Skip to end of metadata
Go to start of metadata

This is a page created for our Direct Simulation Monte Carlo (DSMC) particle transport code which is being developed at Michigan State University. Currently, members of the group are:

Wolfgang Bauer (MSU, iCER), Dirk Colbry (iCER), Jim Howell (MSU), Rodney Pickett (iCER), Irina Sagert (MSU), Alec Staber (MSU), Terrance Strother (LANL)

The page is still under construction. For the latest information concerning our kinetic code see:

http://arxiv.org/abs/1210.8084

http://arxiv.org/abs/1305.4220

http://arxiv.org/abs/1310.3437

 

Above is a simple example of how the DSCM particle code operates. Particles move along straigth lines until they find an interaction partner according to their distance of closest appraoch. The interactions can be various, as long as they are of short range type, that is only binary interactions have to be taken into account. In our case, particles scatter elastically. Due to their interaction they change their velocity and direction of motion, and continue to move until the next interaction occurs.

In the above simulation, the particle density is very low and interactions take place only from time to time. This is a situation we would find for matter in a gaseous state. If we initialized particles with the same absolute velocities and waited, after a while, the particles would equilibrate to the Maxwell-Boltzmann velocity distriubtion. Hereby, the time to reach equilibration depends on how frequent particles scatter with each other. With the DSMC method, we are able to influence this interaction frequency by changing the particles' scattering cross-sections, i.e. the their mean free paths. That way it is possible to simulate systems in equilibrium - when interactions between particles are frequent -, and also out-of-equilibrium, when the interaction rate is low and the system needs a long time, for example, to obtain the Maxwell-Boltzmann velocity distribution. 

If we set the particle density to a large value and ensure frequent interactions, we obtain systems that are highly equilibrated. With that, we can simulate phenomena which usually can only be studied with hydrodynamic simulations, for example shock wave propagation. Below are some examples of our current studies.

Shock wave studies:

Sod shock test

Noh shock test

Sedov test

Shock wave movies:

2D Sod simulation (7 MB) (old)

N = 32,000,000 test-particles, number of timesteps: 350, output every 10 timesteps

2D Sod simulation (1.4 MB) (new)

2D Noh simulation (4.5 Mb) (old)

N = 32,000,000 test-particles, number of timesteps: 500

2D Sedov simulation (old)

 

Fluid instability movies:

Kelvin-Helmholz Instability (high resolution), ca. 200KB

Kelvin-Helmholz Instability (low resolution), ca. 1MB

Rayleigh-Taylor Instability, soon to come