Magnetohydrodynamics with physics informed neural operators
The modeling of multi-scale and multi-physics complex systems typically involves the use of scientific software that can optimally leverage extreme scale computing. Despite major developments in recent years, these simulations continue to be computationally intensive and time consuming. Here we expl...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IOP Publishing
2023-01-01
|
| Series: | Machine Learning: Science and Technology |
| Subjects: | |
| Online Access: | https://doi.org/10.1088/2632-2153/ace30a |
| _version_ | 1797785977646743552 |
|---|---|
| author | Shawn G Rosofsky E A Huerta |
| author_facet | Shawn G Rosofsky E A Huerta |
| author_sort | Shawn G Rosofsky |
| collection | DOAJ |
| description | The modeling of multi-scale and multi-physics complex systems typically involves the use of scientific software that can optimally leverage extreme scale computing. Despite major developments in recent years, these simulations continue to be computationally intensive and time consuming. Here we explore the use of AI to accelerate the modeling of complex systems at a fraction of the computational cost of classical methods, and present the first application of physics informed neural operators (NOs) (PINOs) to model 2D incompressible magnetohydrodynamics (MHD) simulations. Our AI models incorporate tensor Fourier NOs as their backbone, which we implemented with the TensorLY package. Our results indicate that PINOs can accurately capture the physics of MHD simulations that describe laminar flows with Reynolds numbers $\mathrm{Re}\leqslant250$ . We also explore the applicability of our AI surrogates for turbulent flows, and discuss a variety of methodologies that may be incorporated in future work to create AI models that provide a computationally efficient and high fidelity description of MHD simulations for a broad range of Reynolds numbers. The scientific software developed in this project is released with this manuscript. |
| first_indexed | 2024-03-13T01:02:36Z |
| format | Article |
| id | doaj.art-01b8a307bade4d449f0f70bebeb28f40 |
| institution | Directory Open Access Journal |
| issn | 2632-2153 |
| language | English |
| last_indexed | 2024-03-13T01:02:36Z |
| publishDate | 2023-01-01 |
| publisher | IOP Publishing |
| record_format | Article |
| series | Machine Learning: Science and Technology |
| spelling | doaj.art-01b8a307bade4d449f0f70bebeb28f402023-07-06T11:08:55ZengIOP PublishingMachine Learning: Science and Technology2632-21532023-01-014303500210.1088/2632-2153/ace30aMagnetohydrodynamics with physics informed neural operatorsShawn G Rosofsky0https://orcid.org/0000-0002-3319-576XE A Huerta1https://orcid.org/0000-0002-9682-3604Data Science and Learning Division, Argonne National Laboratory , Lemont, IL 60439, United States of America; Department of Physics, University of Illinois at Urbana-Champaign , Urbana, IL 61801, United States of America; NCSA, University of Illinois at Urbana-Champaign , Urbana, IL 61801, United States of AmericaData Science and Learning Division, Argonne National Laboratory , Lemont, IL 60439, United States of America; Department of Physics, University of Illinois at Urbana-Champaign , Urbana, IL 61801, United States of America; Department of Computer Science, The University of Chicago , Chicago, IL 60637, United States of AmericaThe modeling of multi-scale and multi-physics complex systems typically involves the use of scientific software that can optimally leverage extreme scale computing. Despite major developments in recent years, these simulations continue to be computationally intensive and time consuming. Here we explore the use of AI to accelerate the modeling of complex systems at a fraction of the computational cost of classical methods, and present the first application of physics informed neural operators (NOs) (PINOs) to model 2D incompressible magnetohydrodynamics (MHD) simulations. Our AI models incorporate tensor Fourier NOs as their backbone, which we implemented with the TensorLY package. Our results indicate that PINOs can accurately capture the physics of MHD simulations that describe laminar flows with Reynolds numbers $\mathrm{Re}\leqslant250$ . We also explore the applicability of our AI surrogates for turbulent flows, and discuss a variety of methodologies that may be incorporated in future work to create AI models that provide a computationally efficient and high fidelity description of MHD simulations for a broad range of Reynolds numbers. The scientific software developed in this project is released with this manuscript.https://doi.org/10.1088/2632-2153/ace30amagnetohydrodynamicsscientific machine learningphysics informed AIhigh performance computing |
| spellingShingle | Shawn G Rosofsky E A Huerta Magnetohydrodynamics with physics informed neural operators Machine Learning: Science and Technology magnetohydrodynamics scientific machine learning physics informed AI high performance computing |
| title | Magnetohydrodynamics with physics informed neural operators |
| title_full | Magnetohydrodynamics with physics informed neural operators |
| title_fullStr | Magnetohydrodynamics with physics informed neural operators |
| title_full_unstemmed | Magnetohydrodynamics with physics informed neural operators |
| title_short | Magnetohydrodynamics with physics informed neural operators |
| title_sort | magnetohydrodynamics with physics informed neural operators |
| topic | magnetohydrodynamics scientific machine learning physics informed AI high performance computing |
| url | https://doi.org/10.1088/2632-2153/ace30a |
| work_keys_str_mv | AT shawngrosofsky magnetohydrodynamicswithphysicsinformedneuraloperators AT eahuerta magnetohydrodynamicswithphysicsinformedneuraloperators |