NJIT
Neural operator for accelerating coronal magnetic field computations in bifrost MHD model
Abstract
dc:description.abstract<p>The application of the Tensorized Fourier Neural Operator (TFNO) to significantly enhance the computational efficiency of coronal magnetic field calculations within the Bifrost Magnetohydrodynamics (MHD) model is introduced in this study. Leveraging simulated data from the European Sunrise Science Data Center, the TFNO? an extension of the Fourier Neural Operator (FNO) that incorporates tensor decomposition for improved handling of high-dimensional data?is employed to solve time-varying partial differential equations (PDEs) over a 3D domain. The performance of the TFNO is compared with traditional machine learning methods, including Vision Transformer and CNN-RNN (encoder-decoder) architectures, to demonstrate its accuracy, computational efficiency, and scalability. A physics-analysis of the TFNO predictions is also performed to demonstrate the reliability of the method. This advancement not only accelerates the simulation of solar atmospheric phenomena but also provides more reliable prediction capabilities, thus greatly contributing to the understanding of space weather dynamics and its impact on Earth.</p>
Degree
thesis:*- Name thesis:degree_name
- Master of Science in Engineering Management - (M.S.)
- Discipline thesis:degree_discipline
- Mechanical and Industrial Engineering
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Du, Yutao
- Contributors dc:contributor
-
- Bo Shen
- Sanchoy K. Das
- SangWoo Park
Subjects
dc:subject × 5Identifiers
dc:identifier.*- Repository record dc:identifier
- https://digitalcommons.njit.edu/theses/2584