1. Introduction
The decay heat of high-level radioactive waste (HLW) can degrade the integrity of engineered barrier system (EBS) over long periods of time in deep geologic repositories [1]. The diffusion and reaction behavior of dissolved oxygen or hydrogen sulfide, which can corrode the disposal canister, is closely dependent on temperature [2, 3]. In addition, when groundwater vaporizes on the hot surface of the disposal canister, salts can accumulate on the surface and accelerate canister corrosion [4]. It has also been shown that prolonged exposure of the bentonite buffer material to high temperatures transforms the swellable montmorillonite into the non-swellable illite [5]. Therefore, most countries considering HLW disposal are increasing the spacing between disposal boreholes or tunnels to control the maximum bentonite temperature in the repository to below 100°C [6, 7].
If the thermal conductivity of the EBS can be improved to effectively reduce the maximum temperature of the bentonite buffer, the area of the disposal site can be reduced. Therefore, various attempts have been made to reduce the maximum temperature of the bentonite buffer. Several studies have been conducted to increase the thermal conductivity of bentonite buffer since the thermal conductivity of bentonite buffer is low, below 1.0 W·mK−1. Some papers have proposed mixing thermally conductive additives such as graphite to increase the thermal conductivity of bentonite buffer [8, 9]. However, adding additives may have the disadvantage of reducing other buffer properties, such as thermal conductivity or swelling pressure. There have also been attempts to increase the surface area of the disposal canister for active heat transfer [10]. In particular, Wang et al. [11] argued that inserting a copper mesh into a bentonite buffer is a more effective heat transfer method than mixing thermally conductive additives such as graphene or graphite [11]. In addition, Kim et al. [12] showed through computational analysis that introducing a thin copper plate with high thermal conductivity into a bentonite buffer can reduce the maximum buffer temperature very effectively. Kim’s computational analysis showed that a single copper plate with a thickness of 10 mm can reduce the temperature by 4.2°C. However, the work of Wang et al. [11] and Kim et al. [12] was limited to computational analyses and did not lead to real demonstrations.
Therefore, this study aimed to investigate the heat transfer of copper plates in the bentonite buffer proposed by Kim et al. [12] using a small-scale demonstration module. To demonstrate the heat dissipation performance of the copper plate, a small-scale EBS module was designed and fabricated in this study, and the copper plate was designed into two types: simple planar and U-collar shaped. The thickness of the bentonite buffer in the small-scale module was 20 mm, which is about 1/18 of the actual size. The purpose of this demonstration was to verify how much the maximum buffer temperature at the interface between the copper canister and the buffer block is actually lowered by introducing the heat dissipating copper plate. If the copper plate can significantly reduce the maximum bentonite temperature, it is expected that the spacing of disposal boreholes or disposal tunnels can be reduced, which will significantly contribute to reducing disposal site construction costs and securing the long-term integrity of EBS.
2. Experiment
2.1 Design & Manufacturing
Fig. 1 shows the configuration of a small EBS module for the heat dissipation demonstration. The internal space of the EBS module was designed to be D70 × H90 mm. And the size of copper canister was D30 × H60 mm which corresponds to about 1/33 scale based on D1.0 m disposal canister. Since only lateral thermal conductivity was considered in the small module, the top and bottom of bentonite block were insulated with 10 mm thick PVDF blocks in the small module. The thermal conductivity of the PVDF block is about 0.2 W·mK−1, which was much lower than that of the compact bentonite block (0.96 W·m−1⋅K−1 at a dry density of 1.64 kg·m−3). A custom-made cartridge heater (D16 mm × H50 mm) was inserted in the copper canister as a heating source. Two doughnut-shaped bentonite blocks (OD70 × ID30 × H30 mm) and one disk shaped bentonite block (OD 70 × H30 mm) were manufactured using a Ca type bentonite (Clariant KOREA Ltd., Product Name: Bentonil WRK). The dry density of the blocks was about 1.64 kg·m−3, the moisture content was about 13.4%, and the apparent density was about 1.86 kg·m−3. The thermal conductivity of the manufactured bentonite block was measured by QTM-500 (Kyoto Electronics), and the average value was about 0.96 W·m−1⋅K−1. On the other hand, the module had no water supply system to swell the bentonite blocks, since the point at which the temperature of the disposal canister reaches its peak is dry state at the beginning of the disposal period, and when the buffer is saturated with water, the thermal conductivity increases and the thermal problem become not serious.
To measure the temperature at key positions in the module, six K-type thermocouples were installed as shown in Fig. 1. The six positions are the cartridge heater that generates heat, the central contact point of the copper canister and buffer block where the maximum temperature is expected, the upper contact point of copper canister, the inner wall of the test module, the outer wall of the test module, and the internal space of the environmental chamber where the test module is placed. The real-time measurement of the temperature variations over long time was done using a data collection device (Omega USB DAQ module, Model: OM-DAQ-USB-2401). A heat dissipation experiment was conducted in an environmental chamber (Carrier low-temperature refrigerator) so that the small scale EBS module was not affected by ambient temperature change during the test. The environmental chamber can control an inside temperature within 15–20℃. Two power sources were used for the heat dissipation experiment. A power supply (HCS- 2SD10, Capacity: 1 kW) was used to supply a constant electric power, and an automatic temperature controller (MTOPS model: TC200P) was used to maintain the inside of the copper canister at a constant temperature.
