Ti-22Al-24Nb合金热变形行为及本构关系

Hot deformation behavior and constitutive relation of Ti-22Al-24Nb alloy

对等温锻造的Ti-22Al-24Nb合金在变形温度650~850℃、应变速率0.001~1 s^(-1)条件下使用Gleeble-3800热模拟试验机进行了40%变形量的热压缩实验。基于实验数据研究了Ti-22Al-24Nb合金的热变形行为,并分别建立了该合金基于应变补偿的Arrhenius和BP神经网络本构模型。结果表明,Ti-22Al-24Nb合金具有正应变速率和负温度敏感性。在变形开始后,流动应力急剧增大至峰值应力,随后硬化和软化效应逐渐达到平衡,流动应力缓慢下降并达到稳态平衡。应变补偿本构模型相关系数R为0.942,平均相对误差e_(AARE)为14.83%;BP神经网络本构模型的相关系数R为0.992,e_(AARE)为4.46%。建立的BP神经网络本构模型具有比应变补偿本构模型更高的预测精度,更加适用于准确预测Ti-22Al-24Nb合金的流动应力。

The thermal compression experiments with 40%deformation amount of isothermally forged Ti-22Al-24Nb alloy were carried out by Gleeble-3800 thermal simulator at deformation temperature of 650-850℃and strain rate of 0.001-1s^(-1).Based on the experimental data,the hot deformation behavior of Ti-22Al-24Nb alloy was studied,and the Arrhenius based on strain compensation and BP neural network constitutive models of the alloy were established,respectively.The results show that Ti-22Al-24Nb alloy has positive strain rate and negative temperature sensitivity.After the beginning of deformation,the flow stress increases sharply to the peak stress,then the hardening and softening effects gradually reach equilibrium,and the flow stress decline slowly and reaches a steady-state equilibrium.The correlation coefficient R of the strain compensation constitutive model is 0.942,and the average relative error e_(AARE) is 14.83%;and the correlation coefficient R of the BP neural network constitutive model is 0.992,and e_(AARE) is 4.46%.The established BP neural network constitutive model has higher prediction accuracy than the strain compensation constitutive model,which is more suitable for accurately predicting the flow stress of Ti-22Al-24Nb alloy.