2.1 Beishan Underground Research Laboratory (URL)

J. Wang, L. Chen, R. Su, X. G. Zhao

Long-term safe disposal of high-level radioactive waste (HLW) is a challenging task for countries with nuclear power production. Due to its high radioactivity and long half-life, the HLW needs to be permanently isolated from the human environment. Geological disposal has been widely accepted as a feasible approach for safe management of HLW. Deep geological repositories (DGRs) can be built in suitable geological formations at a depth of several hundred meters below ground level. Various rock types such as granite, clay, tuff, and rock salt have been considered as potential rocks for hosting DGRs (Ahn, 2010). However, due to lack of actual implementation experience, the development of DGRs is a long-term and systematic research process. To develop DGRs successfully, many countries constructed underground research laboratories (URLs) to evaluate site suitability, develop and verify disposal concepts, optimize repository system components, and demonstrate long-term safety of DGRs (Wang et al., 2018).

2.1.1 The Strategy for China’s URL

With the rapid development of nuclear industry, China has an urgent and strategic need for safe disposal of HLW. In accordance with the Guidelines on Research and Development (R &D) Planning for Geological Disposal of High-level Radioactive Waste, which was issued by the China Atomic Energy Authority (CAEA), the Ministry of Science and Technology (MOST) and the former Ministry of Environment Protection (MOEP) in 2006, the China’s strategy for HLW disposal is divided into three phases: (1) laboratory-based research and site selection for the disposal facility (2006–2020), (2) underground research and testing (2021–2040), and (3) the disposal facility construction (2041–2050) (Wang, 2010; CAEA, 2006). A milestone is to complete the construction of China’s URL by 2020. Moreover, the 13th Five-year Plan of China (2016–2020) stated that “the construction of China’s URL for HLW disposal should start before 2020”, indicating that the Chinese government attaches great importance to the development of the URL.

Based on an analysis on the development process of URLs worldwide in combination with considerations of current situation in China, a newly defined type of URL (i.e., area-specific URL) was proposed (Wang, 2014). An area-specific URL is located at a site in an area that is considered as a potential area for a DGR. When an area has been selected as the first priority area for a DGR, but a specific site has not been determined, an area-specific URL can be built as long as the site has similar geological conditions to those of a future in-depth DGR site. The area-specific URL has a potential role, i.e., if investigations performed in the URL confirm that the site is suitable for a DGR, the siting process of the DGR will be accelerated. In addition, the acquired data, investigation results, and experience gained from an area-specific URL can be transferable to the future DGR in this area. Therefore, it is also called “Generation 3” URL. With this basic understanding, the major considerations of the China’s URL strategy are as follows (Wang et al., 2018):

  1. 1.

    pgTo build an area-specific URL in a representative granite formation within the area that has been identified as having the greatest potential for a DGR in China.

  2. 2.

    The URL will be a large-scale facility with full functionality.

  3. 3.

    The URL will be about 500 m deep, similar to the depth of the future repository.

  4. 4.

    The URL should be expandable.

  5. 5.

    The URL will serve for technology development and demonstration, site characterization, and public acceptance.

  6. 6.

    The URL will be open to international cooperation in the field of geological disposal.

2.1.2 Site Selection and Site Characterization for China’s URL

2.1.2.1 Site Selection

Site selection for China’s DGR started in 1985 (Wang et al., 2018). Considerable effort has been devoted to the selection of potential sites in granite formations. The attention on granite formations was primarily driven by the widespread occurrence of such rocks in China, coupled with the fact that granitic rocks are suitable for hosting DGRs. Over the past 30 years, significant progress has been made in site selection of the DGR. Six regions were selected as potential regions for the DGR (Wang et al., 2018). Based on a preliminary comparison of the six pre-selected regions, the focus was on site selection and site characterization in northwestern China, Inner Mongolia, and Xinjiang regions, as shown in Fig. 2.1. In 2011, the CAEA and the former MOEP jointly organized an expert meeting. In this meeting, the Beishan area of northwestern China was recommended as the first priority area for China’s DGR (Wang et al., 2018). This decision provided an important basis for selecting the URL site.

Fig. 2.1
A study area map of the U R L site from China with the locations of Xinjiang Uygur and Inner Mongolia Autonomous Regions. It includes Aqishan, Yamansu, Tianhu, Suanjingzi, Xinchang, Jiujing, Shazaoyuan, Tamusu, and Nuorigong.

Location of candidate URL sites in China (Wang et al., 2018)

Based on the achievements obtained from site selection of the DGR, site selection for China’s URL started in 2015. Nine candidate URL sites (Fig. 2.1) were selected for further comparison and demonstration. In the Beishan area, there are four candidate sites. In the Xinjiang and the Inner Mongolia pre-selected regions, there exist three and two candidate sites, respectively. To determine the final site of the URL, comprehensive studies based on surface investigations, borehole drilling, and borehole testing were carried out in parallel at these nine sites. According to the site characterization results and the proposed siting criteria of the URL (Wang et al., 2018), a comparison was conducted among the nine sites. In 2016, a review meeting of senior experts for the URL site recommendation was organized by the CAEA. The result was that the Xinchang site in the Beishan pre-selected region was determined as China’s URL site (Wang et al., 2018). Hence, we call China’s first URL for HLW disposal as Beishan URL.

2.1.2.2 Site Characterization

The Beishan URL site is in the middle of the Xinchang granite intrusion (Fig. 2.2), which is situated in the middle of the Beishan area and has a length of 22 km and a width of 7 km. The topography of the URL site is characterized by a relatively flat landscape with small hills in the Gobi Desert (Fig. 2.3).

Fig. 2.2
A geological map of the U R L site in the Xinchang region. The regions are highlighted for sandy gravel, argillite, arkose, sandstone, phyllite, migmatite, quartz diorite, granite, granodiorite, and biotite schist, along with faults. The scale measures from 1000 meters to 4 kilometers.

Geological map of the Xinchang granite intrusion (Wang et al., 2018)

Fig. 2.3
An aerial view of the landform of the Beishan site in China for disposal of the radioactive wastes.

