2022 Volume 63 Issue 12 Pages 1639-1644
Information on the water permeability of materials for radioactive waste disposal in geological repositories is essential. High strength and ultra-low permeability concrete (HSULPC) is being considered as a material used to package transuranic (TRU) waste. HSULPC would be reinforced by steel fiber to increase its tensile strength and toughness, but the effect of reinforcing steel fiber on permeability is unclear. Permeability tests need to be highly accurate to determine the hydraulic properties of low permeability materials. In this study, the hydraulic conductivity of HSULPC with steel fiber using the transient pulse method. The hydraulic conductivities of HSULPC with/without steel fiber were determined to be around 10−13 to 10−11 m/s under the confining pressures between 2 and 10 MPa and pore pressure of 1 MPa constant. The results further showed that the permeability of these materials had a hysteretic dependence on the effective confining pressure. There was no remarkable difference of hydraulic conductivity between HSULPC with and without steel fiber. SEM observation revealed that HSULPC with steel fiber has isolated voids to some extent included during mixing of concrete but there are few voids on the interface of steel fiber for water channel. However, due to the higher porosity, the hydraulic conductivity of HSULPC with steel fiber is relatively higher than that without fiber. Still, the permeability of fiber reinforced HSULPC is low enough compared with Horonobe mudstone and Toki granite and would enable it to effectively confine 14C radionuclides in TRU waste.
This Paper was Originally Published in Japanese in J. Soc. Mater. Sci., Japan 71 (2022) 228–234.
Fig. 7 Relationships between hydraulic conductivity and effective confining pressure for HSULPC without fiber,15,16) HSULPC with steel fiber, Horonobe mudstone, and Toki granite.17)
Nuclear energy development and utilization result in the generation of radioactive waste. High-level radioactive wastes are liquid waste and vitrified waste generated after reprocessing spent nuclear fuel, while other wastes are classified as low-level radioactive wastes. In Japan, these wastes are disposed of underground in various containers and depths depending on the type of waste. Geological disposal of radioactive wastes consists of multiple barriers: an engineered barrier, which is an artificial structure designed to prevent leakage, and a natural barrier, which is typically an underground rock mass designed to prevent radionuclide migration.1) As a result, the mechanical and transport properties of the barriers must be investigated. Sedimentary and crystalline rocks are common in Japan’s deep underground. Because of their low permeability, mudstone and granite are the most representative of the natural barriers. As a result, studies on radioactive waste processing and disposal have been conducted in Horonobe Town, Hokkaido (for mudstone), and Toki City, Gifu Prefecture (for granite) in Japan.2)
Various studies and developments have been conducted on both engineered and natural barriers. Transuranic (TRU) wastes are low-level radioactive wastes generated during the operation and decommissioning of reprocessing and MOX (uranium–plutonium mixed oxide) fuel fabrication facilities (long half-life low heat generating radioactive waste). Some of them have a high concentration of nuclides, such as 129I and 14C, which are poorly absorbed by the barriers. These wastes, such as high-level radioactive wastes, are expected to be buried deeper than 300 m underground, and alternative technologies are required for their safe disposal.1) High strength and ultra-low permeability concrete (HSULPC) is one of them, and it is being developed as an engineered barrier for the confinement of TRU wastes, including 14C.
Long-term confinement of radionuclides and prevention of their contact with groundwater is required in geological disposal of TRU wastes using HSULPC for the TRU waste package, in which the stainless-steel canister containing radioactive wastes is placed in a steel box3) (Fig. 1). The composition, curing method, and availability of fiber reinforcement are currently being considered to achieve this disposal technique using HSULPC. Steel fibers, in particular, are being considered as fiber reinforcement to improve tensile strength and toughness.
Schematic illustration of alternative concept using HSULPC to radioactive waste packaging.3)
Because HSULPC has low permeability, it has been reported that using conventional methods applied to concrete, such as the output method4,5) and the input method (depth-of-penetration method),6) is difficult and produces inconsistent results.7–9) The impact of steel fibers on permeability is also unknown. When steel fibers are mixed with concrete, the presence of air may influence porosity and permeability. Horiguchi et al.7,10) and Raoof and Sabouraud11) found that fiber inclusion reduced permeability. According to Ali et al.,12) permeability decreased when fiber content was less than 0.5% and increased when fiber content was 1–2% by volume. The permeability (hydraulic conductivity) of steel-fiber reinforced HSULPC, developed to improve tensile strength and toughness, was evaluated in this study, and the effect of fiber addition on hydraulic conductivity was discussed with a focus on the void structure.
