基于正交试验和BP神经网络的轧制差厚板筒形件充液拉深成形性能预测  

Formability prediction of tailor rolled blank cylindrical parts in hydro deep drawing based on orthogonal test and BP neural network

作　　者：张华伟 李渊[2] 吴佳璐 ZHANG Hua-wei;LI Yuan;WU Jia-lu(College of Mechanical and Electrical Engineering,Guangdong University of Petrochemical Technology,Maoming 525000,China;School of Mechanical Engineering and Automation,Northeastern University,Shenyang 110819,China)

机构地区：[1]广东石油化工学院机电工程学院,广东茂名525000 [2]东北大学机械工程与自动化学院,辽宁沈阳110819

出　　处：《塑性工程学报》2024年第12期76-80,共5页Journal of Plasticity Engineering

基　　金：国家自然科学基金资助项目(51475086);茂名市科技计划立项项目(2022025);广东石油化工学院校级科研基金资助项目(2020rc020)。

摘　　要：将充液拉深工艺引入轧制差厚板的成形过程中,以解决其在传统拉深工艺中的缺陷问题。基于LS-DYNA完成了差厚板筒形件充液拉深成形仿真,并通过正交试验讨论了不同工艺参数对差厚板成形性能的影响规律,在此基础上建立BP神经网络模型,对差厚板的充液拉深成形性能进行预测。研究结果表明:影响差厚板筒形件充液拉深成形性能的因素依次为摩擦因数、薄侧压边力、厚侧压边力、厚侧液池压力、薄侧液池压力,即摩擦因数影响最大、液池压力影响最小,基于正交试验分析结果建立了BP神经网络模型对差厚板筒形件充液拉深成形性能进行预测,获得了较高的预测精度。The hydro deep drawing technology was introduced to the forming process of tailor rolled blank(TRB)to solve the defect prob-lems of TRB in the traditional deep drawing process.The hydro deep drawing of TRB cylindrical part was simulated based on LS-DYNA.The influencing laws of different process parameters on the formability of TRB were discussed by the orthogonal tests,and BP neural net-work model was constructed on this basis to predict the hydro deep drawing formability of TRB.The research results show that the influen-cing factors of hydro deep drawing formability of TRB cylindrical part are listed as the friction factor,the blank holder force on the thinner side,the blank holder force on the thicker side,the liquid pressure on the thicker side,the liquid pressure on the thinner side in a de-scending order according to the influencing extent,that is,the friction factor has the maximum influencing extent,and the liquid pressure has the minimum influencing extent.The BP neural network model was constructed based on the orthogonal test results to predict the hydro deep drawing formability of TRB cylindrical part,and higher prediction accuracy was acquired.

关 键 词：轧制差厚板 充液拉深 厚度减薄率 过渡区移动量 BP神经网络 正交试验 筒形件 

分 类 号：TG386[金属学及工艺—金属压力加工]