An adaptive mesh refinement method based on a characteristic-compression embedded shock wave indicator for high-speed flows

Numerical simulation of high-speed flows often needs a fine grid for capturing detailed structures of shock or contact wave, which makes high-order discontinuous Galerkin methods (DGMs) extremely costly. In this work, a characteristic-compression based adaptive mesh refinement (AMR, h-adaptive) method is proposed for efficiently improving resolution of the high-speed flows. In order to allocate computational resources to needed regions, a characteristic-compression embedded shock wave indicator is developed on incompatible grids and employed as the criterion for AMR. This indicator applies the admissible jumps of eigenvalues to measure the local compression of homogeneous characteristic curves, and theoretically can capture regions of characteristic-compression which contain structures of shock, contact waves and vortices. Numerical results show that the proposed h-adaptive DGM is robust, efficient and high-resolution, it can capture dissipative shock, contact waves of different strengths and vortices with low noise on a rather coarse grid, and can significantly improve resolution of these structures through mild increase of computational resources as compared with the residual-based h-adaptive method.