Outputs
Outputs and publications created throughout the project.
A CFD‐based mixing model for vegetated flows
A journal paper published in February 2019, in Water Resources Research.
This paper provides a computational fluid dynamics (CFD)‐based modelling framework for predicting flow field, turbulence, and mixing characteristics within vegetated environments such as ponds and wetlands. The framework has been implemented within a commercial CFD code (ANSYS Fluent 19) via a set of user‐defined functions.
Following the approach outlined by King et al. (2012), the standard k‐ε turbulence closure model has been modified to capture the energy transfer at the vegetation/clear flow shear interface and within the vegetation. The implementation assumes that vegetation is vertical, but nonorthogonal flow in the horizontal plane is accounted for. Values for the drag coefficient and the mixing coefficients are estimated based on the vegetation stem diameter and density.
Following Tanino and Nepf (2008), a switch has been incorporated to account for the fact that the relevant length scale changes, from stem diameter to stem spacing, as stem density increases. A set of model parameters is proposed, based on a re-evaluation of previously published laboratory data and theoretical analysis.
Five different experimental data sets are used to demonstrate that the model is able to predict mixing within fully vegetated systems and due to both vertical and horizontal shear layers.
The framework was developed to provide a practical prediction tool for engineering purposes, in particular for the estimation of residence time distributions in real partially vegetated stormwater management ponds. Its implementation here within a commercial CFD package potentially facilitates application to complex pond geometries, including patches of different types of vegetation with different bulk stem diameter and density characteristics.
A stem spacing-based non-dimensional model for predicting longitudinal dispersion in low-density emergent vegetation
A journal paper published in November 2018, in Acta Geophysica.
Predicting how pollutants disperse in vegetation is necessary to protect natural watercourses. This can be done using the one-dimensional advection dispersion equation, which requires estimates of longitudinal dispersion coefficients in vegetation.
Dye tracing was used to obtain longitudinal dispersion coefficients in emergent artificial vegetation of different densities and stem diameters. Based on these results, a simple non-dimensional model, depending on velocity and stem spacing, was developed to predict the longitudinal dispersion coefficient in uniform emergent vegetation at low densities (solid volume fractions < 0.1).
Predictions of the longitudinal dispersion coefficient from this simple model were compared with predictions from a more complex expression for a range of experimental data, including real vegetation. The simple model was found to predict correct order of magnitude dispersion coefficients and to perform as well as the more complex expression. The simple model requires fewer parameters and provides a robust engineering approximation.
Transverse dispersion in vegetation across a Shear-layer 2016: Artificial, Carex, Typha
This open access dataset, available in October 2018, supports the 'A CFD-based mixing model for vegetated flows' publication.
The data describes transverse concentrations of a neutrally buoyant solute, across a shear layer in water past two measurement locations, within a partially vegetated channel with emergent vegetation. It also describes accompanying transverse profiles of longitudinal velocity.
Measurements were conducted in five types of vegetation
low-density artificial vegetation
high-density artificial vegetation
Carex Acutiformis
winter Typha Latifolia
summer Typha Latifolia
This data set is suitable for estimating transverse dispersion across a shear-layer. It accompanies the journal article submission entitled 'A CFD-based mixing model for vegetated flows'.
Estimating drag coefficient for arrays of rigid cylinders representing emergent vegetation
A journal paper published in September 2018, in Journal of Hydraulic Research.
Flow resistance, due to vegetation, is of interest for a wide variety of hydraulic engineering applications. This note evaluates several practical engineering functions for estimating bulk drag coefficient (CD) for arrays of rigid cylinders, which are commonly used to represent emergent vegetation.
Many of the evaluated functions are based on an Ergun-derived expression that relates CD to two coefficients, describing viscous and inertial effects. A re-parametrisation of the Ergun coefficients based on cylinder diameter (d) and solid volume fraction (ϕ) is presented.
