Numerical Simulation of Horizontal Barrier in Controlling Groundwater and Deformation During Foundation Pit Dewatering

Abstract

In water-rich strata, a traditional vertical barrier exhibits certain limitations when applied to deep foundation pit construction under complex geological conditions, such as it is difficult to completely cut off deep and thick aquifer, which may pose potential risks during pit dewatering. To address the above challenge, this study introduced a mixed barrier system in which the horizontal barrier (HB) was set at the bottom of the foundation pit and was combined with the enclosure wall to collectively retard groundwater seepage into the pit. Based on an actual project in Tianjin, this study established HB models with varying numbers of its layers using ABAQUS 6.14 software. It systematically investigated the effect of HB on groundwater drawdown, ground surface settlement, and enclosure deflection during foundation pit dewatering. The research shows that HB can significantly reduce the magnitude of external water level drawdown by altering groundwater seepage paths while effectively controlling soil settlement. Furthermore, it exhibits favorable overall restraining effects on wall deformation. Varying the number of horizontal barrier layers (L) exhibits an insignificant effect on water-blocking and subsidence-control performance. However, the constraint effect on the enclosure shows a correlation with L.

1. Introduction

With the rapid development of urban construction, the efficient utilization of underground space has increasingly become an important part of modern urban planning. Consequently, deep foundation pit engineering has gradually become a key technical aspect [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17]. The impact on the surrounding environment during the construction of deep foundation pit cannot be ignored [18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39], especially in water-rich areas. Dewatering operations are essential to ensure construction safety and stability of pit [40,41,42,43,44,45,46,47,48,49,50,51,52]. However, due to the relatively high groundwater level and the deep and thick aquifers in water-rich areas, the groundwater seepage paths during dewatering are highly variable, which potentially lead to issues such as water seepage into the pit and ground surface settlement [53,54,55,56,57,58,59,60,61,62,63,64,65,66,67].

How to effectively control the deformation of the foundation pit during dewatering has become a critical issue in the design and construction of engineering in water-rich areas [68,69,70,71,72,73,74,75,76]. As an essential groundwater control measure, a waterproof barrier has been widely adopted, which is mainly by altering groundwater seepage paths to mitigate the environmental impact of dewatering on adjacent areas. In practice, vertical barrier (VB) is commonly used [77,78,79,80,81,82,83,84,85,86,87]. However, in areas with thick aquifers, completely cutting off the groundwater flow would require a significant increase in the depth of the VB, which would pose substantial challenges in terms of construction technology and cost. Although the suspended waterproof barrier [88,89,90,91,92,93,94,95,96] can solve the above problems, it exhibits limitations in effectively blocking groundwater seepage and control foundation pit deformation. Thus, optimizing barrier design to balance hydraulic isolation effectiveness, deformation control, and economic feasibility remains a key research focus in deep pit engineering.

In recent years, it has been proposed to set the horizontal barrier (HB) at the bottom of the foundation pit to control groundwater seepage in an all-round way with the vertical curtain, in order to overcome the limitations of traditional waterproof curtain technology. Many scholars have carried out research on this. Wang et al. [97,98] studied the application of a vertical-horizontal (VH) hybrid waterproof barrier in deep foundation pit precipitation through a transparent soil physical model test and numerical simulation, and explored the influence of the location, thickness, and permeability coefficient of the barrier on the water level change and seepage path inside and outside the foundation pit. Cao, Shi et al. [99,100,101] established numerical models to propose the analysis method for horizontal waterproof curtain, along with computational models for water inflow and water pressure, with the reliability of these methods being validated through laboratory tests and actual engineering applications. These studies have achieved remarkable results in the influence mechanism of horizontal curtain on the water level outside the pit and its design method, but there are still some limitations in the research on the influence of HB on pit deformation induced by dewatering.

In this paper, based on an actual subway foundation pit, dewatering models under different HB configurations are established. The research aims to explore the control effect of HB on groundwater seepage and pit deformation during the process of deep foundation pit engineering dewatering in water-rich areas, while analyzing the influence degree of relevant parameters. The findings provide valuable references for realizing the configuration of HB in practical engineering.

