Analyses 3D des fuites et transfert de contaminant à travers des étanchéités composites comportant plusieurs défauts

International audienceGiven the relatively recent history of geosynthetic clay liners (GCLs), questions remain over their longterm performance in landfills. Defects in the geomembrane (GM) overlying the clayey barrier are likely to develop, providing an advective pathway for contaminants to migrate into the liner and the underlying groundwater. The effects of this process on the integrity of the liner have been usually quantified through hydraulic leakage rates, which does not account for diffusion effects. More recently, efforts have been made to couple leakage with contaminant migration analyses that take into account defects in the geomembrane and evaluate chemical concentration in the groundwater, as well as leakage rates. However, these studies have always been conducted in 1D or 2D, despite the fact that 3D effects are clearly present. We develop a 3D finite-element model and simulate the transport of dichloromethane (DCM) through a typical GCL composite liner system consisting of a GM, a GCL, an attenuation layer and a thin aquifer. We solve the steady-state flow equation, coupled with the reactive diffusion-advection equation through the Soil Pollution Analysis System (SPAS). The GM is either free of defects, or carries one or multiple defects. We run our analyses in 2D and 3D in order to investigate the impact of a number of 3D effects: (a) the extent of the defect in one direction (full or partial length of the landfill base); (b) the direction of groundwater flow (parallel versus normal to the defect); (c) aquifer downstream boundary condition (zero-flux, advective discharge or infinite extent); and (d) different arrangements of multiple defects. Leakage rates and contaminant concentrations are found to increase with defect size as expected. 1D and 2D assumptions about the direction of groundwater with respect to the defect orientation and the mass-transport boundary condition applied downstream in the aquifer can lead to significant underestimation of contamination levels