Estimates of CD are compared to a range of experimental data from previous studies. All functions reasonably estimate CD at low ϕ and high cylinder Reynolds numbers (Rd). At higher ϕ they typically underestimate CD.
Estimates of CD utilising the re-parametrisation match the experimental data better than estimates of CD made using the other functions evaluated, particularly at low ϕ and low Rd.
Use of drag coefficient to predict dispersion coefficients in emergent vegetation at low velocities
A conference paper presented in August 2018 at the 12th International Symposium on Ecohydraulics.
Existing functions for predicting transverse and longitudinal dispersion coefficient within vegetation are dependent on the drag coefficient.
This paper focuses on comparing drag coefficients. It also compares predictions of dispersion coefficient, derived from drag coefficient, to estimates of dispersion coefficient from laboratory measurements.
Three drag coefficients have been considered
drag coefficient estimated from turbulent kinetic energy measurements
drag coefficient predicted from vegetation characteristics
an assumed drag coefficient of one
Predicted dispersion coefficients have been compared to estimated dispersion coefficients in two types of artificial vegetation and two types of real vegetation (Carex and Typha). Dispersion coefficients in Carex were not well predicted.
All three drag coefficients produced reasonable predictions of dispersion coefficients in the artificial vegetation and Typha. They showed the functions that predict dispersion coefficient to be relatively insensitive to drag coefficient. The use of a drag coefficient of one is shown to be reasonable for predicting dispersion coefficient.
A non-dimensional model for longitudinal dispersion in emergent vegetation based on stem-spacing
An extended abstract presented in June 2018 at the fifth IAHR Europe Congress 2018.
Dye tracing in emergent artificial vegetation was used to obtain longitudinal dispersion coefficients. A non-dimensional model based on stem-spacing was developed that predicts longitudinal dispersion coefficient in an emergent vegetation well.
Computational fluid dynamics modelling of residence times in vegetated stormwater ponds
A journal paper published in March 2018, in the themed issue on green infrastructure of ICE Water Management.
Experimental data characterising dispersion within Typha latifolia were previously collected in a laboratory setting. This mixing characterisation was combined with previously proposed computational fluid dynamics modelling approaches. It was combined to predict residence time distributions for vegetated stormwater treatment pond layouts. These layouts, including a wetland, were derived from Highways England design guidance.
The results showed that the presence of vegetation resulted in residence times closer to plug flow, indicating significant improvements in stormwater treatment capability.
The new modelling approach reflects changes in residence time due to mixing within the vegetation. However, it also suggests that it is more important to include vegetation within the model in the correct location than it is to accurately characterise it.
Estimates of hydraulic efficiency suggest that fully vegetated stormwater ponds such as wetlands should function well as a treatment device. More typical ponds, with clear water, need to be designed to be between 50% and 100% larger than their nominal residence times would suggest when designed against treatment criteria.
A CFD based comparison of mixing due to regular and random cylinder arrays
A conference paper presented in May 2017 at the 36th International School of Hydraulics.
Numerous studies have focused on flow and mixing within cylinder arrays because of their similarity to vegetated flows. Randomly distributed cylinders are considered to be a closer representation of the natural distribution of vegetation stems compared with regularly distributed arrays.
In this study, the flow fields associated with two arrays of regularly and randomly distributed cylinders are modelled in two dimensions, using ANSYS Fluent 16.1.
The RSM turbulence model is used to model the turbulence closure, and all the variables are discretised using the second order upwind method. The resulting flow fields are used to run the solute transport model to characterise mixing within each geometry. For the same stem diameter and solid volume fraction, greater dispersion is evident in the random cylinder array compared with the regular array.
Dispersion coefficient values are compared with those reported in the literature and a good agreement is shown. Turbulence length scales estimated from the velocity profiles and optimised dispersion coefficients are close to the cylinder diameter, which is in agreement with theories in the literature.
Transverse and longitudinal mixing in real emergent vegetation at low velocities
A journal paper published in January 2017, in Water Resources Research.
