Materials Data Science and Informatics (IAS-9)

Application of Neural Networks and Machine Learning in Computational Material Modeling / Simulation

Content

1. Introduction
2. Basic concepts from machine learning
3. Artificial neural networks (ANNs)
4. Machine learning in physics and engineering
5. Material modeling in physics and engineering
6. Coupled-field problems (CFPs) based on material models
7. Review of standard numerical solution methods for material-model-based CFPs

Objective

The students will get familiar with the atomic structure of materials and its effect on mechanical properties. The focus will be put on crystalline materials, lattice defects therein as well as their dynamics and contribution to the formation of microstructure. The students are also introduced to relevant concepts from thermodynamics and statistical mechanics as well as state of the art numerical simulation techniques such as molecular dynamics (MD) and the phase-field method.

Recommended previous knowledge

• basic knowledge in physics

Recommended reading

• Lagaris et al., Artificial neural networks for solving ODEs and PDEs, IEEE Trans. Neural Networks 9, 1998
• Li et al., Physics-Informed Neural Operator for Learning PDEs, arXiv:2111.03794v1, 2021

Lecture and exercise dates WS 24/25

• Lecture: Tuesdays, 14:30 - 16:00, in GRS001, Schinkelstr. 2a
• Exercise: Fridays, 13:15 - 14:00, in GRS001, Schinkelstr. 2a

Exam WS 24/25

TBA

Contacts

Administrative Contact: Dr. Katharina Immel
Content: Prof. Dr. Robert Svendsen