Thermal Analysis of GeLs at a Municipal Solid Waste Landfill

This study was conducted to determine long-term thermal regime of landfill liner systems using a field temperature monitoring program and numerical analysis of heat transfer. Temperatures in liner systems that contain geosynthetic clay liners (GCLs) were monitored prior and subsequent to waste placement. Data were collected in three cells at a landfill in Midwestern USA for more than five years. The liner system in one of the cells was left exposed (not covered with waste) for a period exceeding one year subsequent to cell construction. The lowest and highest GCL temperatures were -1°C and 35°C, respectively and the localized temperature gradients ranged from approximately -186°C/m to +134°C/m across the liner system during the exposed period. Seasonal temperature fluctuations were dampened and increasing temperature trends were observed after placement of the first lift of waste over the liner systems. Temperatures in liners reached 30°C under 5-year-old waste with an annual rate of temperature increase of approximately 4°C/a. In general, average temperature gradients decreased, however, high variations in gradients remained subsequent to waste placement. Numerical analysis was used for exposed liner systems to model observed behavior and to predict liner response under varied conditions. Good agreement was observed between measured and predicted temperatures. Temperature distributions in liners were determined for variable thicknesses of sand protective layers in 3 climatic regions. Sand thicknesses of more than 1 m were required to maintain temperatures between 0 and 40°C and thicknesses on the order of 3 m were required to limit seasonal temperature variations to within lOoC in GCLs. Introduction GCLs are used commonly in containment systems for bottom liner systems as well as for cover liner systems. GCLs are subjected to coupled effects of environmental and operational stresses in the field including mechanical, hydraulic, chemical, and thermal stresses. Temperature affects properties and behavior of GCLs (Rowe 1998). In general, detailed spatial or temporal distributions or comprehensive long-term trends for GCL field temperatures are not available. Also, heat transfer in liner systems has not been analyzed extensively in the past. This study was conducted to determine long-term temperature trends in GCL liners in the field. Onset and duration of temperature variations with respect to local air and ground temperatures have been studied. Thermal gradients in a liner system were determined. Heat transfer through GCL liner systems was investigated using numerical analysis. Background Physical, mechanical, and hydraulic properties of GCLs are determined in the laboratory for manufacturing quality control and for conformance and acceptance testing. Baseline mechanical properties, hydraulic properties, and fluid transport characteristics of GCLs have been determined at temperatures near 20°e. The response of GCLs was investigated for variable placement conditions, stress conditions, permeant liquids, and environmental exposure conditions (Rowe 1998). These tests have generally been conducted in the laboratory at standard laboratory conditions even though the test materials may have experienced variable temperatures during certain stages of tests (e.g., cyclic freeze/thaw and wet/dry tests). Properties and response of GCLs are affected by temperature. Freeze/thaw and desiccation of GCLs are controlled by field temperatures. Fluid transport through GCLs is affected by temperature directly due to viscosity and dielectric constant effects and indirectly due to moisture content effects. Also, thermally driven fluid transport (thermoosmosis) occurs in soils. Diffusion is a critical mechanism for fluid transport in low hydraulic conductivity materials (including both clay and geomembrane components of GCLs). The free solution diffusion coefficient, Do, for species common in MSW leachate (Rowe et al. 1995) and the soil diffusion coefficient, D, for a clay soil beneath a landfill (Crooks and Quigley 1984) both increased with temperature. Ion and potential distributions adjacent to charged clay particles, which can affect the microstructure of the soil, are sensitive to temperature (Mitchell 1993). In addition, temperature influences mechanical response of clay soils and it is expected that the same effects will be present in GCLs. Furthermore, durability of geosynthetic components of GCLs is affected by temperature. Several studies of temperatures at the base of landfills (near liner systems) under waste-covered conditions have been reported in geotechnical literature. Yoshida and Rowe (2003) reported temperatures between approximately 20 and 50°C at the base of three landfills in Japan. The temperatures were monitored for 12 to 30 years using sensors that were installed between 7 and 10 years after initial waste placement. Generally, temperatures were high at the onset of the monitoring periods and decreased with time. An average temperature of 14°C was observed for no leachate mounding and increasing temperatures in the range of 20 to 37°C were measured for high leachate mounding at the base of a landfill in Canada over a 7-year period (10 to 17 years after waste placement) [Rowe 1998]. Gartung et al. (1999) reported temperatures between 20 and 40°C underneath wastes that were more than 10 years old and more than 50°C underneath 6 to 8-year old wastes at a German landfill. Dach and Jager (1995) indicated that temperatures in excess of 50°C were measured at the base of landfills in Germany. Koerner (2001) reported temperatures between 20 and 35°C for a liner with GCL in Pennsylvania monitored over 9.5 years. The cell was covered by the end of the third year of monitoring. The temperatures were near 20°C for the first 5.5 years and then increased. Liner temperatures of 10 to 30°C and 20 to 30°C were reported for landfills in California and Florida, respectively (Rowe 1998). Temperatures of bottom liner systems were recorded in winter months to determine extent of freezing of the liners that were