Topography of the Beishan URL site

To understand geological, hydrogeological, and engineering conditions of the Beishan URL site, a series of investigations, such as geophysical surveying, borehole drilling, and hydraulic tests, have been conducted since 2015 (Wang et al., 2018). More than 30 boreholes with various depths were drilled within or around the site, and the borehole locations are shown in Fig. 2.4. In the site characterization phase of the URL, 600-m-deep vertical boreholes, i.e., boreholes BS06, BS28, BS32, and BS33, were used for the exploration of geological conditions of the URL site, especially for the evaluation of rock quality or integrity. In addition, 600-m-deep and 100-m-deep inclined boreholes were used for investigating the features of the faults around the site. For example, boreholes BS38 and BSQ08 were drilled for investigating the northern section of fault F33, while boreholes BS39 and BSQ10 were used for investigating the southern section of this fault. Moreover, in the design phase of the URL, engineering exploration boreholes, i.e., boreholes numbered from ZK01 to ZK07, with a depth ranging from 400 to 600 m were drilled for further investigating the engineering geological conditions of the rock mass around the designed shafts and ramp of the URL.

Fig. 2.4
A study area map of the Beishan U R L region with locations of 1000, 600, and 100 meters of boreholes. It also includes the locations of the fault, U R L site, and engineering exploration borehole.

Borehole locations at the Beishan URL site, modified from Wang et al. (2018)

Geology

The geophysical survey results show that the granitic rocks at the URL site have a depth greater than 2 km. The rock types mainly include monzonitic granite and granodiorite (see Fig. 2.2). Meanwhile, for the drill cores of vertical boreholes, the RQD values larger than 90% accounts for 86.2% of the total drill cores (Wang et al., 2018), indicating very good integrity of rock mass at the site. In addition, the rock cores extracted from seven engineering exploration boreholes drilled in the preliminary design phase of the URL also show the extremely high integrity, as presented in Fig. 2.5. According to the investigation results, a 3D geological model of the URL site before excavation was established (Fig. 2.6) and used for site description and the optimization of the URL design. When more data are available during the URL construction, the geological model will be updated continuously.

Fig. 2.5
A photograph of the extraction of the drill pipe at the U R L site.

Typical drilled cores extracted from the engineering exploration borehole ZK01

Fig. 2.6
A 3 D geological model of the selected Beishan U R L site with the markings for excavation. It also includes regions of granite and faults. The scale measures from 0 to 2000 meters.

A 3D geological model of the Beishan URL site before excavation (Wang et al., 2018)

Hydrogeology

The hydrogeological investigation results show that for all boreholes at the URL site, the water levels vary from 4 to 48 m below ground level. To evaluate the hydraulic properties of intact rocks and fracture zones, injection tests were conducted in vertical and inclined deep boreholes using a developed double packer equipment. The hydraulic conductivity (K) values in the test intervals at different depths were measured. As presented in Fig. 2.7, at a shallow depth less than 50 m, the K values in several test intervals in boreholes BS32, BS35, and BS38 are larger than \(10^{-7}\) m/s. With increasing depth, the K values start to decrease and are not sensitive to depth. For most intervals located in intact rock formations and fracture zones, the K values are less than \(10^{-8}\) m/s and concentrate between \(10^{-12}\) and \(10^{-10}\) m/s (Wang et al., 2018), showing very low permeability of rock mass at the URL site. In addition, only few water-bearing fractured zones with K values ranging between \(10^{-7}\) and \(10^{-6}\) m/s were found in borehole BS33 at depths greater than 300 m.

Fig. 2.7
A simulation model plots depth versus hydraulic conductivity. The y axis ranges from 0 to 600, and the x axis ranges from 1 E 5 to 1 E 14. The plots for B S 06, B S 28, B S 32, B S 33, B S 35, B S 36, B S 37, B S 38, and B S 39 are plotted throughout the graph.

Hydraulic conductivities of the rock masses at the Beishan URL site, modified from (Wang et al., 2018)

Rock mechanics

Laboratory mechanical tests on rock samples, which were collected from six engineering exploration boreholes at the URL site, were carried out. The experimental results indicate that the uniaxial compressive strength (UCS) of the saturated rock samples varies from 108 to 225 MPa with an average value of 162 MPa. To provide far-field stress boundary conditions for optimization of the URL design and stability evaluation of surrounding rocks, a total of 102 hydro-fracturing in situ stress measurements were performed in five boreholes (Fig. 2.4) within the URL site at depths ranging between 60 and 600 m below the ground surface (Wang et al., 2018). The measurement results show that the in situ stresses present an increasing tend with increasing depth, as presented in Fig. 2.8. The magnitudes of the maximum horizontal principal stress (\(\sigma _H\)) are all less than 25 MPa, which is at a low level compared with UCS of the core samples. Meanwhile, the measured \(\sigma _H\) is generally larger than the estimated vertical stress (\(\sigma _v\)), indicating that the in situ stress field at the Beishan underground research laboratory is dominantly controlled by the horizontal tectonic stress. In addition, fracture impression measurements show that the dominant orientation of \(\sigma _H\) is NEE direction with an average value of N55\(^{\circ }\)E, which is in basically agreement with the orientation of crustal velocity vectors of the Beishan underground research laboratory (Zhao et al., 2013).

Fig. 2.8
A scatter plot of depth versus principal stress. The y axis ranges from 0 to 700, and the x axis ranges from 0 to 35. 3 plots for B S, sigma uppercase H, and sigma lowercase h are distributed uniformly on either side of the three fit lines, decreasing from 0 to 700 meters.

Variation of the horizontal stresses with depth at the Beishan URL site, modified from (Wang et al., 2018)

2.1.3 The Beishan URL Construction Project

According to site characterization results of the Beishan URL site, it is concluded that the Xinchang site is very suitable for hosting China’s first URL. On May 6, 2019, the CAEA approved the “Beishan URL Construction” project. The objectives of the project are to construct a world-class URL for HLW disposal at the Xinchang site, to fill the gap in the field research and development platform for China’s HLW disposal, to significantly improve China’s abilities in HLW disposal technology development, and to meet China’s urgent needs for R &D of HLW disposal technologies. The Beishan URL is estimated to cost over CNY 2.72 billion, and the construction period is from 2021 to 2027. A significant milestone is that on June 17, 2021, China kicked off the construction of the Beishan URL, and a ground-breaking ceremony was held at the URL site, marking that China’s efforts on HLW disposal have entered a new phase, i.e., the URL development stage. The Beishan URL will provide an important scientific research platform for the future construction of a DGR, to speed up the process of safe disposal of HLW, and ensure the sustainable development of China’s nuclear industry.