Understanding solute mixing within real vegetation is critical to predicting and evaluating the performance of engineered natural systems such as stormwater ponds. For the first time, mixing has been quantified through simultaneous laboratory measurements of transverse and longitudinal dispersion within artificial and real emergent vegetation.
Dispersion coefficients, derived from a routing solution to the 2D Advection Dispersion Equation (ADE), are presented that compare the effects of vegetation type (artificial, Typha latifolia or Carex acutiformis) and growth season (winter or summer).
The new experimental dispersion coefficients are plotted with the experimental values from other studies. They are used to review existing mixing models for emergent vegetation.
The existing mixing models fail to predict the observed mixing within natural vegetation, particularly for transverse dispersion. This reflects the complexity of processes associated with the heterogeneous nature of real vegetation.
Observed stem diameter distributions are utilised to highlight the sensitivity of existing models to this key length-scale descriptor. This leads to a recommendation that future models intended for application to real vegetation should be based on probabilistic descriptions of both stem diameters and stem spacings.
Transverse and longitudinal dispersion in vegetation 2016: Artificial, Carex, Typha
This open access dataset, available in December 2016, under-pins the 'Transverse and longitudinal mixing in real emergent vegetation at low velocities' publication.
The data describes concentrations of a neutrally buoyant solute in water through both time and space past two measurement locations within full-width fully emergent vegetation.
Measurements were conducted in five types of vegetation
low-density artificial vegetation
high-density artificial vegetation
Carex Acutiformis
winter Typha Latifolia
summer Typha Latifolia
This data set is suitable for estimating longitudinal and transverse dispersion.
Modelling the effects of real vegetation on residence times in ponds and wetlands
A presentation given in October 2016 at the 12th Annual CWA Conference 2016.
View presentation (PDF, 534KB).
Residence times in vegetated stormwater ponds
A project overview presentation given in August 2016 at the HYTECH International Symposium on Interfaces within aquatic ecosystems.
View presentation (PDF, 1.0MB).
Estimating stem-scale mixing coefficients in low velocity flows
A conference paper presented in July 2016 at the fourth IAHR Europe Congress 2016.
Stormwater ponds are SuDS devices intended to moderate the negative environmental impacts of stormwater run-off. A current, joint, research programme is investigating the effects of heterogeneous vegetation distributions in stormwater ponds. The programme is developing CFD techniques to simulate 3D solute transport processes in low velocity flows.
The aim of the project is to generate a unique dataset that describes the influence of different types and configurations of vegetation on the pond’s fundamental flow – and treatment – characteristics. Better characterisation can then be used to better evaluate existing run-off treatment ponds that may be delivering sub-optimal levels of treatment.
This paper presents results from an initial 1D laboratory study. The study had regular uniform emergent artificial vegetation, from which longitudinal dispersion coefficients were evaluated over a range of target flow velocities.
These have been integrated with stem-scale CFD mean velocity predictions, to investigate the ability of a Chikwendu (1986) n-zone type approach to predict transverse dispersion coefficients over a scale suitable for inclusion in future 3D pond models.
Assuming that the small stem-scale fluctuations in velocity form a repeating pattern at the patch scale, and that stem-scale transverse dispersion effects integrate, this approach has been successfully applied to predict mixing in a 2D system solely with parameters estimated from a 1D system.
Quantifying solute mixing and flow fields in low velocity real emergent vegetation
A conference paper presented in February 2016 at the 11th International Symposium on Ecohydraulics.
The treatment efficacy of ponds is dependent on hydraulic residence. In turn, residence times are directly affected by the systems' internal hydraulics and mixing characteristics.
Aquatic vegetation is known to aid water treatment through habitat generation, promoting bio-chemical degradation and through active-uptake. The physical presence of vegetation can significantly alter the residence times.
The heterogeneous distribution of vegetation – in borders or patches – has a strong impact on local mixing via stem and shear scale processes. Lacking is the quantification of the seasonal impact of real, spatially random vegetation on mixing and thus treatment efficacy.
