紊流模型预测街道峡谷流场能力的定量指标评价

Quantitative index evaluation of the predictability of flow field in the street canyons by turbulent models

选取标准k-ε模型、realizable k-ε模型、重整化群(RNG) k-ε模型和剪切应力传输(SST) k-ω模型分别对街道峡谷水槽模型试验的时均速度和脉动强度进行了数值模拟,并依据文献给出的试验数据定性比较了各紊流模型的预测能力;按模型预测值应满足的基本和较高要求共给出6个定量评价指标(IFAC2、IFB、IFAC2-NLV、IFAC20%、IFAC20%-NLV和INMSE)。结果表明:RNG k-ε模型对流动的综合预测能力最好,其次是标准k-ε模型,realizable k-ε模型表现一般,而SST k-ω模型表现相对较差;定量与定性的评价结果基本一致,但定量指标能给出更明确的判断依据,尤其当模型结果差异性不显著时,可利用多个定量指标综合评判。

The standard k-εmodel,realizable k-εmodel,renormalization group(RNG)k-εmodel and shear-stress transport(SST)k-ωmodel were selected to numerically simulate the time-averaged and fluctuating velocity field of street canyons in the water channel experiment,and the predictability of each turbulence model was qualitatively compared based on the experimental data from the literatures.Then a total of six evaluation indices(IFAC2,IFB,IFAC2-NLV,IFAC20%,IFAC20%-NLV,and INMSE)were used to give quantitative assessments of these turbulent models.The results show that RNG k-εmodel is the best model in terms of the comprehensive predictability and follows by standard k-εmodel.The performance of realizable k-εis acceptable while that of SST k-ωmodel ranks the worst.Quantitative and qualitative evaluation results were basically the same.But the quantitative indices can provide a clear judgment criterion.Especially when the difference among the models is minor,multiple indices can be used for comprehensive evaluation.