Study on Deformation Law of Deep Foundation Pit with the Top-Down Method and Its Influence on Adjacent Subway Tunnel

Based on a long-term comprehensive instrumentation program, the performance of an excavation pit constructed by the top-down method in downtown Kunming was extensively examined. The measured excavation responses included the deﬂections of diaphragm wall, vertical column movement, slab axial force, building settlement caused by ground deformation, and the inﬂuence of pit excavation on the adjacent subway tunnel. This paper analyses the monitoring data of the existing construction stage. Based on the analyses on the data of ﬁeld and numerical simulation, the following major ﬁndings were obtained: (1) the relationship between the measured maximum wall deﬂection, δ hm , and excavation depth, H , in this study is δ hm � 0 . 06% H ∼ 0 . 27% H , which is quite diﬀerent compared with the relationship of soft-soil pit δ hm � 0 . 02% H ∼ 1 . 2% H , but closer to the normalized curve of rock-socketed pile δ hm � 0 . 01% H ∼ 0 . 45% H and rock-socketed diaphragm wall δ hm � 0 . 031% H ∼ 0 . 129% H . (2) The relationship between the maximum settlement of column ( δ p ) and excavation depth ( H ) is δ p � − 0 . 09% ∼ 0 . 04% H . The maximum distortion between the diaphragm wall and the column is less than 1/500 of the limit range proposed by Bjerrum. (3) The impact range caused by excavation is about 3.8 times the maximum excavation depth. The ground settlement around the foundation pit is groove type, and the maximum settlement point is located at 2.7 times the maximum excavation depth. (4) The excavation of the foundation pit leads to the maximum vertical settlement of 2mm and maximum horizontal displacement of 5.2mm in the subway tunnel; the maximum change of axial force and bending moment are 8.8kN (the vertical direction) and 6.4kN ·


Introduction
e top-down method has the advantages of little influence on the surrounding environment and saving construction time. It has been widely used since it was put forward in 1933 by Japanese engineers, and many scholars have studied it. Fan and Zhao [1] put forward the theory of superstructurediaphragm-wall-pile-foundation-foundation interaction, which can calculate the stress and deformation of each section of top-down construction. Tan and Li [2,3], Liu et al. [4], Xu et al. [5], Hwang Richard et al. [6], Ou et al. [7], Wong et al. [8], and Leung and Ng [9] combined the engineering case of the top-down method pit to study the deformation law and the influence of pit excavation on the surrounding environment in soft-soil areas (e.g., Shanghai, Singapore, Hong Kong, and Taipei). Compared with the empirical values put forward by Peck [10], Hsieh and Ou [11], and Clough and O'Rourke [12], the deformation law of foundation pits with different construction methods in softsoil areas were analyzed, which provides a good reference for the application of the top-down method in soft-soil areas. With the city's development, the traditional down-top method cannot meet the requirements of the surrounding environment, and the top-down method is more widely used. However, the top-down method in Yunnan is in its initial stage, and the related engineering experience is lacking. Because of the different geological conditions, the deformation law of the top-down method pit in different areas is slightly different, and the engineering experience in the soft-soil area has a limited guiding role for the engineering in this area. Summarizing the engineering experience in this area can make up for this defect.
With the development of urban rail transit, it is common for a subway tunnel to exist around the pit. e existence of a subway tunnel and foundation pit interact with each other, and the subway tunnel affects the soil stress around the foundation pit. In contrast, the excavation of the foundation pit affects the safety of the subway tunnel. Many scholars have studied it. By analyzing the field monitoring data, Kuang et al. [13] found that the pit's excavation will cause the tunnel to move sideways towards the foundation pit and make the tunnel cross-section deform in the shape of a transverse ellipse. After the foundation slab casts, the tunnel's lateral movement is fundamentally controlled to achieve stability. Chang et al. [14] analyzed the field monitoring data and found that a subway tunnel will reduce the ground displacement caused by pit excavation. e longitudinal direction of the tunnel has a great influence on the deformation bearing value of the tunnel. Based on Mindlin solution, Winkler solution, and Boussinesq solution, Jiang et al. [15], Zhang et al. [16], and Zhang et al. [17] put forward analytical solutions for displacement and internal force changes of surrounding subway tunnels caused by pit excavation and explored the influences of tunnel depth, the horizontal distance between tunnel and pit, outer tunnel diameter, foundation soil, excavation area and excavation sequence on deformation, and internal force of tunnels. Ai et al. [18], Qi et al. [19], Hu et al. [20], Lo and Ramsay [21], Wang et al. [22], and Jiang et al. [23] used a numerical method to explore the influence of different construction schemes and reinforcement measures of foundation pit on surrounding subway tunnels. From the above research, it can be found that theoretical method and numerical method are commonly used research methods, but because of the complexity of foundation pit construction boundary, it is difficult to solve the influence of foundation pit excavation on surrounding environment by theoretical method, and numerical method provides a powerful tool for solving this problem [24].
is study investigated the performance of an 18.5 m deep excavation pit beside arterial roads, which was constructed by the top-down method and rock-socketed support structure in the commercial district of downtown Kunming, China. Field measurements of the deflections of diaphragm walls, vertical movements of the column, axial forces of the slab, and the building's settlements caused by ground deformation were reported, analyzed, and established; a three-dimensional model by MIDAS/GTS software is used to explore the influence of excavation on subway tunnels. is project provides a reference for the application of the top-down method and the rock-socketed supporting method in Yunnan.

