HydrodynamicsLessons in this section are a basic introduction to the hydrodynamics module. The preprint of the research paper that will describe the numerical procedure of the hydrodynamics and a first application will soon be available here. In this section we describe the basic setup and the test runs that we have performed so far. The supernova project file can be downloaded here. It is used as the demo project for this series of tutorials. This page also presents standard tests of the hydrodynamic code for comparison with analytical predictions. Note that the Hydro Module in Shape runs 3D simulations on a regular cubic grid. No 1D or 2D simulation with symmetry assumption are done. This is due to the overall concept of how Shape works internally. This puts some restrictions on the simulations that can be done in terms of hardware, resolution and time it takes to run them. For publication grade simulations we recommend a 64 bit system with over 10 GB of RAM (which have to be made available to Shape in the file "Shape_v5.0.vmoptions" that is located in the installation directory). Since the code is parallelized using threading it takes advantage of multiple CPU cores. Hence the more cores you have, the faster the simulation runs. |
|||||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||||
Hydro code numerical scheme and test results Hydrodynamic simulation codes are based numerical procedures for which there are many options, each of which has advantages and disadvantages compared to others. This initial version of the code in Shape is designed to provide a basic tool to simulate the structure and evolution of nebulae that involve shocks produced by winds or explosions with cooling. This was to be achieved with a minimal computational effort, to make it interactively usable on desktop computers and workstations, rather than clusters or supercomputers. Still, the code is fully threaded, which means that it can take advantage of multi-core architectures, with corresponding speed-ups The numerical scheme is a second order upwind shock-capturing Godunov method including a van Leer flux--vector splitting (van Leer, 1982). For more details see the research paper on the code (Steffen et al., 2013, MNRAS, submitted). Cooling is treated as a parameterized three-segment linear (in log-log space of temperature and energy loss) approximation above 10^4 K. Below 10^4 K there is no cooling at this time, the peak is at log(T/K)=5.5 and the free-free starts at log(T/K)=7.5. In this section we present a few tests that illustrate the degree of accuracy of the implementation.
|