2.1.4 The Beishan URL Design and In Situ Test Plan During Its Construction

2.1.4.1 Design of the Beishan URL

Figure 2.9 presents the design of Beishan URL, which is characterized by the layout of three shafts, one spiral ramp and two experimental levels with a maximum depth of 560 m (Wang et al., 2018). The three shafts include one personnel shaft and two ventilation shafts. The conventional drill-and-blast method will be used for excavation of the 6 m diameter personnel shaft, while raise boring machine will be used for excavation of the two 3 m diameter ventilation shafts. The spiral ramp has a length of about 7 km, a cross-sectional diameter of 7 m, and a maximum curve diameter of about 255 m. The ramp will be excavated using the full-face tunnel boring machine (TBM), with the aim to minimize the damage to surrounding rocks. The ramp of Beishan URL is currently the world’s first spiral ramp to be excavated by the TBM. The experimental tunnels will be constructed at two levels, i.e., –280 m and –560 m levels. The ramp and the three shafts are connected to the two experimental levels.

Fig. 2.9
A 3 D design of the Beishan U R L region. It includes the layout of 3 shafts for personnel and ventilation, and T B M, spiral ramp.

A 3D-perspective view of the Beishan URL

2.1.4.2 In Situ Test Plan During Beishan URL Construction

A comprehensive in situ test plan focusing on site characterization and technology development will be performed during URL construction, including geological mapping, geophysical surveying, hydrogeological investigations, rock suitability evaluation, TBM penetration test, in situ stress measurements and excavation damage zone (EDZ) characterization, etc. The test locations along the ramp and at the -560 m level are presented in Figs. 2.10 and 2.11, respectively. In parallel to the above activities, equipment to be used in the URL operation stage will be developed in surface laboratories and then tested in the two experiment levels. This includes excavation equipment for the deposition hole, installation equipment for the buffer material, and equipment for radionuclide migration testing.

Fig. 2.10
A schematic diagram exhibits the construction of the spiral ramp in the Beishan U R L region. It includes the personnel and ventilation shafts with the installation of the auxiliary and man experimental, along with the blasting vibration monitoring, E D Z, T B M, and water inflow monitoring.

Tests to be performed along excavation of the spiral ramp of Beishan URL

Fig. 2.11
A diagram presents the tests performed during the construction of the Beishan U R L site. It features the auxiliary test area of excavation technology, main testing area of excavation technology, disposal process test area, D and B excavation tunnel, sensor installation, and T B M excavation tunnel.

Tests to be performed at –560 m level during construction of Beishan URL

2.1.5 The Progress of Beishan URL

The excavation of the personnel shaft and the ramp of Beishan URL started in June 2022. By the end of December 2022, the personnel shaft has been excavated to a depth of 160 m, while the ramp has been excavated to a length of 495 m by the drill-and-blast method to provide space for assembly and trial operation of the TBM. The quality of surrounding rocks is very good (Fig. 2.12), indicating that the good integrity of rock mass at the URL site is preliminarily verified by the excavation of the shaft and the ramp.

Fig. 2.12
A photograph of the excavation of the longest road tunnel construction.

Very good quality of the rock mass after drill-and-blast excavation of the ramp

As mentioned in the aforementioned section, the ramp of Beishan URL is currently the world’s first spiral ramp excavated by the TBM. In the past few years, the technical feasibility of TBM for tunneling the spiral ramp of Beishan URL has been demonstrated, and the study results show that TBM method is feasible for the Beishan URL project (Ma et al., 2020). In September 2022, the specially designed TBM “Beishan No.1” for the spiral ramp excavation was successfully manufactured at China Railway Construction Heavy Industry Co. LTD in Changsha, China. In November 2022, the TBM was transported to Beishan URL and assembled on site (Fig. 2.13). On December 30, 2022, the TBM tunneling started (Fig. 2.14). Geological and hydrogeological investigations are being conducted following the excavation process of the URL. Currently, the construction of Beishan URL is in smooth progress and coordinated well with in situ tests. It is expected that the achievements to be obtained from Beishan URL will successfully contribute to the development of China’s DGR and similar facilities worldwide.

Fig. 2.13
A photograph of the construction site of the Beishan U R L with the assembly of the T B M to tunnel the spiral ramp.

The assembled TBM “Beishan No. 1” at Beishan URL

Fig. 2.14
A photograph of the assembly of the T B M tunneling under the construction of the spiral ramp.

The TBM “Beishan No. 1” in the ramp

2.2 Chinese Mock-Up Test on GMZ

Y. M. Liu, S. F. Cao, J. L. Xie, L. K. Ma

The shaft-tunnel model with a multi-barrier system located in saturated granite areas is the preliminary concept of the high-level radioactive waste (HLW) repository in China (Wang, 2010). An engineered barrier system (EBS) with buffer materials surrounding the canisters is the main part of the multi-barrier system. The Gaomiaozi (GMZ) bentonite is considered as the potential EBS buffer material, and the Beishan site as the most potential site for HLW disposal project in China (Liu and Wen, 2003).

Figure 2.15 shows the large-scale thermal-hydro-mechanical-chemical (THMC) China-Mock-Up facility. It was designed as a vertical cylindrical tank with an inner diameter of 900 mm and a height of 2200 mm, filled with compacted GMZ-bentonite. The THMC China-Mock-Up facility was built in 2010 and has been in continuous operation at the laboratory of the Beijing Research Institute of Uranium Geology (BRIUG) (Liu et al., 2013). This facility can be used to evaluate key THMC processes in the compacted GMZ-bentonite blocks during the early phase of HLW disposal system, providing a reliable database for numerical modeling and further investigations.

Fig. 2.15
A photograph of the experimental arrangement of the T H M C. It is equipped with a cylindrical tank, rigid cell, water inlet, sample, pressure sensor, porous stone, and force transducer.

External view of THMC China-Mock-Up facility

2.2.1 Experiment Materials

2.2.1.1 Main Properties of GMZ Bentonite

The GMZ bentonite was extracted from the Inner Mongolia autonomous region in northern China, 300 km northwest of Beijing. In this area, the reserves of bentonite are about 160 \(\times 10^6\) tons, and the proven reserves of Na-bentonite are about 120 million tons and the mining area is about 72 km\(^2\). The major bentonite clay layer of the deposit extends about 8,150 m with a thickness of 8.78–20.47 m. This deposit was formed in the later Jurassic period (Liu et al., 2007).

