Geometrical on-the-fly shock detection in smoothed particle hydrodynamics

We present an on-the-fly geometrical approach for shock detection and Mach number calculation in simulations employing smoothed particle hydrodynamics (SPH). We utilize pressure gradients to select shock candidates and define up- and downstream positions. We obtain hydrodynamical states in the up- and downstream regimes with a series of normal and inverted kernel weightings parallel and perpendicular to the shock normals. Our on-the-fly geometrical Mach detector incorporates well within the SPH formalism and has low computational cost. We implement our Mach detector into the simulation code gadget and alongside many SPH improvements. We test our shock finder in a sequence of shock tube tests with successively increasing Mach numbers exceeding by far the typical values inside galaxy clusters. For all shocks, we resolve the shocks well and the correct Mach numbers are assigned. An application to a strong magnetized shock tube gives stable results in full magnetohydrodynamic setups. We simulate a merger of two idealized galaxy clusters and study the shock front. Shock structures within the merging clusters as well as the cluster shock are well captured by our algorithm and assigned correct Mach numbers.