The development of a precise 1D Laser Induced Fluorometry (LIF) and application to a laboratory control of real and artificial vegetation are discussed. Quantifying mixing in low velocity emergent vegetation presents observational challenges when seeking higher precision. Detailed, two-dimensional tracer distributions were visualised.
Flow field and mixing characteristics were quantified for partially (shear) vegetated stems for two artificial and two real vegetation densities. The mixing associated with real vegetation is significantly different to that of artificial vegetation.
The spatial heterogeneity of real vegetation presents a number of challenges when attempting to quantify their geometry. A significant range in stem diameter and local stem density was recorded; further, site selection for trace visualisation has a significant impact on results due to the variability in stem distribution.
The application of LIF to full and partially vegetated real vegetation provides valuable data for practitioners and CFD modellers.
Computational fluid dynamics modelling of a vegetated stormwater pond
A conference paper presented in February 2016 at the 11th International Symposium on Ecohydraulics.
Stormwater treatment ponds are common, but there are few tools to evaluate their performance. Traditional techniques, such as dye traces, take significant amounts of time and are impractical. Computational fluid dynamics is increasingly being used to design and evaluate ponds, but often neglects the effects of vegetation.
This study looks at the CFD modelling of a stormwater pond in Lyby, southern Sweden, both with and without vegetation. The Residence Time Distribution is an established means of evaluating treatment capability of a pond.
CFD modelling of solute transport is carried out to obtain RTDs, which are then compared to experimental RTDs obtained from the pond. The capability and benefits of including vegetation in CFD modelling of ponds are examined.
Representing vegetation within a CFD model pond
A presentation given in September 2015 at the SUDSnet International Conference 2015 (PDF, 1.1MB).
View presentation (PDF, 4.14MB)
Stormwater ponds, as components of urban drainage systems, play an important role in terms of meeting water quality objectives. Vegetation improves these systems' performance in two ways, first by increasing residence time and second by acting as a habitat for organisms.
Efficient design needs basic understanding of vegetation effects on hydrodynamic behaviour of flow, within the open and vegetated parts of the pond. For instance, short-circuiting is one of the common problems in ponds, which causes residence times shorter than the normal residence time and thereby decreasing the pond efficiency. The effects of vegetation on problems such as short-circuiting can be of significant importance.
Computational Fluid Dynamics (CFD) models can be used to investigate and represent these effects. These models allow vegetation stems to be modelled either one-by-one or as a bulk effect through different approaches.
A recent laboratory experiment is modelled using the one-by-one modelling method to investigate and compare mixing coefficients with those determined from the lab data. Two alternative bulk effect approaches will also be compared in this paper.
The first one, which considers vegetation as a porous media, has been recently proposed and evaluated by other researchers. It has shown promise and partial success. The second one is a novel, preliminary, exploration of the potential to apply the lattice Boltzmann method.
Investigating CFD modelling of residence times in vegetated stormwater ponds
A poster presented in June 2015 at the University of Sheffield Engineering Symposium 2015.
View poster (PDF, 33.2MB).
Development of a laboratory system and 2D routing analysis to determine solute mixing within aquatic vegetation
The presentation and a conference paper presented in May 2015 at the 34th International School of Hydraulics.
View presentation (PDF, 2.94MB)
A laser induced fluorometry (LIF) system was developed to quantify mixing within spatially variable aquatic vegetation.
A comparison is made between intrusive fluorometry techniques and the application of LIF. The comparison aims to quantify mixing in real vegetation in the laboratory setting. LIF provides greater spatial resolution when compared to point fluorometry.
Furthermore, LIF is non-intrusive. A two-dimensional routing procedure is used to calculate the longitudinal and transverse velocities and mixing coefficients from a single pulse injection of tracer within a vegetation patch.
