TL;DR
This paper investigates how the distribution and properties of clay layers (aquitards) impact the efficiency of dewatering for deep excavations. Using a 3D groundwater flow model, they found that more extensive clay layers improve dewatering, but higher clay permeability reduces efficiency. The location of these clay layers also matters, with those below the riverbed being more effective than those in the foothills.
What Problem Does This Paper Address?
Deep excavations for civil engineering projects often encounter groundwater, which can compromise stability and safety. Dewatering is crucial to manage this, but its effectiveness is heavily influenced by complex geological conditions, particularly the presence and characteristics of aquitard layers like clay. In river valley areas, these clay layers can be discontinuous due to sedimentation, making it challenging to predict dewatering performance, especially when detailed on-site geological data is limited. This study aims to quantitatively assess how the distribution and permeability of these clay layers affect foundation pit dewatering efficiency, providing a more optimized approach for dewatering designs.
Methods
The researchers employed a three-dimensional, non-steady-state groundwater flow model using the MODFLOW-USG software. This software is notable for its use of unstructured grids, which allows for flexible grid refinement to better represent complex geological features like aquitard layers. The model was calibrated against hydrogeological data and observed groundwater levels from a specific deep excavation site on the Chongjiang River. To analyze the influence of clay layers, they simulated various scenarios by altering the clay layer's coverage, permeability, storage coefficient, and specific yield. The MODPATH module was also used to track groundwater flow directions and understand the impact of cutoff walls.
Datasets
The study utilized hydrogeological data from borehole investigations and pumping tests conducted at the Chongjiang River Inlet Culvert project site. Meteorological data (average annual precipitation and intensity) were also used for rainfall infiltration calculations. Observed groundwater levels from 21 monitoring wells were used for model calibration. No new datasets were created or introduced in this study.
Key Insights and Ideas
- Clay as a Barrier: Increased clay layer coverage acts as a more effective barrier, significantly enhancing dewatering efficiency by reducing groundwater inflow.
- Permeability Trade-off: While more clay is generally good, higher permeability within the clay layer itself can paradoxically reduce dewatering efficiency. This suggests a balance is needed.
- Specific Yield's Minor Role: The specific yield of the clay layer appears to have a less pronounced effect on dewatering efficiency compared to its permeability or coverage.
- Location Matters: The position of discontinuous clay layers is critical. Clay layers situated below the riverbed are more effective at impeding groundwater flow and improving dewatering than those located in the foothills, where continuous recharge from higher ground can occur.
- Cutoff Walls are Crucial: The study reinforces the significant positive impact of cutoff walls in preventing external groundwater from entering the dewatering zone, thereby greatly increasing dewatering effectiveness.
Key Results
The model successfully reproduced actual dewatering processes, with calibration showing errors in water level calculations generally less than 3 meters. The simulations revealed that pumping rates can increase significantly as clay layer coverage decreases. For instance, pumping rates could rise from 18,200 m³/d with 100% clay coverage to 48,000 m³/d with 0% coverage. The study also quantified the impact of hydraulic conductivity and storage coefficient, showing that higher values generally lead to higher pumping rates, with the effect of hydraulic conductivity becoming more pronounced when combined with a higher storage coefficient. The groundwater level in the engineering area was observed to drop from an average of 1832 m to 1790 m within 30 days of dewatering.
Limitations and Future Work
The authors acknowledge that the study area exhibits strong heterogeneity due to the irregular distribution of clay layers. While the model effectively captures the general trends, the precise prediction of dewatering rates can be influenced by the specific spatial continuity of these layers, which may be affected near the main river channel. The study also notes that the influence of the "sealed box" effect of clay layers outweighs location-specific effects when considering total pumping rates. Future work could involve more detailed investigations into the spatial variability of clay layer properties and their impact on dewatering in different geological settings.
Summary updated