The GMZ01 bentonite was mined from the Jiucaigou tunnel located in the east of the GMZ-bentonite deposit, and the GMZ02 bentonite from the middle area which was near the ZK2401 section of GMZ-bentonite deposit (Xie et al., 2018). The natural sodium GMZ bentonite was air dried and its moisture content after drying was about 10%.The naturally dried bentonite was crushed to 200 meshes by a Raymond mill, and some impurities with high hardness and density were removed during this process.

The previous studies have shown that the GMZ01 bentonite has ideal thermal, hydraulic, mechanical, and physico-chemical properties as a buffer material. The main thermal-hydraulic-mechanical (THM) properties of GMZ01 bentonite are summarized below.

The X-diffraction analysis showed that the clay mineralogy is dominated by montmorillonite (75±2%), which is the essential mineral for sealing performance. The GMZ01bentonite also contained varying amounts of quartz (12±1%), cristobalite (7±1%), feldspar (4±1%), calcite and kaolinite (1±1%). The high content of montmorillonite resulted in a high cation exchange capacity (CEC = 77.30 meq/100 g), a large plasticity index (\(I_p\) = 275), and a large specific surface area (\(a_s\) = 570 m\(^2\)/g) (Liu and Wen, 2003). The major exchangeable cations were Na\(^+\) (43.6 meq/100g), Ca\(^{2+}\) (29.1 meq/100g), and Mg\(^{2+}\) (12.3 meq/100g).

The existing studies have suggested that the thermal conductivity of GMZ01 bentonite increases with the increase in dry density and moisture content. With a dry density of 2000 kg/m\(^3\) and a moisture content of 7%, the thermal conductivity of the GMZ01 bentonite is about 1.12 W/mk (Liu and Cai, 2007). Regarding the hydraulic properties, Chen et al. (2006) obtained the water retention curve at room temperature under confined and unconfined conditions. It was found that the coefficient of unsaturated hydraulic conductivity of GMZ01 bentonite for the initial dry density of 1700 kg/m\(^3\) varies between 1.13\(\times 10^{-13}\) m/s and 8.41\(\times 10^{-15}\) m/s during the saturation process (Ye et al., 2009a). Ye et al. (2009b) also investigated the influence of temperature on the water retention capacity of highly compacted GMZ01 bentonite. The results showed that the water retention capacity of the highly compacted GMZ bentonite decreases as the temperature increases.

Table 2.1 Chemical compositions of GMZ-Na-bentonite

The research on mechanical behaviors of GMZ01 bentonite has also been conducted. Ye et al. (2007) performed swelling pressure tests on compacted GMZ bentonite with four dry densities (1.15 g/cm\(^3\), 1.30 kg/cm\(^3\), 1.50 g/cm\(^3\), and 1.75 g/cm\(^3\)) through the constant-volume method, and measured the swelling pressure (about 4.3 MPa) when the GMZ01 bentonite with a dry density of 1.75g/cm\(^3\) (Ye et al., 2007). Based on the experimental data, the regression curve of the relationship between the swelling pressure (\(P_s\), kPa) and dry density (\(\rho _d\), g/cm\(^3\)) of the GMZ01 bentonite at laboratory temperature was derived: \(ln P_s\) = 5.151 \(\rho _d\)–0.618. In view of the limited available data, the further validation of the equation is still necessary.

Cui et al. (2011) systematically investigated the THM behavior of compacted GMZ01 bentonite. In his study, the thermal-mechanical behavior of compacted GMZ bentonite with a dry density of 1700 kg/m\(^3\) was studied using a temperature-suction-controlled isotropic cell, and the significant effect of suction on the compressibility parameters of GMZ bentonite was verified. Based on laboratory experiments, a coupled THM model of GMZ bentonite was proposed to reproduce the main physical-mechanical behavior of GMZ bentonite. However, further validation of the model by the mock-up test and in situ test is still required.

2.2.1.2 Preparation of Compacted Bentonite Blocks and Pellets

The bentonite used for the China-Mock-Up was excavated from Jiucaigou tunnel located in the east of the GMZ-bentonite deposit. The naturally dried bentonite was crushed to 200 meshes (GMZ01, particle size less than 0.07 mm) and 80 meshes (GMZ06, particle size less than 0.18 mm) by a Raymond mill, and some impurities with high hardness and density were removed in this process. Tables 2.1 and 2.2 show the chemical compositions and mineralogical compositions of GMZ01 and GMZ06.

Table 2.2 Mineral compositions of GMZ-Na-bentonite
Fig. 2.16
A photograph of the bentonite blocks in different sizes and shapes. They are in semicircle, rectangular bar, and wedge shapes.

Compacted bentonite blocks

Fig. 2.17
A photograph of the bentonite pellets with different grain sizes, and the scale is placed above the pellets.

Crushed pellets of compacted bentonite blocks

The GMZ-bentonite powder with an average moisture content of 8.7% was compacted into high-density blocks by a computer-controlled triaxial testing machine combined with the self-designed steel molds. Totally, five types of compacted blocks were used in this test, as shown in Fig. 2.16. The GMZ06 was used to compact fan-shaped bentonite blocks, and the GMZ01 was used to compact hemicycle and rectangular bentonite blocks. The dry density of fan and hemicycle-shaped bentonite blocks was 1710 kg/m3, and that of rectangular bentonite blocks was 1930 kg/cm3. The rectangular bentonite blocks were subsequently crushed into small pellets with different grain sizes (see Fig. 2.17) to fill the space between bentonite blocks and the steel tank walls. The average dry density of bentonite pellets was 1300 kg/m3. The total mass of bentonite used in the experiment was about 2058 kg. Once the blocks have swollen and filled all the construction gaps, the dry density of bentonite decreased to an average value of about 1600 kg/cm3.

2.2.2 Structure of China-Mock-Up Facility

The China-Mock-Up facility was constructed as a vertical cylindrical tank. The main components of the testing system were as follows (see Fig. 2.18): a steel tank, a central electrical heater with a temperature control system, a hydration system, an engineered barrier, sensors, a gas measurement and collection system, and a Data Acquisition System (DAS). The steel tank was used to simulate the vertical disposal pit with an inner diameter of 900 mm and a height of 2200 mm. A layer of geotextile was placed between the tank and bentonite. The gaps between the tank and the bentonite blocks were filled with crushed pellets.

