Efficacité d'un seuil artificiel sur l'oxygénation de l'eau et l'élimination de CH4 contenu dans l'eau évacuée par la barrage hydroélectrique de Petit Saut (Guyane française)

Abstract

Quelques mois après le début de la mise en eau du barrage de Petit Saut, la mise en service normale de l'usine conduisait à une désoxygénation de l'eau du tronçon de rivière aval, le rendant incompatible avec la vie aquatique. La solution retenue a été la construction d'un seuil, afin d'apporter de l'oxygène et d'éliminer les gaz réducteurs produits au fond de la retenue, notamment le méthane, consommateur potentiel d'oxygène dissous.Un seuil métallique à deux lames déversantes successives a été construit ; sa configuration prend en compte les principaux critères physiques jouant un rôle significatif sur l'oxygénation de l'eau (hauteur de chute, épaisseur de la lame déversante, le dimensionnement du bassin de réception des chutes, la présence de dispositifs favorisant l'éclatement de la lame d'eau).Placé dans le canal de fuite de l'usine, à une centaine de mètres à l'aval du barrage principal, il est à l'abri des crues et ne crée pas d'obstacle supplémentaire en rivière. L'article chiffre l'effet d'aération de ce seuil pour les deux gaz O 2 et CH 4 dans deux configurations : celles consécutives à l'abaissement partiel de la chute amont réalisé en deux étapes. Après décembre 2001, pour le débit moyen turbiné (près de 200 m 3 /s), l'efficacité d'aération du seuil a baissé de près de 10 % (gain de 80 % en oxygène dissous et élimination de 70 % et 75 % du méthane dissous). Après février 2003, pour un débit de 100 m3/s, 75 % du déficit amont en oxygène dissous est comblé et près de 70 % du méthane dissous éliminé. From the moment tropical reservoirs are impounded, climatic conditions cause rapid (within several weeks) and marked thermal stratification, especially during the dry season. This phenomenon is further exacerbated by the chemical and biochemical processes taking place in the reservoir due to the decomposition of submerged organic matter. In dense tropical forests, the overhead biomass is estimated at roughly 170 t(C)/ha, and the carbon contained in the soil is also not negligible since it is on the order of 100 t(C)/ha. The degree of biodegradability of the different compounds in the flooded biomass is variable, ranging from a few weeks for bacteria to several centuries for tree trunks. The studies carried out at Petit Saut (French Guiana) show that, immediately after impoundment, only the epilimnion (a few dozen centimetres thick) was oxygenated whereas the hypolimnion was characterized by complete anoxia and a very high methane content (about 15 mg/L). Water quality in the river downstream from the reservoir was of course strongly linked to variations in the water quality in the reservoir as well as to its operating mode. The waters passing through the turbines, coming from the bottom layers, were anoxic and loaded with fixed or volatile reducing compounds (e.g., CH 4 , H 2 S), and were responsible for a high immediate or progressive oxygen demand. At Petit Saut, despite an inflow of good quality water, there has been a progressive deoxygenation in the river downstream due to the high methane content (roughly 8 mg/L) of the turbined water. Thus, 40 km downstream from the dam, the oxygen content was less than 2 mg/L and therefore incompatible with most aquatic life. To solve this problem, it was necessary to build an aerating weir capable of reoxygenating the turbined waters and, more importantly, eliminating reducing gases such as methane at the same time. The function of the overflow weir was to entrain air bubbles into the water and to give these bubbles a sufficiently long immersion time to ensure that they dissolve. At the time of its installation, only three examples of oxygenating weirs existed in the entire world, all located in the United States. The weir configuration was tested using a physical model to qualitatively examine the form of the flow both across the weir and downstream from it. The degree to which air bubbles were entrained in the water was also tested, but not the question of evaluating the flux of gaseous exchanges between the air and the water.The system that was finally designed by EDF, in October 1994, was a metallic weir with two consecutive falls, the configuration of which respected the main physical criteria that play a significant role in the oxygenation of water, i.e.:- the height of the falls (roughly 5.40 m, depending on the flow rate);- the thickness of the water stream, the function of which is to entrain air bubbles and keep them in the water for a sufficiently long period of time for the oxygen to dissolve (between 12 and 25 seconds, depending on the flow rate);- the dimensions of the receiving basin of the first waterfall where the air bubbles are held (5 hexagonal alveoli); and- systems to promote the fragmentation of the flow.This structure was placed in the tailrace channel of the plant, approximately 100 m downstream from the main dam. This location protected it from floods and did not create an extra obstacle in the river. In addition, it allowed the water to be re-oxygenated as soon as it left the reservoir. The efficiency of the two waterfalls of the Petit Saut re-aerating weir was tested at two different turbine flow rates: 80 m 3 /s and 230 m 3 /s. In 1996, the results of the measurements showed that for a flow rate of 230 m 3 /s, upstream of the weir the concentrations of CH 4 were around 5 mg/L and dissolved oxygen was 0.8 mg/L. Downstream from the weir CH 4 concentrations were 1.3 mg/L and dissolved oxygen concentrations were 6.8 mg/L. The dissolved methane elimination rate was approximately 75 per cent. At a flow rate of 80 m3/s, upstream of the weir the concentration of CH 4 was 5.5 mg/L and the dissolved oxygen concentration was 0.7 mg/L. Downstream from the weir concentrations of CH 4 and dissolved oxygen were 1.0 mg/L and 7.1 mg/L, respectively. The dissolved methane elimination rate was around 80%. The efficiency of the re-oxygenation was always greater than 90%. These data prove that the efficiency of the Petit Saut weir installation was higher when the turbine flow rate was lower. This could be due to a greater waterfall height, the better entrainment of air bubbles per unit volume and/or a longer air bubble residence time in the downstream flow. Between December 2001 and February 2003, for a flow rate of 200 m 3 /s, the efficiency of the weir decreased by 10%, with the dissolved methane elimination rate at around 70-75%. The level of re-oxygenation was around 80%. Since February 2003, for a flow rate of 100 m 3 /s, the efficiency of the weir has decreased by 10%, the dissolved methane elimination rate was around 70% and the level of re-oxygenation was around 75%. On a local scale, the effect on the quality of the river water has been very positive, as aquatic life has been maintained. Without the weir, the methane contained in the turbined water would have been progressively transformed, along the course of the river, into carbon dioxide. In the absence of significant additions of good quality water and without the weir, a large part of the course of the river would have a dissolved oxygen content of less than 2 mg/L, the critical threshold for the maintenance of aquatic life.At present time, the results of the current ecological survey are used to support studies on biogeochemical processes.