This study’s sample material is HSULPC, developed by the Concrete Committee of JSCE.13) Table 1 shows the composition of the HSULPC. The concrete was placed in the formwork and left in a thermostatic chamber at 20°C for 2 days, followed by steam curing at 90°C for 2 days after unmolding, according to the standard HSULPC curing method. The same composition reinforcing fibers were used to make two types of concretes: with and without reinforcing fibers of the same composition (Figs. 2(a) and (b)). Steel fibers were used for reinforcement in this study, with diameters, lengths, and tensile strengths of 0.1–0.25 mm, 10–20 mm, and 2 × 103 N/mm2 or higher, following the Concrete Committee of JSCE.13) Because spherical lumps of fibers are generated during the mixing, it is suggested that the effect of fiber reinforcement is significantly lost if the amount of fiber added to the concrete exceeds 2%.18) As a result, the volume percentage of fibers in HSULPC was set to an upper limit of 2% to confirm the effect of fiber reinforcement (Fig. 3). Finally, cylindrical specimens with diameters of 50 mm and heights of 25 mm were created. Kato et al.15,16) previously reported the results of permeability tests on HSULPC without fibers.
Enlarged view of the surface of HSULPC with steel fiber.
The physical properties of HSULPC with and without steel fibers, as well as conventional concrete,19) are compared in Table 2. The Brazilian test was used to determine the tensile strength of HSULPC. HSULPC has greater compressive and tensile strengths than conventional concrete. Furthermore, the tensile strength of the HSULPC with steel fibers is greater than that of the HSULPC without fibers.
In this study, the results of the permeability measurements for the mudstone (Fig. 2(c): Wakkanai formation shale) obtained in Horonobe Town, Hokkaido, and granite (Fig. 2(d): Toki granite)17) obtained in Toki City, Gifu Prefecture, as well as those of HSULPC are compared, because they are representative low permeability materials.
This study’s permeability test apparatus (Fig. 4) was the same as that used previously by Kato et al.15,16,20) Because the values of water pressure and differential pressure are temperature sensitive,20) the experimental apparatus was placed in a triple-insulated chamber. During the measurements, the temperature changes around the pressure vessel were kept to within ±0.1°C/h. This permeability test apparatus can be continuously employed for various types of permeability test methods. The transient pulse method,21) which is best suited for low permeability samples, was used in this study.
The transient pulse permeability measurement procedure is based on Kato et al.15,16) However, there are some differences in the procedure for the steel-fiber reinforced concrete specimen. Some thin grooves were generated from the mold on the side of the steel-fiber reinforced concrete specimen, which could be water flow pathways during the permeability measurement. As a result, a silicone sealant was applied to that side of the steel-fiber reinforced concrete specimen (Cemedine Silicone Sealant 8000). It takes 2 days for the sealant to dry and solidify. The specimen was then wrapped in heat-shrinkable tubing across the end caps. To saturate the specimen with distilled water, vacuum deaeration was performed before and after this operation.
The confining pressure was set between 2 and 10 MPa in permeability measurements of steel-fiber reinforced concrete, and the pore water pressure was kept constant at 1 MPa. Measurements were taken with increasing and decreasing confining pressures to investigate the influence of effective confining pressure on permeability and hysteresis as the confining pressure changed. In the transient pulse permeability measurement, a pulse pressure of 40 kPa is applied to the upstream side of the specimen. The measurement was terminated when the applied pulse pressure decreased, and the differential pressure returned to its pre-pulse pressure state.
