Research on Safety Risk Evaluation System for Seepage in Ship Lock Foundation Pit Engineering

Abstract

Ship lock project currently demonstrates a distinct cyclical pattern, accumulating latent hazards that pose a significant threat to project safety. Seepage safety (the condition in which the seepage risk is reduced to an acceptable level) serves as a crucial indicator in the safety risk assessment index system for ship lock project construction, thus necessitating an in-depth analysis of the risk factors impacting seepage safety. Utilizing a ship lock project in China as a case study, this study employs the finite element method (FEM) to analyze the seepage field of the ship lock foundation pit basin and proposes a comprehensive set of methods for risk evaluation and warning models pertaining to seepage safety risks in ship lock engineering. This study reveals that the obstruction of dewatering wells and imperfections in the diaphragm wall are the primary factors contributing to seepage damage. The investigation conducted a quantitative analysis of the impact of these two factors on the seepage field of the ship lock pit, considering pore pressure, water head, gradient, and flow velocity. A comprehensive set of evaluation indicators for seepage safety was formulated, drawing on the principles of multi-objective optimization, and a method for delineating the safe range of ship lock pit excavation under seepage action was proposed. Subsequently, an integrated seepage safety risk assessment system for ship lock pit excavation engineering was established. These research findings offer a scientific foundation for the management of seepage safety in ship lock pit excavation engineering and provide valuable references and guidance for the development of anti-seepage systems.

1. Introduction

Inland water transportation, as a crucial element of the national integrated transportation system and comprehensive utilization of water resources, offers numerous advantages, such as high transport capacity, minimal land usage, low energy consumption, limited environmental impact, and cost-effectiveness. It plays a pivotal role in driving economic and social development [1]. Ship locks serve as essential navigation structures for overcoming river water level differences and are vital for ensuring the smooth operation of inland water transportation [2,3,4,5]. However, the construction process of ship lock engineering involves intricate procedures and frequent vertical cross-over operations, significantly increasing operational safety risks. Among these risks, seepage presents the most prominent challenge. Seepage failure can result in soil instability or ground collapse, posing a threat to the safety of nearby structures and significantly impacting the overall construction of the project. Research indicates that seepage failure incidents constitute a significant proportion of deep foundation pit construction [6,7], with base leakage, sudden surge, and lateral soil erosion being the primary forms of seepage failure. Additionally, seepage failure may precipitate subsequent engineering quality issues in construction structures [8,9]. An appropriate anti-seepage system is typically necessary to mitigate the risk of seepage failure.

Effectively managing the seepage process within excavated foundation pits presents a significant technical challenge in the construction of ship locks. The foundational work of French hydraulic engineer Darcy, who derived the Darcy flow law through experimental research, has been instrumental in advancing seepage studies [10]. Terzaghi [11] and Rendulic [12] integrated the Darcy flow law into the three-dimensional consolidation theory, elucidating the principles governing pore water pressure distribution during seepage. Analytical methods [13,14,15,16] and discrete element techniques [17,18,19] are extensively employed to investigate basin seepage issues; researchers such as Wei [20], Hu [21], Yu [22], and Yuan [23] have utilized these approaches to address seepage challenges associated with ship lock foundations. However, it is evident that these methodologies primarily focus on specific case studies without addressing widespread issues related to seepage damage or establishing a comprehensive safety evaluation framework for seepage conditions. In contrast, the finite element method offers several advantages: it accurately captures complex constitutive relationships among rock and soil materials, excels in modeling intricate boundary conditions, and benefits from established methodologies and computational procedures. Notably, Tony [24], Bi [25], Wu [26], and others [27,28,29] have developed high-fidelity numerical models of ship lock foundation pits to analyze their respective seepage problems.

