摘要
以6个热处理工艺参数作为输入层参数,以热处理后定向凝固钛合金的屈服强度作为输出层参数,构建6×48×1结构的定向凝固钛合金热处理工艺优化神经网络模型。结果表明,神经网络模型预测误差小于3%,具有较强的预测能力和较佳的预测精度。与试验优化工艺参数相比,采用神经网络模型优化参数热处理后,定向凝固Ti-6Al-4V-0.5Ce钛合金的抗拉强度和屈服强度分别提高178、186 MPa。
The neural network model of heat treatment process optimization of the directional solidification titanium alloys was built with three layers by 6×48× 1, which was with six heat treatment process parameters as input parameters, and with yield strength of the directional solidification titanium alloy as output parameter. The results show that prediction error of the neural network model is lower than 3%, which is with good prediction ability and high precision. Compared with the optimized process by experiment, the tensile strength and yield strength of the directional solidification titanium alloy Ti-6Al-4V-0.5Ce with the optimized heat treatment process based on the neural network model increase by 178 MPa and 186 MPa, respectively.
出处
《兵器材料科学与工程》
CAS
CSCD
北大核心
2017年第4期30-33,共4页
Ordnance Material Science and Engineering