超塑变形往往具有空洞敏感性，对空洞的研究引起国内外学者的重视并取得较大进展，但现有描述起塑变形时空洞损伤行为的力学模型普遍存在精度问题。利用神经网络对超塑变形时的空洞损伤程度进行预测，不仅可提高精度，同时亦能充分反映超塑变形工艺参数对损伤的影响规律。因此，这就为研究超塑变形时的空洞损伤提供了一种新方法，同时也为神经网络的应用开辟了一个新领域 Superplastic deformation tends to have cavity sensitivity. The study on cavity attracts more and more attention from scholars both at home and abroad. However, the accuracy of the existing mechanical models that describe the damage behavior of cavity during plastic deformation generally exists. Using neural network to predict the degree of void damage in the process of superplastic deformation can not only improve the accuracy, but also fully reflect the impact of the parameters of the superplastic deformation process on the damage. Therefore, this provides a new method for studying hollow damage in superplastic deformation and also opens up a new field for the application of neural networks