中文摘要：

以800H合金的热压缩实验为基础,分析800H合金在不同温度和应变速率下800H合金的流动应变行为。基于800H合金变形温度、应变率、应变和应力的实验数据,建立关于800H合金热变形的GRNN神经网络预测模型。依据GRNN神经网络训练结果,选择平滑因子为0.2的网络。应力预测值和实验结果的相关性分析表明,建立的800H合金热变形行为GRNN神经网络模型稳定性高、泛化能力很强,可应用于其他合金的热变形行为预测。

英文摘要：

The thermal compression test of 800H alloy was used to analyze the flow strain behavior of 800H alloy at different temperatures and strain rates. The deformation temperature, strain rate, strain and stress of the alloy obtained through thermal compression experiment were used to establish the hot deformation GRNN network prediction model for 800H alloy. Based on the results of GRNN neural network training, the prediction results of selecting a network smoothing factor of 0.2 are best. The correlation analysis of stresses predicted values and experimental results show that the prediction model of 800H alloy thermal deformation behavior through GRNN neural network has high stability and strong generalization ability, which can be applied in hot deformation behavior prediction of other alloys.