利用马氏距离判别法准确实现对裂纹的识别

Crack identification with mahalanobis distance discrimination method

马氏距离可以有效地对拉深件初始裂纹声发射信号进行识别,通过采集拉深件畚斗在不同润滑条件下声发射信号、选用Matlab软件编程对采集到的声发射信号进行局域波分解得到自适应的本征模函数分量,有效提出各组IMF的能量值组成声发射信号的特征参数,计算出无裂纹、有微裂纹和明显裂纹三种状态下的马氏距离作为标准值,通过其他组的马氏距离与标准值进行比较,准确实现对拉深件裂纹状态的识别,也反映了板料在拉深过程中摩擦性能。

Mahalanobis distance can be used to identify effectively initial cracks with acoustic emission signals. Here, the acoustic emission signals of a drawing bucket were acquired under different lubricating conditions. Then, MATLAB software was adopted to do the local field wave decomposition of the collected acoustic emission signals to obtain adaptive intrinsic mode functions （IMFs）. The acoustic energy of each IMF was extracted to form the characteristic parameters of the collected acoustic emission signals. Mahalanobis distances for 3 statuses of no crack, crack and obvious crack were calculated as the standard values, the other groups of Mahalanobis distances were compared with the standard values to realize the crack identification of a drawing part and to reflect the friction behavior of a sheet metal in a drawing process.