Experimental and numerical characterization of roughness for bi-dimensional numerical models in extreme flood conditions.

Présentation

Durée : 3 ans

Experimental and numerical characterization of roughness for bi-dimensional numerical models in extreme flood conditions.

Laboratories : EDF R&D – LNHE and Saint-Venant Hydraulics Laboratory (LHSV), located at Chatou, France
Periods : 3 years (2025-2028)

Context and objectives

The simulation of flows induced by extreme floods (e.g., to map flood-prone areas during a 100-year return period flood as part of Flood Risk Prevention Plans) currently relies mainly on two-dimensional hydrodynamic models. A simulation requires that the models have been verified (from a numerical point of view) and then validated (i.e., tested on real events already observed). In particular, it is essential to validate the numerical value of parameters such as the roughness coefficients of the area impacted by the flood. This step of the study, called "model calibration," is based on two elements:

• The availability of in situ observations (e.g., flood marks, aerial photographs of flooded areas, hydrometric stations in a watercourse, etc.): in this case, the objective of the calibration is to determine the roughness coefficient(s) that best reproduce one or more observed events (the flood marks).
• Expert advice: in this case, roughness coefficients are generally assigned based on the type of soil potentially impacted by the flood event (e.g., meadows, fields, urban areas, wetlands, etc.).

Currently, the calibration of the minor bed (before overflow) is carried out using hydrometric data to test the selected roughness coefficients on several flood events (not or slightly overflowing). For the floodplain, the model calibration is based on a small number of observations (e.g., flood marks), sometimes very distant in time (compared to the time of the study) and which do not cover the entire numerical domain affected by the phenomenon studied (e.g., when the overflowing floods are less intense than those retained in the study), which forces the user to define the roughness parameters based on expertise. Thus, the floodplain is divided into zones based on land use (e.g., zoning from Corine Land Cover). Then, for each land use zone, a roughness coefficient (e.g., Strickler coefficient) is defined based on the ranges of values available in the literature (e.g., Chow, 1959 and 1984; Fisher and Dawson, 2003). However, the state of the art prior to the definition of roughness is mainly based on "one-dimensional" studies (Strickler laws, composite beds, etc.) and on flood events (observed or simulated) of lower intensity than those used for the project floods of operational studies (100-year return period flood or higher).

Indeed, it is well known that the vegetation in floodplains, channels, and riverbanks affects the vertical velocity profile with consequences on the hydraulic resistance. Accordingly, the impact of vegetation on hydraulic resistance is crucial in different fields such as nature-based solutions, sediment transport and flooding studies (Nikora et al., 2008). This field is still a challenging task and has not been completely understood. As a result, different approaches have been proposed in literature for different types of vegetation (rigid and flexible), and for different conditions (emergent, just submerged and submerged). Furthermore, in some cases, the classical friction formulations of depth-averaged hydraulic model are unsuitable for simulating the effects of vegetation (Folke et al. 2019a).

For these reasons nine vegetation laws have been already implemented in TELEMAC-2D and tested with experimental data measured in flume experiments focusing on the model’s ability to reproduce the vegetation effects on hydraulic resistance (Folke et al. 2019a, Dallmeier and Rüther 2023, Folke et al. 2020). Recently, Cui et al. (2023) presented a vegetation law which relies on the velocity superposition model developed by Nikora et al. (2013). This model was tested by Cui et al. (2023) for a wide range of vegetation densities and different vegetation conditions including artificial, natural, flexible, and rigid vegetation providing promising results. Considering the great flexibility showed by the model, the vegetation friction law proposed by Cui et al. (2023) has been implemented in TELEMAC-2D following the same methodology proposed by Folke et al. (2019a) for the 9 vegetation laws already implemented in TELEMAC-2D (Farina et al., 2024).

In this context, the objective is to develop a new methodology for characterizing roughness in flood zones, considering (i) the extreme floodings (in terms of submersion level) and (ii) the land use conditions typical of the prone areas near to EDF facilities (peri-urban and agricultural areas, or urban environment for submersion waves).

Work Program

Approach

The thesis work aims to improve the understanding and the modeling/parametrisation of roughness in bi-dimensional numerical modelling during extreme floods. The approach is mainly based on laboratory experiments coupled with the improvement of existing numerical tools.