Conversely, the construction of the anti-seepage system represents a critical component in ensuring the seepage safety of ship lock projects. An analysis of seepage failure incidents at ship locks both domestically and internationally reveals that most failures stem from deficiencies within the anti-seepage system. Consequently, it is imperative to perform comprehensive qualitative and quantitative analyses on the uncertain indicators associated with this system to safeguard the seepage integrity of ship lock projects [30,31,32]. However, there exists a lack of stringent definitions and evaluative criteria for assessing the seepage safety of ship lock foundations, alongside an absence of relevant codes or standards for guidance. Risk evaluation models developed through empirical modeling [33] and machine learning techniques [34] typically necessitate extensive datasets; they cannot often be distilled into specific indicators. Multi-objective optimization is a methodology for addressing mathematical problems characterized by multiple potentially conflicting objective functions, thereby requiring trade-offs to achieve optimal solutions [35,36]. The primary aim of multi-objective optimization algorithms is to identify a set of equilibrium solutions—known as Pareto optimal solutions [37]—derived from balancing and reconciling various sub-objectives. Therefore, once the factors influencing seepage safety are identified, a comprehensive index reflecting these influences can be established using multi-objective optimization methods.

This study focuses on a ship lock project and employs the finite element method to analyze the seepage field of the ship lock basin. It investigates the impact of defects in the anti-seepage system on the seepage characteristics of the ship lock project, and it performs a sensitivity analysis of seepage parameters such as water head and slope. Drawing from the calculation results and principles of multi-objective optimization, it proposes a set of methods for evaluating seepage safety risks in ship lock projects, offering a scientific foundation for ship lock design and construction.

2. Seepage Calculation of Ship Lock Foundation Pit Engineering

2.1. Project Overview and FEM Model

The hub project features a regulating sluice gate and a Class IV ship lock. The hydraulic structures of the project are designed to meet Grade 2 standards. The designed discharge capacity of the regulating sluice gate is 3910 m³/s, with a flood control capacity of 4770 m³/s. In consideration of navigation planning and current traffic volume requirements, a double-lane ship lock has been planned on the left side of the existing ship lock. Both upper and lower lock gates for both new and old ship locks align with the regulating sluice gate, while the approach channel is asymmetrically arranged. Figure 1 illustrates the layout of the ship lock. The upper and lower lock gates for the double-lane ship lock are monolithic concrete structures resembling human figures, while the lock chamber takes on an inverted “Π” shape in monolithic concrete construction. The water distribution system follows a dispersed configuration.

Multiple instances of seepage were observed during the excavation of the ship lock foundation pit. This study focuses on studying the third occurrence, which occurred at the location of the upstream right bank guide wall. The inflow point was identified at the connection between the steel sheet pile of the upstream steel cofferdam and the old navigation wall, as illustrated in Figure 2.

2.2. FEM Model

A three-dimensional seepage analysis model was constructed utilizing field measurement data from the ship lock pit. The ship lock pit measures 43.8 m in width and 300.0 m in length, with a bottom elevation of 16.90 m. Its anti-seepage system comprises a diaphragm wall and dewatering wells. The left side of the foundation pit is sloped and excavated, featuring a concrete diaphragm wall with a thickness of 0.8 m, a bottom elevation of 7.0 m, and a top elevation of 34.5 m. On the right side of the pit slope, dewatering wells are arranged with a bottom elevation of 8.0 m, a length of 8.85 m, a diameter of 0.5 m, and an interval of 20 m between adjacent dewatering wells. The upper lock head water level stands at 32.50 m, and the lower lock head water level is at 27.90 m. A three-dimensional seepage calculation model was developed using ABAQUS 2022 software, employing the C3D8P element type. The finite element mesh comprises 441,939 elements and 468,776 nodes, as illustrated in Figure 3a. This chapter focuses on the impact of defects in the anti-seepage system on the seepage safety of the ship lock project, primarily considering indicators such as pore water pressure and gradient. Therefore, the construction process of the initial excavation of the foundation pit is not taken into account, i.e., the deformation index during the seepage process is disregarded. The specific arrangement of the anti-seepage system can be seen in Figure 3b. The parameters of the FEM model are detailed in

Table 1. Parameters of FEM model.

