Mourad BOUTOUCHENT

Today, access to energy is essential for the development of populations. However, in some cases, the geographical configuration does not allow the installation of such an electricity network or its extension to isolated sites. In such cases, local authorities opt for diesel generators to electrify isolated settlements, even in tropical zones, despite the exorbitant price of fuel oil and the difficulty of transporting it to isolated sites.

Local authorities should do enough to promote the development of renewable energies, particularly in the tropics where hydropower dominates over wind and solar energy. However, despite the fact that the tropics have a very large rainfall regime and powerful hydrographic networks, the hydrology of this area and the way in which these watercourses function are still poorly understood and little studied. Indeed, most of the studies carried out transpose findings and data from the temperate zone directly to the tropical zone without a clear analysis of their relevance.

Rural and/or isolated populations are generally located on the banks of large rivers in tropical zones, so hydropower should be considered as an attractive alternative to diesel generators for supplying these populations with electricity. Indeed, photovoltaic and wind power systems should be considered only after protection and maintenance measures have been taken against extreme humidity and lightning strikes. In addition, micro-hydro plants require a certain head and very expensive civil engineering structures. Consequently, tidal turbines are the natural choice, given their very low basic and maintenance costs compared with other renewable energy systems, and the fact that turbines are available to operate in rivers with low currents.

Usually, the installation of river tidal turbines requires the prior identification and characterisation of sites where the flow velocity remains above 1 or 1.5 m/s and the water depth remains above 2m. This in turn requires the prior determination of hydraulic variables such as flow velocity, water depth, flow rate, pressure, turbulent energy and particularly the flow of hydrokinetic energy in the vicinity of isolated populations.

Tropical rivers, on the other hand, are supplied by very high rainfall and flow over very rugged terrain. This allows them to develop complex morphologies such as anabranches and meanders. Understanding the temporal evolution of the above variables along these rivers is a challenge in itself, as it can only be achieved by numerical modelling.

Numerical modelling of flows in a large anabranched river in turn requires a high-resolution Digital Terrain Model (DTM) and knowledge of the flow variables at the boundaries of the computational domain. This is where the work begins, with a mixture of objectives, questions and challenges. Here are just a few of them:

• Most tropical rivers suffer from a lack of hydrological data, so how do you go about producing a DTM that is accurate enough for hydrodynamic simulations?
• What do we expect from hydrodynamic modelling in anabranchs outside the areas of interest? Is it really essential to obtain accurate results in these areas?
• Which model is best suited to tropical river morphologies - 1D, 2D or 3D - and why?

This thesis is divided into four chapters. The first is a general introduction to the issues addressed in the thesis. It begins by outlining the state of the art in the electrification of isolated sites and presents a more or less exhaustive overview of tidal turbine technology. Secondly, we review the methods used to evaluate the continental tidal energy resource, based on specific studies carried out in the United States, Canada, Europe and Brazil. The criteria for pre-selecting sites potentially suitable for tidal turbine development in the Maroni, as well as the broad outlines of the method chosen for in-depth characterisation of these sites, are then defined in the light of the river's specific morphological and hydraulic features.

The second chapter describes the state of the art in the mathematical modelling of river flows and presents some comparative studies between 1D, 2D and 3D models based on real cases and/or on what is agreed by the scientific community. This has made it possible to justify the choice of the Saint-Venant 2D model and the TELEMEC2D resolution model, given the hydromorphological context of the Maroni.

The third chapter focuses on the collection of existing data in the Maroni and the field campaign carried out in May 2015. Bathymetric interpolation methods in rivers with complex morphology are then reported, analysed and criticised. This enabled us to develop a new bathymetry interpolation technique adapted to the complexities of anabranched rivers and the lack of data. This method combines GIS processing with geometric and hydrological computer algorithms. It enables anisotropic interpolation to be carried out in the direction of flow and the transverse direction. It is a robust method because it is capable of estimating the data missing from its operating prerequisites. The result of this chapter is the first DTM of the Maroni over a total length of 105.5km (from Maripasoula upstream to Grand Santi downstream). In this DTM, anabranchs with no bathymetric data, if they are located outside the survey perimeters, are considered as porous environments. The islets of these anabranches are represented by a porosity coefficient in the hydrodynamic model.

The fourth chapter is entirely devoted to hydrodynamic simulations in the Maroni. The available hydrometric data are analysed, criticised and corrected, and new rating curves are drawn up and compared with their DEAL and IRD counterparts. The new curves provide a better fit for the gauging than the old curves. Next, the friction and numerical parameters of the hydrodynamic model were rigorously calibrated. A friction map was then obtained for the entire calculation domain, for three separate grids and two DTMs. In the first DTM, all the LiDAR ground-level data were assimilated into the geometric model. In the second DTM, on the other hand, LiDAR data from the porous reaches is excluded from the assimilation process. The results clearly show that the simulations carried out on the second DTM run more quickly than those carried out on the first, without any loss of accuracy.

The results were processed and analysed using a JAVA code that we developed. It automatically applies the criteria imposed on water height and flow speed and thus delimits the sites that are favourable for the installation of tidal turbines. In some sites of interest, flow speeds exceed 2m/s and the tidal energy density remains above 4 kW/m².

Key words : Wide rivers, bathymetry, ADCP, anabranches, meanders, DTM, TELEMAC2D, porous media, tidal power.