Numerical Simulation Of DEM-SPH Coupling Based On Local Average Technique And Parallel Computation Of GPU

The coupling interaction between solid granules and fluids is often encountered in daily life,industrial production and natural environment,such as fluidized bed,mineral processing,grinding,debris flow,sandstorm,river slurry transport,etc.Numerical simulation is an effective means to study such problems.Accurate numerical simulation is beneficial for understanding the complex physical phenomena in the above process and provides a strong scientific basis for the optimal design of related engineering machinery and the prevention of natural disasters.The length-scale of interest about research questions determines the method of simulation for granule–fluid systems.Granule-fluid coupling numerical simulation can be divided into the following three types of models: granule-fluid Direct Numerical Simulation(DNS),Discrete Particle Model(DPM),Two-Fluid Model(TFM).At present,in the numerical simulation of granule-fluid coupling under these three types of frameworks,the traditional CFD method based on grids is used for the simulation of the fluid part,which is a very effective method for a class of problems that do not include free liquid surface.The features of the problem that studied in this paper include the free surface of large deformations,and focuses on the movement behavior of individual granules and granule groups.Therefore,SPH is adopted to simulate the dynamic behavior of fluid,and develop the coupling model of DEM-SPH.Smooth particle hydrodynamics(SPH),as a mesh-free particle method,has advantages of convenience and stability compared to traditional grid-based CFD methods,when the fluid has large deformation and complex free surface flow problems.Discrete Element Method(DEM)is a numerical method that widely used to analyze and calculate the dynamic characteristics of discrete granules.Therefore,according to the features of the research problem,in this paper we use DEMSPH coupling to perform granule fluid coupling simulation.SPH is used to solve the free-surface fluid-liquid motion,DEM simulates collisions between discrete solid granules,and each DEM granule represents a point source or sink of momentum,thereby achieving momentum exchange between DEM and SPH.This paper mainly adopts the local averaging technique of particle form discretization to develop the numerical coupling model of DEM-SPH that between granular materials and fluid.In order to ensure the rationality of coupling between solid and liquid phases,using the SPH momentum equation that derived from Lagrange equation,the momentum conservation between DEM-SPH solid and liquid phases under the DP model is effectively solved.In order to improve the efficiency of DEM-SPH coupled computational program,we adopted GPU parallel computing technology.Compared with the serial computational method,GPU parallel computing effectively improves the coupling computational efficiency and computational scale.On the basis of the above research,the process of single granules falling into the water and the solid-liquid dam break were simulated using the DEM-SPH coupling model and the results of the DEM-SPH coupling simulation are close to the experimental results,which verifies the correctness of the DEM-SPH coupling model used in this paper and the reliability of the computational program.At last,In this paper,the deficiency of DEM-SPH numerical model for simulating granular materials and fluid coupling is discussed,and its solution and application prospect are discussed.