摘要
为了减小铝合金U形件弯曲成形的回弹角,提出了基于三黑洞系统粒子群算法的成形工艺优化方法。介绍了U形件弯曲成形的工艺流程和回弹角定义方法,以最小化U形件回弹角建立了目标函数。选择加热温度、压边力、模具间隙、冲压速度作为优化参数,确定了优化空间。使用最优拉丁超立方抽样法设计了实验,基于自适应神经元网络拟合了目标参数与优化参数之间的函数模型。为了提高粒子群算法的粒子多样性,提出了三黑洞理论粒子群算法。使用改进算法求解回弹角的优化模型,得到最优参数为:加热温度为240℃、压边力为50 k N、模具间隙为1.2 mm和冲压速度为800 mm·s^-1。经实验验证,优化后的回弹角均值比优化前减小了9.37%,标准差也略有下降,说明经过优化后,U形件回弹角有一定减小,且质量稳定性有一定提高。
In order to reduce the springback angle of U-shaped workpiece for aluminum alloy during bending,the optimization method for forming process based on three-black-hole system particle swarm algorithm was proposed.Then,the bending process flow and the definition method of springback angle for U-shaped workpiece were introduced,and the objective function was built for minimizing the springback angle of U-shaped workpiece.Furthermore,taking heating temperature,blank holder force,die clearance and stamping speed as the optimization parameters,the optimizing space was confirmed,the experiments were designed by optimal Latin hypercube sampling method,and the function model between the target parameters and the optimization parameters was fitted by adaptive neural network.Finally,in order to improve the particle diversity of particle swarm algorithm,three-black-hole theoretical particle swarm algorithm was put forward,and the optimization model of springback angle was solved by the improved particle swarm algorithm to obtain the optimal parameter combination with the heating temperature of 240℃,the blank holder force of 50 k N,the die clearance of 1.2 mm and the stamping speed of 800 mm·s^-1.The verified experiments show that the mean value of springback angle after optimization is reduced by 9.37%compared with that before optimization,and the standard deviation is also slightly decreased.Therefore,after optimization,the springback angle of U-shaped workpiece is reduced to a certain extent,and the quality stability of U-shaped workpiece is improved.
作者
胡丽华
王涛
任少蒙
傅旻
Hu Lihua;Wang Tao;Ren Shaomeng;Fu Min(Department of Mechanical Engineering,Hebei Institute of Mechanical and Electrical Technology,Xingtai 054000,China;College of Mechanical Engineering,Tianjin University of Science&Technology,Tianjin 300222,China)
出处
《锻压技术》
CAS
CSCD
北大核心
2020年第11期60-67,共8页
Forging & Stamping Technology
基金
邢台市科技计划项目(2019ZC060)。
关键词
弯曲成形
U形件
自适应神经元网络
三黑洞系统
粒子群算法
bending
U-shaped workpiece
adaptive neural network
three-black-hole system
particle swarm algorithm