2. Numerical Model

2.1. Problem Setup

Figure 1 shows the typical soil profile and the position of HB in a deep foundation pit project located in Tianjin [56], which serves as the basis for this numerical simulation study. The dimensions of the pit were specified as 155 m long and 40 m wide; the enclosure wall was set at 32.5 m depth and 0.8 m thickness; multiple dewatering wells with 24.9 m depth were evenly arranged in the pit; and the bottom of HB was set at 32.5 m depth. The soil profile consists of alternating layers of highly permeable and low-permeability strata, with five silty sand-dominated layers (Aq0-AqⅣ) and four clay-dominated layers (AdⅠ-AdⅣ) arranged sequentially within 50 m from top to bottom, exhibiting a characteristic alternating distribution. The key physical parameters of the soil layers are shown in Figure 1, in which the horizontal permeability coefficient (K_h), vertical permeability coefficient (K_v), and elastic modulus (E) were determined through inversion calculation by a comprehensive approach combining field measurements and computed results [62]. The rest of the parameters were from laboratory tests.

All the parameters of the numerical simulation are shown in

Table 1. Values of parameters in the numerical model, including the thickness of HB (t_HB) and the number of HB (L).

Parameter	Value
tHB	0, 1, 0.5, 0.3 m
L	0, 1, 2, 3 layer

, with the permeability coefficient of the HB set to 0.001 m/d. Additionally, t_HB is the thickness of HB and L is the number of HB. The study focuses on the key parameter of L. A total of four numerical models were established, including the operating conditions without the HB.

This study primarily explores the response characteristics of soil and the enclosure wall during pre-excavation dewatering. In accordance with current support requirements [102], the first-level support is incorporated in the numerical model prior to dewatering operations. Considering the HB system is positioned within the AqІІ aquifer, its presence will:

(1)

Obstruct groundwater seepage from outside to inside pit during dewatering;

(2)

Modify hydraulic gradients and pore pressure distribution patterns;

(3)

Alter deformation mechanisms of both surrounding soil and enclosure wall.

2.2. Model Setup

Based on geomechanically theories [103], this study developed coupled hydromechanical finite element models using the ABAQUS software to analyze the soil and wall response problems described in Section 2.1.

As illustrated in Figure 2a, the dimensions of the soil were constructed according to actual engineering geological conditions (Figure 1) and the dewatering influence radius theory [104]. The specific size parameters are longitudinal (length direction) 1640 m, transverse (considering symmetry) 877.5 m. The C3D8P solid element was used to simulate soil behavior, which can effectively reflect the fluid-structure interaction effect in the dewatering process. The selection of an appropriate constitutive model is crucial for accurately characterizing soil behavior [105,106,107,108,109,110,111]. Considering the elastic deformation characteristics observed in the Tianjin soil during dewatering, the Mohr–Coulomb constitutive model was adopted in this study to accurately simulate this behavior.

Figure 2b shows the remaining structural components of the numerical model, including struts, dewatering wells, enclosure wall, and HB. To accurately reflect their mechanical responses during dewatering, these components were simulated using linear elastic models [112,113,114], that is, assuming a linear stress-strain relationship within the small deformation range. The dewatering wells were modeled using an S4 shell unit, and its elastic modulus was set to 210 GPa according to the standard specification [102]. The struts, enclosure wall, and HB were simulated by B31 beam elements, C3D8I solid elements, and C3D8P solid elements, respectively, all assigned an elastic modulus of 30 GPa. Extensive research has demonstrated that the compressive strength and impermeability of concrete directly influence the durability of walls [115,116,117]. During dewatering operations, enclosure wall are subjected to coupled mechanical actions from both surrounding soils and groundwater. To accurately simulate this interactive behavior, zero-thickness elements were used to simulate the contact characteristics at both enclosure wall-soil and dewatering well-soil interfaces. Based on the Coulomb’s friction theory and existing research [118,119], the ultimate shear slip (γ_crit) was set 5 mm and the friction coefficient (μ) was set 0.3.

