Abstract

Contributed Talk - Splinter Computational

Thursday, 12 September 2024, 14:30   (S26)

Hybrid radiation hydrodynamics scheme with tree-based pseudo-SPH particles

Cheryl S. C. Lau [1], Maya A. Petkova [2], Ian A. Bonnell [1]
[1] University of St Andrews, UK; [2] Chalmers University of Technology, Sweden

Computational methods for modelling stellar photoionization remains an active area of development, particularly for star formation simulations where feedback from massive stars is crucial. We present a hybrid radiation hydrodynamics (RHD) scheme that couples SPH to grid-based Monte Carlo Radiative Transfer (MCRT) codes, as first developed by Petkova et al., for incorporating the effects of ionizing radiation. This scheme employs the Exact mapping method for transferring fluid properties between SPH particles and Voronoi grids on which the MCRT simulation is carried out. The mapping, however, can become computationally infeasible with large numbers of particles or grid cells. In this talk, I will present a novel optimization method which adaptively converts gravity tree nodes into pseudo-SPH particles. These pseudo-particles act in place of the SPH particles when being passed to the MCRT code, allowing fluid resolutions to be temporarily reduced in the neutral regions. A smoothing length solver and a novel neighbour-finding scheme dedicated to tree nodes have also been implemented. I will show that our tree-based RHD scheme is capable of producing results in strong agreement with the benchmarks, and achieves a speed-up that scales with the reduction in the number of particle-cell pairs being mapped.