Study On Neighborhood Particle Search Algorithm And Its Application In Fluid Simulation Based On SPH Method

Neighborhood particle search algorithm is widely applied in SPH method, discrete element method and other meshless methods. Neighborhood particle search is a method which can find the neighborhood particles of the current particle from the problem domain. The paper studied the different neighborhood particle search algorithms, and a local related-area particle search algorithm was designed, several fluid simulation applications based on the SPH method were had been made to verify the algorithm. In the paper, the main contents are as follows:Firstly, we studied the existing neighborhood particle search algorithms, summarized the searching principle. A more efficient particle search algorithm was designed. This process contains the following three points of work.(1) Studied the Neighborhood particle searching algorithms including global particle search matching algorithm, the linked-list search algorithm, the tree search algorithm and other search algorithms such as Bucket algorithm. The searching process and computational complexity of these algorithms were analyzed.(2) Summarized the existing search algorithms, reduced the searching areas and improved the data utilization, designed a fast neighborhood search algorithm based on local related neighbor areas. The basic idea of this algorithm is:Taking full advantage of the obtained data, to get the searching areas for the current particle, then to find the neighborhood particles from the neighborhood areas. During designing the algorithm, for the problem of the surface particles'neighborhood failed to be found, compared the number density of neighborhood particles and location information, to extract the surface particles and the search process of the surface particles were dealt with separately. For the problem of searching overlapping areas, boolean variable was applied to mark the searched particles, so as to eliminate the repeated searching. The algorithm does not need to add mesh or to re-establish complex data structure, it not only reduced the search areas, but also improved the data utilization.(3) The particle information and algorithm function interface were packaged while designing the search algorithm, so it is easy to compare and to validate the algorithm with other different search algorithms in fluid experiments.Secondly, A fluid simulation system platform base on SPH method was designed and implemented by using C++programming language and OpenGL graphics technology, the system applied three different kinds of searching algorithms which are global search algorithm, linked-list search algorithm and the paper's search algorithm to do three different fluid simulation experiments which are dam break, fluid free dripping and single screw extruder. After experiments data was had been compared and fitted, we could verify that the algorithm in my paper is performed much better.