Performance of Diaphragm Walls during Ultra-Deep Excavations in Karst Areas: Field Monitoring Analysis

Deep foundation pit excavations have become more extensive for the construction of underground spaces with rapid urbanization. Diaphragm walls are commonly used to support deep excavations. However, due to the complex geological conditions in karst areas, construction accidents frequently occur during the excavation of foundation pits. This study aims to investigate the performance of diaphragm walls in karst areas through ﬁ eld monitoring analysis. A kick-in deformation mode of the diaphragm wall is revealed during the foundation pit excavation. Furthermore, the results show that the diaphragm walls present multiple deformation modes rather than a single mode. Additionally, this study proposes a method to calculate the lateral displacement of the diaphragm walls at different depths. It is found that the karst caves have a considerable impact on the stability of diaphragm walls, as demonstrated by their lateral displacement. The hidden karst caves reduce the bearing capacity of the bedrock, rendering it insuf ﬁ cient to resist the active earth pressure. As a result, the bottom of the diaphragm wall is kicked into the foundation pit, causing signi ﬁ cant lateral displacement and posing risks during excavation. The ﬁ ndings of this study contribute to the design and construction of similar excavations in karst areas.


Introduction
The rapid development of urbanization has prompted the large-scale construction of underground spaces for highrise buildings and subway transportation.As a result, deep excavation engineering has become more extensive.The shape of the excavation is typically designed as a rectangle [1][2][3][4][5], circle [6,7], or irregular polygon [8,9] based on the superstructure of the buildings and the geological conditions.Construction accidents frequently occur during the excavation of foundation pits, especially in karst areas, due to the complex geological conditions.Therefore, it is crucial to investigate the performance of foundation pits in order to reduce the risk associated with these excavations.
Numerous studies have been conducted to investigate the performance of excavations [10][11][12][13][14][15][16][17][18].These studies utilized theoretical analysis, physical model tests, numerical simulations, and field monitoring.Compared with the other three methods, field monitoring analysis can reflect the realistic stress of the construction sites, and hence, the performance of the excavations can be obtained accurately [19][20][21].For instance, Clough and O'Rourke [22] examined the movements of in-situ walls and clarified ground movement patterns by updating the existing database with information from both conventional and new systems.They revealed three typical profiles of movement for braced and tied-back walls.Wang et al. [23] investigated the displacement of retaining walls and the settlements of the ground based on 300 cases of deep excavations in Shanghai soft soils.They found that the ratio of the maximum ground settlement to the maximum lateral displacement of walls ranges from 0.4 to 2.0.Lin et al. [24] proposed a high-risk target model that incorporated the TOPSIS method with hybrid fuzzy sets.They employed a case study of an excavation in Tianjin province, China, to demonstrate the capabilities of the model.Diaphragm walls, also known as D-walls, are commonly used to provide support for deep excavations due to their high stiffness, strong integrity, and excellent impermeability.The deformation and displacement of diaphragm walls are key parameters that reflect the stability of the excavations.Consequently, extensive investigations have been conducted on the performance of diaphragm walls [25][26][27][28][29][30][31][32][33].Karst formations are widely distributed in southern China, including Guangdong, Guangxi, and Hunan.Numerous accidents have occurred during the excavation of foundations due to the presence of karst caves.However, there have been few studies investigating the performance of diaphragm walls in excavations located in karst areas.Karst areas are characterized by the presence of dissolved rock formations, such as limestone, that can create karst caves, potentially leading to instability of supporting structures and ground collapse during deep foundation pit excavation.The impact of karst caves on the stability of diaphragm walls remains unclear.
This study conducted a field monitoring analysis of diaphragm walls in karst areas to assess the performance of D-walls.First, the conditions of the project, including the location and size of the excavation, geological condition, supporting form, and excavation scheme, are presented.Subsequently, the monitoring system for the excavation is described.A method is proposed to calculate the lateral displacement of diaphragm walls by utilizing the deformation and horizontal movement of the top wall.Lastly, the discussion focuses on the deformation and displacement of diaphragm walls to illustrate the effects of karst caves.The data from this study can assist in calibrating theoretical methods and numerical models.Furthermore, the findings of this study can provide valuable insights for the design and construction of similar foundation pit excavations.As depicted in Figure 1, the excavation was designed in the shape of an irregular polygon.The excavation has a length of 588 m and a width of 168 m.The irregular shape and large size of the excavation heighten the risk associated with the foundation pit excavation.In order to induce the risk of excavation, the foundation pit was divided into five zones (i.e., Zone 1-Zone 5 in Figure 1   Limestone belongs to soluble carbonate rocks that are easily dissolved by the rapid downward movement of groundwater.Subsequently, karst caves are formed within the limestone layer, as depicted in Figure 3. Based on the geology profiles, limestone is widely distributed throughout the excavation site.Furthermore, Guangzhou experiences abundant rainfall, resulting in a high water table.Consequently, karst caves are extensively formed within the excavation site, as shown in Figure 2, thereby increasing the risk associated with the excavation.

