基于渐近式权值小波降噪和Adaboost算法的液压泵故障诊断

Fault Diagnosis for Hydraulic Pump Based on Gradual Asymptotic Weight Selection of Wavelet and Adaboost

为了解决液压泵早期故障诊断难的问题,提出了一种基于渐近式权值小波降噪和Adaboost算法的液压泵故障诊断方法。针对早期故障特征难以有效提取的问题,根据最优化理论,通过对传统小波分析方法得到的信号进行渐近式权值的选择,得到了信噪比较好的降噪信号,并从中选取了最优特征集。同时,针对神经网络过学习和欠学习的现象,采用Adaboost算法对最优特征进行训练,实现了对不同故障类型的识别。实验结果表明,渐近式权值小波降噪能有效地去除噪声,提高信噪比,较为有效地提取最优故障特征;与BP神经网络相比,Adaboost算法具有更高的故障识别精度。

In order to solve the problem of identifying incipient faults of a hydraulic pump,a novel method of fault diagnosis based on gradual asymptotic weight selection of wavelet and Adaboost ensemble was proposed.Aiming at the features of incipient faults not Abstracted effectively,based on optimization theory,selection of gradual asymptotic weight by using the signals from traditional wavelet was to get the higher SNR factor of denoised signals.The denoised signals were used to select the optimal features.Then,aiming at the problem of neural network＇s over-learning and under-learning,the optimal features were trained with Adaboost algorithm to identify the different fault cases.Testing results show,compared with traditional wavelet,gradual asymptotic weight selection of wavelet can denoise,improve SNR factor,and Abstract the optimal fault features effectively.Adaboost algorithm has a higher classification success rate than the BP neural network.