LS-SVM回归算法在刀具磨损量预测中的应用

发布时间：2018-04-16 19:20

本文选题：刀具磨损量预测 + 最小二乘支持向量机　；参考：《中国机械工程》2015年02期

【摘要】：提出了基于最小二乘支持向量机回归算法的刀具磨损量预测方法。该方法首先利用经验模态分解算法对非线性、非平稳的声发射信号进行平稳化处理,得到了若干个固有模态函数;然后建立了每个固有模态函数的自回归模型,并提取模型系数构造特征向量;最后采用最小二乘支持向量机回归算法实现了刀具磨损量的预测。该方法与神经网络预测算法相比,具有更高的预测准确率,可有效预测当前切削状态下10s后的刀具磨损量。
[Abstract]:A tool wear prediction method based on least squares support vector machine (LS-SVM) regression algorithm is proposed.Firstly, the empirical mode decomposition algorithm is used to stabilize the nonlinear and non-stationary acoustic emission signals, and some inherent modal functions are obtained, and then the autoregressive model of each inherent mode function is established.The model coefficients are extracted to construct the eigenvector and the least square support vector machine (LS-SVM) regression algorithm is used to predict the tool wear.Compared with the neural network prediction algorithm, this method has higher prediction accuracy and can effectively predict the tool wear after 10 seconds in the current cutting state.
【作者单位】：东北电力大学;吉林石化工程设计有限公司;
【基金】：东北电力大学博士科研启动基金资助项目(BSJXM-201115)
【分类号】：TG71;TH117.1

【参考文献】