Construction Phases and Retaining
Structures.The construction phases of the excavation are presented in Tables 3  and 4. The excavation of the foundation pit lasts approximately 2 years, from June 12th, 2020, to May 8th, 2022.It should be noted that March 26th, 2019, the commencement date of the excavation monitoring, is designated as day 0 in this study for the purpose of presenting the results conveniently and succinctly in the subsequent sections.
Figure 2 displays the schematic cross-section views of the excavation's retaining structures.The first struts of Zone 1-Zone 4 are located at a depth of 5.0 m, while the first strut of Zone 5 is located at a depth of 1.3 m.The second struts of Zone 1-Zone 4 are located at a depth of 10.0 m, while the strut of Zone 5 is at a depth of 6.85 m.The third struts of Zone 2-Zone 4 are located at a depth of 16.5 m, while the third struts of Zone 5 are at a depth of 13.15 m.The fourth struts of Zone 3-Zone 5 are located at a depth of 20.45 m.The last struts of Zone 4 and Zone 5 are located at a depth of 26.45 m.

Methodology of Monitoring
3.1.Monitoring Equipment.The horizontal displacement at the top of the D-walls was measured using a TS16 electronic total station with a precision of AE1.0 mm.A positive displacement indicates movement toward the excavation, while a negative displacement indicates movement away from the excavation.The deformation of the D-walls was measured using SINCO50302510 clinometers with a precision of 0.02 mm per 500 mm.The location of the monitoring points is illustrated in Figure 4.
This paper presents the theory of clinometer measurement to propose a method for calculating the lateral displacement of D-walls.As depicted in Figure 5(a), guide casing is pre-embedded in the diaphragm wall.When measuring the deformation of the D-walls, the inclinometer probe is inserted at the bottom of the D-walls and then lifted upward.The inclination angle (θ i , i = 1, 2, …, n) of the D-walls can be measured successively (as shown in Figure 5).The deformation of the D-walls at each depth (Δs i ) can be calculated using the following equation: where L is the distance between successive readings.It should be noted that the calculation of deformation assumes that the bottom of the D-walls is fixed.
3.2.Method to Calculate the Lateral Displacement of the D-Walls.The deformation of the D-walls is equal to the lateral displacement only when the bottom of the D-walls is fixed.However, if there are unknown karst caves in the foundation pit, the bearing capacity of the D-walls may not be sufficient to withstand the active earth pressure.Consequently, the bottom of the D-walls may be pushed inward during the excavation of the foundation pit.The lateral displacement of the D-walls will not be equal to the deformation.Therefore, a method is proposed to calculate the lateral displacement of the D-walls.Figure 6 illustrates the six statuses of the deformed Dwalls based on the position of their top and bottom, encompassing all possible scenarios.In cases (A) and (B), both the top and bottom of the D-walls move toward the foundation pit.In case (A), the top of the D-walls is positioned on the right side of the vertical line, while in case (B), it is located on the left side of the vertical line.In case (A), the top of the D-walls is positioned on the right side of the vertical line, while in case (B), it is located on the left side of the vertical line.In case (A), the top of the D-walls is positioned on the right side of the vertical line, while in case (B), it is located on the left side of the vertical line.In contrast to case (E), the bottom of the D-walls moves away from the foundation pit, while the top of the D-walls moves toward the foundation pit in case (F).
The lateral displacement of the bottom of the diaphragm walls (δ b ) can be calculated by utilizing the horizontal displacement and the deformation at the top of the D-walls, as follows:  The horizontal displacement at different depths of the D-walls (δ i ) can be calculated by utilizing the lateral displacement at the bottom of the D-walls and the deformation at the corresponding depth (Δs i ), as follows:     7(a), the horizontal displacement of the D-walls in Zone 1 gradually increases during the excavation of the foundation pit.At day 950, the horizontal displacement of A181 (monitored at point S62) and A191 (monitored at point S66) experienced a sharp increase as the struts were removed.The displacement reached its maximum values within approximately 10 days (51.6 mm for A181 and 33.0 mm for A191).The removal of struts is a construction process associated with high risk.Therefore, timely replacement of supports is crucial for mitigating construction risks.As depicted in Figure 7(b), the horizontal displacement of A123 (monitored at point S44) in the external corner of the foundation pit rapidly increased and reached its maximum value of 19.0 mm during excavation.In contrast, the horizontal displacement of A140 (monitored at point S49) in the internal corner of the foundation pit was minimal, with a maximum value of only −1.5 mm.This is because the D-walls experience higher pressure in the external corner compared to the internal corner.

