABP的滚动轴承故障诊断分析

Fault Diagnosis of Rolling Bearing Based on Wavelet Packet and GABP

为提高诊断滚动轴承故障的效率和准确率,本文将小波包变换、BP神经网络和遗传算法三者相结合,提出了一种基于小波包和GABP神经网络的故障诊断模型。由小波包的分解与重构在滚动轴承故障原始信号中提取有效的故障特征向量,并利用遗传算法优化BP神经网络,然后训练和诊断滚动轴承信号的故障类型。同时,运用Matlab软件把采集的数据进行仿真分析。仿真结果表明,相对于传统BP神经网络,利用遗传算法优化的神经网络对故障的诊断正确率更高,并且收敛速度较快,说明由遗传算法优化的BP神经网络在故障诊断方面具有较好的效果,而且遗传算法的引入使轴承故障诊断的适应度和准确率更高。该研究为滚动轴承的故障诊断提供了理论基础。

In order to improve the efficiency and accuracy of rolling bearing fault diagnosis, a system of fault di-agnosis which combines wavelet packet with genetic algorithm and back propagation neural network is designed in this paper, and a fault diagnosis model based on wavelet packet and GABP neural network is put forward. The fault characteristic vectors in rolling bearing fault signal are effectively obtained by wavelet packet decom-position and reconstruction, and back propagation neural network is optimized by genetic algorithm, and then the fault types of rolling bearing are trained and diagnosed. The Matlab software is used to simulate the collect-ed data in the same time. The result of simulation shows that in contrast to the traditional back propagation neural network, the precision of bearing fault diagnosis is higher, and the convergence rate is faster. It shows that the back propagation neural network optimized by genetic algorithm has good effect in fault diagnosis, and the introduction of genetic algorithm makes the fault diagnosis of bearing more adaptable and accurate. The study provides a theoretical basis for fault diagnosis of rolling bearing.