During the implementation phase of this study, the placement of temperature and moisture sensors followed the layout depicted in Fig. 8. The sensors were embedded and buried simultaneously with the foundation soil, which was placed and compacted in layers. Four rows of temperature sensors were arranged at intervals of 10 cm, and three rows of moisture sensors were positioned at distances of 10 cm, 25 cm, and 40 cm from the bottom of the channel. The sensor layout of the left and right channel models shall be consistent. A total of 32 temperature sensors and 18 moisture sensors were installed. Once the channel model was formed, a sprinkler maintenance procedure was carried out to replenish the water lost during the filling process. After the hemp tiles were laid on the surface of the channel, the model was essentially complete. Subsequently, a bracket and a movable beam were positioned above the box to secure the displacement sensors. The displacement sensors were fixed to the bracket using tape to ensure that no relative movement occurred between the sensor and the bracket. The beam, equipped with the support, was then placed above the box, and both ends were weighted down to prevent any relative displacement between the box and the beam. The axis of the displacement sensor was oriented perpendicular to the surface of the channel to measure the normal displacement of the channel lining. A total of four displacement sensors were installed: one on the top of the soil slope and one at the bottom of the channel.
Figure 9 illustrates the frost heaving and deformation process of the top and bottom of the channel for both in-situ silt soil and modified soil. As shown in Fig. 9, the frost heave amount of the soil exhibited a trend of initially increasing, followed by a decrease, and finally stabilizing within each freeze-thaw cycle. This pattern corresponded with the fluctuations in ambient temperature. In the first freeze-thaw cycle, the maximum frost heave volume at the top of the in-situ silt soil channel reached 4.47 mm. In the subsequent second to fourth freeze-thaw cycles, the maximum frost heave volumes were 2.94 mm, -0.35 mm, and - 1.56 mm, respectively. Similarly, the maximum frost heave volume at the top of the modified soil channel also occurred during the first freeze-thaw cycle, measuring 2.13 mm. Notably, in the third and fourth freeze-thaw cycles, the maximum frost heave volumes for both soils were negative, indicating soil compression. This phenomenon can be attributed to two main factors. First, thaw subsidence deformation exceeded frost heave deformation during the freeze-thaw cycle, resulting in a final frost heave volume lower than the initial channel top height. Second, during the freeze-thaw cycles, the water at the top of the channel dissipated. Due to the air-cooled cooling method employed in the model box, the cold air carried away some of the water in the upper layers of the soil during the melting stage, leading to a decrease in unfrozen water content. As a result, the frost heave in the subsequent freezing stages was slightly lower than in the previous cycles.
The frost heave deformation at the bottom of the channel follows a similar trend to that at the top, initially increasing, then decreasing, and eventually stabilizing. The maximum frost heave of the in-situ silt soil occurred during the first freeze-thaw cycle, measuring 5.52 mm. Similarly, the maximum frost heave of the modified soil also occurred in the first freeze-thaw cycle, reaching 3.89 mm. In each freeze-thaw cycle, the deformation due to thaw subsidence slightly exceeded the frost heave deformation, resulting in a final negative frost heave, indicating a slight compression of the soil. This phenomenon can be attributed to two primary factors. First, the loss of unfrozen water -- although the bottom of the channel is covered with hemp tiles, which significantly reduces water loss, a small amount of unfrozen water still escapes through the gaps between the tiles and the border. Second, the compression effect caused by the weight of the tiles on the soil also contributes to the final frost heave result. Both factors together influence the observed slight compression at the bottom of the channel.
Further research, in conjunction with the studies of Han and Zhang, demonstrates that under the influence of multiple freeze-thaw cycles, the internal migration of water and temperature fluctuations within the soil mass lead to a reshaping of its structure. Specifically, during the freezing process, the water in the soil freezes into ice crystals, causing displacement and rearrangement of soil particles along with volume expansion. In the melting stage, the ice crystals transform into liquid water, altering the stress state between the soil particles and leading to soil compression. This continuous cycle of expansion and compression results in dynamic changes in soil porosity and compactness, which significantly influence its mechanical properties. When analyzing the modified soil, it is observed that its frost heave deformation is generally lower than that of the in-situ silt soil. A comparative analysis indicates that the stability of the modified soil under freeze-thaw cycles is superior to that of in-situ silt soil.
A comparison of frost heave deformation across different soil types at the same channel position, as shown in Fig. 9, reveals significant differences. Data analysis indicates that the maximum frost heave value of the modified soil is consistently lower than that of the original silt, both at the top and bottom of the channel. This result further supports the conclusion that the addition of nano-ZnO to silt can effectively mitigate soil frost heave. For example, at the bottom of the channel, the maximum frost heave of the original silt is 5.52 mm, while that of the modified soil is only 3.89 mm, representing a substantial reduction. The improvement measures reduced the frost heave by 29.5%, demonstrating the significant impact of nano-ZnO in enhancing the freeze-heave resistance of the soil.
