摘要
以浇注温度、模具初始温度、浇注速度、铸造压力、浇注方式作为输入层函数,以力学性能作为输出层函数,采用5×30×1的三层拓扑结构构建了铝合金汽车轮毂低压铸造工艺神经网络优化模型,并对其进行了训练、预测验证和产线应用。结果表明,铝合金汽车轮毂低压铸造工艺神经网络优化模型具有较强的预测能力,精度较高,相对预测误差2.87%~4.31%,平均相对误差值为3.48%。神经网络优化工艺使试样的抗拉强度、屈服强度比产线原工艺分别增大了7.1%和6.9%,力学性能得到了显著提升。
Taking pouring temperature, initial mould temperature, pouring speed, casting pressure and pouring mode as input layer function and mechanical properties as output layer function, the neural network optimization model of low pressure casting process for aluminum alloy wheel hub was constructed by using a three-layer topology structure of 5 ×30×1. And its training, predictive verification and production line application were carried out. The results show that the neural network optimization model of low pressure casting process for aluminum alloy automobile hub has strong prediction ability and high accuracy. The relative forecast error is 2.87%-4.31%, and the average relative error is 3.48%. Compared with those of the original process parameters of the production line, the tensile strength and yield strength of aluminum alloy automobile hub samples processed by low pressure casting with the neural network model are increased by 7.1% and 6.9% respectively, and the mechanical properties are improved significantly.
作者
于文涛
张思祥
YU Wentao;ZHANG Sixiang(School of Automotive Engineering,Tianjin Vocational Institute,Tianjin 300410,China;School of Mechanical Engineering,Hebei University of Technology,Tianjin 300401,China)
出处
《热加工工艺》
北大核心
2020年第3期89-91,95,共4页
Hot Working Technology
基金
天津职业大学校级基金项目(20171107)
天津市教育科学规划课题(VE3141).
关键词
汽车轮毂
铝合金
神经网络
低压铸造工艺
力学性能
automobile hub
aluminium alloy
neural network
low pressure casting process
mechanical properties
作者简介
于文涛(1988-),男,天津人,讲师,硕士,从事汽车控制技术、汽车材料技术研究,电话:18722605875,E-mail:yuwentvip@163.com
