数控滚齿机工作台热-力变形分析及预测建模  

Thermal-force deformation analysis and prediction modeling of CNC gear hobbing machine workbench

作　　者：王四宝[1,2] 郭忠政[1,2] 马驰 王时龙 WANG Si-bao;GUO Zhong-zheng;MA Chi;WANG Shi-long(College of Mechanical and Vehicle Engineering,Chongqing University,Chongqing 400044,China;State Key Laboratory of Mechanical Transmissions,Chongqing University,Chongqing 400044,China)

机构地区：[1]重庆大学机械与运载工程学院,重庆400044 [2]重庆大学机械传动国家重点实验室,重庆400044

出　　处：《吉林大学学报（工学版）》2023年第10期2761-2772,共12页Journal of Jilin University:Engineering and Technology Edition

基　　金：国家重点研发计划项目(2019YFB1703701);国家自然科学基金重点项目(51635003);重庆市留学人员回国创业创新支持计划项目(cx2021035);重庆市科委自然科学基金项目(cstc2019jcyj-msxm0058);重庆市自然科学基金创新群体科学基金项目(cstc2019jcyj-cxttX0003)。

摘　　要：为了减小滚齿机工作台变形对加工精度的影响,对工作台热、力变形进行了研究,提出一种基于子种群自适应思维进化算法优化反向传播(SAMEA-BP)神经网络的滚齿机工作台热-力变形预测方法。通过SAMEA对BP神经网络的初始值、权重和阈值等参量进行调整,有效提升了基于神经网络的热-力变形预测准确度。结合K均值聚类策略和灰色关联分析(GRA)对影响热误差的温度测点进行耦合性和关联度分析,将热误差输入变量从8个测点减少到3个;针对滚齿加工中切削力导致的工作台变形,利用机床主轴电流表征切削力,并作为预测模型的输入变量。试验结果表明:本文模型平均预测精度为95.1%,与其他模型进行的对比分析验证了本文SAMEA-BP模型的有效性和泛化性。In order to reduce the influence of hobbing workbench deformation on machining accuracy,the thermal and mechanical deformation of the workbench were studied.A Subpopulation Adaptive Mind Evolution Algorithm-Back Propagation Neural Network was proposed for thermal-mechanical deformation prediction of hobbing workbench.The initial value,weight and threshold of BP Neural Network were adjusted by SAMEA,which effectively improved the prediction accuracy.Combining K-means clustering and grey correlation analysis(GRA),the thermal error input variables were reduced from eight to three.Aiming at the table deformation caused by cutting force in gear hobbing,the cutting force was represented by the current of machine tool spindle and used as the input variable of the prediction model.Experimental results show that the average prediction accuracy of the proposed SAMEA-BP model is 95.1%,and comparison with other models verifies the validity and generalization of the model.

关 键 词：热致误差 力致误差 SAMEA-BP神经网络 均值聚类算法 灰色关联分析 

分 类 号：TH161[机械工程—机械制造及自动化]