The heater in the China-Mock-Up was made of 1-ton carbon steel with a diameter of 300 mm and a height of 1600 mm. It only simulated the thermal emission, dimensions, and weight of the reference canister. The weight pressure of the heater was 1.24 MPa, and the temperature of the heater was automatically controlled by a temperature-monitoring system.

The hydration system was employed to simulate the water penetration from the host rock. In the hydration system, the water supply to the barrier was realized by four vertical tubes (\(\varnothing \) 10 mm) installed on the interior boundary of the cylindrical tank. As presented in Fig. 2.18, the water tank was connected to the injection tubes and an argon tank in the experiment. By controlling the gas pressure in the argon tank, the water was injected from the bottom of the mock-up facility. Simultaneously, the injection rate was measured by a mass flowmeter, and the water sample tank was weighted constantly by a mass balance to quantify the injected water mass. Table 2.3 shows the main chemical compositions of water from borehole BS05 of the Beishan site used in the hydration system.

The engineered barrier was composed of compacted GMZ-bentonite blocks and pellets surrounding the heater. The blocks were arranged in 44 sections: 32 sections were set in two concentric rings around the heater and the left 12 sections in two rings and a core. Bentonite pellets and bentonite powder were filled in the installation space between the heater and the bentonite block, or steel tank.

More than 160 sensors, including the temperature, relative humidity (RH), stress, and displacement, were installed inside and outside the facility to monitor the evolution of bentonite and heater. The inside sensors were distributed in bentonite blocks and pellets in 7 sections (sections I–VII) vertically. Different types of sensors were placed within each section to investigate the temperature change, hydration process, and the behavior of buffer materials under complex coupling conditions. Table 2.4 lists the sensors installed in the mock-up test and associated parameters. The sensors were placed within the grooves cut in the compacted bentonite blocks or in the surrounding pellets.

Fig. 2.18
A schematic of the cross section of the mock up facility setup. It is equipped with a temperature control system, gas measurement and collection system, data acquisition system, displacement sensors, bentonite blocks, heater, steel tank, bentonite pellets, hydration system, and R H and T sensors.

General layout of the china-mock-up facility

Fig. 2.19
A graph plots temperature and water injection versus time. The hydration line begins at (0, 0), remains stable up to (190, 0), follows an increasing trend, and ends at (1490, 350). The heating line begins at (0, 23), follows an increasing trend with some fluctuations, and ends at (1490, 92).

Temperature and water consumption with time

Table 2.3 Chemical compositions of underground water from Beishan (unit: mg/L)
Table 2.4 Summary of the sensors installed in the mock-up test

2.2.3 Test Procedures

The China-Mock-Up facility was assembled completely on September 10, 2010. The data acquisition and monitoring system automatically recorded all the measurement data every 10–30 min from January 1, 2011. In the data recording, January 1, 2011 was identified as Day 0 on the time scale after pre-operation.

On January 19, 2011, the heater was switched on and the temperature of the heater reached 30\(^{\circ }\)C and remained constant. From Day 188 to Day 400, the temperature raised to 90\(^{\circ }\)C gradually at a speed of 1\(^{\circ }\)C/d. Finally, the constant temperature control mode was automatically activated by the computer, maintaining a temperature of 90\(^{\circ }\)C. Figure 2.19 shows the heating phase with time.

The real THMC experiment with the water injection began on July 8, 2011 (Day 188). The water injection rate was gradually increased from 400 g/day to 1500 g/day, and the water injection rate was controlled artificially to avoid potential damage to the sensors by the rapid saturation in the first stage. From August 25, 2013 (Day 967), the water injection pressure was gradually controlled from 0.2 MPa to 1 MPa. Figure 2.19 illustrates the water consumption with the time. To be mentioned here, there were no water supplies artificially sometimes due to holidays and some maintenance.

2.2.4 Analysis of Experimental Data

The sensors placed in the bentonite provided reasonable and consistent outputs. In this report, the experimental results recorded from January 1, 2011 to December 31, 2015 are analyzed, including the temperature change, relative humidity (RH), the stress in the bentonite, and heater displacement.

2.2.4.1 Temperature Evolution

There are four temperature sensors installed in each measurement profile. The measurement range of temperature sensor is from 0 to 300\(^{\circ }\)C, with an accuracy of 0.1 \(^{\circ }\)C. Figure 2.20 presents the temperature evolution in bentonite in sections II, III, V, and VI. It can be noticed that the temperature continuously increases with time at the first stage of temperature increase (30 \(^{\circ }\)C– 90 \(^{\circ }\)C). The trends of the temperature change also vary with the four seasons of environmental conditions. Moreover, the distribution of temperature is non-uniform vertically, and the temperature is much higher in the central part. Even in section V, the temperature is still below 70\(^{\circ }\)C. In the experiment, the lower temperature can be partly attributed to the existence of installation space between the heater and compacted GMZ blocks, which is 5cm in width and filled with pellets. This installation method may reduce the thermal conductivity of the barrier in the area. In addition, the effective heating length in the center of the heater with a height of 1.6 m is only 1.2 m.

Fig. 2.20
4 graphs plot temperature versus time for sections 2 to 6 with the respective layout and top view of the China-mock-up facility model. 4 temperature lines follow a zig-zag pattern between 0 and 80, with some fluctuations.

Temperature evolution in sections II, III, V, and VI

This is an important factor causing the higher temperature in the central part of the barrier. Besides, the temperature distribution is also influenced by a complex coupling mechanism. Considering that the saturation changes thermal conductivity, the temperature distribution also depends on the saturation process in the compacted bentonite. Due to the interruptions of the electric power supply, some fluctuations of temperature are also recorded.

Figure 2.21 shows the temperature distribution in the China-Mock-Up facility on July 8, 2012.

Fig. 2.21
A contour plot of the mock-up facility with the distribution of the temperature. The center region of the tank has the highest distribution, and the corner region has the lowest distribution. The temperature ranges from 30 to 90 degrees Celsius.

Temperature distribution in the China-Mock-Up on July 8, 2012

2.2.4.2 Relative Humidity (RH)

A total of 24 RH sensors are placed into holes drilled into the bentonite blocks at seven sections (sections I–VII) vertically. The RH and temperature in the bentonite can be measured by the RH sensors. Figure 2.22 shows the RH and temperature evolution in bentonite at sections II, III, V, and VII.

