摘要
以连杆材料、模具预热温度、始锻温度、终锻温度、锻压速度为输入层参数,以磨损体积为输出层参数,构建了5×25×1三层拓扑结构的连杆锻压工艺优化神经网络模型,并对该模型进行了预测和验证。还对45钢连杆和40Cr钢连杆的磨损性能进行了测试分析。结果表明,连杆锻压工艺神经网络模型的平均相对训练误差为4.1%,平均相对预测误差为4.5%,具有较精准的预测能力和精度。应用BP神经网络模型优化锻压工艺的45钢连杆和40Cr钢连杆的磨损体积分别较产线现用工艺减小了38.2%、44%。神经网络优化的连杆锻压工艺的最佳参数为:材料40Cr钢、模具预热温度450℃、始锻温度1240℃、终锻温度810℃、锻压速度32 mm/s。
Taking connecting rod material, die preheating temperature, initial forging temperature, final forging temperature and forging speed as input layer parameters and wear volume as output layer parameter, the neural network model for optimizing forging process of the connecting rod with 5×25×1 three-layer topological structure was constructed, and the model was predicted and validated. The wear property of 45 steel connecting rod and 40Cr steel connecting rod was also tested and analyzed. The results show that the average relative training error is 4.1% and the average relative prediction error is 4.5% in the neural network model of connecting rod forging process, which has accurate prediction ability and precision. The wear volumes of 45 steel connecting rod and 40Cr steel connecting rod under the optimized BP neural network model forging process are reduced by 38.2% and 44% respectively compared with that of the existing production line process. The optimum parameters of forging process of the connecting rod based on neural network are material of 40 Cr steel, die preheating temperature of 450 ℃, initial forging temperature of 1240 ℃, final forging temperature of 810 ℃ and forging speed of 32 mm/s.
作者
齐莉
于晓鹏
QI Li;YU Xiaopeng(Jilin Engineering Vocational College,Siping 136001,China;Computer College,Jilin Normal University,Siping 136000,China)
出处
《热加工工艺》
北大核心
2020年第13期96-99,共4页
Hot Working Technology
基金
吉林省高等教育学会2016年度高教科研重点课题(JGJX2016B33)
吉林省教育厅2016年度职业教育与成人教育教学改革课题(2016ZCY054)。
关键词
BP神经网络
连杆
锻压工艺
工艺优化
磨损性能
BP neural network
connecting rod
forging process
process optimization
wear performance