摘要
结合离散单元法(DEM)和计算流体力学(CFD)的CFD-DEM被广泛应用于流化床的数值模拟中,但受DEM计算工作量大的限制,难以达到实际流化床的模拟要求,因此开展对DEM并行算法的研究具有重要意义。针对流化床内颗粒运动的特点,基于DEM建立了气固两相流动中颗粒相的并行化计算模型,在OpenMP并行技术的基础上,实现了颗粒受力计算、颗粒速度和位置更新的并行化。采用Fortran语言开发了DEM并行数值模拟程序,模拟了三维流化床内颗粒堆积的过程,并与串行程序的模拟结果进行了对比。结果表明:基于OpenMP技术并行程序可正确模拟流化床内的颗粒堆积过程,而且具有良好的加速比,能够显著地提高计算效率、缩短计算时间;当计算线程数增加时,加速比随之增加,但增幅相对减少。
The CFD-DEM is widely used in numerical simulation of fluid beds,but restricted by computing workload of DEM,it is difficult to meet the requirements for simulation of actual fluidized beds.Therefore,conducting researches on parallel algorithms for DEM is particularly important.According to the characteristics of particle motion in the fluidized bed,parallel computing model of particle- phase in gas solid twophase flow based on the method of DEM was established,which can achieve the parallelization between force calculation of particles and update for velocity and location of particles on the basis of the OpenMP parallel technology.A 3-D parallel DEM code for simulating the particle deposited process in a fluidized bed was developed based on the OpenMP standard in this paper.The computation results which used parallel program were compared with that used serial program.The results of parallel program show that it has a better speedup than the serial program,and the computational efficiency is improved greatly with the parallel program.As the number of threads increasing,the speedup ratio increases gradually,while the increase rate is relatively reduced.
出处
《热力发电》
CAS
北大核心
2016年第1期42-48,共7页
Thermal Power Generation
关键词
流化床
颗粒堆积
离散单元法
计算流体力学
气固两相流
并行计算
OPENMP
加速比
fluidized bed
particle accumulation
discrete element method
computational fluid dynamics
gas solid two phase flow
parallel computation
OpenMP
speedup ratio