As illustrated in the figures, the compacted bentonite is progressively saturated from section I to section VII, and the distance to the heater has a significant influence on the saturation velocity. Due to the temperature fluctuation, there are also some fluctuations in the RH evolution. The heating process generally leads to the increase of RH, which is probably related to the generation of the vapor phase.

Fig. 2.22
4 graphs plot relative humidity and temperature versus time for sections 2, 3, 5, and 7 with the respective layout and top view of the mock-up facility model. Several lines for R H and temperature lines follow a zig-zag pattern in an increasing manner between 0 and 120, with some fluctuations.

RH and temperature evolution in sections II, III, V, and VII

The RH evolution in sections III and V are much more complex. With the increase of temperature, the bentonite near the heater has a decrease in RH due to the drying effect of the heater, particularly in the inner rings (H10, H16, and H18). In the inner part, the following stages can be noticed: (a) the stable RH stage: there is a stable RH stage with some fluctuations; (b) the decreased RH stage: when the temperature is kept at 90\(^{\circ }\)C, continuous heat transfer from the heater leads to the drying effect, and then RH decreases; (c) the increased RH stage: with the increase of water injection rate, hydration gradually overcomes the drying effect and then the RH increases. This wetting tendency is tightly related to the accelerated saturation process caused by the increased water injection rate. On the contrary, the desiccation is not noticed (H7 and H15) in the outer part where the drying effect is insignificant. This findings are consistent with the previous results obtained by Villar et al. (2012).

Due to the non-uniform water supply in the vertical direction, the saturation process in section VII is less significant than that in section II at the bottom. The desiccation phenomenon was also observed from the sensor H22 located in the inner ring. In addition, the fluctuation of RH induced by the heating interruptions is particularly evident in section VII. It indicates that the generated vapor phase moves in both radial and longitudinal directions.

In conclusion, the RH variation in the inner rings is the result of the drying effect of the heater and the water penetration. The experimental results indicate that, due to the low permeability of the compacted bentonite, the drying effect was dominant in these sections at the beginning of the test. This finding is consistent with the results reported in other research works (Villar et al., 2012).

2.2.4.3 Stress Evolution

The swelling stress was measured by stress sensors installed in two or three different directions, i.e., x-, y-, and z-directions in each measurement profiles. Moreover, there were another three stress sensors directly contacting the inner top, bottom, and side walls of the steel tank, respectively. The CYG-1712 stress sensors were used, and its measurement principle was the Wheatstone bridge from SQSENSER (China). The stress sensor was capable of withstanding temperatures up to +100 \(^{\circ }\)C with a stress measurement ranging from 0 to 20 MPa, as shown in Fig. 2.23. The sensors were placed within the grooves cut in the compacted bentonite blocks in three directions, including vertical, radial, and hoop directions.

Fig. 2.23
Left. A photograph of the sensor attached to the groove cut and the measurement of 30-centimeter is highlighted on the scale. Right. A schematic diagram of the bentonite with the location of hoop, radial, and vertical sensors.

Stress sensor and its location in the compacted bentonite

(1) Total pressure evolution

Under the THMC-coupled condition, the stress evolution in the compacted bentonite can be influenced by several factors, including gravity, the thermal expansion induced by high temperature, and the swelling pressure generated by bentonite saturation with water penetration.

Fig. 2.24
4 graphs plot swelling pressure versus time for sections 2, 3, 5, and 7 with the respective layout and top view of the mock-up facility model. Several lines for swelling pressure lines follow an increasing trend with some fluctuations.

Stress evolution at sections II, III, VI, and VII of China-Mock-up

Figure 2.24 shows the stress evolution in bentonite at sections II, III, VI and VII. It can be seen that the variation of the total stress is rather limited or negligible in the initial stage, which is caused by the initial gaps between sensors and blocks, and bentonite blocks and pellets. The stress inside the bentonite increases from the bottom to the top and from the area near the water tube to the interior gradually. After more than 2 years of water injection, the stress in the inner part increases quickly, and becomes higher than the outer part gradually. This indicates the water seeps into the inner part.

The highest vertical stress of 3.8 MPa is recorded by sensor 36 in the middle of section VI located on the top of the heater after 1512 days. At the bottom of the heater, the highest vertical stress of 2.3 MPa is also recorded by sensor 15 in section II after 1512 days. The swelling pressure of GMZ bentonite tested by this facility is 3.8 MPa, while the conventional swelling pressure under the dry density of 1600 kg/m3 is 3.17MPa. This can be explained as follows: as the heater moves up by 7.57 mm, the gap between the bentonite blocks as well as the gap between the sensor and the blocks at the top of the heater is squeezed, indirectly leading to the increase of local dry density of bentonite. Besides, the thermal expansion of the heater may be another reason.

In the higher level of bentonite blocks above the heater, the highest vertical stress of 2.4 MPa is recorded by sensor 44 in section VII after 1403 days. This can be explained as follows: the saturation process of whole barrier has not finished yet; stress release is induced by the initial gaps between the bentonite blocks and pellets, and the gaps between the sensors and the blocks.

The highest hoop stress of 3.26 MPa is recorded by sensor 18 within the inner ring in section II located on the bottom of the heater after 1284 days. The highest radical stress of 2.73 MPa is recorded by sensor 17 within the inner ring in section II located on the bottom of the heater after 1365 days. The highest radical stress is lower than the vertical stress and the hoop stress.

(2) Stress distribution

According to the data of the stress sensor in three directions, the stress distribution on July 8, 2012 (1 year after operation) is obtained, as shown in Fig. 2.25. After the water injection, water is concentrated at the bottom due to the gravity effect, and the initial gaps between bentonite and steel tank, and bentonite and heater. Besides, the saturation process is vertically non-uniform which brings difficulties to the definition of boundary condition in numerical studies. Because of the low hydration rate and inhomogeneous saturation process in the vertical direction, the stress variation is relatively limited and inhomogeneous. The stress is higher at the bottom of the heater which may be related to the gravity of the heater. Since the mass of the heater is 1000 kg, the pressure under the heater by its weight is 1.24 MPa, and the radical stress and hoop stress around the heater are lower, indicating that water has not penetrated to the inner part of the bentonite.