In order to accurately simulate soil deformation and groundwater flow behavior during foundation pit dewatering, the following boundary conditions were introduced:

(1)

Displacement boundary: three-directional constraints were set on the bottom surface and asymmetrical sides of the model, and only normal-direction constraints were applied to the symmetrical surface;

(2)

Hydraulic boundary: the bottom and symmetrical surfaces of the model were set as impervious boundaries, the asymmetrical sides were set as constant water head boundaries, and the soil contact surface around the dewatering wells was defined by fixed water head boundaries.

2.3. Model Validation

Based on the actual engineering field pumping test, the effectiveness of the numerical model is verified. The test lasted for 3.2 days, with 25 wells installed inside the pit. Among these, 22 wells were activated for dewatering, while the remaining three wells were used to monitor groundwater level changes. Multiple observation wells, deformation monitoring points were arranged outside the pit to observe the water level changes of different aquifers, as well as the deformation degree of soil and enclosure.

The comparison of the measured and simulated values is shown in the figures below. Figure 3 shows the time-dependent of groundwater level; when the time of dewatering reaches 30h, the water level is basically stable. The maximum observed water level fluctuation reached 15 m compared to the initial water level (2 m), with Aquifer AqⅠachieving stabilization earlier than Aquifer AqⅡ. The trend of the measured value is generally consistent with the calculated value. Figure 4 shows the temporal variation of ground surface settlement. Both numerical simulation results and field monitoring data demonstrated a continuous increase in settlement magnitude. The discrepancy remained within 30%, indicating that the numerical model effectively simulates dewatering-induced ground settlement. Figure 5 further compares the computed and measured values of the enclosure deflection. The results indicate relatively minor deviations at the center point, while the error at the corner point is slightly larger. These variations primarily stem from simplified modeling assumptions regarding structural components, soil strata, and the dewatering process. Nevertheless, the built model remains considered reasonable because it is a good reflection of the actual situation.

3. Results and Discussion

3.1. Water Level Drawdown

Figure 6 shows the temporal and spatial distribution curves of groundwater drawdown outside the pit. In Figure 6a, the water level drawdown at 27 m depth under different L exhibits rapidly initial reduced and then stabilized, with all stabilization process completing within 24 h. In Figure 6b, the drawdown magnitude gradually decreases with increasing distance from the wall after dewatering stabilizes. The results show that the presence of HB (L = 1, 2, and 3 layers) can reduce the groundwater drawdown, and the maximum reduction range is more than 8m, which is validated across both temporal and spatial distributions. This shows that the HB can effectively block the seepage of groundwater into the foundation pit, so as to control the influence range of dewatering.

3.2. Ground Surface Settlement

Figure 7 presents the temporal and spatial distribution of ground surface settlement outside the pit. In Figure 7a, all cases show rapid initial ground surface settlement followed by asymptotic stabilization, and HB (L = 1 to 3 layers) accelerate stabilization by 24−48 h compared to the case of L = 0. Figure 7b shows that after reaching steady-state conditions, as the distance from the wall increases, the surface settlement pattern first increases, then decreases, and finally stabilizes. Similar to the effect of HB on groundwater drawdown, the presence of HB can reduce the surface settlement, with a maximum reduction exceeding 10 mm. This indicates that HB control ground surface settlement by effectively regulating groundwater level change.

3.3. Enclosure Wall Deflection

Figure 8 shows the variation curve of the enclosure deflection with time and the depth of the soil. Figure 8a shows that the deflection at the symmetry surface exhibits a trend of rapid initial increase followed by stabilization. Notably, when L = 0, the deflection decreases significantly after increasing to the peak, which may be attributed to stress redistribution in the soil caused by large deformations, leading to a rebound of the wall. In Figure 8b, the enclosure deflection increases first and then decreases with the increase of soil depth under steady-state dewatering conditions. Within the depth range of 20–32.5 m, the deflection with HB is generally less than that in the L = 0 case, because the HB is connected to the bottom of the wall to effectively limit its deformation. However, the comprehensive comparison shows that the wall deflection is basically greater in the presence of HB than in the L = 0 case. This phenomenon can be explained as follows: the HB restricts the groundwater drawdown and soil settlement, leading to the increase of water and soil pressure outside the pit, which in turn aggravates the deformation of the wall.

