基于BiLSTM-LSSVM的螺杆转子铣削加工廓形预测  

Profile Prediction of Screw Rotor Milling Based on BiLSTM-LSSVM

作　　者：李佳 孙兴伟[1,2] 赵泓荀[1,2] 穆士博 刘寅 杨赫然 LI Jia;SUN Xingwei;ZHAO Hongxun;MU Shibo;LIU Yin;YANG Heran(School of Mechanical Engineering,Shenyang University of technology,Shenyang 110870,China;Key Laboratory of Numerical Control Manufacturing Technology for Complex Surfaces of Liaoning Province,Shenyang University of technology,Shenyang 110870,China)

机构地区：[1]沈阳工业大学机械工程学院,沈阳110870 [2]沈阳工业大学辽宁省复杂曲面数控制造技术重点实验室,沈阳110870

出　　处：《组合机床与自动化加工技术》2024年第9期153-156,162,共5页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：辽宁省应用基础研究计划项目(2022JH2/101300214);国家自然科学基金项目(52005346);2022年度辽宁省教育厅高等学校基本科研项目面上项目(LJKMZ20220459)。

摘　　要：针对螺杆转子盘铣刀加工过程中的轮廓预测问题,提出了基于双向长短时神经网络-最小二乘支持向量机(BiLSTM-LSSVM)的螺杆廓形预测方法。首先,对加工过程中的振动信号进行采集并进行降噪预处理,降噪后的信号进行降采样处理随后输入BiLSTM中进行时序预测;其次,对时序预测后的信号进行特征提取,将提取后的特征向量输入LSSVM进行廓形预测;最后,以五头螺杆为例通过正交实验对BiLSTM-LSSVM模型进行试验验证,并对预测廓形进行误差补偿实验。实验结果表明,提出的基于BiLSTM-LSSVM的螺杆廓形预测模型可对螺杆转子盘铣刀加工螺杆廓形进行准确预测,进而为螺杆转子加工廓形补偿提供支持。Aiming at the problem of profile prediction in the machining process of screw rotor disc milling cutter,a screw profile prediction method based on Bi-directional long short-term memory and least squaresupport vector machines(BiLSTM-LSSVM)was proposed.Firstly,the vibration signals in the processing process are collected and pre-processed for noise reduction.The signals after noise reduction are down-sampled and then input into BILSTM for time series prediction.Secondly,feature extraction is carried out on the signal after time series prediction,and the extracted feature vector is input into LSSVM for profile prediction.Finally,the BiLSTM-LSSVM model is verified by orthogonal experiment,and the error compensation experiment is carried out for the predicted profile.The experimental results show that the proposed screw profile prediction model based on BiLSTM-LSSVM can accurately predict the screw profile machining of the screw rotor disc milling cutter,and then provide support for the screw rotor profile compensation.

关 键 词：螺杆转子 长短时神经网络 最小二乘支持向量机 廓形预测 

分 类 号：TH162[机械工程—机械制造及自动化] TG54[金属学及工艺—金属切削加工及机床]