摘要
6系铝合金接头锻压工艺是当下工业制造中常用锻压工艺,为进一步推动工业加工制造,提高工作效率,文章在的神经网络技术的基础上提出对6系铝合金接头锻压工艺进行改进优化,提出采用四层拓扑结构(5x30x6x1结构),将铝合金牌号、模具预热度数、起始锻温、终止锻温及锻压速度等为输出参数,以抗拉强度为输出参数,构建出完整的6系铝合金接头锻压工艺神经网络模型,并投入到实践中调试分析。最终发现,该模型误差小、预测性突出,实践测试中平均相对预测误差最小达到3.3%,和生产线传统加工工艺相比,以神经网络技术为基础的6系铝合金接头锻压工艺抗拉强度提高近16%,拉伸性能突出,对优化工业加工生产有重要意义。
The forging process of 6-series aluminum alloy joint is a common forging process in current industrial manufacturing.In order to further promote industrial processing and manufacturing and improve work efficiency,this paper proposes to improve and optimize the forging process of 6-series aluminum alloy joint on the basis of neural network technology,and proposes to adopt four-layer topological structure(5x30x6x1 structure)to make the aluminum alloy brand,die preheating degree and start.Forging temperature,ending forging temperature and forging speed are output parameters,and tensile strength is the output parameter.A complete neural network model of 6 series aluminum alloy joint forging process is constructed and put into practice to debug and analyze.Finally,it is found that the error of the model is small and the prediction is outstanding.The average relative prediction error in the practical test is as low as 3.3%.Compared with the traditional processing technology of the production line,the tensile strength of the 6 series aluminum alloy joint forging process based on the neural network technology is increased by nearly 16%,and the tensile property is outstanding,which is of great significance to optimize the industrial processing and production.
作者
王大为
雷艳惠
WANG Da-wei;LEI Yan-hui(Xianyang vocational Technical College,Xianyang 712000,China)
出处
《世界有色金属》
2019年第19期166-168,共3页
World Nonferrous Metals
基金
2019年度咸阳职业技术学院科研基金项目“基于神经网络的铝合金冲锻工艺优化”(2019KYC03)
关键词
神经网络
铝合金接头
锻压
工艺
拉伸性能
neural network
aluminum alloy joint
forging
process
tensile properties
