基于GA优化的VMD-SVM识别角度头故障特征

提出一种基于遗传算法优化的变分模态分解（variational mode decomposition，VMD）-支持向量机（support vector machine，SVM）方法来识别机床角度头故障特征。首先采用遗传算法对VMD算法的输入参数进行优化，将优化后的VMD算法用于振动信号的分解，得到各本征模态函数（IMF）后，求得对应的能量熵；然后通过SVM算法筛选出有效故障数据，再利用峭度和相关系数相结合的方法将其中的IMF筛选出来并重构信号；最后，对该信号作频谱分析，分析相关特征信息，识别并诊断出故障。根据仿真和实验结果，所提方法对于故障角度头的有效信号筛选正确率高，对于噪声抑制效果良好，特征提取快速有效，可用于机床故障诊断领域。

A VMD-SVM algorithm based on genetic algorithm optimization for angle head fault feature recognition

This paper presents a variational mode decomposition (VMD)-support vector machine (SVM) approach based on genetic algorithm optimization to identify angle head fault feature. A genetic algorithm is used to optimize the input parameters of the VMD algorithm, and the optimized VMD algorithm for the decomposition of vibration signals is then used to obtain the energy of each instrinsic mode function (IMF) and its respective energy entropy. An SVM algorithm is used to select the effective fault data. The IMF is filtered out, and the signal is reconstructed according to a combination of kurtosis and correlation coefficient. Spectral analysis is then performed to extract the relevant feature information and identify the fault. The results of the simulations and tests show that the proposed method has high accuracy for effective signal selection of the fault angle head, a good effect on noise suppression, and fast and effective feature extraction. The method proposed in this work can be used in machine tools fault diagnosis.