Scientists at Tokamak Energy Ltd and the Laboratory for Scientific Computing (LabSC) have teamed up to work on the development of advanced high performance computational multiphysics algorithms for the numerical simulation of nuclear fusion reactors.
Tokamak Energy (TE) is a leading global commercial fusion energy company who are developing compact spherical tokamaks with the aim of reaching commercial fusion power generation. TE are pioneering the development of high-temperature, superconducting magnets for tokamaks, which makes the compact, spherical design a viable option for future power stations.
The project aims to generate computational models that can capture the complex interacting physical processes inside a fusion reactor across a broader range of temporal and spatial scales than current models, thus moving closer to an operational predictive modelling capability. This would be the first step towards a multiphysics computational software framework for integrated whole-system simulations of tokamak reactors.
Initial work is on the development of all-Mach number algorithms for plasma disruption physics and the mathematical and computational assimilation of plasma boundary physics. These are solved in a high performance computing platform which can accommodate the detailed vessel geometry and the disparate length scales of the system by means of embedded boundary methods and hierarchical adaptive mesh refinement respectively.
The figure shows a section of Tokamak Energy’s ST40 compact spherical tokamak (left) and a numerical solution for magnetic flux field using the embedded boundary adaptive mesh refinement model under development at LSC (right).