Volumetric evaluation of meshless data from smoothed particle hydrodynamics simulations

Abstract

As an alternative to conventional Eulerian methods in the field of computational fluid dynamics (CFD), smoothed particle hydrodynamics (SPH) has been developed. Its mesh-free method is useful in problems, especially, where a free surface is present. Although researchers are currently able to simulate up to hundreds of millions of particles in a volume, the analysis and visualization of the particle datasets are still limited especially, due to the lack of connectivity between particles and the number of particles. In this paper, we present a volumetric evaluation technique of SPH particle data using hierarchical particle data structures. Our approach does not resample or triangulate the meshless particle data on grid structures, instead, the evaluations are performed in a fragment program with the original SPH kernels used in the CFD simulations. The SPH particle information and the hierarchical data structures are stored in 2D and 3D textures respectively and the 3D texture stores the access pointers, such as texture coordinates, to the 2D textures. To achieve interactive frame rates during interaction we suggest to control the kernel radii while generating the hierarchical data structures. This approach allows us to visualize over a million of SPH particles interactively.