The stress distribution on October 8, 2013 (2 years after the operation) is obtained, as shown in Fig. 2.25. The maximum stress in the area below the heater may be due to the gravity effect of the heater itself, as well as the full penetration of water in the bottom area of the facility, resulting in the maximum stress. The stress around the heater is lower, which may be due to the stress release caused by the initial gaps between the bentonite blocks and pellets with the heater and steel tank, and the gaps between the sensors and blocks. It also indicates that the water has not penetrated to the bentonite near the heater (the bentonite has very low permeability and a long time is required for water to seep into the inner parts).

The stress distribution on July 8, 2014 (more than 3 years after operation) is shown in Fig. 2.25. It indicates that bentonite at the bottom of the heater is nearly saturated. The stress is still higher at the bottom of the heater due to the gravity. The lower dry density of the bentonite pellets results in low radical stress and hoop stress near the steel tank and heater.

Fig. 2.25
3 sets of contour plots of the mock-up facility with the stress distribution for vertical, radial, and hoop stresses. The region near the heater has the low stress zone, and the bottom region has the high stress area. The scale ranges from 0 to 2.5 megapascals.

Stress distribution in the China-Mock-Up on July 8, 2012 (top), October 8, 2013 (mid), July 8, 2012 (bottom)

2.2.4.4 Displacement Evolution of Heater

According to the reference concept for HLW disposal in China, the canister is completely supported by the surrounding buffer material. Therefore, the mechanical performance of buffer material may affect the overall stability of the canister. To evaluate this potential influence, six LVDT sensors were installed on the top and bottom of the heater to monitor its vertical displacement.

As shown in Fig. 2.26, the Linear Variable Differential Transformer (LVDT) sensors HTGGA-20 from KING SENSOR (China) are used. The selected displacement sensor has a fairly small size, which can reduce the interference to the whole system as much as possible. The lengths of LVDT sensors at the bottom and top of the heater are 250 mm and 300 mm, respectively. The measuring range of the displacement sensor is ±20 mm and the accuracy is more than 0.1%. The diameter of the main part of the LVDT sensor is 22 mm. The LVDT sensors were specifically constructed of steel 316L with a rubber sleeve on top to improve its corrosion resistance and water tightness.

Fig. 2.26
Left. A photograph of the selected displacement sensor. Right. 2 schematic diagrams of the bentonite with the locations of L V D T 1 to 3 and L V D T 6 to 8.

Displacement sensor and its location in measuring sections

Figure 2.27 shows the vertical displacements measured by six LVDT sensors. The positive value recorded by the LVDT6 to LVDT8 indicates that the sensor is in the compression state and the heater moves upward. The negative value recorded by the LVDT1 to LVDT3 indicates that the sensor is in the tensile state and the heater moves upward. Because the electric components of the LVDT6–LVDT8 sensors were directly connected to the heater, coupled with the THMC harsh environment, LVDT6–LVDT8 failed one after another after water injection and the temperature of the heater gradually reached 90oC.

Four phases of movement evolution of the heater can be observed: (a) phase I: the heater moves downward and upward slowly with a maximum downward displacement of 0.06 mm and a maximum upward displacement of 0.29 mm in 188 days; (b) phase II: when the water injection begins, and the heater temperature increases to 90\(^{\circ }\)C in 346 days, the heater moves upward 2.6 mm quickly; (c) phase III: when the temperature is kept at 90\(^{\circ }\)C with continuous water injection, the heater moves upward continuously, while the movement rate gradually decreases to a stable displacement of 5.6 mm with some fluctuations in 730 days; and (d) phase IV: the heater moves upward slowly again with the progress of water injection. A maximum upward displacement of the heater (8.64 mm) is recorded by LVDT3 at day 1590. The six LVDT sensors obtained similar data in different directions from phase I to phase III. The difference between LVDT6 and LVDT7 maybe caused by the bentonite penetration into the gap between the steel plate screwing the LVDT6 to LVDT8 and the top lid of the steel tank. The difference among LVDT1, LVDT2, and LVDT3 maybe caused by the bentonite penetration into the gap between the steel plate screwing the LVDT1 to LVDT3 and the bottom of the heater. The penetration point of bentonite is probably near the LVDT6, LVDT1, and LVDT2, respectively.

The heater is completely supported by the compacted bentonite in China-Mock-Up. Therefore, the mechanical performance of compacted bentonite may affect the movement evolution of the heater. When the weight and load of the heater are 1 ton and 1.24 MPa, the compacted bentonite under the heater is consolidated and subject to volumetric/deviatoric creep. During the heating, the thermal expansion of the heater and bentonite occurs. The swelling of bentonite may be the main cause of the upward movement of the heater.

Fig. 2.27
A graph plots displacement versus time. 3 lines for L V D Ts 1, 2, and 3 begin at (0, 0) and follow a decreasing trend between negative 8 and 0. 3 lines for L V D Ts 6 to 8 begin at (0, 0) and follow an increasing trend between 0 and 7.

Displacement evolution of the heater

Note: The positive value indicates that the LVDT is in the compressed state, and the negative value indicates that the LVDT is in the tensile state.

(1) Effect of Temperature and RH on the Heater Displacement

The temperature of heater and the evolution of compacted bentonite recorded by different sensors indicate that the displacement of the heater is directly related to the evolution of temperature and RH in bentonite (Fig. 2.28). RH sensors (H2 and H4) are located in the bentonite blocks at sections I and II under the heater, respectively. The RH sensor of H20 is located in the bentonite block at section VI above the heater.

In phase I, when the temperature increases to 30\(^{\circ }\)C and remains constant, the heater moves upward slightly and remains stable with some fluctuations due to the thermal expansion on the heater and bentonite.

In phase II, when the temperature increases to 90\(^{\circ }\)C gradually, the heater moves upward quickly because of the continuous thermal expansion of heater and bentonite. The slow increase in RH can be observed by the RH sensor of H2. This suggests that the bentonite under the heater is wet and the swelling of the bentonite occurs. The decrease of RH can be observed by RH sensor of H20. It can be seen that heat transferred from the heater leads to the drying effect on the bentonite above the heater and the volume of the bentonite may be reduced.

In phase III, the temperature is kept at 90\(^{\circ }\)C, and the rapid increase of RH is observed by RH sensors of H2 and H4. It suggests that the bentonite under the heater is wet and swelling quickly. The displacement curves have the same trend with the RH curves of H2. As the bentonite under the heater becomes saturated, the heater gradually stops moving upward. The decrease of RH can be observed by the RH sensor of H20. It indicates that the bentonite above the heater is dried constantly. The upward movement of the heater at phase III is mainly caused by the swelling of bentonite under the heater.

