The Xiluodu Dam is a concrete double-curvature arch dam with a crest elevation of 610 m and a height of 285.5 m. Since the impoundment of the Xiluodu reservoir, remarkable river valley contractions (RVCs) have been observed upstream and downstream of the reservoir, potentially threatening the safety of the dam. However, the cause of these RVCs remains unclear. Based on an analysis of hydrogeological conditions, the RVCs were determined a result of the expansion of the aquifer, within which the effective stress decreased due to an increase in the hydraulic head after reservoir impoundment. Referring to the hydrostatic seasonal time (HST) model, a groundwater hydrostatic seasonal (GHS) model is proposed for simulating and predicting the development of the RVCs. Unlike the HST model, the GHS model can provide information on aquifer hydraulic diffusivity. The calibration results illustrate that the GHS model can accurately fit the observed RVCs data. The calculation results revealed that the RVCs were mainly affected by the hydraulic head of the confined aquifer, and that seasonal effects gave rise to less than 10% of the total RVCs. Finally, the development of RVCs were predicted using the GHS model. The prediction results demonstrated that the RVCs of most monitoring lines in the Xiluodu reservoir would gradually approach a convergence condition after 6 February 2021. Until the deadline of the prediction on 1 May 2035, there is still one monitoring line that has not reached a convergence condition (whose RVCs are 157.6 mm, and where the RVC growth rate will decrease to 0.005 mm/d by that time). Considering the large amount of RVCs, we think the safety of the dam requires closer consideration.
The response laws of groundwater dynamics on the riverbank to river level variations are highly dependent on the river level fluctuation process. Analytical solutions are widely used to infer the groundwater flow behavior. In analytical calculations, the river level variation is usually generalized as instantaneous uplift or stepped, and then the analytical solution of the unsteady groundwater flow in the aquifer is derived. However, the river level generally presents a complex, non-linear, continuous change, which is different from the commonly used assumptions in groundwater theoretical calculations. In this article, we propose a piecewise-linear approximation to describe the river level fluctuation. Based on the conceptual model of the riverbank aquifer system, an analytical solution of unsteady groundwater flow in an unconfined aquifer under complex boundary conditions is derived. Taking the Xiluodu Hydropower Station as an example, firstly, the monitoring data of the river level during the period of non-impoundment in the study area are used to predict the groundwater dynamics with piecewise-linear and piecewise-constant step approximations, respectively, and the long-term observation data are used to verify the calculation accuracy for the different mathematical models mentioned above. During the reservoir impoundment period, the piecewise-linear approximation is applied to represent the reservoir water level variation, and to predict the groundwater dynamics of the reservoir bank.
The external water pressure and seepage quantity are crucial factors in tunnel construction design, especially under complex geological conditions. Both analytical solutions and numerical simulations can be used to calculate the external water pressure in a shallow circular cross-section tunnel. Analytical formulas are usually derived based on homogeneous surrounding rock, which seems somehow unrealistic. In reality, the engineering geological conditions are complex and the surrounding rock is heterogeneous. Therefore, the results of analytical formulas may deviate from the actual results. This study takes the No. 1 shallow-buried tunnel in the north bank of the Xiaolangdi irrigation area as its research background. Based on heterogeneous and anisotropic analyses, the horizontal permeability, the vertical permeability, the parameters of the lining and grouting circle, and the position head are the main factors affecting the stability of the external water pressure. The result shows that the external water pressure increases as the ratio of the horizontal permeability coefficient to the vertical permeability coefficient increases. Moreover, when a grouting ring is used to block the water, the lining is guaranteed to have a certain degree of permeability, which helps to better achieve the purpose of reducing the external water pressure.
Groundwater is a critical natural and strategic economic resource, and the accurate prediction of groundwater depth dynamics is essential for the rational development and utilization of water resources. However, under the combined influence of climate variability, human activities, and complex hydrogeological conditions, groundwater level time series exhibit strong nonlinear and non-stationary characteristics, posing great challenges to the accurate prediction of groundwater level dynamics. Most existing prediction models rely on sufficient hydro-meteorological and exploitation data that are difficult to obtain in water-scarce regions, or fail to effectively decouple the multi-scale features of non-stationary groundwater level signals, resulting in limited prediction accuracy and insufficient generalization ability. To address these research gaps, this study takes Zhengzhou, a typical water-deficient city in the Yellow River Basin, as the study area, and proposes a hybrid deep learning framework combining Variational Mode Decomposition (VMD) and Long Short-Term Memory (LSTM) neural network for predicting shallow and intermediate-deep groundwater level changes. Kolmogorov–Arnold Networks (KANs) and Gated Recurrent Units (GRUs) are selected as benchmark models to verify the superior performance of the proposed framework. In this framework, the non-stationary groundwater level signal is adaptively decomposed into Intrinsic Mode Functions (IMFs) with distinct frequency characteristics via VMD. An independent LSTM model is constructed for each IMF to capture its unique temporal variation pattern, and the final groundwater level prediction is obtained by linearly reconstructing the predicted results of all IMFs. The results show that the coefficient of determination (R2) of the VMD-LSTM model exceeds 0.90 for all monitoring datasets, with low Mean Absolute Error (MAE) and Mean Squared Error (MSE). It significantly outperforms the benchmark models in handling nonlinear and non-stationary time series features. Using only historical groundwater level data as input, the proposed framework effectively overcomes the limitation of insufficient driving variables in data-scarce regions and fully explores the multi-scale evolution of groundwater dynamics through the synergistic effect of multi-scale decomposition and deep learning. The method presented in this study provides a novel and reliable technical approach for groundwater level prediction in water-deficient and data-limited areas, and also offers scientific support for the rational management and sustainable utilization of regional groundwater resources. Future research will incorporate driving factors such as meteorology and exploitation to further improve the model’s ability to capture abrupt changes in groundwater level dynamics.