Results and Discussion
Figure 7(c) illustrates that the top of the majority of D-walls in Zone 3 exhibited movement away from the foundation pit during excavation, which contradicted the findings of Wang et al. [34].The top horizontal movement of A9 (monitored at point S3), A16 (monitored at point S5), and A211 (monitored at point S72) exhibited rapid increases following the completion of the final level excavation.To A Figure 7(e) illustrates the horizontal displacement at the top of D-walls in Zone 5.The displacement in Zone 5 was comparatively smaller, and the D-walls exhibited greater stability than that observed in the D-walls of other zones.This is attributed to the limited presence of karst caves in the bedrock of Zone 5, which provides sufficient bearing capacity to withstand the earth's pressure.The displacement of A104 (monitored at point S38) and A122 (monitored at point S43) increased between day 567 and day 657 due to the substantial construction load.Following the removal of the construction load, the horizontal displacement of the diaphragm walls decreased.

Deformation of the D-Walls.
The identification of distinct deformation modes of the D-walls during excavation is crucial for design purposes.Clough and O'Rourke [22] observed three distinct deformation modes in the in-situ walls, as depicted in Figure 8.It is evident that the walls deformed toward the foundation pit in all modes as a result of the active earth pressure.Figure 9 illustrates the deformation of the  Advances in Civil Engineering representative D-walls during the excavation of the foundation pit in this study.Multiple deformation modes can be observed within the same D-walls during excavation, which differs from the single parabolic deformation mode reported by Wang et al. [34] in soft ground.Furthermore, an unexpected deformation mode can be observed, as depicted in Figure 10.The lower portion of the D-wall deforms toward the excavation, while the upper portion deforms toward the outside of the foundation pit.For instance, A146 (monitored at J53) and A157 (monitored at J55) exhibited this deformation mode.This deformation mode is referred to as the kick-in mode in this study.
To provide a clear observation of the D-wall deformation process during excavation, examples are illustrated in Figure 11.The deformation process of D-wall A166 (monitored at J58) is illustrated in Figure 11(b).Following the completion of Level 1 excavation (at Stage 1), A166 exhibited minimal deformation.Upon completion of the Level 2 excavation (at Stage 2), A166 displayed cantilever deformation.The maximum deformation, located at the top of A166, reached 32.5 mm.Upon completion of the first strut, the maximum deformation decreased to 27.3 mm.After a period of 312 days, the parabolic deformation mode was observed.The maximum deformation, at a depth of 14.5 m, reached 33.4 mm.

