离散统一气体动理学算法(Discrete Unified Gas Kinetic Scheme,DUGKS)是一种多尺度计算流体力学方法,该方法适用于所有流域的流体力学模拟.在三维不可压缩热流体的仿真中,原始双分布DUGKS的算法结构较为复杂,且时间步长受限于较小的CFL数(CFL＜1),其计算效率亟待提升.在对原始DUGKS的研究中发现,可以通过改变原始算法中对微通量的求解方式来优化算法结构,发展一种三维热流体简化DUGKS.理论上,简化DUGKS可以使计算量节约50％左右.三维自然对流的数值模拟结果则进一步表明:在相同的计算网格和时间步长下,简化DUGKS的数值精度与原始DUGKS相当,其计算速度接近原始DUGKS的2倍;简化DUGKS 在 CFL数为1.7时仍能获得精确的预测结果,采用较大的时间步长可以使计算成本进一步节约70％.算法结构的简化和数值稳定性的提高使DUGKS的计算效率大幅提升.
Simplified discrete unified gas kinetics scheme for three-dimensional thermal flows
The discrete unified gas kinetics scheme(DUGKS)is a multi-scale computational fluid dynamics method,which can capture the flow physics of all flow regimes.In the simulation of three-dimensional(3D)thermal flows with the original DUGKS,the complex structure will lead to enormous computational cost,and the time step is always limited by a small CFL number(CFL＜1).Hence,it is urgently necessary to improve the computational efficiency of DUGKS.In the study of DUGKS,a simplified 3D DUGKS is developed for thermal flows by changing the method of solving the micro flux in the original algorithm.Theoretically,simplified DUGKS can save about 50％of the calculation.While the numerical simulation results of 3D natural convection further show that,the numerical accuracy of the simplified DUGKS is coincident with the original DUGKS,but its computational efficiency is close to twice as that of the original DUGKS under the same calculation grid and time step.In addition,the simplified DUGKS is still accurate when the CFL number is up to 1.7,which further improves the calculation efficiency by about 70％.In short,the simpler structure and the superior numerical stability make the present scheme much more efficient than the original DUGKS.
discrete unified gas kinetics scheme3D natural convectionsimulation of thermal flowsimplified scheme
NETL
NSTL
万方数据
