为探寻金属材料热变形流变行为建模的高效方法, 运用Gleeble-3500型热模拟机对6013铝合金进行了热压缩试验, 基于试验数据和三种元模型方法(克里金、径向基函数与多元多项式)分别构建了该合金的热变形流变行为模型, 并就各方法建模的预测能力进行了分析与对比。结果表明: 三种元模型方法均可建立较高精度的预测模型, 可以描述各热力学参数之间高度非线性的复杂关系; 在模型预测精度和泛化能力方面, 克里金方法较径向基函数和多元多项式方法均具有显著优势, 是元模型方法中可用于金属热变形流变行为建模的一种较为准确有效的方法。

In order to look for an effective approach to model the thermal flow behavior of metal materials, the hot compression test of 6013 aluminum alloy was conducted using Gleeble-3500 thermal simulator. Based on the experimental data and three types of metamodeling method (Kriging, radial basis function and multivariate polynomial), the hot deformation flow behavior models of the alloy were built respectively and their prediction abilities were analyzed and compared. The results show that the prediction models constructed from the three metamodeling methods were all of a relatively high accuracy and can be used to describe the complicated nonlinear relationship among thermodynamic parameters. The modeling accuracy and generalization capability of the Kriging method were much better than those of the other two methods, indicating that Kriging method is a relatively accurate and efficient method for modeling hot deformation behavior of metals among the metamodeling methods.