The hydraulic conductivity is evaluated by the least-squares method for the temporal change of the differential pressure obtained in the permeability measurement using the following equation, which is the solution proposed by Kato et al.22) based on Brace et al.21) to investigate the permeability characteristics:
| \begin{equation} \frac{\Delta h(t)}{H} = \exp \left\{ -\frac{KAt}{l}\left(\frac{1}{S_{u}} + \frac{1}{S_{d}} \right) \right\} \end{equation} | (1) |
The steel-fiber reinforced concrete was tested with the silicone sealant applied to the specimen’s side, as previously described. Following the measurements, it was discovered that the silicone penetrated significantly between the specimens and the end caps in the apparatus because of the confining pressure. The cross-sectional area of the water flow decreased in this case. As a result, if the original cross-sectional area of the specimen is used, the hydraulic conductivity value is underestimated. Given this challenge, photos of both sides of the specimens were taken with a digital camera after the measurements. Then, using image processing software (Adobe Photoshop), the area where no silicon penetrated was calculated. As a result, the percentages of the silicon-free area at both ends were 55% and 60%, respectively. These were considered to correct the cross-sectional area used to calculate hydraulic conductivity. It should be noted that the method used in this study to measure the hydraulic conductivity of the HSULPC differs from that described in the Concrete Committee of JSCE.13)
As seen in Fig. 5, the transient pulse approach generates a differential pressure decay curve. The hydraulic conductivity was calculated using the least-squares approximation of eq. (1) from the slope of the relationship between the differential pressure and the elapsed time in the single-logarithmic diagram (Fig. 6).
Temporal changes of the upstream and downstream hydraulic pressure difference for HSULPC specimen with steel fiber at a confining pressure of 2 MPa and pore pressure of 1 MPa obtained by the transient pulse method with a pressure pulse of 34 kPa.
Data analysis using least squares method with the data of Fig. 5 on a semilog graph.
Figure 7 depicts the obtained hydraulic conductivity for each sample. The hydraulic conductivity obtained under increasing confining pressure (the loading process in Fig. 7) is depicted as solid symbols, whereas the hydraulic conductivity obtained as the confining pressure decreases (the unloading process in Fig. 7) is shown as open symbols. In Fig. 7, solid and dotted lines represent the loading and unloading processes, respectively. Due to a lack of data, the hydraulic conductivity of the Toki granite for the unloading process was not included.
The hydraulic conductivity of the HSULPC is on the order of 10−13 to 10−12 m/s for both steel-fibered and non-steel-fibered specimens, with steel-fibered specimens having relatively higher values. The hydraulic conductivity of Horonobe mudstone and Toki granite is on the order of 10−11 m/s and 10−12 m/s,17) respectively, for low permeability rocks.
Figure 7 shows that for the HSULPC and the Horonobe mudstone, the effective confining pressure dependence of hydraulic conductivity decreases with increasing confining pressure, and the hysteresis phenomenon, in which the hydraulic conductivity remains low and does not recover during the unloading process, is remarkable.
First, we explain why the differential pressure’s temporal change is slightly convex downward in the single-logarithmic diagram in Fig. 6. This is due to the penetration of silicone, which reduces the cross-sectional area of the specimen. Furthermore, the sample’s compressive storage cannot be neglected in this case. When pore water migrates into the specimen, silicone penetration creates a smaller cross-sectional area of the upstream surface and a substantially greater cross-sectional area of the inner part of the specimen, resulting in water flow not only in the axial but also in the radial directions. As a result, a large drop in differential pressure occurs immediately after the pulse pressure is delivered. Pore water, on the other hand, migrates from the downstream surface to the exterior of the specimen in the latter half of the measurement time. The pressure drop; however, is suppressed during this moment. Furthermore, the analytical solution of Brace et al.21) utilized in this investigation is valid when the sample’s compressive storage is negligibly small compared to that of the apparatus. Nevertheless, the specimen compresses and stores a small amount of water even when this assumption is met since the sample’s porosity is not zero. Because of these influences, the differential pressure’s temporal change is slightly convex downward. However, when the complete temporal change of the differential pressure is used in the study of hydraulic conductivity evaluation, this effect disappears.
Following that, SEM (scanning electron microscope) observations on the samples were performed to investigate the structure of the HSULPC with steel fibers that maintained low permeability. Figures 8(a) and (b) show SEM images of the HSULPC without and with steel fibers, respectively. The aggregate and cement components are represented by the dark and light gray colors in both figures. The white and black colors in Fig. 8(b) depict steel fibers and voids, respectively. When Figs. 8(a) and (b) are compared, the specimen with steel fibers contains more voids. This is attributed to the participation of air during the concrete mixing process.
SEM images of HSULPC (a) without fiber and (b) with steel fiber.
