建筑扰动条件下大气流动与扩散的CFD模拟

用FLUENT模式对中性大气、单个建筑的气流扰动情况进行模拟,并以风洞试验数据检验模拟效果；将模拟方法应用于类似城市建筑阵列条件的大气污染扩散问题,并且与现场示踪试验比较. 结果表明:FLUENT对建筑扰动条件的平均风场模拟效果良好,FAC2(模拟值与试验值之比在0.5~2之间的比例)在水平与垂直风速下分别达到77.9%与61.0%；对湍流特征量的模拟偏差稍大,K(湍流动能)虽总体偏小,但FAC2仍达到了54.6%. 选择湍流闭合的标准K-ε(ε为湍流动能耗散率)方案、重整化群K-ε方案和雷诺应力模型方案对结果的影响均不大. 采用FLUENT模拟了类似城市街区建筑阵列条件的大气扩散个例, 模拟结果反映了建筑扰动导致的扩散烟流轴线相对于平均风向的非常规偏移,并且扩散浓度与示踪试验结果相符较好,下风向32与63m处的侧向模拟浓度峰值的相对误差分别为72.5%与36.9%. 相比于高斯模式ISC3,FLUENT对复杂建筑阵列条件的扩散模拟结果更符合实际,如污染物向上风向扩散以及在建筑物周围堆积与绕流的现象. FLUENT扩散模拟还显示:近源处相邻建筑街道峡谷中的最大浓度沿下风向“阶跃”式减小,排放源所在街道峡谷中的最大浓度可比相邻街谷中的高几倍甚至1个数量级以上. 

CFD Simulations of Flow and Dispersion under Construction Disturbance Conditions

Researches of flow and dispersion in urban area are important to meet the practical demands of studies of vehicular exhaust pollution and accidental leakage. Numerical simulations with field observations and physical simulations (e.g., wind tunnel simulation) are often used to address atmospheric flow and dispersion problems in cities. In general, numerical simulations can provide specific detailed informations of flow, turbulence and dispersion. To deal with small-scaled flow and dispersion, computational fluid dynamics (CFD) models are widely applied and are capable of reproducing small-scaled characteristics of flow and dispersion. A CFD model-FLUENT was adopted to simulate the disturbance of a single building on atmospheric flow under the neutral stability condition. Wind tunnel data published by Hamburg University was collected to validate the modeled results and assess the choices of parameters and parameterization schemes. This technique was applied to address atmospheric dispersion problems in quasi-urban building arrays. Concentration data of MUST and a tracer experiment were used to evaluate the dispersion modeling. The results indicated that FLUENT well performed in mean flow simulation under construction disturbance condition. Simulated results for FAC2of horizontal and vertical modeling wind speed were 77.9% and 61.0% respectively. However, thre was a relatively larger discrepancy between the modeling and measured turbulence characteristics, when the FAC2of modeling turbulent kinetic energy was 54.6%. The turbulent closure scheme options (i.e., standard K-ε scheme, Re-normalization Group K-ε scheme and Reynolds Stress Model scheme) had trivial effects on simulated results. The dispersion simulation case within quasi-urban buildings was satisfactory. The result of the dispersion case correctly reproduced the irregular plume shifting caused by building disturbance compared to the average wind direction. The modeling concentrations were generally in accordance with measured results, and the relative tolerance of the modeling maximum concentration in lateral profile at 32and 63m downwind were 72.5% and 36.9% respectively. Relative to the Gaussian plume model-ISC3, FLUENT can provide more practical concentration distribution under complex conditions. FLUENT simulation also indicated that the maximum lateral concentration decreased along the downwind direction in “step change” pattern between the adjacent street canyons near the release source. The maximum lateral pollutant concentrations in the same street canyon as the source were several times or even one order of magnitude higher than that in the next canyon. FLUENT results can reproduce the phenomena of dispersion towards upwind area and also of accumulation and detouring flow around the constructions. Further researches are required to analyze the phenomena.