摘要
本文采用纳秒脉冲激光铣削方法对尺寸小于1000μm,特征尺寸小于100μm的微锥体进行加工,激光铣削过程中脉冲激光与材料相互作用复杂,因此加工结果难以预测。本文首次以微圆锥的形状误差为预测目标,包括锥角误差(ATα)、直径误差(TDS)和对称性误差(ES),首先采用正交试验对显著影响加工结果的参数进行选择,然后利用采用不敏感损失函数的支持向量回归机(ε-SVR)建立激光微烧蚀加工预测模型,并且根据参数对预测结果的影响对参数进行择优选取,实验结果表明,较优参数的ε-SVR模型预测得到ATα=3.8°,TDS=5.35μm,ES=1.15μm。因此,ε-SVR适用于激光铣削加工微锥的误差预测,并对激光铣削加工其他微三维结构提供了参考。
In the paper,nanosecond pulsed laser milling method is used to process micro cones with size less than 1000μm and feature size less than 100μm,the interaction between pulsed laser and material is complex in the process,hence,the processing results are ambitious to predict.The present paper takes the form deviation of cone as the research object for the first time,including cone angel error(ATα),diameter error(TDS)and symmetry error(ES).Firstly,orthogonal experiments are used to select the parameters that significantly affect the processing results,then,a prediction model of laser micro ablation processing is established by using support vector regression machine(SVR)with insensitive loss function(ε-SVR),and according to the influence of the parameters on the prediction results,the parameters are selected optimally,results show thatε-SVR model with selected parameters receive forecasted values:ATα=3.8°,TDS=5.35μm,ES=1.15μm.Therefore,ε-SVR is suitable for the prediction of laser milling micro-cone processing,it could provide a reference for parameter optimization of laser milling other micro-3 d structures.
作者
邵琰杰
陈扬枝
谢雄敦
肖小平
Shao Yanjie;Chen Yangzhi;Xie Xiongdun;Xiao Xiaoping(School of Mechanical and Automotive Engineering,South China University of Technology,Guangzhou,Guangdong510640,China)
出处
《应用激光》
CSCD
北大核心
2020年第4期730-734,共5页
Applied Laser
基金
国家自然科学基金(项目编号:51575191)
清华大学摩擦学国家重点基金项目(项目编号:SKLTKF17BB04)
中央高校项目(项目编号:2017B0071,2018PY12)
关键词
激光铣削
加工预测
支持向量回归
形状误差
微加工
laser milling
micro machining
form deviation
support vector regression
processing prediction
作者简介
邵琰杰(1997-),男,硕士研究生,研究方向:激光减材微纳制造。E-mail:913704036@qq.com;通信作者:陈扬枝(1965-),男,博士,教授,研究方向:线齿轮、摩擦学、微机械。E-mail:meyzchen@scut.edu.cn