Figure 9(a) shows a higher magnification SEM image of the HSULPC with steel fibers. Figure 9(b) is an enlargement of the rectangular portion in Fig. 9(a), demonstrating that voids are concentrated near the fiber-cement interface. From the hydraulic conductivity and SEM observation results, it is concluded that although there are numerous isolated voids with no significant connectivity in the HSULPC with steel fiber, water flow paths do not emerge along the steel fiber. As a result, adding steel fibers does not influence the hydraulic conductivity of HSULPC.
SEM images of HSULPC with steel fiber. Figure (b) shows the enlarged view of rectangular region in Fig. (a).
The influence of hydraulic conductivity on effective confining pressure is then explored. The coefficients (ratios) of the effective confining pressure dependency of hydraulic conductivity were computed for HSULPC with and without steel fibers and Horonobe mudstone using the following equations.
At loading, the effective confining pressure dependency coefficient is:
| \begin{equation} R_{l} = \frac{K_{i} - K_{f}}{K_{i}} \end{equation} | (2) |
Effective confining pressure dependence coefficient at unloading:
| \begin{equation} R_{u} = \frac{K_{f} - K_{b}}{K_{f}} \end{equation} | (3) |
Hysteresis coefficient:
| \begin{equation} R_{h} = \frac{K_{i} - K_{b}}{K_{i}} = R_{l} + \frac{K_{f}}{K_{i}}R_{u} \end{equation} | (4) |
Dependency of hydraulic conductivity on effective confining pressure. Rl, Ru and Rh represent the coefficients of confining pressure dependency under loading and unloading and hysteresis.
The effective confining pressure dependency and hysteresis phenomenon observed in this work are similar to those observed in sandstones,25) where it is believed that voids and cracks are closed with increasing confining pressure and do not reopen after unloading. The effective confining pressure dependency of hydraulic conductivity is identified in all three samples, as illustrated by Rl in Fig. 10. The significantly small value of Rl for the HSULPC with steel fibers is thought to represent the HSULPC’s great resistance to deformation and the minimal number of cracks with low aspect ratio. While both concretes have positive Ru values, Horonobe mudstone indicates a negative value. This suggests that the HSULPC’s hydraulic conductivity decreased considerably upon the release of the confining pressure. It is probable that the hydration reaction occurred during the permeability measurement in the non-hydrated region of the HSULPC specimen under water-saturated9) conditions and promoted the closure of voids and cracks. The interaction of these events is expected to be responsible for the exceptional hysteresis observed in HSULPC, as evidenced by Rh. To evaluate the hysteresis of hydraulic conductivity while excluding the effect of the hydration reaction, specimens of adequate age must be used.
Following the placement of the TRU waste packages in the repository, the entire disposal facility will be backfilled with buffer and backfill materials.26) As a result, swelling pressure of buffer and backfill materials, an increase in overburden pressure owing to backfilling, and an increase in pore pressure due to groundwater level rise are expected. Because of these phenomena, it is projected that the hydraulic conductivity of HSULPC would stabilize in the order of 10−13 m/s after being buried underground and subjected to pressure, even if steel fibers are added. It is concluded that HSULPC with steel fibers has low permeability and is effective for radionuclide confinement.
The transient pulse method was used to assess the permeability of HSULPC, which was created as an engineered barrier material. The influence of steel fibers on permeability was investigated. The hydraulic conductivity of HSULPC was in the order of 10−13 to 10−12 m/s for samples with and without steel fibers. This is lower than the permeability of low permeability rocks like mudstone and granite. SEM observation of HSULPC with steel fibers revealed that water flow pathways are not developed along the steel fiber, and numerous voids with poor connectivity are formed due to the participation of air during concrete mixing. The hydraulic conductivity of HSULPC with steel fibers is slightly higher than that of HSULPC without fibers due to the large voids, yet their values are in the same order. The dependence of the hydraulic conductivity on effective confining pressure and hysteresis was also observed in HSULPC. This study introduced and statistically assessed the effective confining pressure dependence coefficient. According to the findings, HSULPC with steel fibers exhibited good resistance to deformation and a low number of cracks with a low aspect ratio. It is also indicated that during the permeability measurements, hydration reactions could occur in the non-hydrated sections of HSULPC, promoting the closure of voids and cracks. The combination of these factors resulted in HSULPC’s considerable hysteresis. The hydraulic conductivity of HSULPC is expected to stabilize in the order of 10−13 m/s once the TRU waste package is buried underground and subjected to pressure. It is concluded that HSULPC with steel fibers has low permeability and is effective for radioactive confinement.
