A GPU accelerated dynamic p-adaptation for simulation of EM-plasma interaction

As a dominant physical mechanism in the highly complicated nonlinear interaction between high-power electromagnetic (EM) fields and plasmas at microwave frequencies, the ionization-diffusion mechanism results in a plasma front with an extremely high density gradient. To capture such a high density gradient and the resulting highly localized secondary EM fields, a numerical method with a high spatial resolution is required. In this paper, a dynamic p-adaptation method is proposed based on the discontinuous Galerkin time-domain (DGTD) algorithm with higher-order interpolatory basis functions by changing the basis order wherever and whenever needed. The proposed dynamic p-adaptive DGTD method is further accelerated using graphics processing units (GPUs) to allow a highly efficient parallel simulation of the EM-plasma interaction problems.