基于变量敏感度筛选的回归型支持向量机的数控机床热误差预测  

Thermal Error Prediction of CNC Machine Tools Based on Regression Support Vector Machine with Variable Sensitivity Screening

作　　者：李铁军[1] 崔尚仪 张义民[1] LI Tie-jun;CUI Shang-yi;ZHANG Yi-min(Equipment Reliability Institute,Shenyang University of Chemical Technology,Liaoning Shenyang 110142,China;Machinery and Power Engineering College,Shenyang University of Chemical Technology,Liaoning Shenyang 110142,China)

机构地区：[1]沈阳化工大学装备可靠性研究所,辽宁沈阳110142 [2]沈阳化工大学机械与动力工程学院,辽宁沈阳110142

出　　处：《机械设计与制造》2024年第9期41-43,50,共4页Machinery Design & Manufacture

基　　金：大型重载滚动轴承的可靠性和寿命预测的理论与方法研究—NSFC-辽宁联合基金(U1708254)。

摘　　要：随着机械制造行业的迅猛发展,对于数控机床的定位精度要求越来越高。为了提高机床定位精度,建立了基于变量敏感度筛选与回归型支持向量机(SVR)混合模型,并将其用于数控机床热误差预测方法。该方法基于对变量敏感度分析,筛选掉敏感度低的干扰自变量。本方法与基本SVR模型对数控机床热误差预测值进行对比,结果表明基本SVR受到敏感度低的干扰自变量影响,预测结果与实测热误差结果偏差较大;经过变量敏感度筛选之后的SVR混合模型预测值具有更高的准确度,验证了此模型的可行性。With the rapid development of machinery manufacturing industry,the positioning accuracy of CNC machine tools are more and more demanding.In order to improve the positioning accuracy of CNC machine tools,a hybrid model based on variable sensitivity screening and regression support vector machine(SVR)was established to predict the thermal errors of CNC machine tools.This method is based on the sensitivity analysis of variables,and the interference independent variables with low sensitivity are screened out in time.Compared with basic SVR model for predicting the thermal errors of CNC machine tools,the result shows that the basic SVR is affected by the interference independent variables with low sensitivity,and the predicted results devi⁃ate greatly from the measured thermal error results.After the sensitivity screening of variables,the predicted values of the SVR model have higher accuracy,which verifies the feasibility of this model.

关 键 词：数控机床 回归型支持向量机 变量敏感度筛选 热误差 

分 类 号：TH16[机械工程—机械制造及自动化] TH133.33