基于神经网络的高强镁合金铸造工艺优化

Casting Process Optimization of High Strength Magnesium Alloy Based on Neural Network

以静置温度、静置时间、熔化时和浇注时的机械振动辅助搅拌的振动频率为输入层参数,以抗拉强度为输出层参数,构建4×16×8×1四层结构的高强镁合金铸造工艺优化神经网络模型,并对神经网络模型进行了预测验证。此外对优化铸造工艺制备的高强镁合金Mg-10Gd-2.3Y-1.2Zn-0.3Zr-0.2Co进行了显微组织、物相组成和力学性能测试与分析。结果表明,该神经网络模型的平均相对预测误差为1.9%,具有较高的预测能力和预测精度;用优化铸造工艺制备的该高强镁合金抗拉强度达548 MPa、屈服强度达412 MPa、伸长率达14.7%。

The neural network model of the casting process optimization of high strength magnesium alloy with 4×16×8×1 four layers structure was established, in which the input layer parameters were made of holding temperature, holding time,the vibration frequencies of mechanical vibration assisted stirring in melting and casting, and the output layer parameter was tensile strength. The neural network model was forecasted and verified. Besides the microstructure, phase composition and mechanical properties tests of high strength magnesium alloy Mg-10Gd-2.3Y-1.2Zn-0.3Zr-0.2Co prepared by the optimization process were carried out. The results show that the average relative prediction error of the neural network model is 1.9%, and the model has higher prediction ability and prediction accuracy; the tensile strength of high strength magnesium alloy prepared by using the optimized casting process is 548 MPa, the yield strength is 412 MPa, and the elongation is 14.7%.