Part	Elastic Modulus/Mpa	Poisson Ratio	Osmotic Coefficient (cm/s)	Pore Ratio	Density/g/cm3	Critical Gradient
silty clay	8.0	0.15	3.0 × 10−6	0.30	1.8	1.3
silt loam	5.5	0.25	8.0 × 10−6	0.50	2.1	0.8
Silty-sandy interlayer	2.5	0.25	3.0 × 10−4	0.40	2.5	0.6
Sandy loam	6.5	0.35	4.0 × 10−4	0.65	1.9	0.6
silt loam	7.5	0.30	5.0 × 10−6	0.60	2.2	1.2
Fine sand	7.5	0.30	1.0 × 10−3	0.55	1.8	0.5
silt loam	10.0	0.15	6.0 × 10−6	0.30	3.2	1.2
dewatering well	12.0	0.32	6.0 × 10−1	0.30	2.5	-
diaphragm wall	7.5	0.30	7.3 × 10−8	0.60	2.1	-

, while Figure 4 depicts the distribution of soil layers and some pressure monitoring points’ locations in the field.

2.3. Principles of Multi-Objective Optimization

The fundamental approach to evaluating seepage risk (the probability of seepage failure due to various factors) involves transforming the practical engineering issue into an optimization problem and determining the critical risk value through numerical computation. Based on the principles of multi-objective optimization and numerical analysis, two key variables are chosen as independent factors for risk assessment: firstly, the well clogging rate, calculated as (standard working efficiency − actual working efficiency)/standard working efficiency; secondly, the proportion of defective area in the diaphragm wall, defined as the ratio of defect area to the total selected sectional area. These quantitative indicators enable effective determination of the critical seepage damage value, facilitating prediction and control of seepage risk. The objective function is formulated by the following equation:

$$d\left(K_i\right)=={\displaystyle \sum _{i=1}^n{\lambda }_{i}d}\left[F\left(X_i,K_i\right)\right]$$

(1)

In the formula, d(K_i) represents the objective function, d [F(X_i, K_i)] represents the effectiveness function of each evaluation index, X_i denotes the number of evaluation indicators, and λ_i signifies the weight of the effectiveness function. The independent variables K₁ and K₂ represent the rate of clogging of dewatering wells and the proportion of defective areas in the diaphragm wall, respectively. The technical route for seepage risk assessment is illustrated in Figure 5.

3. Safety Risk Assessment System for Seepage in Ship Locks

3.1. Seepage Calculation Result

This study examined the seepage barrier system under three distinct calculation scenarios: standard conditions, clogging of dewatering wells, and defects in diaphragm walls. Three representative cross-sections were chosen from the 3D model, as depicted in Figure 6. Cross-sections 1-1 and 3-3 correspond to horizontal and vertical views of dewatering wells, while cross-section 2-2 corresponds to the location of the diaphragm wall defect.

Figure 7 illustrates the comparison between the observed and calculated pore pressure values at monitoring points within the barrier system under normal operating conditions. The figure demonstrates that the model proposed in this study effectively captures the trend of pore pressure decline at each monitoring point, with a high degree of agreement between observed and calculated values.

Figure 8 illustrates the pore pressure distribution in various sections, clearly demonstrating the dissipation of pore pressure due to rainfall wells and diaphragm walls. The selected sections indicate that the ship lock’s waterproof system has effectively lowered the zero-pore pressure surface to a depth of less than 2 m below the slab, meeting construction requirements. The intersection of the wetted line with the slope represents the overflow point. Figure 9 depicts the water head flow network, revealing no outflow across the entire cross-section. As shown in Figure 10, under the normal operation of existing waterproof systems and rainfall wells, maximum water head gradient reduction during excavation is measured at 0.46, which falls below the critical gradient reduction value of 0.60 for soil layers, indicating a relatively low risk of hydraulic failure.

