FLASH was originally developed by the University of Chicago under the auspices of the DOE/ASCI program and now maintained by the Flash Center for Computational Science (link) at the University of Rochester. FLASH is a multi-physics, finite-volume, Eulerian code whose capabilities include Adaptive Mesh Refinement (AMR) on a block-structured mesh; state-of-the-art hydrodynamics and magnetohydrodynamics solvers; implicit solvers for diffusion using the HYPRE library that include thermal conduction, radiation diffusion, resistivity, and viscosity; and a generic, highly scalable, parallel particles framework including Lagrangian tracer particles.It uses various split and unsplit hydrodynamic solvers including 5th order Weighted essentially non-oscillatory (WENO) methods and the piecewise parabolic method (PPM). FLASH is a professionally software-engineered code (Fortran based) with a wide user base and has been applied to a variety of astrophysics problems, high energy density physics experiments, and to fundamental problems such as combustion, fluid instabilities, and turbulence. FLASH is compiled efficiently with unique physics for each problem and is highly parallel scaling to 100K cores. The AMR capabilities and simple block-structured mesh make it a highly efficient code.
The Eulerian-Lagrangian particle-in-cell (PIC) method was implemented in FLASH by our research group, and includes phase change and breakup for droplets. In this method particles are represented as meshless Lagrangian points which interact with the continuous phase represented on an Eulerian mesh. For computational efficiency, the Lagrangian points are virtual particles and can represent groups of particles (up to 1000s of particles per point). The discretized conservation equations in the split Piecewise Parabolic Method (PPM) solver are modified to include source terms for the mass, momentum, and energy to introduce the effect of active particles. The momentum source appears due to the drag force experienced by the particles, where the coefficient of drag is calculated using various drag models. The energy and mass source terms are found using various models including the Ranz-Marshall correlation and Spalding transfer model. Mass transfer significantly modifies the energy source term due to the latent heat of vaporization. Phase change properties are calculated from thermodynamic curve fits. A new species where the particle vapor phase is stored after evaporation is initialized with the mass transfer model. These models have been validated against analytical solutions and known values. Breakup models (for various breakup regimes) have also been implemented, and validated against past experimental work from our group. Various reaction models have also been implemented including a generic N-species, M-step framework for multistep models.Together with the breakup models, the code allows us to simulate complex reacting mutliphase flows.