Advances in Civil Engineering
The deformation process of D-wall A181 (monitored at J62) is illustrated in Figure 11(c).The cantilever deformation mode was observed at both Stage 1 and Stage 3. Upon completion of the first and second struts, the cantilever deformation mode transitioned to the parabolic deformation mode at Stage 5. A181 experienced a maximum deformation of 20.3 mm at a depth of 13 m.Upon removal of the angle strut, the parabolic deformation mode reverted back to the cantilever deformation mode.The deformation of the D-wall rapidly increased.At Stage 6, A181 experienced a significant maximum deformation of 64.5 mm, resulting in the development of cracks.
The deformation process of A1 in Zone 3 is depicted in Figure 11(d).The deformation process of A1 exhibited similarities to that of A146.A maximum deformation of −40.5 mm was recorded at Stage 5. Upon completion of the third strut and excavation of the foundation pit to Level 4 (at Stage 7), the maximum deformation of A1 decreased to −5.6 mm.However, upon excavation of the foundation pit to Level 5, the deformation mode transitioned to the kick-in deformation mode, resulting in increased deformation.The deformation at the top of A1 escalated to −33.9 mm.The lower portion of the D-walls deformed toward the foundation pit, with a maximum deformation of 21.4 mm at a depth of 26 m.
In summary, the excavation of foundation pits in karst areas reveals four distinct deformation modes of D-walls.Prior to the construction of any struts, the D-walls exhibit a cantilever mode during excavation, as depicted in Figure 8(a).Once the first strut is constructed, the D-walls exhibit a parabolic mode due to the constraint at the top, resulting in minimal deformation (Figure 8(b)).As the excavation depth increases and additional struts are constructed, the D-walls transition into a combined mode characterized by increased deformation and deformation toward the foundation pit (Figure 8(c)).In the presence of hidden karst caves within the foundation pit and the absence of any surcharge along its edges, the deformation mode of the D-wall exhibits the kick-in deformation mode (Figure 10).During the excavation process, if monitoring data indicate the kick-in deformation mode of D-walls, it is likely that hidden karst caves are present in close proximity to the D-walls.Boreholes should be conducted to identify the presence of karst caves, and appropriate treatment measures should be implemented to mitigate significant displacement of the D-walls and ensure safe excavation.

4.3.
Lateral Displacement of the D-Walls.As mentioned in previous sections, the lateral displacement of the D-walls is equivalent to the lateral deformation, but only when the bottom of the D-walls is fully constrained, resulting in zero displacement.However, the bottom of the D-walls frequently shifted toward the foundation pit due to the presence of unknown karst caves in karst areas.Therefore, the lateral displacement at the bottom of the D-walls should be computed using Equation (2), while the lateral displacement at different depths of the D-walls should be calculated using Equation (3).
Figure 12 illustrates the variations in the lateral displacement of representative D-walls (A146, A166, A181, and A1) throughout the excavation of the foundation pit. Figure 12(a) displays the lateral displacement of A146.It is evident that the D-wall uniformly shifted toward the foundation pit due to the active earth pressure.At Stage 3, the maximum displacement, located at a depth of 10.0 m, was recorded at 6.3 mm.Following the completion of Level 3 excavation, the displacement at the lower position of the D-wall exhibited a rapid increase, with a maximum displacement of 23.6 mm at a depth of 16.0 m.Additional boreholes were conducted near A146 in the foundation pit to investigate the cause of this response.Unexpected karst caves were observed (as depicted in Figure 13(a)).The active earth pressure compressed and caused the collapse of the karst caves, thereby reducing the bearing capacity of the bedrock and resulting in a rapid increase in displacement at the lower position of the D-wall.After a period of 312 days (at Stage 6), the maximum displacement increased to 62.9 mm at a depth of 17.0 m.The D-walls uniformly shift toward the foundation pit under the active earth pressure and surcharge.The presence of unknown karst caves results in a larger displacement at the lower position of the D-walls compared to the upper position.In this study, this behavior exhibited by the D-walls is classified as a Type A response, as depicted in Figure 13(b).Advances in Civil Engineering A166 (as depicted in Figure 11(b)) due to the complete fixation of its bottom in the bedrock.Following the completion of Level 1 excavation at Stage 1, A166 exhibited a maximum displacement of merely 1.7 mm.The displacement of the Dwall increased as the foundation pit excavation progressed.At Stage 5, the maximum displacement of the D-walls reached 22.9 mm at a depth of 8.5 m.After a period of 312 days (at Stage 6), the maximum displacement of the D-walls measured 33.1 mm at a depth of 14.5 m.
Figure 12(c) illustrates the lateral displacement of A181 throughout the excavation of the foundation pit.The displacement of the D-wall was relatively negligible when the  The bearing capacity of the bedrock proved inadequate in withstanding the active earth pressure.