In the phase IV, the rapid increase of RH can be observed by the RH sensor of H20 after 700 days, indicating that the bentonite above the heater is wet and swells quickly. Because the total mass of bentonite in the test tank remains unchanged and the volume of the bentonite above the heater swells, the bentonite above the heater pushes the bentonite around the heater to the bottom of the heater and induces the upward movement of the heater again.

Fig. 2.28
A graph plots displacement and temperature and relative humidity versus time. 7 lines for L V D Ts 1 to 3, T heater, H 2, H 4, and H 20 follow an increasing trend from 0 to 10 and from 0 to 100.

Displacement evolution of the heater with temperature and RH in the bentonite

(2) Effect of Stress in Bentonite on the Heater Displacement

Figure 2.29 shows the displacement evolution of the heater with the stress change in bentonite. Stress sensor S15 is placed under the heater in section II. Stress sensor S36 is placed above the heater at section VI. Stress sensor S32 is placed around the heater in hoop direction at section V. A possible reason is that the bentonite under heater might be consolidated and has volumetric/deviatoric creep under the load of 1.24 MPa caused by the self-weight of the heater.

In phase I, a maximum downward displacement of the heater (0.06 mm) is recorded at day 46. A possible reason is that the bentonite under heater might be consolidated and has volumetric/deviatoric creep under the load of 1.24 MPa caused by the self-weight of the heater.

In phase II, the stress recorded by S15 located in section II under the heater and the stress recorded by S36 located in section VI above the heater increase slightly with some fluctuations. The upward movement of the heater at phase II is partly caused by the bentonite swelling.

In phases III and IV, the stress recorded by S15 indicates that the stress in bentonite under the heater increases quickly from Day 700 to Day 1060. After that, the stress is kept at 2 MPa with a fluctuation. The stress recorded by S32 which is located in section V around the heater increases quickly from Day 646 to Day 1426. This is because the location of S32 is near the water tube than S15 and S36. Based on the stress and RH, it is deduced that the bentonite under the heater is saturated after Day 1060. The stress under the heater adjusts with the stress increase around the heater and the weight of heater. After Day 700, the heater moves upward slowly again with the saturation of bentonite under the heater and around the heater. The data fluctuation of LVDT1 and LVDT2 has the same trend as that of stress sensors S15 and S32.

Fig. 2.29
A graph plots displacement and stress versus time. 6 lines for L V D Ts 1 to 3, S 15, S 32, and S 36 follow an increasing trend from 0 to 10 millimeters and 0 to 4 megapascals with several fluctuations. The line for S 36 has the highest value.

Displacement evolution of the heater and stress in the bentonite blocks

The stress recorded by S36 increases during phase III, then fluctuates and increases quickly after 1200 days. This indicates that the heater moves upward and induces increased pressure in the bentonite above heater. The stress fluctuates with the heater movement. The stress above the heater exceeds the stress under the heater after 1296 days, then the stress above the heater pushes the heater to move downward. With the downward movement of the heater, the stress above the heater decreases. As the water penetrates to bentonite and the bentonite swells continuously, the stress in the bentonite increases again. The bentonite above the heater pushes the bentonite around the heater to the bottom of the heater and induces the upward movement of the heater again. The fluctuation of LVDT3 has the same trend as the stress sensor S36 after 1296 days. A maximum upward displacement of the heater (8.64 mm) is recorded after 1590 days. With the saturation of bentonite and stress redistribution in bentonite, the heater remains stable with a minor fluctuation. Therefore, the upward movement and fluctuation of the heater at phase IV are mainly caused by the bentonite swelling and stress redistribution.

After the bentonite in the disposal pit is fully saturated, the canisters may move downward due to the weight of canisters as phase I. However, with the in situ stress, the canisters probably remain stable with a minor fluctuation. More tests about the movement of canisters in the disposal pit should be conducted in the future.

2.2.5 Summary

As a large-scale mock-up facility, the China-Mock-Up was used to investigate the THMC behavior of GMZ bentonite based on a preliminary concept of the HLW repository in China. This facility was built in 2010 and has been operated continually in the laboratory of BRIUG. Through the China-Mock-Up test, a large number of data about the evolution of compacted bentonite and the suitability evaluation of buffer material under THMC-coupled conditions have been obtained for the first time. Based on the currently recorded results, several preliminary conclusions can be drawn as follows:

(1) The temperature within the bentonite increases and varies with time and environmental conditions. Considering that the saturation process may change the thermal conductivity, the temperature distribution is influenced by the coupling mechanism between the thermal conduction and the saturation process in the mock-up test.

(2) The saturation process of the compacted bentonite is strongly influenced by the drying effect of the heater and the wetting effect of the water penetration. The bentonite has a low permeability, and a long time is required for water to penetrate into the inner parts. The RH fluctuations are generated by a complex mechanism, including saturation process, vapor generation and drying effect. Prior to 1600 days, the relative humidity of all RH sensors in the bentonite is 100%.

(3) The stress evolution in the compacted bentonite is affected by the gravity of heater and water, the thermal expansion induced by high temperature, the swelling pressure of bentonite generated by water penetration, stress release induced by the initial gaps between the bentonite blocks and pellets, the gaps between the sensors and the blocks, and also the upward displacement of the heater. The highest vertical stress of 3.8 MPa is recorded in the middle of section VI located on the top of heater after 1512 days. The highest hoop stress of 3.26 MPa and radical stress of 2.73 MPa are recorded within the inner ring in section II located at the bottom of heater after 1284 days and 1365 days, separately.

(4) The movement evolution of heater is affected by the consolidation and volumetric/deviatoric creep under the load of 1.24 MPa caused by the self-weight of the heater, the thermal expansion of heater and bentonite, the swelling of bentonite under the heater, around the heater, and above the heater in turn. A maximum downward displacement (0.06 mm) and upward displacement (8.64 mm) of the heater are recorded on Days 46 and 1590. With the saturation and stress redistribution of bentonite, the heater remains stable with a minor fluctuation.

Based on the analysis of the existing experimental data, it is considered that the China-Mock-Up test is a valuable data source for improving the understanding of the THM process in EBS, and establishing the reliable numerical method for predicting the long-term THM-coupled behavior of EBS. With the progress of the experiment, the conclusions will be further examined and refined.