基于改进EMD和LS-SVM的刀具磨损状态识别

针对经验模态分解(empirical mode decomposition, EMD)的端点效应、停止准则和虚假分量作了改进处理, 通过对仿真信号的对比验证证明了改进方法的可行性.采集切削加工中的声发射(acoustic emission, AE)信号并对AE信号运用改进EMD方法分解为若干个固有模态函数(intrinsic mode function, IMF)分量, 利用IMF分量和原始信号的相关系数作为判断依据, 剔除分解中产生的虚假分量, 然后提取IMF分量的归一化能量值并将其作为特征向量.将提取的特征向量分为2组:一组用于对最小二乘支持向量机(least squares support vector machine, LS-SVM)训练;另一组用于识别刀具磨损状态.实验结果表明该方法可有效地表征刀具的磨损状态.

State Recognition of Tool Wear Based on Improved Empirical Mode Decomposition and Least Squares Support Vector Machine

End effect, stop standard and artificial components of empirical mode decomposition (EMD) are improved in this paper, and the contrast test of simulation signal verifies the feasibility of the improved method. Acoustic emission signals were collected in the cutting process and the improved empirical mode decomposition method was used to decompose them into a series of intrinsic mode function components. Artificial components that were produced in the decomposition were then removed according to correlation coefficients of the IMF components and original signal, and normalized energy of the IMF components was extracted as characteristic vector. The characteristic vector was divided into two groups: one group was used to train the least squares support vector machine and the other was used to identify the tool wear state. The experiment result shows that it can effectively characterize tool wear states.