Physics-Informed Deep Learning for Time-Domain Electromagnetic Radiation Problem
- Resource Type
- Conference
- Authors
- Ge, Yingze; Guo, Liangshuai; Li, Maokun
- Source
- 2022 IEEE MTT-S International Microwave Biomedical Conference (IMBioC) Microwave Biomedical Conference (IMBioC), 2022 IEEE MTT-S International. :114-116 May, 2022
- Subject
- Bioengineering
Components, Circuits, Devices and Systems
Fields, Waves and Electromagnetics
Deep learning
Interpolation
Conferences
Neural networks
Electromagnetic radiation
Microwave theory and techniques
Electromagnetic fields
computational electromagnetics
electromagnetic radiation
machine learning
physics-informed nerual networks
- Language
We explore the application of physics-informed deep learning to solve time-domain electromagnetic problems. This method takes advantage of the differentiability of neural networks and fully integrated with first principles. Compared to traditional approach, there is no need of discretization. Numerical experiment verifies the accuracy of this scheme.