Feasibility of the porous zone approach to modelling vegetation in CFD
The presentation and a conference paper presented in May 2015 at the 34th International School of Hydraulics
View presentation (PDF, 1.08MB)
Vegetation within stormwater ponds varies seasonally. Its presence affects the flow field, which in turn affects the pond’s Residence Time Distribution, and its effectiveness at pollutant removal. Vegetated flows are complex and, as a result, few suitable tools exist for evaluating realistic stormwater pond designs.
Recent research has suggested using a porous zone to represent vegetation within a CFD model. This paper investigates the feasibility of this approach using ANSYS Fluent.
One of the main benefits of using a porous zone is the ability to derive the relevant parameters from the known physical characteristics of stem diameter and porosity, using the Ergun equation. A sensitivity analysis on the viscous resistance factor 1/α and the inertial resistance factor C2 has been undertaken by comparing model results to data collected from an experimental vegetated channel.
Best fit values of C2 were obtained for a range of flow conditions including emergent and submerged vegetation. Results show the CFD model to be insensitive to 1/α but very sensitive to values of C2. For submerged vegetation, values of C2 derived from the Ergun equation are under-predictions of best-fit C2 values, as only the turbulence due to the shear layer is represented.
The porous zone approach does not take into account turbulence generated from stem wakes such that no meaningful predictions for emergent vegetation were obtained. C2 values calculated using a force balance show better agreement with best-fit C2 values than those derived from the Ergun equation. Manually fixing values of k and ε within the porous zone of the model shows initial promise as a means of taking stem wakes into account.
A brief review of pond residence time studies
A conference paper presented in April 2015 at the 2015 Annual Postgraduate Research Student Conference.
Vegetation has a significant effect on the purifying processes in storm-water ponds. As pond hydrodynamics are also strongly affected by vegetation, it is necessary to understand flow and mixing processes in vegetated ponds for effective treatment design.
In this paper, the Residence Time Distribution (RTD) and related indices are discussed. Selected recent field and laboratory studies on vegetated ponds are reviewed. Selected RANS CFD (Reynolds-averaged Navier–Stokes Computational Fluid Dynamics) studies, which investigate pond geometry, are outlined.
Those studies which investigate the effects of vegetation are divided into two types. First are those which consider it as bed roughness or a retarding force, and second are those which consider it as a flow zone. The conclusions from the first approach are that it is sufficient to predict average hydraulic features for flood studies but insufficient for investigating mixing effects of vegetation.
Recent publications have suggested that the second approach, such as considering vegetation as a porous media, may be more suitable. However, the most recent study highlighted certain limitations, as well as providing some preliminary suggestions for overcoming them.
This discussion reinforces the need for further comprehensive research into this, and alternative, approaches.
Quantifying mixing in low velocity emergent vegetated ponds
A poster presented in October 2014 at the 10th Annual CWA Conference 2014.
View poster (PDF, 3.88MB)
The effects of vegetation on the hydraulic residence time of stormwater ponds
A conference paper presented in October 2014 at the 19th IAHR-APD Congress 2014.
View conference paper (PDF, 1.06MB)
Stormwater ponds treat polluted run-off from urban areas, highways and agricultural land. Vegetation plays a key role in water treatment, but further understanding is required to identify how vegetation density and spatial distribution within a pond affect the residence time. This is an important parameter with respect to water treatment.
This paper presents results from a preliminary study, where the residence time distribution and discharge of a water treatment pond were measured. They were measured two stages within the vegetation’s seasonal growth cycle, representing the minimum and maximum states of the vegetation's density.
The results show clear and significant differences between the residence time distribution for the two cases, and highlight the need for further work on the topic.
Residence times in vegetated stormwater ponds
A poster presented in June 2014 at the 2014 IAHR UK Chapter Annual Technical Meeting.
View poster (PDF, 1.99MB)
Pollution in the environment - Aquatic Vegetation
A poster presented in June 2014 at the 2014 IAHR UK Chapter Annual Technical Meeting.
View poster (PDF, 2.21MB)