Figure 10 illustrates the dislocation and deformation of the lining plate after the test, clearly showing the displacement and joint misalignment of the lining panels. This indicates that frost heave significantly impacts the canal lining structure. Frost heave can lead to the fracture and deformation of lining panels, and, in extreme cases, cause the failure of the entire structure, presenting a serious threat to the stability of the engineering system. Furthermore, with the continuous action of multiple freeze-thaw cycles, the water within the soil undergoes repeated freezing and melting processes, leading to the destruction and reshaping of the soil structure. This phenomenon is particularly prevalent in water delivery channels in cold regions, especially during the transitions between extreme winter cold and spring thaw. The migration of water and the formation of ice crystals result in significant changes to the soil's pore structure, which not only negatively affects the structural integrity of the channel but also diminishes its water transport capacity, thereby increasing the complexity of operations and maintenance.
Studies have shown that unfrozen water remains present in frozen soil, even at extremely low temperatures, such as - 70 °C. The unfrozen water content plays a crucial role in the thermodynamic characteristics of the soil and, consequently, has a profound impact on the safety of engineering structures. Therefore, the dynamic changes in unfrozen water content are not only a key index for evaluating the stability of a project but also a core parameter for preventing freezing damage to the structure. Figure 11(a) illustrates the trend in the volume of water content of the in-situ silt soil model at 10 cm and 40 cm from the bottom of the channel, while Fig. 11(b) shows the corresponding changes in the volume water content of the modified soil model at similar positions. By comparing these two figures, it is evident that, at different depths, the soil volume water content follows a similar pattern during each freeze-thaw cycle: it decreases initially, then increases, and eventually stabilizes. This behavior can be attributed to the effects of temperature fluctuations during the freeze-thaw cycle. When the soil temperature drops below the freezing point, the freezing front advances downward, causing the water above the channel to freeze and reducing the liquid water content, thus lowering the volume water content. As the ambient temperature rises, the ice melts into liquid water, resulting in an increase in volume water content. As the freezing process progresses from top to bottom, the volume water content at 40 cm below the channel decreases more rapidly than at 10 cm, and the unfrozen water content at 10 cm remains higher than at 40 cm due to the lower temperature at the top of the channel. Additionally, Fig. 11 shows that the content of unfrozen water in frozen soil decreases as the temperature drops, which is consistent with the findings of Li et al.. This further confirms that the content of unfrozen water is significantly influenced by the soil temperature gradient.
Figure 12 presents a comparison of the volume water content between in-situ silt soil and modified soil at the same depth from the channel bottom. Specifically, Fig. 12(a) and Fig. 12(b) correspond to changes in water content at 10 cm and 40 cm from the bottom of the channel, respectively. A comprehensive analysis of the two figures reveals a consistent trend: the volume water content of the modified soil is significantly lower than that of the original silt at every depth. Despite having the same initial water content, the content of unfrozen water in the modified soil is notably reduced, and the volume water content of the modified soil remains consistently lower than that of the in-situ silt soil throughout the experiment. These results indicate that nano-ZnO plays a significant role in reducing the content of unfrozen water in the soil. Further analysis of Fig. 12(c) reveals that the volume water content of in-situ silt soil at 40 cm is lower than that of the modified soil at 10 cm. This phenomenon highlights the significant influence of water evaporation on the distribution of water content, particularly in the surface soil layer. The top of the soil sample is more susceptible to environmental factors, such as airflow and temperature gradients, which leads to more intense water evaporation. As a result, the water content in the shallow soil decreases significantly. This finding underscores the importance of considering the effect of water evaporation in freeze-thaw experiments, especially the dynamic changes in water content in shallow soil, which should be carefully addressed in future studies and practical engineering applications.
Figure 13 illustrates the variation in average frost depth under different soil conditions, along with a comparative analysis. Figure 13(a) shows the changes in frost depth of the in-situ silt soil under the influence of freeze-thaw cycles, while Fig. 13(b) corresponds to the changes in the frost depth of the modified soil. A common trend can be observed when comparing the frost depths of the two soils: in the first freeze-thaw cycle, both soils exhibit larger frost depths and a significant freezing rate. This is primarily due to the low initial temperature field, which results in a frost depth and growth rate during the first freeze-thaw cycle that is notably higher than in subsequent cycles. Further analysis of the data in Fig. 13; Table 2 reveals that, in terms of maximum frost depth, the in-situ silt soil reaches 39.95 cm, while the modified soil reaches 39.09 cm, indicating that the frost depth of the in-situ silt soil is slightly greater than that of the modified soil. During the freezing process, both the top surface of the channel and the top of the channel slope experience two-way freezing. Specifically, the freezing front at the top of the channel advances downward in the vertical direction, while the freezing front at the top of the channel slope moves downward in the direction perpendicular to the slope. The formation and migration of the freezing front are primarily driven by water movement within the soil. The downward migration of the freezing front provides the driving force for water migration toward the frozen layer. As this process occurs, the content of unfrozen water decreases, the matrix potential increases, and the matrix potential gradient drives water in the unfrozen area to migrate to the freezing front, where it subsequently freezes.
