Parallelized Radiative Transport and Phase Space Distributions in Heavy Ion Collisions
Abstract
Numerical solutions of the Boltzmann transport equation (BTE) present a framework for modeling non-equilibrium dynamics in heavy ion collisions. However, the computational power required to solve the seven-dimensional integro-differential equation reaches impractical levels for realistic, high-statistics simulations involving radiative 2 to 3 and 3 to 2 scattering processes with sequential (single-processor) algorithms. This thesis presents a new parallelized MPC/Grid code that was developed to enable such simulations. The code was tested extensively for correctness, and speedups of up to about 30x were seen relative to single-processor execution. The parallelized code was then used in a study that required high-statistic simulations, to address the ambiguity in the conversion from a fluid dynamical description to a particle description of a system. Such conversion is necessary in all comparisons of hydrodynamic simulation results to experimental data. Four existing fluid-to-particle conversion models for shear viscous fluids were assessed based on their ability to reconstruct, using hydrodynamic variables alone, the full transport phase space density for a massless one-component gas undergoing 2 to 2 scatterings in a 0+1D boost-invariant Bjorken scenario. Besides establishing the regions of validity of the four models, novel improvements are proposed that greatly increase the reconstruction accuracy of these models (by about 10x relative to the most commonly used model). Analytical simplifications of the BTE in the near-free-streaming regime are also presented, in order to gain insight into the functional form of phase space densities in the presence of interactions. These will enable the construction of yet more accurate, theoretically well-founded fluid-to-particle conversion models in the future.
Degree
Ph.D.
Advisors
Molnar, Purdue University.
Subject Area
Nuclear physics|Particle physics
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.