In order to explore the horizontal bearing characteristics of pile foundations in large-thickness artificial fill foundation and the applicability of the m-method in engineering practice, the field tests of horizontal bearing capacity of single piles and the theoretical calculations of the m-method are carried out. The variation patterns of displacement and bending moment under horizontal loads are obtained, and the deformation characteristics of pile-soil under horizontal loads are presented. The variation of the proportional coefficient m of the horizontal resistance coefficient of the foundation soil with the load and displacement is further obtained. The results show that with the increase of loading, the horizontal displacement increases continuously, and after reaching the critical value of 480 kN, the horizontal displacement and displacement gradient of pile foundations change drastically. Under the same loading, the bending moment of pile foundations increases first and then decreases with the increase of depth. The bending moment diagram shows the distribution pattern of “small at both ends and large in the middle”. At the same depth, the bending moment increases with the increase of load, and negative bending moment appears at a certain depth. The influence range of bending moment along the depth is about 10 m. The proportional coefficient m of the horizontal resistance coefficient decreases exponentially with the increase of load and displacement. When the load and displacement are small, the m values of the two piles are quite different. When the load and displacement increase to a certain value, the m values of the two piles are close and finally stabilized near a specific value. The calculation results of the m-method are verified and improved by experimental data. It is found that the maximum displacement calculated by the m-method is close to the measured results. When the load is small, the calculated maximum bending moment is close to the measured results. After exceeding the critical load, the difference is large, indicating that when the load exceeds the critical value, the maximum bending moment does not show a linear elastic increase. It is necessary to correct the bending moment calculation result by the correction coefficient function β. The corrected bending moment is more consistent with the measured value, indicating that the m-method works well in engineering practice.

In order to study the adsorption characteristics of water vapor on unsaturated loess, isothermal adsorption experiments under different humidity conditions were carried out by vapor equilibrium method. The adsorption behavior of water vapor on the surface of unsaturated loesswas analyzed, and the effects of temperature, mineral composition and content, dry density on the adsorption property of soil were discussed. The experimental results show that the water vapor adsorption capacity of unsaturated loess increases with the increase of relative humidity, and the entire process consists of three stages: monolayer adsorption, multilayer adsorption and capillary condensation. The GAB model can describe the water vapor adsorption process of unsaturated loess. There is a significant negative correlation between water vapor adsorption capacity and temperature. When the relative humidity is constant, the adsorption capacity of water vapor decreases with the increase of temperature. Water vapor adsorption of unsaturated loess is closely related to mineral composition, and clay mineral content directly affects its water vapor adsorption capacity. In addition, the effect of dry density on water vapor adsorption capacity can be divided into two stages. When the relative humidity RH<80%, water vapor adsorption capacity increases with the increase of dry density. For capillary condensation stage, with the increase of dry density, the amount of water vapor adsorption no longer increases but decreases.

Water diversion aqueducts are highly susceptible to freezing and blockage at low temperatures in winter, particularly in the cold and arid regions of northern China. There are also significant declines in the water conveyance capacity and structural performance. A great threat has then been posed to the long-term safe operation of water diversion projects in these regions. Therefore, it is crucial to the patterns of winter water temperature and the mechanisms of icing evolution in aqueducts. However, it is still lacking in the winter operational characteristics of aqueducts. This study aims to investigate the spatiotemporal variations of water temperature and icing behavior under the influence of multiple coupled factors. A three-dimensional coupled numerical model was developed for the non-isothermal flow heat transfer in a closed aqueduct. Some parameters were then considered, including air temperature (T_a), flow velocity (u), inlet water temperature (T₀), solar radiation, and wind speed. The characteristic parameters were selected as the temperature drop value (ΔT) and temperature drop rate (Δf_u) of the water flow in the aqueduct. A systematic analysis was implemented to explore the temporal and spatial variation patterns of winter water temperature. Additionally, a predictive model was also established for the water temperature and freezing points. The average water temperature was then calculated at the outlet section of the aqueduct under various conditions. The icing locations of the water flow were predicted under specific temperature scenarios. A two-dimensional transient icing model was established to consider the effect of the ice-water phase transition on heat transfer in the water flow in an aqueduct. The icing evolution was also obtained in the aqueduct water flow. The correctness of the model was verified to compare the indoor test data with the simulation. The results show that there was a decrease in the water temperature of the aqueduct over time in winter, especially with the increasing water delivery length. Along the length of the aqueduct, the temperature drops of water flow exhibited an overall trend of rapid decline followed by a slower reduction. The ΔT value decreased with the increasing u under certain meteorological conditions, as T₀ rose. While there was an increase. The primary influencing factor on water temperature was the flow velocity u, with the largest temperature drop rate in the range of 0-1m/s. The temperature drops near the solid walls of the aqueduct were approximately 2 to 4 times greater than that at the center of the cross-section. Solar radiation caused a greater decrease in the water flow temperature near the aqueduct wall at night than during the day. In contrast, the water temperature at the center of the aqueduct cross-section was less affected by solar radiation. According to the water flow freezing in aqueducts, the release of latent heat during condensation shared a compensatory effect on the water temperature. The bank ice width was used as a quantitative indicator of icing. The amount of ice formation increased over time. There was a shorter critical length for the water flow near the shaded side wall of the aqueduct to reach the freezing point. The ice formation on the shady side was approximately three times that on the sunny side. This finding can also provide a strong reference for the safe operation of aqueducts at low temperatures in winter.

In order to study the water infiltration and self-weight collapse deformation characteristics of Jingyuan loess with large thickness under the condition of immersion, a field immersion test without water injection holes was carried out in the self-weight collapsed loess site of Jingyuan North Station along the Zhongwei−Lanzhou Railway. The surface and underground collapsible deformation, cracks, water content and vertical stress in the soil around the test pit were monitored and analyzed. The water diffusion, self-weight collapsible characteristics and vertical stress in the soil were studied, and the regional correction coefficient β₀ value and wetting angle were discussed. The results showed that: the change of volumetric water content was divided into four stages: immersion stabilization (two), rapid increase (one) and slow increase (one). In the immersion process, the vertical infiltration of water was accelerated and the radial diffusion was slowed down at 21 m, and the final shape of the wetting front was presented as elliptical. According to the water content test results of exploratory wells and boreholes, the maximum wetting angle was calculated to be 41°. The self-weight collapse process of loess in the site went through three stages: severe collapse, slow collapse and consolidation stabilization. At the end of the test, a total of 13 ring cracks were developed, and the farthest point of the cracks was 26 m from the edge of the test pit. According to the laboratory test and field test results, it was suggested that the regional correction coefficient should be corrected along the depth of the soil layer, and the β₀ value was taken as 1.05 within 0−10 m and 0.95 within 10−27 m. In the depth range from the surface to 21 m, the foundation soil was saturated and fully collapsed. The vertical stress in the soil increased linearly along the depth, and the vertical stress in the soil was close to the saturated self-weight stress. The foundation soil below 21 m failed to collapse entirely, and the vertical stress in the soil decreased gradually. The research results could be applied to the later construction of Zhongwei−Lanzhou Railway and provide a reference for other regional engineering projects.