The objectives will be achieved by following the two work axes below:

1. Definition of an experimental set-up representative of flood prone areas and extreme flooding
2. Analysis of experimental data with a digital twin (bi-dimensional model) for the improvement of existing numerical model TELEMAC-2D

The methodological questions of work axis A can be broken down as follows:

To answer these questions, EDF R&D (LNHE department) as part of the MOISE project (Modeling of External Flooding) has carried out preliminary to design, using existing literature and a TELEMAC-2D numerical model, an experimental channel for conducting a test campaign to evaluate the roughness on a vegetated plain for different submersion levels (see Figure 1).

The key objectives for the future are:

• To define the experimental conditions (discharge, slope, materials, …) adapted to reproduce field conditions;
• To define the experimental devices necessary for measurements of the velocity field (drones, flow velocimeter, LSPIV, …);
• To define the experimental materials adapted to reproduce “urban vegetations” (crops, meadows, wetlands…).
• To carry out the experimental tests and to deeply analyze the measurements.
• To propose a new methodological approach to characterize roughness coefficients for all flood regimes (from very low to very high overflow), for two-dimensional models (TELEMAC2D) and the land use conditions typical of the valleys where EDF sites are located.

The methodological questions of work axis B can be broken down as follows:

• Are the numerical values of the roughness coefficients of the floodplain usually employed in practical applications adapted to extreme floods (high submersions of the floodplain) and two-dimensional modeling?
• Which roughness modeling approaches are most suitable for highly overflowing floods and the land use conditions typical of prone areas near to industrial and urban facilities (urban environment)?

To this purpose it can be worthwhile to remember that TELEMAC-2D considers the effects of vegetation by the principle of linear superposition and the total friction coefficient is decomposed into two different contributions: the friction term of the bottom and the friction generated by the vegetation.

The vegetation laws available in version v8p5r0 are reported in Table 1.

Additionally, the vegetation law recently implemented in TELEMAC 2D by Farina et al., (2024) will also be considered.

The calibration of these laws through numerical observations could be useful for a better understanding of physical phenomena related to roughness.

Available Tools

The doctoral student can rely on the experiences of the LNHE teams on the use of specific devices and experimental facilities for measuring flood velocities and friction coefficients. Numerical modeling will be carried out with the TELEMAC2D code of the open-source TELEMAC-MASCARET system developed, among others, at LNHE (www.opentelemac.org/).

Schedule

The thesis will start at the end of 2025 or early 2026.

The first year of the thesis (2026) will be devoted to:

• Bibliography review on roughness in bi-dimensional numerical models. The student will benefit of a literature review already available at EDF R&D LNHE,
• Definition of the experimental set-up (flood conditions, measurements, …),
• Construction of a numerical twin of the flume (using TELEMAC-2D),
• Beginning of the experimental runs with a given material.

The second year of the thesis (2027) will focus on:

• Experimental runs with different flood conditions and materials,
• Assessment of roughness through physical (measurements) and numerical simulations,
• Test of different friction laws and beginning of the calibration of frictions law’s parameters.

The third year of the thesis (2028) will aim to:

• End of the experiments with all materials selected during the 1^st year,
• Proposition of an experimental set-up for extensive physical simulations (with other materials not tested during the PhD),
• Proposition of a new friction law adapted to a real study case,
• Numerical simulations on a real site and comparison with classical approach.
• Write the thesis manuscript.

The work will be valorized in at least two articles for international conferences (e.g., River Flow, AIRH conferences, Telemac User Club…) and two articles for scientific journals (e.g., Journal of Hydraulic Engineering, Journal of Hydrology, Water Resources Research).

Profile, Supervision, and Collaborations

Supervision

The thesis will be carried out in collaboration between EDF and LHSV. The thesis grant is funded by EDF and ANRT (CIFRE grant). The thesis will be supervised by Sébastien Boyaval (LHSV). The thesis supervision will consist of Vito Bacchi, Magali Jodeau and Yvan Bercovitz (EDF LNHE & Laboratoire d’Hydraulique Saint-Venant).

