Coupled DEM-SPH Method for Interaction between Dilated Polyhedral Particles and Fluid

The discrete element method (DEM) and smoothed particle hydrodynamics (SPH) can be adopted to simulate the granular materials and fluid media respectively. The DEM-SPH coupling algorithm can be developed for the dynamic interaction between the two media. When the particle material is simulated by polyhedral element, a fluid-solid coupling interface would lead to the complex geometry between the granular particle and the fluid. The boundary particle method is traditionally used for the fluid-solid interface but with low computational efficiency. In this paper, the dilated polyhedral element is constructed based on Minkowski sum theory, while the contact force between the elements is calculated by Hertzian contact model. Accordingly the dilated polyhedra based DEM is established. The weakly compressible SPH is adopted to simulate the fluid medium, while the interaction on the geometrically complex fluid-solid interface is evaluated with the repulsive force model which can be determined by the contact detection between SPH particles and solid particles in geometry. This method avoids the storage and calculation of a large number of boundary particles, which can be potentially applied for the complex fluid-solid boundary. In order to improve the computational efficiency, a GPU-based parallel algorithm is employed to achieve high performance computation of SPH. The acceleration of the parallel algorithm is evaluated by the cases of dam break. The numerical simulation of the impact of dam break on cubes is implemented. The simulation results are verified with the corresponding experimental and simulation results. Therefore, the rationality and accuracy of the DEM-SPH coupling method for numerical simulation of the interaction between granular materials and fluid media are illustrated. This method is then adopted for the impact of falling rocks on underwater pipeline. The force of water and rocks on the pipeline is analyzed. This method can be further applied for real engineering problems.