针对数控机床的主轴故障,将经验模态分解(EMD)方法和支持向量机(SVM)相结合,用于故障诊断。采用EMD将信号分解成具有不同特征尺度的本征式分量IMF,分析各IMF,通过求取均方根值提取各特征向量,然后将各特征向量输入支持向量机,建立故障分类器进行状态识别。实验结果表明,预测结果完全正确,该方法有效。
The empirical mode decomposition (EMD)method and support vector machine (SVM) are combined together and applied to fault diagnosis of numerical control machine tool spindle in this investigation. The original signal will be decomposed into intrinsic mode functions(IMF) within various features of scale by the EMD. Thus the eigenvector is estimated by analyzing the IMF and is input into the support vector machine classifier to distinguish the conditions. The obtained experimental results achieve ideal effect with the right prediction results.