Three-dimensional finite-element analyses of seepage and contaminant transport through composite geosynthetics clay liners with multiple defects
Introduction
Geosynthetic Clay Liners (GCLs), made in most cases of bentonite sandwiched between two geotextile layers, have become widely used in landfills around the world. Industrially manufactured in the form of rolls, they are easier to handle than Compacted Clay Liners (CCLs) and, when combined with thin geomembranes (GM) as composite liners, provide powerful insulation against waste leachate. However, given that the liner operates in an aggressive hydraulic, mechanical, thermal and chemical environments, questions remain about the GCLs long-term performance and their ability to maintain their insulation function over their intended design life (Rowe et al., 2004).
Wrinkles and other defects often develop in the GM, in the short- and long-terms, for various reasons, including diurnal temperature variations and poor quality assurance during placement (Nosko and Touze-Foltz, 2000). Owing to the swelling properties of bentonite, GCLs develop good contact with the overlying GM, reducing the impact of advective contaminant transport generated by the defects. Extensive research has been undertaken into leakage rates and contaminant migration through clay liners (Giroud et al., 1992, Giroud, 1997, Touze-Foltz et al., 1999, Foose et al., 2001, Touze-Foltz and Barroso, 2006, Touze-Foltz and Giroud, 2003, Touze-Foltz and Giroud, 2005, Çelik et al., 2008). Simplified formulae for calculating leakage rates, in CCLs and GCLs, have been developed for various types of defects and the effects of wrinkles size and frequencies on the hydraulic performance of GCLs has been widely studied (Wilson-Fahmy and Koerner, 1995, Giroud and Touze-Foltz, 2005, Rowe, 2005). Research has also been undertaken into contaminant levels within underlying aquifers due to defects in the geomembrane (Rowe, 2005, Abuel-Naga and Bouazza, 2009, El-Zein and Rowe, 2008, El-Zein and Touze-Foltz, 2010).
While the leakage rate is an important performance indicator, it only reflects the strength of advective flow through the liner. A more robust groundwater risk assessment must consider other contaminant fate processes, such as molecular diffusion in the clay and the attenuation layer, mechanical dispersion in the groundwater, sorption and decay. For example, the Ontario regulations recognize the two necessities of: (a) considering both advection and diffusion; and (b) defects in the geomembrane: “The design must consider both advective and diffusive contaminant transport and must include examination of the effect of the failure of any engineered facilities when their service lives are reached.” (Ontario Landfill Regulations, 2010).
A number of studies have performed such combined hydro-chemical analyses in 1D and 2D (Rowe and Booker, 2005, El-Zein and Rowe, 2008, El-Zein and Touze-Foltz, 2010). El-Zein and Rowe (2008) and later El-Zein and Touze-Foltz (2010) identified specific conditions under which leakage rates does not reflect the true magnitude of contamination taking place.
1D and 2D analyses are relatively simple to conduct but they suffer from significant limitations. They are based on simplifying assumptions such as: (a) the leak must extend to the full-length of the landfill cell, in one dimension; (b) no diffusion or mechanical dispersion occur in that dimension; and (c) the groundwater must flow in the direction normal to the leak in the aquifer. A three-dimensional approach makes it possible to avoid these assumptions. However, to the best of our knowledge, no paper has yet addressed the coupled effect of hydraulic flow and chemical transport of contaminants through defective liners in 3D.
In this paper, we develop a 3D finite-element model of a typical GCL system. We use the steady-state flow equation, coupled with the reactive diffusion-advection equation and apply them to the multi-layered medium through the Soil Pollution Analysis System (SPAS). We simulate the transport of DCM through a system consisting of a GM, a GCL, an attenuation layer and a thin aquifer. The GM is either free of defects, or carries one or multiple defects. We run our analyses in 2D and 3D in order to quantify the effects of three commonly encountered sets of assumptions: (a) the defect extend the full-length of the landfill in one direction (as opposed to partial length); (b) the groundwater flow is normal, rather than parallel, to the direction of the defect; (c) a zero-flux boundary condition downstream from the aquifer, rather than an advective discharge or infinite boundary condition; and (d) a single defect, rather than multiple ones.
Fig. 1 is a graphic illustration of points (a), (b) and (d). In the remainder of the paper, we briefly present the theory of seepage and contaminant transport on which our models are based. Next, we describe the computer code developed and the landfill models we built, including the material properties used. Finally, we present our results and discuss our findings.
Section snippets
Theory of seepage and contaminant transport
The transport of water and dissolved chemicals in the water-saturated subsurface can be represented by the semi-coupled steady-state seepage and time-dependent reactive diffusion-advection equations. Water flow is assumed to be at steady-state because the time scale of chemical transport is much larger than that of hydraulic seepage in clayey soils. Soil is assumed to be a continuum. Chemical transport is governed by a Fickian diffusion process (which represents molecular diffusion and
Finite-element implementations and CONFEM3D
Two finite-element formulations of Eqs. (1), (2), (3), (4), (5), (6), (7), (8), (9), (10), (11), (12), (13), (14), (15), (16), (17), (18), (19), (20) were implemented in a computer program CONFEM3D with a graphic interface SPAS3D. The first formulation, LTFEM, solves a Laplace-transform version of the equation then inverts the solution numerically back into the time domain (El-Zein et al., 2005, El-Zein et al., 2006, El-Zein and Balaam, 2008). This formulation is computationally efficient
Study of the migration of DCM with SPAS3D
Building on a previous study by El-Zein and Rowe (2008), the case of the migration of DCM through a GCL composite liner with a single or multiple rectangular leaks, into the underlying aquifer, has been simulated. DCM is an organic contaminant with a relatively high diffusion coefficient through high-density polyethylene geomembranes.
