radiation hydrodynamics
The bulk of my research is spent performing fluid simulations with large high-performance codes like Athena++. While Athena++ is open-source and has an active development community, the necessary physics can be sufficiently problem-specific that user development is warranted. In problems related to planet formation the evolution and observational signatures are largely governed by principles of non-relativistic radiative transfer. As a necessity, I have coupled the hydrodynamics with a non-relativistic radiative transfer module using the short-characteristics method. By virtue of the method and design choices made during the implementation, my radiative transfer module offers several advantages:
- solves the true transfer equation thereby properly capturing any anisotropies
- parallelized in-line with the Athena++ meshblock framework
- second-order discretization
- capable of simulating isotropic scattering through an iterative method
- extensible to multi-frequency transport
Unfortunately, like most methods of radiative transfer, its use adds an extra 1-2 orders of magnitude in computation time and is only compatible with cartesian coordinates. Nevertheless it has been tested against several standard test problem and proven to be extremely useful in my work on planet formation.
Along similar lines, I have also taken the time to implement an exact Riemann solver for the Athena++ code. Approximate Riemann solvers tend to be adequate for the majority of problems, but my exact solver can be used to confirm their validity for a given problem and has proven to be rather useful in distinguishing numerical effects from physical ones. While these modules are not currently merged with the Athena++ master branch, either of these modules can be made available upon reasonable request.