数控机床超声振动切削表面粗糙度预测RNN评估  

RNN Evaluation of Surface Roughness Prediction for Ultrasonic Vibration Cutting of CNC Machine Tools

作　　者：刘喜庆 宋佳佳 张亚辉 Liu Xiqing;Song Jiajia;Zhang Yahui(School of Electronic Information Engineering,Zhengzhou City Vocational College,Zhengzhou Henan 452370,China;College of Intelligent Manufacturing,Zhengzhou City Vocational College,Zhengzhou Henan 452370,China)

机构地区：[1]郑州城市职业学院电子信息工程学院,河南郑州452370 [2]郑州城市职业学院智能制造学院,河南郑州452370

出　　处：《机械管理开发》2024年第8期28-29,32,共3页Mechanical Management and Development

摘　　要：超声振动切削是目前数控机床机械制造中的热点,保证工件的表面质量是关键的目标。为此设计了一种基于回归神经网络(RNN)的数控机床超声振动切削预测评估方法,并开展铣削试验测试分析。研究结果表明:预测结果与实测结果的相关性高达0.99,表明获得了良好预报效果,最大偏差仅为0.05,表明设计的算法可以保证获得理想处理结果。相比较卷积神经网络(CNN)、支持向量机(SVM)与高斯过程回归(GPR)方法,RNN获得较好预测效果,磨损和表明粗糙值分别是3.685和2.216,R^(2)值达到了0.975。该研究有助于适合于高精度制造领域,具有很好的开发价值。Ultrasonic vibration cutting is a hot spot in CNC machine tool manufacturing,and ensuring the surface quality of the workpiece is the key goal.Therefore,a prediction and evaluation method of ultrasonic vibration cutting for CNC machine tools based on regression neural network(RNN)is designed,and the milling test is carried out.The research results show that the correlation between the predicted results and the measured results is as high as 0.99,indicating that a good prediction effect is obtained,and the maximum deviation is only 0.05,indicating that the proposed algorithm can guarantee the ideal processing results.Compared with convolutional neural network(CNN),support vector machine(SVM)and Gaussian process regression(GPR),RNN obtained better prediction results.The wear and roughness values were 3.685 and 2.216,respectively,and the R^(2)value reached 0.975.This research is helpful to be suitable for high-precision manufacturing field and has good development value.

关 键 词：振动切削 表面粗糙度 回归神经网络 磨损 质量评估 

分 类 号：TH137[机械工程—机械制造及自动化]