In summary, the addition of nano-ZnO to soil effectively reduces the frost depth and significantly enhances its freeze-thaw stability. The improved performance of modified soil during the freezing process offers a promising material alternative and a theoretical foundation for frozen soil engineering in cold regions. Moreover, this research provides valuable practical insights and technical support for engineering design and construction in such environments.
In this study, the temperature field of the middle section of two soil models (in-situ silt soil and modified soil) is compared and analyzed. Figure 14 illustrates the location distribution of analysis nodes during the entire freeze-thaw cycle, while Fig. 15 presents detailed changes in the temperature fields of the two soil models under different ambient temperatures. Figure 15(a) shows the initial temperature field of the middle section of both soil models. The results reveal that the initial temperature distributions of the two models are highly similar. Figure 15(b) depicts the temperature field distribution at the lowest ambient temperature. Significant differences in the internal temperature field of the model are observed. The soil temperature at the bottom of the channel is markedly lower than at the top, primarily due to the substantial difference in thermal properties between the ceramic tiles and the soil. Soil generally has a larger heat capacity and lower thermal conductivity compared to ceramic tiles, which have a smaller heat capacity and higher thermal conductivity. This discrepancy causes the tiles to be more sensitive to changes in ambient temperature, cooling rapidly at low temperatures and thereby lowering the temperature of the soil beneath them. As a result, the temperature at the bottom of the channel drops significantly. In contrast, under the highest ambient temperature, Fig. 15(c) shows the temperature field distribution relative to the low-temperature environment. Under high-temperature conditions, ceramic tiles heat up first due to their smaller heat capacity, which in turn transfers heat to the soil below via conduction. This heat transfer causes the temperature at the bottom of the channel to be substantially higher than at the top. This phenomenon can again be explained by the thermal properties of the tiles: their high thermal conductivity and low heat capacity enable them to respond rapidly to ambient temperature changes and conduct heat to the soil in contact with them, thus increasing the temperature at the bottom of the channel. The soil at the top of the channel, exchanging heat with circulating air in the test chamber, experiences partial heat dissipation, slowing its warming. This creates a temporary temperature inversion (lower layer warmer than upper layer), which reflects the real-world role of canal linings as heat bridges during thawing-accelerating heat transfer to underlying soil and influencing frost heave dynamics. Overall, the distribution of the temperature field demonstrates that the temperature difference between the top and bottom of the channel is not only directly influenced by ambient temperature but also significantly impacted by the thermal properties of the covering materials, such as ceramic tiles. The results underscore the importance of material selection in engineering design, particularly in cold regions or extreme climates, where the choice of materials can substantially alter the temperature field distribution, affecting the thermodynamic stability and frost resistance of engineering structures.
A comprehensive analysis of the temperature variations in the original silt model and the modified soil model, as shown in Figs. 15(b) and 15(c), reveals that the temperature variation range of the modified soil is slightly smaller than that of the in-situ silt soil under both the lowest and highest ambient temperature conditions. This phenomenon can be attributed to the significant reduction in the sensitivity of the soil to ambient temperature changes after the addition of nano-ZnO. The thermal resistance provided by nano-ZnO plays a key role in this behavior by reducing the soil's thermal conductivity and impeding heat transfer within the soil. The core driving force behind soil temperature changes is closely related to the soil's thermal properties and the molecular movement of soil particles and water molecules. The experimental results demonstrate that the addition of nano-ZnO slightly increases the heat capacity of the soil while reducing its thermal conductivity. This effect occurs because nano-ZnO promotes the bonding of water molecules with soil particles, forming stable polymer structures. Within these polymers, some water molecules become bound, which restricts their ability to move freely and nearly prevents their participation in the heat transfer process. Since water molecules typically have high thermal conductivity, unfrozen water plays a crucial role as a medium for heat transfer in soil. Consequently, when these water molecules are bound, the overall thermal conductivity of the soil decreases. Additionally, the presence of nano-ZnO increases the heat exchange demand per unit volume of soil. In other words, the modified soil requires more heat input or output to achieve the same amplitude of temperature change during heating or cooling processes. This mechanism makes the modified soil less responsive to fluctuations in ambient temperature, resulting in reduced sensitivity to temperature changes. To validate the claim that nano-ZnO alters soil heat capacity and thermal conductivity, we conducted supplementary experiments using a thermal property analyzer (Hot Disk TPS 2500 S) on both in-situ silt soil and modified soil. The results are shown in Table 3.
The results showed that the thermal conductivity of the modified soil decreased by about 19%; In terms of heat capacity, the modified soil increased by about 15%. These data directly confirmed that nano ZnO reduced the thermal conductivity and increased the thermal capacity, which supported our mechanism interpretation.
In summary, the introduction of nano-ZnO not only effectively reduces the soil's responsiveness to ambient temperature fluctuations but also enhances its thermal stability by inhibiting heat transfer within the soil. This property offers significant advantages for the use of modified soil in cold region engineering. It helps mitigate the deterioration of soil properties caused by temperature variations and improves the thermal stability and long-term durability of engineering structures.