采用天鹰优化卷积神经网络的精密数控机床主轴热误差建模  

Thermal Error Modeling of Spindle for Precision CNC Machine Tool Based on AO-CNN

作　　者：李国龙[1] 陈孝勇 李喆裕 徐凯 唐晓东 王志远 LI Guolong;CHEN Xiaoyong;LI Zheyu;XU Kai;TANG Xiaodong;WANG Zhiyuan(State Key Laboratory of Mechanical Transmission,Chongqing University,Chongqing 400044,China)

机构地区：[1]重庆大学机械传动国家重点实验室,重庆400044

出　　处：《西安交通大学学报》2022年第8期51-61,共11页Journal of Xi'an Jiaotong University

基　　金：重庆市技术创新与应用发展专项重点资助项目(cstc2019jscx-mbdxX0034)。

摘　　要：针对数控机床因主轴热误差而严重影响加工精度等问题,结合求解最优解能力强的天鹰优化算法(AO)以及自学习和自适应能力强的卷积神经网络(CNN),提出一种采用AO-CNN的数控机床主轴热误差模型。根据磨齿加工过程特点,总结磨齿机主轴系统热变形规律,确定了X方向热误差为影响齿轮加工的主要因素;利用模糊C均值聚类(FCM)和相关系数法筛选出关键温度点;利用AO算法优化CNN结构的卷积核,并且建立AO-CNN的数控机床主轴X方向热误差预测模型。在2种不同转速的工况下对所建立模型的性能进行了验证,结果表明,采用AO-CNN进行热误差建模,数控机床X方向的热变形预测精度相比于CNN模型提高了15%,具有更加优越的预测精度。To accurately predict the thermal error of the spindle of a CNC machine tool and avoid serious impacts of such thermal error on gear machining accuracy,a thermal error model of grinding machine spindle based on AO-CNN is proposed by combining the convolutional neural network(CNN)with strong self-learning and self-adaptive abilities and the aquila optimizer(AO)with strong ability to solve the optimal solution.Firstly,the thermal deformation principle of the spindle and the grinding process were analyzed,and it is found that the thermal error in X direction is the main factor affecting the machining accuracy.Then,the key temperature points were selected by use of the fuzzy C-means clustering(FCM)algorithm with relevant coefficients.The convolution kernel of CNN structure was optimized by AO algorithm and the X-direction thermal error prediction model was established for the spindle of the CNC machine tool based on AO-CNN.Finally,the performance of the model was verified under two experimental conditions at different speeds.And the results show that the thermal error prediction accuracy of AO-CNN model in X direction of the CNC machine tool improved by 15%compared with the CNN model providing superior prediction accuracy.

关 键 词：数控机床主轴 热误差建模 卷积神经网络 天鹰优化器 模糊C均值聚类 

分 类 号：TH161[机械工程—机械制造及自动化]