超声滚挤压轴承套圈表面性能预测模型建立  

Construction of surface performance prediction model of ultrasonic roll extruded bearing ring

作　　者：王晓强[1,2] 张彪 崔凤奎[1,2] 刘志飞 王排岗[1,2] WANG Xiao-qiang;ZHANG Biao;CUI Feng-kui;LIU Zhi-fei;WANG Pai-gang(School of Mechanical Engineering,Henan University of Science and Technology,Luoyang 471003,China;Collaborative Innovation Center of Machinery Equipment Advanced Manufacturing of Henan Province,Henan University of Science and Technology,Luoyang 471003,China)

机构地区：[1]河南科技大学机电工程学院,河南洛阳471003 [2]河南科技大学机械装备先进制造河南省协同创新中心,河南洛阳471003

出　　处：《塑性工程学报》2022年第6期25-32,共8页Journal of Plasticity Engineering

基　　金：国家自然科学基金资助项目(U1804145);国家重点研发计划(2018YFB2000405)。

摘　　要：通过正交试验研究了工件转速、进给速度、振幅和静压力4个工艺参数对超声滚挤压轴承套圈表面性能的影响,得到了不同工艺参数组合下轴承套圈的表面粗糙度、表面残余应力和表面硬度。根据正交试验所得的轴承套圈表面性能数据,分别通过BP神经网络、径向基神经网络和多元回归法建立了超声滚挤压表面性能预测模型,并分别对试验值和预测值进行了对比。结果表明,通过多元回归法建立的超声滚挤压表面性能预测模型所预测的表面粗糙度、表面残余应力和表面硬度的平均相对误差最小,分别为5.18%、5.08%和3.38%。The influence of four process parameters of workpiece speed,feeding speed,amplitude and static pressure on surface performance of ultrasonic roll extruded bearing ring was studied by orthogonal tests,and the surface roughness,surface residual stress and surface hardness of bearing ring with different process parameter combinations were obtained.According to the surface performance data of bearing ring obtained by orthogonal tests,the prediction models for ultrasonic roll extruded surface performance were established by BP neural network,radial basis function neural network and multiple regression method,respectively,and the test values were compared with the predicted values.The results show that the average relative error of surface roughness,surface residual stress and surface hardness predicted by the ultrasonic roll extrusion surface performance prediction model established by multiple regression method is the smallest,which are 5.18%,5.08%and 3.38%,respectively.

关 键 词：正交试验 BP神经网络 RBF神经网络 多元回归法 

分 类 号：TG115[金属学及工艺—物理冶金]