基于LMDPE与SVM机床刀具磨损故障诊断

2019-07-11

作者：冯胜单位：中海石油（中国）有限公司天津分公司

摘要：针对机床刀具磨损故障信号具有信号噪声大、频带混叠以及信噪比低的问题，提出了一种基于局部均值分解(Local Mean Decomposition，LMD)—排列熵(Permutation Entroy，PE)与支持向量机(Support Vector Machines，SVM)的机床刀具磨损故障诊断方法。首先对刀具磨损故障信号进行LMD分解，再根据相关系数去除噪声信号以及由于分解误差所带来的冗余信号后，选取合适的乘积分量(Product Function，PF)进行信号重构，然后将重构后的信号计算排列熵并通过标量量化处理后得到特征向量，最终将特征向量输入到已训练完成的支持向量机中来判别刀具的磨损状态，试验结果验证了该方法对机床刀具磨损故障诊断的有效性和实用性。

关键词：刀具磨损；局部均值分解；排列熵；支持向量机；故障诊断

中图分类号：TG707；TH117文献标志码：ADOI：10.3969/j.issn.1000-7008.2019.08.027

Fault Diagnosis of Tool Wear for Machine Tool Based on LMDPE and SVM

Feng Sheng

Abstract:Aiming at the characteristics of machine tool wear signal with large signal noise，frequency band aliasing and low signaltonoise ratio，a machine tool wear fault based on local mean decomposition (LMD)permutation entropy (PE) and support vector machine (SVM) is proposed diagnosis method.The method firstly performs local mean decomposition on the tool wear signal.After removing the noise signal and the redundant signal generated by the decomposition，the appropriate product function (PF) is selected for signal reconstruction，and then the reconstructed signal is calculated and the entropy value is calculated.The scalar quantization is used to obtain the feature vector，and finally the feature vector is input into the support vector machine to determine the wear state of the tool.The test results verify the validity and practicability of the method for discriminating the tool wear failure of the machine tool.

Keywords:tool wear；local mean decomposition；permutation entropy；support vector machine；fault diagnosis