New dynamic two-layer model for predicting depth-averaged velocity in open channel flows with rigid submerged canopies of different densities
Introduction
The presence of vegetation is of great importance in the overall conveyance capacity of natural channels (Darby, 1999; Stone and Shen, 2002). Numerous studies have been conducted to explore the flow characteristics of vegetated open channels, with many of them focusing on explicit velocity distribution (Huai et al., 2009; Nikora et al., 2013; Tang et al., 2014). However, in natural rivers depth-averaged velocities are used more frequently than distinct velocity profiles in solving practical engineering problems (Li et al., 2015b). Moreover, canopy densities in actual situations are usually variable. This means depth-averaged velocity formulae that have a unified basis for vegetated flow with different canopy densities are still needed. Using rigid cylinders to mimic vegetation, we attempt to develop a new model of depth-averaged velocity in open channel flows with submerged vegetation of different densities. The aim of this model is to explore the physical effects of vegetation stems without the additional biological factors, such as leaves and root systems.
Two major methods, namely, analytical and experimental methods, are currently utilized to address the issue of depth-averaged velocity. The two-layer division approach based on geometric factors is most commonly used (Stone and Shen, 2002; Baptist et al., 2007; Huthoff et al., 2007; Yang and Choi, 2010; Cheng, 2011; Shi et al., 2019). For instance, in the study of Yang and Choi (2010), submerged vegetated flow is divided into two layers, namely, vegetation and surface layers, with the top of the canopy as the boundary, which is shown in Fig. 1(a). In the vegetation layer, the velocity distribution is assumed to be uniform, whereas in the surface layer, a logarithmic velocity profile is adopted. In these cases, the average velocity in the surface layer is calculated by taking the height average scheme. Meanwhile, the depth-averaged velocity in the entire cross section is calculated by using the weighted combination method. Verification by experimental data or simulated data from numerical models shows that these simplified models are generally in good agreement when used to predict depth-averaged velocity in open channel flows with submerged vegetation (Baptist et al., 2007; Huthoff et al., 2007; Yang and Choi, 2010; Cheng, 2011; Li et al., 2015b).
Nonetheless, as stated by Li et al. (2015b), a simple layer division method merely based on geometric factors cannot reflect the turbulence features of vegetated flows, especially when canopy densities change. Luhar et al. (2008) and Nepf (2012) drew attention to the difference in the internal flow mechanisms for vegetated flow with sparse and dense canopies. They demonstrated that the difference depends on the relative importance of bed shear stress and canopy drag force. For sparse canopies, bed shear stress plays a more important role than canopy drag force in the vegetation layer. Thus, the original momentum balance between potential gradients and canopy drag force no longer exists. Hence, the velocity profile in the vegetation layer is close to that shown in open channel flows without vegetation (Fig. 2a and b). For dense canopies, the discontinuity in drag force occurs at the interface of submerged vegetated flow. This phenomenon causes the Kelvin–Helmholtz (K–H) instability, which leads to a region of strong vertical momentum exchange (Fig. 2d). In addition, the transition between sparse and dense canopies takes place at CDahv ≈ 0.1 (Fig. 2c), where CD is the drag coefficient, a is the canopy frontal area per volume, and hv is the height of vegetation (Belcher et al., 2003).
For submerged vegetated flow, the influence of vertical momentum exchange is not limited to the surface layer as it also penetrates into the vegetation layer. This characteristic calls for a layer division method that can show the internal turbulence properties of submerged vegetated flow. Herein, we apply the dynamic boundary based on the concept of auxiliary bed, which is initially proposed by Li et al. (2015b). The boundary divides the flow into an upper layer of vertical turbulent exchange (called the suspension layer) and a lower layer of longitudinal turbulent exchange (called the basal layer), as shown in Fig. 1(b) (Nepf and Vivoni, 2000). For vegetated flow with sparse canopies, this boundary overlaps with the channel bottom because the basal layer of longitudinal momentum exchange vanishes. As for vegetated flow with dense canopies, the dynamic boundary that splits the vertical and longitudinal momentum exchange can easily be observed within the canopies. Therefore, the dynamic division method provides a unified physical basis for vegetated flow with both sparse and dense canopies by effectively representing the turbulence structures.
Applied as a useful formulae derivation tool in previous studies (Baptist et al., 2007; Tinoco et al., 2015; Shi et al., 2019), genetic programming (GP) algorithm is combined with the dynamic boundary to derive the depth-averaged velocity in vegetated flow with rigid submerged canopies. Note that data in this study are collected from flume experiments where canopies are arranged in aligned or staggered forms. The average velocity in each layer is considered separately. In the basal layer, the average velocity is derived from the momentum balance between potential gradients and canopy drag force. In the suspension layer, the average velocity is deduced from a mass of data by using the GP algorithm. Then, data grouping and pre-processing is conducted in terms of the physical basis for the auxiliary bed, which is mainly about the generation and formation of K–H vortices. The maximum dissimilarity algorithm (MDA) is applied to data selection (Camus et al., 2011; Goldstein et al., 2013). The final formulae of depth-averaged velocity for the entire cross section are obtained through weighted combination. The physical meaning and performance of the proposed two-layer model are illustrated in this study.
