【摘要】 在变形温度为800～950℃和应变速率为0.001～1 s^-1的条件下对Ti₂AlNb合金进行了等温恒应变速率压缩实验，研究其流动应力行为，基于实验数据，构建了修正J-C和BP神经网络两种本构模型，分析了两种本构模型预测精度的差异。结果表明：Ti₂AlNb合金的流动应力具有负温度敏感性和正应变速率敏感性；修正J-C本构模型具有较好的预测精度，其相关系数R和平均绝对相对误差AARE分别为0.982和7.37%,但该模型的预测精度随着变形条件变化存在较大差异；BP神经网络本构模型具有更高的预测精度，其R和AARE分别达到了0.998和1.82%,且该模型的预测精度随着变形条件变化仅存在较小的差异；因此，BP神经网络本构模型比修正J-C本构模型更能准确地反映Ti₂AlNb合金流动应力与热力参数之间的非线性关系，更好地表征该合金的流动应力行为。更多还原

【Abstract】 Under the conditions of deformation temperature of 800-950 ℃ and strain rate of 0.001-1 s^-1, isothermal constant strain rate compression experiments were conducted on Ti₂AlNb alloy to study its flow stress behavior. Based on the experimental data, two constitutive models, modified J-C and BP neural networks, were constructed, and the difference in prediction accuracy between the two constitutive models was analyzed. The results show that the flow stress of the Ti₂AlNb alloy exhibits negative temperature sensitivity and positive strain rate sensitivity. The modified J-C constitutive model exhibits good prediction accuracy, with a correlation coefficient R and an average absolute relative error AARE of 0.982 and 7.37%, respectively. However, the prediction accuracy of this model varies greatly with changes in deformation conditions. The BP neural network constitutive model exhibits higher prediction accuracy, with R and AARE reaching 0.998 and 1.82% respectively, and the prediction accuracy of this model only has a small difference with changes in deformation conditions. Therefore, the BP neural network constitutive model can more accurately reflect the nonlinear relationship between the flow stress and thermal parameters of the Ti₂AlNb alloy than the modified J-C constitutive model, and better characterize the flow stress behavior of the alloy.更多还原