LES和DNS在平坡异重流数值模拟中的比较研究

Comparative Study of LES and DNS on Numerical Simulation of Gravity Current over a Flat Bed

摘要: 本文利用FLUENT软件模拟了开闸式异重流的演变特性。使用二维大涡模拟(LES)和直接数值模拟(DNS)，对异重流的形态变化、瞬时掺混系数、势能转变，动能剖面变化等进行了相关的对比分析。结果表明，LES和DNS对于异重流头部位置的预测都具有较好的准确性，DNS的结果略微精确，但需要较多的计算网格数。此外，在异重流与环境流体的交界面发展过程中，相比于LES模拟的结果，DNS的结果具有较少的K-H不稳定性产生。DNS模拟的异重流基本上集中于头部，而且DNS模拟的结果与实验异重流的形态较为接近。LES和DNS模拟计算得出的异重流与环境流体的掺混程度相近，但DNS的交界面较为光滑。LES和DNS对于异重流沿途势能变化的结果相近，但DNS的结果具有较小的背景势能。对于特征断面处动能逐时变化，相较于LES，DNS模拟较少出现多次波峰的震荡现象。本文的结果可为后续模拟异重流运动时，提供选择数值模型的参考依据。

Abstract: In this paper, the FLUENT software was used to simulate the evolution characteristics of the lock-exchange gravity currents over a flat bed. Two-dimensional large eddy simulation (LES) and direct numerical simulation (DNS) were utilized to compare the morphological change, instantane-ous mixing coefficient, potential energy transformation, and kinetic energy at the characteristic section. The results show that both LES and DNS have good accuracy in predicting the head position of the gravity currents, but the results of DNS are slightly more accurate with more computational grids required. For simulating the development process of gravity currents, the DNS results show less K-H instability occurring at the interface of current and ambient fluid than the LES simulations do. For DNS simulations, heavy fluid basically concentrates on the current head. In addition, the DNS results are closer to the gravity current morphology from experimental observations. The mixing degree of density flow with ambient fluid estimated by LES and DNS simulations is similar, but the result of DNS shows smoother current-ambient fluid interface. LES and DNS results obtain similar potential energy changes along the course of density flow, but the results of DNS have smaller background potential energy. For the temporal variations of kinetic energy at characteristic cross-sections, compared with LES, DNS performs fewer oscillating peaks. The results of this paper can provide a reference for the selection of numerical models on gravity current simulations in the future.