Transient temperature and water distributions in compacted MX80 bentonite under high temperature gradients

Abstract

Numerous element-scale and full-scale tests have been carried out on compacted bentonite to investigate its coupled thermo-hydro-mechanical (THM) behavior when used as a buffer material in high-level nuclear waste geological repositories, such as the Full-scale Engineered Barriers Experiment (FEBEX) and its forerunner laboratory mock-up test [5]. In these past studies the applied temperatures were below 100°C, while new repository conditions being explored in the recently started HotBENT project at the Grimsel Test Site involve waste canister temperatures up to 200°C [2]. Recently, Lu and McCartney [3] conducted a tank scale test where a temperature of 200°C was maintained by a cylindrical heating element having a diameter of 12.5 mm at the center of a compacted MX80 bentonite layer. A notable temperate drop in temperature was observed with distance from the small diameter heating element. A similar phenomenon was also reported by Gens et al. (2020) for heating tests on MX80 pellets during heating to 140°C. Measurements from the FEBEX project indicate a more distributed temperature decay away from the heater [3]. This discrepancy between laboratory and field tests provided the motivation for better understanding the temperature and water distribution near a heater under high temperature gradients to better guide the design of laboratory-scale experiments in the future.

        A one-dimensional heating test was performed in this study to investigate the coupled THM response of a MX80 bentonite layer in unconstrained conditions during basal heating. Measurements include the transient temperature and volumetric water content redistribution at different vertical distances from the flat plate heater under high-temperature gradients as well as global volumetric strains. Three dielectric sensors (Decagon 5TM) were placed into a compacted MX80 layer having an initial gravimetric water content of 12.2% and a dry density of 1.15 Mg/m³ within a PVC modified Proctor mold (inner diameter of 152 mm and height of 178 mm), as shown in Figure 1(a). Low-density polyethylene sheeting was stretched across the top and bottom of the soil layers to minimize any global loss of water from the compacted bentonite layer during heating. A temperature-controlled hot plate was used as the basal heat source, and a dial indicator was mounted on the top of the soil layer to measure volume changes. Fiberglass insulation was wrapped around the cell to minimize lateral heat loss from the side boundary. A basal temperature of 128°C was maintained during a heating stage lasting 4000 h to ensure THM equilibrium, after which three short-term (96 h for each) cooling-heating cycles were applied. Results indicate that soil temperatures increase rapidly in the first 20 h of heating then gradually stabilize (Figure 1(b)). The volumetric water content increases at the beginning of heating then slowly decreases (Figure 1(c)). This increasing-decreasing trend is due to the upward movement of a wetting front away from the heater due to coupled heat transfer and water flow. An initial increase in volumetric strain (swelling) was followed by a decrease (contraction) as heating progressed (Figure 1(d)). This is consistent with the observed thermal response of normally consolidated expansive soils (e.g., [4]). A sharp monotonic decrease in volumetric strain occurred during the three short-term cooling-heating cycles. The temperature and water content follow nearly identical trajectories in each cycle. The temperature distribution in the bentonite with distance from the heating element after 4000 h of heating is smoother than the sharp drop-off in temperature from the small-diameter heater observed by Lu and McCartney [3], as shown in (Figure 1(f)). The main implication of the results presented in this study is that the surface area of the heater in contact with the surrounding bentonite plays a major role in the temperature distribution and the associated coupled water flow. This is a critical finding that may help guide the geometric design of small-scale laboratory mock-up experiments used to study THM processes in controlled laboratory conditions.