基于FOA-LSSVM混合优化的9Ni钢本构模型

Establishment of Constitutive Model of 9Ni Steel Based on FOA-LSSVM Hybrid Optimized Algorithm

针对单一具体的数学方程式模型难以准确描述9Ni钢材料整个热塑性加工过程中材料对外部参数的力学响应问题,在9Ni钢热模拟试验机压缩试验数据基础上,采用新的果蝇算法(Fruit Fly Optimization Algorithm FOA)对最小二乘支持向量机(LSSVM)模型惩罚因子c和核宽度λ进行寻优,构造了FOA-LSSVM混合优化的9Ni钢本构模型。模型预测值与实验值的对比结果表明:FOA-LSSVM9Ni钢本构模型具有良好的拟合性,能反映9Ni钢热加工过程中各个阶段的流变行为,其预测值与实验值之间的最大、最小、平均绝对百分比误差分别为6.21%、0.19%和2.64%;模型具有很高的预测精度和鲁棒性,可描述9Ni钢的高温流变力学行为。

For the mechanical response of 9Ni steel material throughout the thermoplastic processing cannot be accurately described by a specific mathematical equation model. On the basis of experimental data produced by 9Ni isothermal compression experiment, our study utilize fi＇uit fly optimization algorithm （FOA） to optimize LSSVM model penalty factor c and core width A and build up a 9Ni steel constitutive relation modified by combination of FOA and LSSVM. The FOA-LSSVM model predicted values were used to compare with experimental values. The results show that the model has a good fit of experimental data. The theological behavior in thermoplastic processing of 9Ni steel can be accurately described by the model, MAXAPE, MINAPE and MAPE between FOA-LSSVM predict values and the experimental value is 6.24%, 0.19% and 2.64%, respectively. To summarize, the model is accuracy and robustness, and the dynamic mechanical response of 9Ni steel in thermoplastic processing can be precisely described by FOA-LSSVM model.