As shown in Fig. 2, two types of copper plates were designed: a flat type and a U collar type. The outer diameter of the copper plate was designed to 69.5 mm to minimize the gap with other barriers. The thickness of the copper plate was set to 0.3 mm which was about 1/7 scale regarding to the 2.0 mm of previous study [12]. The collar height of the U collar copper plate was set to 10 mm. The function of the collar is to secure the close contact of copper plate to other barriers like the copper canister and module inner wall. The copper plate was placed between the two doughnut-shaped bentonite blocks at the middle of the copper canister that generates heat.
2.2 Operation of Small Module
Fig. 3 shows the installation figures of small-scale EBS module for the demonstration of heat dissipation of a copper plate. The small module was operated in two conditions: (a) constant temperature and (b) constant electric power of the copper canister. In the constant temperature test, the environmental chamber was maintained at 18°C and the core temperature of copper canister was adjusted to 50, 70, and 90°C in stages, and then the temperature gradient of key parts in the module was measured. (b) In the constant power test, the environmental chamber was also maintained at 18°C and the electric power supplied to the cartridge heater in the copper canister was kept constant. The supply current was fixed at 4.16 A, but the supply voltage was adjusted to 30-40-50-55-60 V in stages. Those were corresponded to 125-166-208-229-250 W based on electric power.
The experiment was conducted in three different heat dissipation configurations as shown in Fig. 4: (a) no copper plate, (b) flat copper plate, and (c) U collar copper plate. Since this test aimed to verify the effectiveness of the heat dissipation of copper plate in the small module, it was necessary to calibrate each thermocouple before the test whether all thermocouples exhibit the same figure at the same temperature condition. For the purpose, the environmental chamber was controlled at constant temperature without any heating for 3 days to reach an equilibrium thermal state, and the temperatures recorded at six sensors were compared, and corrected based on the core temperature of the copper canister.
3. Result and Discussions
3.1 Temperature Measurement
It was intended to verify the heat dissipation performance of a copper plate by tracking the temperature change in key parts of a small module. In the constant temperature test, the environmental chamber was maintained at 18°C, and the core temperature of the copper canister was controlled at three constant temperature of 50-70-90°C. The temperature change at key parts in the module was recorded. Fig. 5 shows the graphs of the temperature changes in key parts under the constant temperature mode where the flat copper plate was applied. The equilibrium temperatures of key parts were collected by drawing a blue vertical line at the stabilized point on the graph of Fig. 5.
In the constant power mode, the environmental chamber was also maintained at 18°C, and the equilibrium temperatures of key parts were measured by changing the electric power. Fig. 6 shows the graphs of the temperature changes of key parts under constant power mode where the flat copper plate was applied.
3.2 Temperature Gradient
The equilibrium temperatures of key parts in the small module were collected from the constant temperature test and constant power experiment, and the temperature gradient from the core of the copper canister to the environmental chamber were evaluated in three heat dissipation conditions.
3.2.1 No heat dissipation plate
Fig. 7 shows the equilibrium temperature gradient in a small module without a heat dissipating copper plate under (a) the constant temperature mode and (b) the constant temperature mode. The graph patterns shown in the constant temperature mode and the constant power mode were almost the same. This was understood as a natural result because the external temperature was fixed at 18°C, and the core temperature is also equilibrium in the both mode. And as the core temperature or the power supplied to the core increased, the temperature gradient became steep. The temperatures at the contact points (B1 and B2) between the copper canister and the bentonite buffer were not different. And as the temperature difference (B2 and B3) between the two sides of the buffer is rather steep, it meant that heat transfer through the bentonite buffer was slow.
3.2.2 Flat copper plate
Fig. 8 shows the equilibrium temperature gradient in a small module with a flat copper plate under (a) the constant temperature condition and (b) the constant power condition. At a high power supply of 253 W, the temperature of the environmental chamber rose to 20°C, slightly more than the controlled 18°C, due to the insufficient cooling capacity of the environmental chamber. Even in this graph, there was little temperature difference between the two contacts points (B1 and B2), and as the core temperature increased, the temperature difference between the two sides (B2 and B3) of the buffer block became larger. The overall temperature gradient when the flat copper plate was applied did not differ significantly from that without the copper plate, but the buffer temperature (B1) was slightly reduced.
