中文摘要：

利用计算流体力学软件FLUENT，采用七种不同湍流模型（标准k-ε、RNGk-ε、realizablek-ε、标准k-ω、SSTk-ε、RSM和Spalart-Allmaras）对城市街道峡谷内的气流运动和污染物扩散进行了数值模拟。计算获得的峡谷内气流场和污染物浓度分布表明，虽然七种湍流模型都预测出峡谷内生成一个控制污染物输移的顺时针大旋涡（主旋涡），但它们在次生涡的模拟上存在差异。通过将峡谷迎风面和背风面上的无量纲污染物浓度分布模拟结果与风洞试验数据进行对比分析，揭示出：标准k-ε模型的总体预测效果最好，RNGk-ε模型、realizablek-ε模型以及RSM模型的预测效果次之，而标准k—ω模型、SSTk-ω模型以及Spalart-Allmaras模型的预测效果较差。在环境评价和基于环境容量约束的交通配流方案优选中，本研究成果对于如何合理选择湍流模型来模拟预测机动车排放污染物在峡谷内的扩散分布具有重要指导意义。

英文摘要：

Numerical calculations were performed using the FLUENT code and the seven different turbulence models （i.e. a series of standard, RNG and realizable k-ε models, the standard and SST k-ω models, the Reynolds Stress Model, and the Spalart-Allmaras model） to simulate the airflow and pollutant dispersion inside an urban street canyon. The simulated wind fields and concentration contours show that all the studied turbulence models produce a similar main vortex structure with clock-wise that controls pollutant transport inside the canyon but they display differences in the prediction of secondary eddies. The calculated pollutant concentration profiles along the lee- and windward walls of the canyon were compared with the wind tunnel measured data. It is revealed that （1） the standard k-ε model gives the best overall performance, whereas the performance of the RNG and realizable k-ε models as well as RSM is secondly good, （2） the performance of the Spalart-Allmaras model and the standard and SST k-ω models is less good since the pollutant concentrations on the lee-and windward walls of the canyon are underestimated by the Spalart-Allmaras model and overestimated by the standard and SST k-ω models.