Section snippets
Genetic programming
GP is a specific evolutionary technique that is frequently used to solve hydraulic engineering problems (Babovic et al., 2001; Huai et al., 2018; Goldstein and Coco, 2014). GP has gradually evolved into a useful and practical method for deriving data-driven predictors since its development by Schmidt and Lipson (2009). In contrast to other techniques, such as linear regression or artificial neural networks, GP is particularly suitable for solving physical problems, since it requires no previous
Results
The depth-averaged velocity in the entire cross section (U) is computed by taking the weighted average of the average velocity in each layer. The equation can be described as follows:
Then, the formulae for depth-averaged velocity in different situations are obtained.
For condition CDahv > 0.1,
Meanwhile, for condition CDahv < 0.1,
Nikora et al. (2001) derived the relationship between
Comparison with previous models
The following is a list of previously proposed models for estimating the depth-averaged velocity in the entire cross section.
- (1)
Baptist et al. (2007). Baptist et al. obtained the following formula by applying the dimensionally aware GP to the simulated data of the one-dimensional vertical (1-DV) model for submerged vegetation:
Summary
In this study, we combine the concept of auxiliary bed proposed by Li et al. (2015b) with the GP algorithm to predict the depth-averaged velocity in submerged vegetated flow. We segregate the vegetated flow into the basal and suspension layers while considering the auxiliary bed as the boundary. The average velocity in the basal layer is derived from the momentum balance. For the suspension layer, the GP algorithm is applied to obtain two sets of predictors for open-channel flow with sparse and
CRediT authorship contribution statement
Fan Yang: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Writing - original draft, Writing - review & editing, Visualization. Wen-Xin Huai: Conceptualization, Methodology, Writing - review & editing, Supervision, Project administration, Funding acquisition. Yu-Hong Zeng: Writing - review & editing, Supervision.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgement
This work is financially supported by the National Natural Science Foundation of China (Nos. 11872285, 11672213, and 51439007). The authors thank V. Pasquino for providing the experimental data to help with the manuscript. Comments and suggestions made by the Editor Dr. D'Odorico and Reviewers have greatly improved the quality of the paper.
References (55)
- et al.
Analysis of clustering and selection algorithms for the study of multivariate wave climate
Coast. Eng.
(2011) - et al.
Prediction of wave ripple characteristics using genetic programming
Cont. Shelf Res.
(2013) - et al.
Three-layer model for vertical velocity distribution in open channel flow with submerged rigid vegetation
Adv. Water Resour.
(2009) - et al.
New formulation for the effective relative roughness height of open channel flows with submerged vegetation
Adv. Water Resour.
(2015) - et al.
Turbulence structure and coherent motion in vegetated canopy open-channel flows
J. Hydro-Environ. Res.
(2008) - et al.
Predicting the bulk average velocity of open-channel flow with submerged rigid vegetation
J. Hydrol.
(2019) - et al.
Determining drag coefficients and their application in modelling of turbulent flow with submerged vegetation
Adv. Water Resour.
(2014) - et al.
Physically sound formula for longitudinal dispersion coefficients of natural rivers
J. Hydrol.
(2017) - Babovic, V., Keijzer, M., Aguilera, D.R., Harrington, J. 2001. An evolutionary approach to knowledge induction: genetic...
Modelling floodplain biogeomorphology
(2005)
On inducing equations for vegetation resistance
J. Hydraul. Res.
Adjustment of a turbulent boundary layer to a canopy of roughness elements
J. Fluid Mech.
The wind in the willows: flows in forest canopies in complex terrain
Annu. Rev. Fluid Mech.
Representative roughness height of submerged vegetation
Water Resour. Res.
Hydraulic radius for evaluating resistance induced by simulated emergent vegetation in open-channel flows
J. Hydraul. Eng.
Effect of riparian vegetation on flow resistance and flood potential
J. Hydraul. Eng.
Mean flow and turbulence in a laboratory channel with simulated vegetation
Hydraul. Eng. Ser. No. 51. Hydrosystems Laboratory
Mixing layers and coherent structures in vegetated aquatic flows
J. Geophys. Res.-Oceans.
The limited growth of vegetated shear layers
Water Resour. Res.
Shallow flows over a permeable medium: the hydrodynamics of submerged aquatic canopies
Transp. Porous Med.
A machine learning approach for the prediction of settling velocity
Water Resour. Res.
Use of conventional flow resistance equations and a model for the Nikuradse roughness in vegetated flows at high submergence
J. Hydrol. Hydromech.
Experimental study on flow velocity characteristics and pollutant transport in vegetation channel
Estimating the transverse mixing coefficient in laboratory flumes and natural rivers
Water Air Soil Pollut
Analytical solution of the depth-averaged flow velocity in case of submerged rigid cylindrical vegetation
Water Resour. Res.
Atmospheric Boundary Layer flows: Their structure and Measurement
Genetic programming: On the Programming of Computers By Means of Natural Selection
Cited by (8)
Flow velocity adjustment in a channel with a floating vegetation canopy
2024, Journal of HydrologyEffect of Multiple Layered Vegetation on the Velocity Distribution of Flow in An Open Channel
2022, Proceedings of the IAHR World CongressComparison of flow velocity distributions in open channel with continuous and uniform submerged vegetation
2021, Shuili Fadian Xuebao/Journal of Hydroelectric EngineeringDischarge and velocity variation of flows in open channels partially covered with different layered vegetation
2021, E3S Web of Conferences