A Fully Implicit Parallel Solver for MHD Instabilities in a Tokamak

Qinghe Yao,Zichao Jiang,Zhuolin Wang,Junyang Jiang,Zhiwei Ma
DOI: https://doi.org/10.1007/s10894-023-00369-5
IF: 1.793
2023-01-01
Journal of Fusion Energy
Abstract:Aiming at the long-term and high-precision simulation of the magnetohydrodynamic (MHD) instabilities in the tokamak model, we developed a parallelized solver based on a fully implicit difference scheme. A 4th-order precision difference scheme and the Newton–Krylov method are employed in the proposed solver for both the flow and the electromagnetic field. To achieve high parallel efficiency, we adopt a strategy based on the spatial domain decomposition to partition the large Jacobian matrices in the iteration, and a buffer area based on the grid density is utilized to minimize the memory and time consumption. The accuracy of the methodology is verified, and the numerical results are validated by comparison with recognized results. The numerical results of the tearing mode instability in the tokamak model have demonstrated the precision and reliability of the algorithm, and the high parallel efficiency has been proven by the scalability test on the platform with up to 1280 threads, showing significant potential in the large-scale simulation of MHD problems.
What problem does this paper attempt to address?