Project Introduction.
is study project is located in downtown Kunming (Figures 1(a)-1(c) show the aerial views of the project). e excavation pit was approximately 113 m by 170 m in the plan, and there are four stories of the basement with an excavation depth of 18.5 m. e east side of the pit is Beijing road, the main traffic road to the city. e south side of the pit is the commercial building with 20-floor concrete pile foundation, the bank of the 25-storey reinforced concrete structure with pile foundation is located next to the commercial building and two residential buildings with 7 floors of natural foundation (reference numerals 1 and 2 in Figure 2). ere are 17-storey and 34-storey pile foundations, residential buildings, and apartments with frame structure on the north side. On the west side are the community and residential buildings (reference numerals 3-6 in Figure 2) with 7 floors of brick and concrete with a shallow foundation. e investigated excavation consisted of three sections (i.e., the bottom-up excavation sections A and B and the top-down excavation section C). Considering (1) limited construction space and (2) mitigation of potentially adverse effects on adjacent buildings and facilities as much as possible, the top-down method was adopted for this deep excavation. Table 1 briefly illustrates the top-down excavation sequence in this study. Unfortunately, for some reasons, after the completion of B2 slab (10.8 m), some areas were excavated to 12 m, and the foundation pit stopped construction. It started with the construction of a 1 m thick diaphragm wall. e diaphragm wall was toed through the silty clay layers into the underlying moderately weathered limestone to provide lateral stability and effectively cut off groundwater. e load-bearing elements consisted of augercast-in-place (ACIP) piles, and the diameter is 1 m, which penetrated the strongly weathered limestone layers. Interior H-section steel columns (460 mm × 460 mm) were constructed on the deep-seated ACIP piles to sustain construction loads and function as permanent structures for later use. After the supporting structure was completed, the bottom-up excavation sections, A and B, started excavation. When the underground parts of A and B completed construction, superstructure construction to 7th floor, the topdown excavation section C started excavation. Considering the vast area of section C (about 19210m 2 ) and the facility of unearthed passage, section C was divided into six regions for excavation to reduce the impact of excavation (as shown in Figure 2). Figure 3 shows the section of A-A′, which shows the relationship between the foundation pit and the surrounding subway tunnel, the foundation's main structure, and the stratigraphic distribution.

Geology and Soil
Parameters. Before excavation, soil conditions at the site were explored by a series of field exploration programs (e.g., boreholes and standard penetration test). e strata concerned in this project site consisted of seven layers with total depth up to 79.7 m. e first upper layer was an artificial fill, and beneath it was the silt clay layer. e cobble layer, silty clay, and limestone underlie the silt clay. e observed long-term groundwater table at the site was around 0.3 to 0.5 m. e specific information of the main soil layer is shown in Table 2.