A cross-section of the base model, with 1 leak is shown in Fig. 2. Material properties are shown in Table 1. The landfill is taken to be 40 m wide and 100 m long,
Contours of velocities and concentrations
Fig. 5 shows typical results generated by SPAS3D. The build-up of high seepage velocities around the defects can be seen in Fig. 5a. Fig. 5b depicts the sudden spike in downward Darcy velocities around the defect. The strength of leakage and the width of the leakage area clearly depend on the width of the defect and the transmissivity of the GM-GCL interface. Finally, Fig. 5c provides a view of the lower surface of the aquifer and the way contaminants spread from an area below the leak. The
Conclusions
The development of SPAS3D allows more accurate numerical analyses, with fewer simplifying assumptions, of multi-layered liner systems with multiple defects to be conducted. Our results have three implications. First, the assumption made in 2D analyses of a long leak, does not seem to incur more than about a 30% increase in groundwater concentration of the organic contaminant, on the conservative side. This is an acceptable level of inaccuracy in 2D analyses given the computational benefits that
References (29)
- et al.
Rate of leakage through a composite liner due to geomembrane defects
Geotextiles and Geomembranes
(1992) - et al.
Numerical characterization of advective gas flow through GM/GCL composite liners having a circular defect in the geomembrane
Journal of Geotechnical and Geoenvironmental Engineering
(2009) - et al.
Lapack User’s Guide
(1999) - et al.
One dimensional advective-dispersive transport into a deep layer having a variable surface concentration
International Journal of Numerical and Analytical Methods in Geomechanics
(1987) - et al.
Effect of vadose zone on the steady-state leakage rates from landfill barrier systems
Waste Management
(2008) - et al.
A case study of wrinkles in a Textured HDPE geomembrane on a slope
Proceedings Geoamericas
(2008) - El-Zein A., Balaam N., 2008. SPAS: A New Finite-Element tool for modelling transport through damaged landfill liners,...
- et al.
Multiple-porosity contaminant migration by finite-elements
International Journal of Geomechanics ASCE
(2005) - et al.
Three-dimensional finite elements for the analysis of soil contamination using a multiple-porosity approach
International Journal of Numerical and Analytical Methods in Geomechanics
(2006) - et al.
Impact on groundwater of concurrent leakage and diffusion of dichloromethane through geomembranes in landfill liners
Geosynthetics International
(2008)
Equivalency of Landfill Liner Designs, including Clay Liners, Geosynthetics Liners and Polymer-Enhanced Mineral Barriers, under French Legislation
Predicting leakage through composite landfill liners
Journal of Geotechnical and Geoenvironmental Engineering
Equations for calculating the rate of liquid migration through composite liners due to geomembrane defects
Geosynthetics International
Equations for calculating the rate of liquid flow through geomembrane defects of uniform width and finite or infinite length
Geosynthetics International
Cited by (34)
Numerical estimation of the equivalent hydraulic conductivity for canal concrete lining with cracks
2024, Journal of HydrologyNumerical modeling of contaminant advection impact on hydrodynamic diffusion in a deformable medium
2022, Journal of Rock Mechanics and Geotechnical EngineeringCitation Excerpt :Contaminant transport models are generally derived from the assumption that the clay or composite liners are of an incompressible material, in which only solute advection or dispersion transport occurs (Ogata, 1970; Bear, 1972; Freeze and Cherry, 1979; Chen et al., 2007; Ciftci, 2017). Relevant solutions of partial differential equations are solved by finite element methods (e.g. Javadi and Al-Najjar, 2007; El-Zein et al., 2012; Ma et al., 2021), boundary element methods (e.g. Tan et al., 1994; Šarler and Kuhn, 1998a, b; Al-Bayati and Wrobel, 2021), spectral methods (e.g. Zhao and Liu, 2017; Chen et al., 2018) and meshless methods (e.g. Gingold and Monaghan, 1977; Liu et al., 1995; Babuška and Melenk, 1997; Liu and Gu, 2002; Xiao, 2004). In particular, (Lin et al., 2020, 2021) developed a backward substitution-based meshless method to solve 1D and 2D linear and nonlinear advection–diffusion–reaction problems.
Influence of pore water chemistry on GCL self-healing with hydration from silica sand
2021, Geotextiles and GeomembranesPerformance issues of barrier systems for landfills: A review
2021, Geotextiles and GeomembranesStochastic evaluation of leakages through holes in wrinkle networks of composite liners
2020, Geotextiles and Geomembranes