四种紊流模型模拟弯道水流精度的对比分析

1)长江科学院,武汉 430010; 2)三峡地区地质灾害与生态环境湖北省协同创新中心,湖北宜昌 443002; 3)武汉大学水资源与水电工程科学国家重点实验室,武汉430072; 4)黄河勘测规划设计有限公司,郑州 450003

水利工程; 紊流模型; 弯道水流; 数值模拟; 水面线; 流速分布; 紊动能; 环流强度

Comparative analysis of simulation accuracy for flow in bends using four turbulence models
Wu Huali1,2, Zhang Xiaofeng3, Jin Zhongwu1,2, and Chen Cuixia4

1)Yangtze River Science Research Institute, Wuhan 430010, P.R.China2)Collaborative innovation Center for Geo-Hazards and Eco-Environment in Three Gorges Area, Yichang 443002, Hubei Province, P.R.China3)State Key Laboratory of Water Resources and Hydropowe

hydraulic engineering; turbulence model; bend flow; numerical simulation; water surface profile; velocity distribution; turbulent energy; circulation intensity

DOI: 10.3724/SP.J.1249.2015.05506

备注

采用标准k-ε模型、RNG k-ε模型、可实现k-ε模型和Reynolds应力模型等4种经典的紊流模型,以连续弯道水流为对象进行数值模拟.根据水槽试验资料,从弯道水流纵比降、横比降及纵向流速分布等方面,对比分析了4种模型模拟结果的精度.认为Reynolds应力模型能更好地再现弯道中复杂的三维紊流特征,具有较高的精度,模拟结果优于标准k-ε模型、RNG k-ε模型和可实现k-ε模型.以单弯道水流为对象,从沿程水面线、纵向流速沿垂线分布、紊动能沿水深分布及环流强度分布特性等方面,检验了Reynolds应力模型的精度. 发现在模拟岸边界变化引起的二次流现象时,Fluent软件中的Reynolds应力模型效果优于其他3种紊流模型.

We carry out numerical simulations for the turbulence characteristics of meandering channels by four different turbulence models as standard k-ε model, RNG k-ε model, realizable k-ε model and Reynolds stress model, respectively, by using the Fluent software. We compare the accuracies of the four models and analyze multiple aspects of longitudinal surface slope, transverse slope of water surface, longitudinal velocity distribution and so on. The results show that the Reynolds stress model is the best in accuracy for bend flow among the four turbulence models. Then we use the Reynolds stress model to simulate the flow characteristics of a single-bend channel and further test the accuracies in water surface profile, longitudinal velocity distribution along the vertical, turbulent kinetic energy distribution along the depth and circulation intensity distribution characteristics. Finally, we find that the Reynolds stress model can better simulate the secondary flow caused by changes of land boundary than the other three models.

·