The doctoral student will be hosted at EDF R&D - LHSV (Chatou). The doctoral student will be enrolled at the IP Paris doctoral school (https://www.ip-paris.fr/education/doctorat)

An annual thesis committee will include scientific experts.

Candidate Profile

The candidate sought to meet the project's objective must have at least basic knowledge in the following disciplines:

• River hydraulics,
• A first experience in experimental science is recommended,
• Numerical hydraulic modeling,
• Programming (FORTRAN, PYTHON)

Références

Chow, V.T. (1959). Open channel hydraulics, New York McGraw-Hill, 680 pages.

Fisher, K, Dawson, H (2003) Reducing Uncertainty in River Flood Conveyance – Roughness Review. Project W5A-057. DEFRA / Environment Agency – Flood and Coastal Defence R&D Programme.

Folke, F., R. Koopman, G. Dalledonne and M. Attieh. 2019. “Comparison of different vegetation models   using TELEMAC-2D", XXVIth TELEMAC-MASCARET User Conference, 2019.

Dallmeier A. and Rüther N., 2023. Advanced representation of near-natural vegetation in hydrodynamic modelling. XXIXth TELEMAC-MASCARET Users Conference 2023.

Folke, F., S. Niewerth, J. Aberle. 2020. Modelling of just-submerged and submerged flexible vegetation. In: Kalinowska, Monika (Hg.): Abstract Book, 6th IAHR Europe Congress Warsaw Poland 2020. Warschau: IAHR. S. 263-264.

Cui, H., S. Felder, and M. Kramer. 2023. “Predicting flow resistance in open-channel flows with submerged vegetation.” Environmental Fluid Mechanics, 23 (4): 757–778. https://doi.org/10.1007/s10652-023-09929-x.

Nikora, N., V. Nikora, and T. O’Donoghue. 2013. “Velocity Profiles in Vegetated Open-Channel Flows: Combined Effects of Multiple Mechanisms.” Journal of Hydraulic Engineering, 139 (10): 1021–1032. https://doi.org/10.1061/(asce)hy.1943-7900.0000779.

Farina G, Bacchi V, Pilotti M (2024) Implementation and validation of a new friction vegetation law in TELEMAC-2D, XXXth TELEMAC-MASCARET Users Conference, Chambery, France

Lindner K (1982) Der Strömungswiderstand von Pflanzenbeständen, Braunschweig

Pasche R, Rouvé G (1985) Overbank flow with vegetatively roughended flood plains, Journal of Hydraulic Engineering, 111(9)

Järvelä J (2004) Determination of flow resistance caused by non‐submerged woody vegetation. International Journal of River Basin Management, 2(1), 61–70

Whittaker P, Wilson C, Aberle J (2015) An improved Cauchy number approach for predicting the drag and reconfiguration of flexible vegetation, Advances in Water Resources, 83, 28-35

Baptist MJ, Babovic V, Rodríguez Uthurburu J, Keijzer M, Uittenbogaard RE, Mynett A, Verwey A (2007) On inducing equations for vegetation resistance, Journal of Hydraulic Research, 45(4), 435–450

Huthoff F, Augustijn DCM, Hulscher SJMH (2007), Analytical solution of the depth-averaged flow velocity in case of submerged rigid cylindrical vegetation, Water Resources Research., 43, W06413

Van Velzen EH, Jesse P, Cornelissen P, Coops H (2003) Stromingsweerstand Vegetatie in Uiterwaarden, 2003.029. RIZA, Arnhem

Luhar M, Nepf H (2013) From the blade scale to the reach scale: A characterization of aquatic vegetative drag, Advances in Water Resources 51, 305–316

Västilä K., Järvelä J (2014) Modeling the flow resistance of woody vegetation using physically based properties of the foliage and stem”, Water Resources Research 50, 229–245

Dallmeier A, Folke F, Rüther N (2023) Advanced representation of near-natural vegetation in hydrodynamic modelling, XXIXth TELEMAC-MASCARET Users Conference, Karlsruhe, Germany

Box W, Järvelä J. Västilä K (2022) New formulas addressing flow resistance of floodplain vegetation from emergent to submerged conditions, International Journal of River Basin Management. 22(3), 333–349