3.2.3 U collar copper plate
Fig. 9 shows the equilibrium temperature gradient in a small module with U collar copper plate under (a) the constant temperature mode and (b) the constant temperature mode. The temperature gradient from the core to the environmental chamber showed a significant difference from the previous two cases. The temperature difference between both sides of the buffer (B2, B3) was greatly reduced, and instead, the temperature difference between the module surface and the environmental chamber (M surface, Air) increased. This result meant that heat transfer in the buffer layer became activated when a U-collar copper plate was inserted into the bentonite blocks.
3.3 Degree of Heat Dissipation
3.3.1 Constant temperature conditions
The heat dissipation effect of the copper plate was analyzed when the core temperature of the copper canister was constant at 90°C (Fig. 10). It was confirmed that there was a steep temperature drop through the buffer layer without any plate or with flat copper plate, and the temperature drop in the buffer layer was about 17.7℃ and 15.1℃ for the no plate and copper flat plate respectively. However, when the U-collar copper plate was adopted, the temperature drop in the buffer layer decreased to 8.1℃, which was about half of the others. Therefore, it can be said that heat transfer was much better when the U collar copper plate was adopted. From this result, it was found that the contact state of copper plate with other barriers had a great influence on the core-heat dissipation.
3.3.2 Constant power conditions
The heat dissipation effect of the copper plate was analyzed when the power supply to the copper canister was similar near at 250 W (Fig. 11). When a copper plate was inserted, there was little temperature reduction at the interface (B1) between the copper canister and the buffer, and it was only 1.8℃ lower than that of no copper plate. However, when a U collar copper plate was adopted, the temperature reduction was increase to 9.1℃. Therefore, it has been experimentally proven that the maximum buffer temperature can be significantly lowered when the U collar copper plate is applied to the EBS for a HLW disposal.
3.4 Comparison With Computational Analysis
This small module heat dissipation tests were carried out with a focus on the heat transfer in the lateral direction of the bentonite buffer. The side thickness of the buffer layer is 20 mm, corresponding to 1/20 of the real thickness of 350 mm [12]. In addition, the electric power supplied to the copper canister was in a range of 125–250 W which is in a range of 1/16~1/8 scale considering the 1.91 kW decay heat of the disposal canister loaded with four high burn-up PWR spent nuclear fuels [13]. Therefore, this small module heat dissipation test might be a conservative experiment in which the temperature difference between the two sides of the buffer may be lower because the thickness of the buffer is relatively thinner compared with the heat scale.
The maximum buffer temperature in the absence of a copper plate was as high as 81.6℃ in the previous computational analysis of single full-scale borehole [12]. However, when one flat plate with a thickness of 10 mm was inserted in the buffer layer, the maximum surface temperature decreased to 77.4°C, resulting in 4.2°C reduction. And when one U collar copper plate was inserted, the maximum buffer temperature was about 75.6°C, resulting in 6.0°C reduction, making no big difference from that of the flat plate [12]. In the case of computational simulation, unlike the real experiment, the contact between the copper flat plate and the other barriers was assumed to be good. Therefore, it was understood that there was no significant difference in temperature reduction between in case of the U collar copper plate and flat copper plate. Even considering the conservative results, the temperature reduction of 1.8°C for the adoption of flat copper plate was unexpectedly lower than the 4.2°C reduction in the computational analysis [12]. The reason was judged to be that there was some serious gap between the copper plate and the other barriers. However, in the case of a U collar copper plate, a big temperature reduction of 9.1°C occurred as shown in Fig. 11, which was greater than the temperature reduction of 6.0°C derived from the computational analysis [12]. Factors that affect heat dissipation efficiency include plate thickness, collar height, and the aspect ratio of canister, can made it difficult for the accurate comparison of experimental results with the computational analysis. In this demonstration, the aspect ratio of the copper canister is 2:1, but the aspect ratio of real disposal canister is about 5:1, so it is believed that two to three copper plates may be needed in the real EBS if the similar temperature reduction is wanted like the small module test.
In conclusion, the maximum temperature reduction of a bentonite buffer using a copper heat dissipation plate suggested in a previously computational simulation [12] was well verified using a small EBS module, in which a heat dissipation copper plate was applied.
4. Conclusions
In this study, a small-scale module demonstration was performed to verify the thermal transfer efficiency of copper plates as a way to reduce the maximum temperature of the bentonite buffer in contact with hot disposal canister at HLW disposal repository. The demonstration was performed using three cylindrical buffer blocks with a thickness of 2 cm and one copper plate with a thickness of 0.3 mm. The heat dissipation test was performed in the conditions of no copper plate, flat copper plate, and U-collar copper plate. By verifying the heat transfer effect of copper plates in a small module, it was found that for the flat copper plate, the maximum temperature reduction of the bentonite buffer under 250 W power supply was only about 1.8°C. However, for the U-collar copper plate, the maximum temperature reduction reached 9.1°C. Therefore, it was found that the contact of copper plate with other barriers are very important for the heat dissipation. In conclusion, it was found that the application of copper plates can be a very effective way to increase the density of HLW disposal while keeping the maximum temperature limit of the HLW disposal repository.