3.2. Evaluation Index

3.2.1. Silting of Dewatering Well

When the dewatering well operates under normal conditions, its water pumping rate is 960 m³/day. Considering variations in operational efficiency, the dewatering rates decrease by 10%, 20%, and 30%, respectively. Figure 11 illustrates the pore pressure distribution within the dewatering well at varying degrees of clogging. It is evident that a decrease in dewatering efficiency leads to a significant increase in pore pressure. When the efficiency decreases by more than 20%, effective dissipation of pore pressure becomes unattainable, causing the zero-pore pressure boundary to rise to the bottom of the excavation and pose a threat to seepage stability.

Figure 12 illustrates the hydraulic network diagram of the dewatering wells under varying degrees of clogging. With the decreasing efficiency of the dewatering wells, the water headline noticeably increases. The points where the infiltration line intersects with the bottom and slope of the excavation are identified as overflow points. It is evident that when the dewatering efficiency decreases by more than 20%, overflow points emerge at both the bottom slab and the right slope of the excavation.

Figure 13 illustrates the variation in water head gradient for a well experiencing different degrees of clogging. It is observed that layer ⑥₂ exhibits the highest gradient ratio, indicating that clogging significantly impacts the drainage efficiency of this layer, leading to delayed discharge of porous water and consequently resulting in reduced water level difference and a smaller gradient. Therefore, assessing the influence of well clogging on seepage safety solely based on gradient may not accurately reflect real conditions.

Figure 14 illustrates the variation in flow velocity (flvel) distribution under varying degrees of groundwater well clogging. It is evident that with an increase in the degree of groundwater well clogging, the outflow capacity diminishes. Inadequate discharge of groundwater may lead to potential overflow at the slope position adjacent to the well, thereby inducing seepage damage.

In summary, the velocity distribution map surrounding the dewatering well indicates that blockages within the well hinder the timely discharge of accumulated groundwater. This obstruction directly leads to an increase in pore water pressure and hydraulic head within the foundation pit. Conversely, such blockages also result in a reduced water level differential in the foundation pit, which effectively diminishes the gradient present therein.

3.2.2. Defect in Diaphragm Wall

Select model sections on the left and right of section 2-2 within a 0.5 m length to quantitatively analyze the impact of diaphragm wall defects on the safety of basement pit seepage. Evaluate the effects of diaphragm wall defect areas measuring 1.5 m², 3 m², and 4.5 m² on basement pit seepage safety. Figure 15 illustrates the distribution of pore pressure under varying degrees of diaphragm wall defects, indicating a significant increase in pore water pressure with defective walls. When the area of the diaphragm wall defect exceeds 3 m², the zero-pore pressure boundary rises above the bottom of the basement pit, elevating the seepage failure risk.

Figure 16 illustrates the hydraulic network diagram depicting water head flow under varying degrees of defects in the diaphragm wall. As the area of the diaphragm wall defect increases, there is a noticeable elevation in the water headline. It is evident that when the area of the diaphragm wall defect exceeds 3 m², overflow occurs at the bottom sluice gate and on the right slope of excavation, significantly elevating the likelihood of seepage damage in these areas.

Figure 17 illustrates the gradient distribution in the vicinity of the defect location under varying degrees of diaphragm wall defects. The presence of a defect in the diaphragm wall leads to an influx of water through the defective area, resulting in an elevation of the water level difference at that location and, consequently, an increase in the gradient. Upon reaching an area of 4.5 m², the value of the gradient exceeds the critical threshold for the soil layer.

Figure 18 illustrates the flow velocity distribution under varying degrees of diaphragm wall defect conditions. It is evident that with a smaller defect area, water flow ascends around the defective position. As the defect area expands, the flow velocity around the defective position increases and exhibits an upward impact trend, potentially leading to water overflow onto the slab and consequent seepage damage.

