摘要
传统的导流模设计主要建立在经验的基础上,具有主观性、不可靠性.本文通过正交法把由经验公式和工艺要求确定的初步模具参数设计成实验样本,基于此,以UG为平台进行模具CAD造型,然后导入软件SuperForge模拟,得到SDV值作为优化目标,最后用神经网络建立模具参数的数学模型,遗传算法优化模具参数,由此设计出的模具一次试模成功.本文为设计导流模提供了一种行之有效的方法.
Traditional design of the flow guided die is mostly set up on the basis of experience.The design has subjectivity and fallibility.In this paper,initial design die parameters gained from experience (formula) and technological requirement are taken as experiment samples.Based on these draw CAD modelling of the dies with UG.Then,simulate the extrusion process with SuperForge and Get the SDV which is regarded as goal of (optimizing.)Finally,set up the mathematics model of die parameters with neural network and optimize the (parameters) with the genetic algorithm.The experiment of using the designed die which is based on the system is successful for the first time.The paper puts forward a kind of effectual means for designing the flow guided die.
出处
《南昌大学学报(工科版)》
CAS
2004年第3期15-18,共4页
Journal of Nanchang University(Engineering & Technology)
基金
江西省教育厅重大项目
南昌大学"211工程""九五"标志性资助项目
关键词
导流模
模拟
神经网络
遗传算法
优化
the flow guided die
simulate
neural network
genetic algorithm
optimizing