Influence of SPH Regularity and Parameters in Dynamic Fracture Phenomena

Abstract

The Smoothed Particle Hydrodynamics (SPH) method can be used with advantage in the field of fracture mechanics, which is especially true when quasi-brittle materials are involved. The advantages of the SPH method are more evident when loading speed increases and dynamic material fractures start to occur. Since the SPH method is a meshfree method, the large deformation and eventual fragmentation of material during simulations can be solved without major complications. This happens because of a phase of the SPH method in which a search is made for neighbouring particles and the constraints created between them within a chosen time interval. The number of neighbouring particles depends on the size of the area where the search takes place. This area – the support domain – may therefore be considered as one of the key control elements in simulations using the SPH method. The influence of the number of particles and their initial distribution on the results is also a question. Particle clusters (areas with increased particle concentration) may be formed in cases of poor regularity. Consequently, false (numerical) cracks which bypass these clusters may appear in the simulation. The article describes an experiment concerning the dynamic loading of concrete L-specimens simulated by the SPH method. Different density distributions and initial particle distribution regularities are chosen in the simulation. The results show that it is especially necessary for the initial configuration to exhibit regular particle distribution if simulations are to be executed successfully. False cracks tend to occur more frequently with increasing particle distribution irregularities. A certain degree of compensation can be achieved via the appropriate choice of support domain size with its variations during the simulation.The Smoothed Particle Hydrodynamics (SPH) method can be used with advantage in the field of fracture mechanics, which is especially true when quasi-brittle materials are involved. The advantages of the SPH method are more evident when loading speed increases and dynamic material fractures start to occur. Since the SPH method is a meshfree method, the large deformation and eventual fragmentation of material during simulations can be solved without major complications. This happens because of a phase of the SPH method in which a search is made for neighbouring particles and the constraints created between them within a chosen time interval. The number of neighbouring particles depends on the size of the area where the search takes place. This area – the support domain – may therefore be considered as one of the key control elements in simulations using the SPH method. The influence of the number of particles and their initial distribution on the results is also a question. Particle clusters (areas with increased particle concentration) may be formed in cases of poor regularity. Consequently, false (numerical) cracks which bypass these clusters may appear in the simulation. The article describes an experiment concerning the dynamic loading of concrete L-specimens simulated by the SPH method. Different density distributions and initial particle distribution regularities are chosen in the simulation. The results show that it is especially necessary for the initial configuration to exhibit regular particle distribution if simulations are to be executed successfully. False cracks tend to occur more frequently with increasing particle distribution irregularities. A certain degree of compensation can be achieved via the appropriate choice of support domain size with its variations during the simulation.