摘要
以11个差压铸造工艺参数作为输入层参数,以抗拉强度作为输出层参数,构建了11×55×5×1四层结构的铝合金差压铸造工艺优化神经网络模型。结果表明,神经网络模型的平均相对训练误差2.8%,平均相对预测误差为2.9%,模型具有较强的预测能力和较佳的预测精度。
Taking 11 differential pressure casting process parameters as input layer parameters, and taking the tensile strength as the output layer parameter, the neural network model（four layers structure of 11×55×5×1） for differential pressure casting process optimization of aluminum alloys was built. The results show that the average relative training and prediction errors of the neural network model are 2.8% and 2.9%, respectively. The model has stronger prediction ability and better prediction accuracy.
作者
陈新林
胡帅
董燕飞
CHEN Xinlin;HU Shuai;DONG Yanfei(Henan Vocational College of Agriculture, Zhengzhou 451450, China;Henan University of Urban Construction, Pingdingshan 467036, China)
出处
《热加工工艺》
CSCD
北大核心
2018年第7期99-101,共3页
Hot Working Technology
基金
河南省教育厅项目(15A413008)
关键词
神经网络算法
铝合金
差压铸造
工艺优化
抗拉强度
neural network algorithm
aluminum alloy
, differential pressure casting
process optimization
tensile strength
