Particle flow simulation system and method

Abstract

The invention discloses a particle flow simulation system and method on the basis of GPUs. The method includes: generating particle information according to particle modeling information input by a client, and generating geometry information; receiving the particle information and the geometry information, determining that which GPUs in which computational nodes are used according to the number of particles and the number of idle GPUs in the computational nodes, determining that which particles are calculated by which GPUs in which computational nodes according to the determined number of the GPUs and distribution condition of the particles in space, and performing distribution according to determined results; calculating stress of each particle caused by impact of the particles in a plurality of GPUs in a parallel mode, and working out accelerated speed further so as to simulate particle flow; showing simulation results. According to the particle flow simulation system and method, high-density particle virtual experiment simulation can be achieved, energy consumption is reduced, and meanwhile calculation efficiency is improved.

Description

technical field [0001] The invention relates to the technical field of particle flow simulation, in particular to a GPU-based particle simulation system and method applicable to the research of granular matter or solid structure. Background technique [0002] Particle system has always been the research content of people's attention. In industrial fields such as: food control, chemistry, civil engineering, oil and gas, mining, pharmaceuticals, powder metallurgy, energy and other fields have a large number of applications; in terms of theoretical research, how to accumulate the most dense accumulation, and under what circumstances the sand pile collapses To study issues such as avalanches. In order to study related particle systems, people need to build large-scale experimental particle systems, which is time-consuming and labor-intensive. Moreover, due to the high cost, some particle systems need to operate under extreme conditions, so it is impossible to build them throug...

AI Technical Summary

Problems solved by technology

At present, the implementation of DEM methods is mainly based on CPU. Due to the insufficient computing power of CPU, these methods can only calculate small spatial and temporal dimensions within an acceptable computer time; or due to high construction costs, it is necessary to build Large-scale or even ultra-large-scale CPU computer clusters consume too much power, and the cost of use and maintenance is extremely high

In addition, the current DEM method implemented by CPU either has a small number of particles or a low-density particle collision, which cannot fully realize the high-density and large-scale particle collision simulation

Images