3.4. Effect of HB Layers (L)

Figure 9 shows the variation of the maximum water level drawdown (D_m) and ground surface settlement (δ_vm) outside the pit with the number of horizontal curtain layers (L). It can be observed that when L increases gradually, both D_m and δ_vm initially decrease and then increase, but the overall change amplitude is small, which indicates that different L values have a limited effect on D_m and δ_vm. This phenomenon may be attributed to the heterogeneity of the soil layers in which the multi-layer barrier is located. The varying permeability and other properties of different soil layers weaken the overall water-blocking and soil-retaining effectiveness of the barrier. The relatively small variation of D_m with L in Figure is primarily attributed to the following factors: the permeability coefficient of the horizontal barrier is already low, which significantly weakens the hydraulic connectivity inside and outside the pit. Consequently, changes in the number of the barrier layers have a minimal impact on altering the hydraulic connectivity. These findings suggest that increasing L provides only marginal improvements in controlling water level drawdown and ground surface settlement.

Figure 10 presents the variation curve of the maximum enclosure deflection (δ_hm) with L. As L increases, δ_hm shows a monotonically decreasing trend, with the reduction magnitude gradually diminishing. This indicates that while increasing L enhances the control effect on wall deformation, the marginal benefit of this effect progressively weakens. The observed phenomenon can be attributed to the following mechanisms: (1) optimization of seepage paths, where each additional barrier layer reduces the hydrodynamic pressure gradient; (2) effective redistribution of horizontal earth pressure to adjacent soil layers through layered configuration (resulting in lower stress concentration compared to single-layer systems); and (3) enhanced overall stiffness of the barrier system due to multi-layer arrangement, which further suppresses the deformation of the wall. From an engineering perspective, these mechanisms demonstrate that rational layering can improve lateral displacement control while maintaining constant material consumption, providing valuable insights for deformation control in deep excavation projects.

3.5. Maximum Surface Settlement and Wall Deflection

Figure 11 illustrates the correlation between δ_hm and δ_vm. To eliminate the influence of dewatering depth (H_d = 22 m), the normalization method is used to express both δ_hm and δ_vm as ratios to H_d, while comparing with research data from other deep foundation pit projects. The results demonstrate that the data points in this study are mainly distributed in the range of δ_hm = 0.46δ_vm and δ_hm = 1.74 δ_vm, whereas the δ_hm to δ_vm ratios from literature data concentrate within the range of 1.32–2.22. This significant discrepancy primarily stems from the dual mechanism of HB: (1) the soil settlement is significantly reduced by effectively controlling the seepage of groundwater; (2) it structurally constrains the lateral displacement of the enclosure wall.

4. Conclusions

This study investigates the control effects of HB on groundwater and foundation pit deformation, with particular focus on the influence of varying the barrier layers on groundwater drawdown, ground surface settlement, and enclosure deflection. The main findings are summarized as follows:

(1)

Control effects groundwater drawdown and ground settlement: HB significantly suppresses water level decline and associated soil settlement. Notably, increasing the number of barrier layers does not yield substantial additional control benefits, indicating that a single-layer HB is sufficient to achieve optimal water-blocking and settlement-reducing effects. Further increasing the barrier layers exhibits diminishing marginal utility.

(2)

Deformation characteristics of the enclosure wall: HB can enhance the overall stiffness of the foundation pit structure, and its influence on enclosure deflection shows an obvious depth effect. In shallow soil layers, the barrier alters the soil and water pressure distribution outside the pit, leading to increased wall deflection. In the deep soil layers, the synergy interaction between HB and the enclosure wall dominates, effectively controlling wall deformation.

(3)

Applicability analysis: while multi-layer HB improve the overall water-blocking capacity and structural stiffness, their higher construction complexity and cost must be considered. It is more suitable for large-scale foundation pit engineering or projects with high stringent groundwater control demands due to its construction demand index.