It is evident that when a defect occurs in the diaphragm wall, water infiltrates through the breach and impacts the diaphragm. As the size of the defect enlarges, there is a marked increase in pore water pressure within the excavation, leading to a significant rise in water head. Furthermore, the gradient at the location of the defect also intensifies due to an augmented difference in the water head on either side.

3.3. Performance Function Based on the Efficiency of a Precipitation Well

The calculation results in Section 2 indicate that the left lower corner of the excavation is particularly vulnerable to permeability failure when defects are present in the impermeable system. Consequently, a monitoring point located 3 m below the left lower corner was selected. The relationship between the water head H at this monitoring point and the clogging rate K₁ of the dewatering well is depicted on the σ_p~K₁ plane, as illustrated in Figure 19a. As K₁ increases, there is a continuous rise in the water head at the monitoring point. When this exceeds 3 m, it signifies that the water level has surpassed the bottom slab of the excavation, indicating a potential risk of permeability failure. The evolving relationship between H and K₁ aligns with an exponential growth pattern.

The relationship between the gradient ratio J_C at the monitoring point and the efficiency of well K₁ is depicted in the J_C~K₁ plane, as illustrated in Figure 19b. With an increase in K₁, inadequate dissipation of pore water pressure leads to a marginal change in water level difference at the monitoring point, resulting in a lower gradient compared to normal operation. Overall, as K₁ increases, the gradient at the monitoring point does not surpass the critical threshold but exhibits a declining trend following a quadratic function.

In order to standardize the performance function and remove its dimensional units, we normalize it using the following calculation equation:

$$d\left[F\left(X_i,K_i\right)\right]={\displaystyle \frac{F\left(X_i,K_i\right)-F{\left(X_i,K_i\right)}_{\min }}{F{\left(X_i,K_i\right)}_{\mathrm{m}ax}-F{\left(X_i,K_i\right)}_{\min }}}$$

(2)

The normalized function for the objective function, with the head and gradient as targets, is expressed by Equations (3) and (4):

$$d\left[F\left(H,K_1\right)\right]=0.123\ast \left[1-\exp \left(K_1/0.451\right)\right]$$

(3)

$$d\left[F\left(J_C,K_1\right)\right]=2\ast \exp \left(-K_1/1.372\right)-1$$

(4)

Figure 20 illustrates the comparison of the performance function regarding the clogging rate of dewatering well K₁. As the clogging rate of dewatering well K₁ increases, the water head at the monitoring point rises. Upon reaching a certain degree of clogging, the water head exceeds 3 m, potentially leading to overflow at the base or above excavation, following along the direction of permeable water flow and possibly forming a seepage channel, resulting in pipe burst and soil erosion. Conversely, dewatering well clogging constrains water level decline, leading to a reduced water level difference within the soil layer and decreased gradient; thus, a gradient at monitoring points is maintained consistently lower than critical levels. Specifically, when K₁ reaches 0.273, water head H surpasses its critical value; henceforth, the clogging rate of dewatering well K₁ is selected as its performance function based on changes in water head H.

3.4. Performance Function Based on Ground Wall Defects

By selecting the ratio of defect area to section area (the height of diaphragm wall x section length) as the target function K₂, a plot is generated to illustrate the relationship between water head H and K₂ at the monitoring point, as depicted in Figure 21a. As K₂ increases, there is a corresponding rise in water head, leading to an increased risk of permeability damage when the water level exceeds the bottom plate height. The relationship between H and K₂ follows a quadratic function. Furthermore, a plot illustrating the relationship between the gradient J_C and K₂ at the monitoring point is presented in Figure 21b. The defect of the diaphragm wall results in localized elevation of water levels, contributing to an augmented difference in water head at the monitoring point and an increase in gradient. Reaching a certain degree of defect within the diaphragm wall leads to exceeding the critical gradient at the monitoring point.

