Numerical simulation of three-dimensional hypervelocity impact using smoothed particle hydrodynamics

The searching of neighbor particles is the key to the computational efficiency of smoothed particle hydrodynamics (SPH). An algorithm of multi-level dynamic grid-linked list search was presented. All the background grids are squares or cubes with the side length of one level of grids being whole number times of the lower level. Only limited amount of memory was consumed to store dynamically the virtual coordinates of background grids and particles. The number of grid levels and the multiple of different levels of grid sizes can be selected optimization ally according to the particle number. The optimized values of grid parameters were obtained through verification of searching efficiency. The optimized results show that the searching efficiency is dominated directly by the number of grid levels and influenced by the multiple of different levels of grid sizes to a certain extent. The searching algorithm was validated by the simulation of two oblique hypervelocity impact examples.