Magnetically confined nuclear fusion (MCF) is a highly multidimensional problem in both space and time. As such, modeling MCF through predictive numerical simulations has been representing one of the most challenging tasks for the numerical and high-performance computing (HPC) community for decades.
Due to the complexity of both the physics involved and the technology of modern HPC infrastructures, relying on open-source software (OSS) represents a tremendous opportunity for development acceleration. Besides, it constitutes an impelling necessity to align the time scales of science understanding and industrial advancement.
Renaissance Fusion (RF) is pursuing the design and deployment of a compact high-field stellarator with key novel features such as its liquid metal first wall and laser-engraved coils. The implementation of such an ambitious program requires the prompt availability of highly sophisticated and validated numerical codes to deploy a massive modelling program and carry out large multidimensional parameter investigations.
Inspired by the pioneering vision of the Numerical Tokamak Project [1], which aimed to generate reliable predictions of machine performance through physics-based high-performance computing, RF put efforts to select a comprehensive suite of open-source software and combine them into a package that supports and accelerates RF’s high-field stellarator nuclear fusion project. We will discuss the present state of this initiative and the plans for its expansion.
Moreover, we describe a reactor-relevant OSS study directed atthe identification of possible plasma heating scenarios in high-field stellarators. This investigation has been carried out through the BEAMS3D [2] code and has allowed us to address the role of the different fast ion loss and energy deposition channels as well as their dependence on beam energy and injection parameters.
[1] The Numerical Tokamak Project (NTE), Simulation of Turbulent Transport in the Core Plasma: A Grand Challenge in Plasma Physics, LLNL Report, December 1993 [2] M. McMillan and S. A. Lazerson, BEAMS3D Neutral Beam Injection Model, PPCF, 095019, 56, 2014