报告题目:Application of physically informed neural networks for multiscale-multiphysics modelling of functional materials: A case study on perovskite ferroelectrics(物理信息神经网络在功能材料多尺度多物理场建模中的应用——以钙钛矿铁电体为例)
报 告 人:Michael Zaiser 教授、博导 长江学者讲座教授
报告时间:9月2日(星期三),10:00-11:30
报告地点:明德楼B305
报告人简介:
Michael Zaiser,德国埃尔朗根-纽伦堡大学材料系讲席教授(Chair Professor),材料模拟研究所所长。1994年毕业于德国斯图加特大学,获理论物理博士学位,历任英国爱丁堡大学讲师、教授。四川省“天府学者”特聘专家,教育部“海外名师”。主要研究领域为微纳米材料力学及高性能材料,应用统计物理、材料科学、固体力学等理论,研究材料的微结构和缺陷的无序性和随机性,及对其材料宏观力学性能的影响。在Science,Nature Physics,Nature Communications,Advanced Energy Materials,Physical Review Letters等国际顶尖期刊上发表学术论文200余篇,累计引用达8000余次,出版专著2部,参与编写专著5部。
报告摘要:
Microstructural processes and macroscopic response of many functional materials are characterized by couplings between multiple fields which are mediated by complex atomic-scale processes. Predictive physical modelling therefore requires a multiscale approach integrating information from quantum mechanics, atomistic dynamics, and microstructure physics. We discuss such an approach for a paradigmatic ferroelectric material (BaTiO3) with perovskite structure. We show how physically informed neural networks can be used to construct differentiable representations of atomic-scale data which serve to parameterize complex multiphysics models within a continuum phase-field framework. We discuss the conceptual problems and technical issues encountered and demonstrate the predictive power of the obtained model in simulating domain microstructure evolution under mechanical stimulation.
相关成果:
Wang, X, Wendler, F., Azuma, H., Zaiser, M., Ogata, S., Kobayashi, R., Lei, Z., Tan, X. (2026). Multiscale Modelling of Ferroelectrics using a Physics-Informed Neural Network Driven by Molecular Dynamics Data: Parameter Identification and Field Reconstruction. npj Computational Materials, accepted.
主办单位:新能源与材料学院/光伏新能源现代产业学院
四川省玄武岩纤维复合材料开发及应用工程技术研究中心
四川省页岩气高效开采先进材料制备技术工程研究中心
能量转换与储存先进材料四川省国际科技合作基地
氢能绿色制取川渝重点实验室
科学技术发展研究院