Summary and Conclusions
This study investigates the performance of diaphragm walls during the excavation of ultra-deep foundation pits in karst areas through field monitoring analysis.The horizontal movement at the top and the deformation of the diaphragm walls were monitored.The proposed method enables accurate calculation of the lateral displacement of the diaphragm walls at different depths.
The study yields the following conclusions: (1) In addition to the three typical deformation modes (i.e., cantilever model, parabolic mode, and combined mode), this study identifies a new kick-in deformation mode of the diaphragm walls in the karst area.Furthermore, during the excavation of deep foundation pits, the diaphragm walls exhibit multiple deformation modes instead of a singular mode.(2) The embedded depth of diaphragm walls should be increased during the design of the excavation to prevent kick-in failure in the karst area.During the construction of the excavation, if the kick-in deformation mode is observed in the monitoring data, borehole drillings should be conducted to identify hidden karst caves.Subsequently, appropriate treatment measures should be implemented to mitigate significant displacement of the diaphragm walls and ensure a safe excavation.(3) Further investigation through physical model experiments and numerical simulation analysis is required to uncover the mechanism of how karst caves affect the performance of diaphragm walls.Despite its limitations, the findings of this study can offer valuable insights for the design and construction of deep excavations.
(a)).The width of each zone is provided in Figure 1.The depths of Zone 1 and Zone 2 are 15.9 and 24.56 m, respectively.Both of the depths of Zone 3 and Zone 4 are 34.68 m.The depth of Zone 5 is 35.18 m.

FIGURE 1 :
FIGURE 1: Schematic and layout map of the excavation: (a) shape and size of the excavation and (b) layout map of the excavation.

4
Advances in Civil Engineering line whereas Δs ð−Þ represents the top of the D-walls is on the left side of the vertical line.

4. 1 .
Horizontal Displacement at the Top of D-Walls.
Figure 7   illustrates the horizontal displacement at the top of the

FIGURE 6 :
FIGURE 6: Schematics of the deformation and the displacement of the diaphragm walls: (a) Case A, (b) Case B, (c) Case C, (d) Case D, (e) Case E and (f) Case F.

Figure 12 (
b) illustrates the lateral displacement of A166 throughout the excavation of the foundation pit.The displacement of A166 closely aligned with the deformation of

FIGURE 10 :
FIGURE 10: Kick-in deformation mode of D-walls found in Karst areas.

Figure 12 (FIGURE 12 :
Figure 12(d)  illustrates the lateral displacement of A1 throughout the excavation of the foundation pit.It is observed that the top of the D-wall shifted outward from the foundation pit after the completion of the first level excavation at a depth of −5.0 m.The D-wall started to tilt.As the foundation excavation progressed, the angle of inclination of the D-wall continued to increase.This is attributed to the presence of karst caves in the foundation pit, as depicted in 2.1.Location and Size of the Excavation.Guangzhou is situated in southern China, specifically in the lower reaches of the Pearl River.The excavation investigated in this study is situated in the northwest of Guangzhou, near the Shijin River.

2
Advances in Civil Engineering 2.2.Geological Condition in Karst Area.The geological profiles of these five zones are depicted in Figure2.The subsoils comprise fill, muddy clay, medium-coarse sand, and silty clay.Calcareous shale and limestone are underlying the soil layers.The thickness of soils and rocks varies significantly among different excavation zones.Table1displays the soil properties, including natural density, water content, void ratio, cohesion, friction angle, coefficient of compressibility, modulus of compressibility, and deformation modulus.Table2presents the properties of the calcareous shale and limestone, including the coefficient of compressibility, modulus of compressibility, deformation modulus, and standard value of uniaxial ultimate compressive strength.

TABLE 1 :
Soil properties of clay and sand in the excavation site.

TABLE 2 :
Properties of rock in the excavation site.

TABLE 3 :
Schedule of construction phases of Zone 1-Zone 3.

TABLE 4 :
Schedule of construction phases of Zone 4 and Zone 5.