摘要
以镁合金牌号、固溶温度、固溶时间、时效温度和时效时间为输入层参数,以抗拉强度为输出层参数,采用5×40×15×1四层拓扑结构构建了体育器材用细晶镁合金热处理工艺优化神经网络模型。结果表明,模型具有较佳预测能力和较高预测精度,经过5972次迭代计算后收敛,预测误差在-3%～3%。与采用原工艺处理的细晶镁合金的抗拉强度相比,使用神经网络模型优化的热处理工艺明显提高了细晶镁合金的抗拉强度。
Taking alloy grade, solid solution temperature, solid solution time, aging temperature and aging time as input parameters, and taking tensile strength as output parameters, the neural network model of the heat treatment process optimization of magnesium alloys with fine-grain for sports equipment was built by using the four layer topology with 5 ×40 × 15 ×1 structure. The results show that the model has better prediction ability and higher prediction accuracy, and compute is convergent after 5972 iterations. The prediction error of the model is between -3% and 3%. Compared with the tensile strength of fine-grain magnesium alloy treated by the original process, the heat treatment process optimized by neural network model can obviously improve the tensile strength of fine grain magnesium alloy.
作者
龙羽
张建
LONG Yu;ZHANG Jian(College of Physical Education and Health Sciences, Yangtze Normal University, Chongqing 408100, China;School of Materials Science and Engineering, Hebei University of Technology, Tianjin 300401, China)
出处
《热加工工艺》
CSCD
北大核心
2018年第12期234-236,240,共4页
Hot Working Technology
关键词
神经网络
工艺优化
细晶镁合金
抗拉强度
neural network
process optimization
fine-grain magnesium alloy
tensile strength
