arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2201.09723cond-mat.mtrl-scicond-mat.dis-nnphysics.data-an
收藏

二维石英玻璃的失效预测

Predicting the failure of two-dimensional silica glasses

Francesc Font-Clos, Marco Zanchi, Stefan Hiemer, Silvia Bonfanti, Roberto Guerra, Michael Zaiser, Stefano Zapperi

摘要:

能够根据结构信息预测材料的失效是一个基本问题,对于设备和组件的监测具有巨大的实际和工业意义。由于深度学习的最新进展,即使对于高度无序的固体,准确的故障预测也变得可能,但该过程中使用的参数数量之多使得对结果的物理解释变得不可能。在这里,我们解决这个问题,并使用机器学习方法来预测失败的模拟二维二氧化硅玻璃从其初始未变形的结构。然后,我们利用加权类激活映射(梯度CAM)建立注意力地图与预测,我们证明,这些地图是服从物理解释的拓扑缺陷和局部势能。我们表明,我们的预测可以被转移到具有不同形状或大小的样本,而不是训练中使用的样本,以及实验图像。我们的策略说明了人工神经网络训练的数值模拟结果可以提供可解释的预测实验测量的结构的行为。

英文摘要:

Being able to predict the failure of materials based on structural information is a fundamental issue with enormous practical and industrial relevance for the monitoring of devices and components. Thanks to recent advances in deep learning, accurate failure predictions are becoming possible even for strongly disordered solids, but the sheer number of parameters used in the process renders a physical interpretation of the results impossible. Here we address this issue and use machine learning methods to predict the failure of simulated two dimensional silica glasses from their initial undeformed structure. We then exploit Gradient-weighted Class Activation Mapping (Grad-CAM) to build attention maps associated with the predictions, and we demonstrate that these maps are amenable to physical interpretation in terms of topological defects and local potential energies. We show that our predictions can be transferred to samples with different shape or size than those used in training, as well as to experimental images. Our strategy illustrates how artificial neural networks trained with numerical simulation results can provide interpretable predictions of the behavior of experimentally measured structures.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