基于GPGPU的LBM障碍绕流的实时模拟  

Real-Time Flow around Obstacles Simulation Based on Lattice Boltzmann Method on GPGPU

作　　者：张强[1] 张卓鹏[1] 王长波[1] 夏正冬[1] 崔军营[1] 

机构地区：[1]华东师范大学软件学院,上海200062

出　　处：《东华大学学报（自然科学版）》2011年第4期408-411,共4页Journal of Donghua University(Natural Science)

基　　金：国家自然科学基金资助项目(61070128;60603076);中央高校基本科研业务费专项资金资助项目

摘　　要：基于显卡的通用计算(GPGPU)是近年来并行计算和快速绘制的热点.格子Boltzmann方法(LBM)作为流体动力学的新方法,其并行性好,常常用于基于物理的流体模拟,且具有适用于复杂边界障碍的特性,但计算较为复杂.利用GPGPU技术来加速LBM的流体计算模型,构建了基于图形处理器(GPU)的流体计算框架,实现了格子Boltzmann计算的D2Q9和D3Q15模型,并用于实时的障碍绕流模拟.General purpose graphics process unit（GPGPU） which based on the graphics card has become a hotspot in the area of parallel computing and fast rendering.Lattice Boltzmann method（LBM）,a newly introduced method for computer fluid dynamics with high parallelism,is commonly applied to physics-based fluid simulation with expandability to complex boundaries and obstacles model but complex computation.The fluid simulation is accelerated by GPGPU and the framework of fluid simulation based on graphics process unit（GPU） is presented.And the LBM D2Q9 and D3Q15 model used in real-time fluid flow around obstacles simulation is realized.

关 键 词：图形处理器(GPU) 格子BOLTZMANN方法 OpenGL着色语言 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]