exposed for several years at 2 landfills in Germany (Dullmann and Hilpusch 1997). Temperatures below O°C were measured within the clay liners in both landfills, with higher frost penetration depths observed for clay overlain by a geomembrane than clay overlain by gravel. Remedial measures included excavation of clay and use of polystyrene insulation. Rodatz and Voigt (1997) recommended site-specific heat transfer analysis and use of insulation to prevent freezing of soil liners in Central Europe based on laboratory and field test plot investigations of compacted clay liners. They also indicated that a I-m-thick waste layer would provide insulation to the underlying liners. Numerical analysis was used to investigate heat transfer in waste-covered soil liners. Doll (1997) and Heibrock (1997) used one-dimensional nonlinear differential equations of heat transfer to assess thermally driven moisture flow and desiccation cracking of compacted clay liners. The models had provisions for heat transfer as well as water and water vapor transfer for unsaturated conditions in the soil liner. Yoshida and Rowe (2003) used one-dimensional nonlinear differential equations of heat transfer to predict temperatures in a landfill and the underlying soil. The model had provisions for heat flow due to conduction and water transport and for heat generation due to decomposition of the waste. Detailed modeling of heat transfer in exposed liners (before waste placement) has not been commonly reported. Limited data is available on long-term field temperatures of GCLs including onset and duration of temperature variations and thermal gradients. Also, heat transfer analysis has not been reported for GCLs. The study described in this paper was conducted to investigate the thermal regime of liner systems using field monitoring and numerical analysis. The field monitoring portion of the study is an ongoing investigation. Even though preliminary data were reported elsewhere by Yesiller and Hanson (2003), all of the liner temperature data obtained at the site is reported herein. Temperatures are reported for exposed and waste-covered conditions in the liners. A numerical model was developed to predict temperatures within an exposed bottom liner system. The model was used to provide recommendations for thickness of protective cover layers (insulating layers) to be placed over liner systems to limit excessive seasonal temperature and thermal gradient fluctuations in the liner systems. Field Investigation This study is conducted at a municipal solid waste landfill located in Michigan, U.S.A. The site is in an area that has a humid continental temperate climate. The climatological conditions at the site are reported for the period 1999 to 2003. The annual normal average high and low temperatures were 15.1°C and 5.5°C, respectively. The annual normal precipitation and the average snowfall at the site were 868 mm and 1068 mm, respectively. The average barometric pressure was 103 kPa and the average relative humidity was 70.6%. The seasonal ground surface temperatures at the site typically varied between -1°C and 30°C. The mean annual earth temperature measured at depth at the site was 11.7°C. The landfill is located in an area with silty clay soils to typical depths of 21 m underlain by hardpan (stiff layer consisting of cemented sand and fine gravel) and bedrock at a depth of approximately 23 m. The water content of the clay ranges from approximately 10 to 12% and the unit weight of the clay is 22 kN/m . A confined aquifer is located at the hardpan/bedrock elevations. Excavations for the construction of individual cells extend to depths of approximately 10 to 13 m below original ground surface. The landfill consists of 11 cells: 8 existing cells and 3 future cells. The liner systems of 3 existing cells (Cells J, I, and D) were instrumented for this study. The cells have similar single composite liner systems, which consist of (from top to bottom): 450­ mm-thick protective sand layer, nonwoven needle-punched PP geotextile / HDPE geonet (GT/GN), 1.5-mm-thick HDPE geomembrane (GM), and needle-punched laminated GCL (mass/unit area of 3660 g/m). Cell J (area 3.2 ha) was constructed in 1998 and filled with waste from 1999 to 2001. Cell I (area 3.5 ha) was constructed in 1999 and filled from 2000 to 2002. Cell D (area 5.2 ha) was constructed in 2001 and has been filled since December 2002. Waste is generally placed in constant height lifts that cover the entire base area of a cell. Such a lift generally has an accumulated thickness of 4 m and is placed over a period of approximately 40 days. The average waste column heights in Cells J, I, and Dare 44 m, 31 m, and 17 m, respectively. Temperatures are measured using thermocouple sensor arrays. An array consists of a series of wires beginning at a monitoring box and terminating at various points along a linear path. Type K thermocouple wires consisting of Nickel alloys (Ni-Cr/Ni-Mn­ AI) were selected for their resistance to chemical environments. The configurations of the sensor arrays are presented in Figure 1. Arrays were placed within the sand layer in all three cells (Type A). In Cell D, two additional arrays were used: Type B array was placed on the proof-rolled subgrade immediately below the GCL, parallel to the array within the sand layer, and Type C array was placed vertically below the GCL in the underlying subgrade approximately 95 m away from the edge of the cell. The long arrays start near the mid-section of a perimeter edge of a cell and extend towards the center of a cell. Type A arrays were installed in Cells J and I in March and November 1999, respectively followed by waste placement within 2 to 3 months. The liner system in Cell D was constructed between September and December 2001. Types Band C arrays were installed below the GCL in September 2001 and a Type A array was installed in the sand in December 2001. The liner system was left exposed for more than a year until waste placement started in December 2002. A Arrays '\ / Cell J: 183-m array, 10 sensors Cell I: 169-m array, 11 sensors Cell D: 186-m array, 9 sensors D Sensor Array