Optimization and stabilization of LBM free surface flow simulations using adaptive parameterization

We present a method to speed up and stabilize free surface simulations with the lattice Boltzmann method (LBM). This is done by adaptively changing the parameterization of the simulation in a way that corresponds to a different size of the simulation time step. This means that the Mach number changes as well, and requires a rescaling of all distribution functions. Hence we only perform the rescaling when the velocities in the simulation become too large or small. We will demonstrate the effect of this procedure for two and three-dimensional test cases. In addition to a reduction of the necessary LBM steps, this method can also be used to stabilize gravity driven simulations, where the maximum velocities are not known a priori.