Author(s): Wen-Hao Pan; Hua-Kang Zheng
Linked Author(s):
Keywords: K-ε model; Vegetated flow; Wake kinetic energy; Double averaging
Abstract: Rivers, wetlands and lakes with vegetation are of great research interest in flood control, contaminant transport, and biological conservation. While the near bed turbulent structures at the plant scales have significant effects on momentum and mass transport processes, resolving the flow field at individual blade scale still remains computationally impractical. To address quantitatively the complex issues imposed by a high spatial variability of this kind of flow, the double averaging approach was proposed by many researchers as an integrative up-scaling framework. However, there still exists problems while transferring the characteristics of individual plant scales into a path scale, especially for second order flow field characteristics. To improve the double averaging procedure in describing second order characteristics, this study developed a k-ε model by modifying the modeling of wake turbulence kinetic energy generation term. In this model, the conservation equations for emergent and submerged vegetated flows are averaged both spatially (horizontally) and temporally. The turbulent kinetic energy (TKE) was split into shear kinetic energy and wake kinetic energy (WKE). We propose that the WKE is generated not only by the vegetation stem but also the vertical shear at the water-vegetation interface, which can be proved by the power spectrum of flow velocity. The model was validated against experiments of flow through emergent and submerged vegetation for mean velocities, turbulent intensities, Reynolds stresses, and different terms in the TKE budget. Results show that this model provides very good representation of the mean velocity and Reynolds stress. Additionally, it outperforms exiting k-ε models in the terms of TKE and dissipation rates, which are highly dependent on the stem scale both in submerged and emergent conditions. This model was also used to predict vegetation-induced mechanical energy loss in the term of Manning roughness coefficients, which show reasonable accordance with certain field observations. The proposed model is limited to flows with sufficiently large stem Reynolds number.
Year: 2024