基于能量特征的刹车片内部缺陷检测方法

Method of Internal Defect Detection of Brake Pads Based on Energy Features

为解决刹车片内部缺陷检测的难题,提出了一种通过分析敲击信号对刹车片内部缺陷进行检测的方法。首先,截取通过敲击获得的声音信号,然后利用小波包分解和变分模态分解(VMD)提取能量特征,根据可分性判据选择最优特征作为分类特征,最后运用K最近邻(KNN)算法对故障特征进行识别,获得准确的类别信息,准确率达到了96%以上,对刹车片内部是否存在缺陷做出有效判断。通过对实际信号进行实验验证,表明该方法能够准确的识别刹车片的好坏,提高检测效率和精度。

In order to solve the the brake pads internal defect detection problem, a detection method based on knocking signal is presented. First, we intercept the knocking signal, then extracting energy features by wavelet packet decomposition and Variational mode decomposition, select the optimal feature according to the separability criterion as the classification feature, finally, using KNN algorithm to identify fault feature and obtain accurate classification information, the accuracy is up to 96% , and effective judgment is made on whether there are defects in the brake pads. The experimental results show that the method can identify the brake pads accurately and improve the detection efficiency and accuracy.