The normalized function for the objective function with water head and gradient as targets is expressed by Equations (5) and (6), respectively:

$$d\left[F\left(H,K_2\right)\right]=3.01K_2^2+0.249K_2$$

(5)

$$d\left[F\left(J_C,K_2\right)\right]=1.462\ast \left[1-\exp (-(K_2/1.159))\right]$$

(6)

Figure 22 illustrates that, at K₂ values of 0.192 and 0.205, respectively, the functions (F(H,K₂)) and (F(J_C,K₂)) attain their critical values. In general, in comparison to changes in water head, the gradient is more likely to reach its critical value. Therefore, for safety considerations, the efficiency function for K₂ should be based on the gradient at the measurement point.

3.5. Objective Function for Risk Assessment

The above analysis indicates that the primary deficiencies in the anti-seepage system stem from the imperfections of the diaphragm wall and the clogging of dewatering wells. Furthermore, damage to the anti-seepage system typically occurs concurrently rather than independently, thus allowing for the derivation of a comprehensive risk evaluation index for seepage damage based on these two factors. Weighting factors can be allocated to each factor according to their respective proportions in actual engineering projects, resulting in an expression for the comprehensive evaluation index as follows:

$$d\left(K\right)={\lambda }_1\left[0.123\ast \left(1-\exp \left(K_1/0.451\right)\right)\right]+{\lambda }_2\left[1.462\ast \left[1-\exp (-(K_2/1.159))\right]\right]$$

(7)

In this approach, K represents the comprehensive risk assessment index, d(K) denotes the comprehensive evaluation function, and λ₁ and λ₂ are the respective weights assigned to these two factors. Figure 23a illustrates the variation of the objective function d(K) under different weight combinations, clearly delineating the critical value range Kp corresponding to changes in the evaluation function. Figure 23b depicts alterations in safety margins across various weight allocation scenarios, categorized based on the critical value Kp.

Consequently, the water head H and the gradient H are identified as critical indicators for assessing the safety range. In scenarios involving silting of dewatering wells, the water head H should be treated as a functional variable. Conversely, when defects in the diaphragm wall are present, the gradient J_C should be regarded as a functional variable. Factors K₁, associated with well plugging, and K₂, related to continuous wall defects, are analyzed alongside field monitoring data to derive functions for each factor along with their respective weights. This process culminates in a comprehensive risk function that effectively delineates the safety interval concerning seepage. Should the calculated comprehensive risk assessment index K exceed this established safety threshold, appropriate anti-seepage measures should be implemented—such as utilizing reinforced materials for continuous walls or establishing redundant drainage systems.

4. Conclusions

This study developed a FEM model to comprehensively analyze the seepage field in the ship lock pit post-dewatering, with a specific focus on evaluating the seepage safety of the ship lock pit under conditions of anti-seepage system defects. A comprehensive seepage safety assessment system was established. The key findings are as follows:

(1)

The numerical analysis indicates that when the anti-seepage system is intact, the seepage resistance of the anti-seepage wall remains within normal parameters, and drainage in the basement operates effectively. The zero-pressure boundary is situated 2 m below the bottom slab, with each soil layer’s gradient remaining below the critical threshold, thereby satisfying design specifications.

(2)

Clogging of dewatering wells and defects in the diaphragm wall significantly influence the seepage characteristics within the pit. Increased severity of clogging in dewatering wells correlates with a larger area of the diaphragm wall defects, resulting in heightened water head elevation. Furthermore, assessing seepage failure based on gradient metrics proves inadequate under conditions characterized by dewatering well clogging.

(3)

An effectiveness function was derived from two risk factors: dewatering well clogging and diaphragm wall defects; distinct weights were assigned to these factors to formulate a comprehensive evaluation function for seepage failure risk. This approach enables an effective delineation of safe ranges concerning seepage actions.

The findings presented herein provide valuable insights for constructing anti-seepage systems and conducting safety evaluations related to seepage in ship lock pit engineering.