Instrumentation.
A long-term comprehensive instrumentation program was conducted in situ to monitor the deep excavation performance for securing the safety of this project. Figure 4 presents the instrumentation employed to analyze deformation rules. e deflections of the diaphragm walls were monitored by 11 inclinometer tubes (designated as SC1 to SC7 and SC13 to SC16), and the length of the inclinometer is consistent with the height of the diaphragm wall. Vertical column movements were surveyed by a level instrument (C1-C56). Axial forces in floor slabs measured by vibrating wire stress meters which connected with the steel bars in the slab, and the axial force change of the slab is obtained by monitoring the steel bar meter. A level instrument surveyed critical building settlements at 56 locations (D1-D56). Road settlement was observed by a level instrument at 8 sites (L1-L8). ere are 12 groundwater tables outside the pit (SW1-SW12). To ensure the safety of the subway tunnel, deformation observation points are set on the sidewall and vault of the subway tunnel and monitored by the level instrument. e distance between monitoring points is 20 m, and the distance between monitoring points is reduced to 10 m in the section with large deformation of foundation pit supporting structure. e positive and negative values in the diagram represent the diaphragm wall moving outside the pit and inside the pit, respectively. e deflection profiles developed into a bulged profile inward as excavation depth increased, and the locations of maximum wall deflections moved downward as excavation went deeper. Due to the diaphragm wall embedded rock, the deflection is different from the soft-soil; near the bedrock, the deflection is gradually concentrated and decreased; after entering the rock stratum, the deformation of diaphragm wall decreases sharply due to the high stiffness of rock. Instead of the deformation form in soft-soil pit, the deformation of diaphragm wall decreases gradually in soft-soil pit, so there will be obvious inflection point at the rock-soil interface [25]. With the casting of B2 slab, the wall stopped moving towards the excavation site. On the contrary, it moved backward some of the retained sides. It indicates that the B2 slab can suppress the inward wall deflection movements behind. e deflection of SC7 is significantly smaller than the others because it is very close to the cylindrical shaft of section A. e maximum deflections of SC3 and SC7 located at the short side (South-North) of the foundation pit were 20.8 mm and 14.5 mm, respectively, which are smaller than the deflections of SC4 (32.2 mm) and SC14 (27.2 mm) at the long side of the foundation pit (West-East). e relationship between the measured maximum wall deflection, δ hm , and excavation depth, H, is δ hm � 0.06%H ∼ 0.27%H, which is quite different from the normalized deformation curve of the existing soft-soil foundation pit δ hm � 0.02%H ∼ 1.2%H [3], and compared with the normalized curve of the rock-socketed supported foundation pit, δ hm � 0.01%H ∼ 0.45%H (rock-socketed pile) [26], and δ hm � 0.031%H ∼ 0.129%H (rock-socketed diaphragm wall) [27] is closer (as shown in Figure 6). e different normalized deformation between the soft-soil pit and rocksocketed supported pit is due to rock and soil's different stiffness. e area, the ratio of length to the fineness, and the form of the supporting structure of pit lead to the difference of normalized range between this study and other rock-socketed supporting foundation pits. Advances in Civil Engineering 3

Vertical Column Movement.
In the top-down method, the column's deformation is closely connected with the pit's safety and quality. Following excavation, the column might move upwards or downwards, which depends on basal heaves and vertical loads applied on the column (e.g., the weights of the cast floor slabs, friction force applied by  ground soils, construction equipment, and heavy vehicles for material transportation). Figure 7 presents the typical development of the vertical column movements during the construction of stages 3 to 5. It was noticed that the columns experienced undulated vertical movements throughout the construction. at can interpret this phenomenon during each excavation level. e columns were lifted first by the basal heave due to immediate stress relief (soil removal) and then were pushed downwards under the cast floor slabs' weight. After the casting of B1 slab, the column shows a continuous sinking trend, and the settlement tends to be stable until the casting of B2 slab. It can be seen that the stress release caused by excavation after the casting of B1 slab is less than the weights of the slabs and the construction load, and the settlement does not continue to develop because the foundation pit stops construction after the casting of B2 slab. e ratio of the maximum settlement of the column pile δ p to the excavation depth H is δ p � −0.09% ∼ 0.04%H. e maximum differential settlement of column pile deformation and diaphragm wall deformation is less than 1/500 of the differential settlement limit value of the supporting structure proposed by Bjerrum in 1963 [28].  Advances in Civil Engineering earth pressures released due to excavation were sustained by the floor slabs above the excavation surfaces, and the deeper excavation has little influence on the axial force of the slab. For example, the maximum axial force change of the B0 slab on the first floor of the basement excavation is 93 kN, and with the casting of the B1 slab, the system's stiffness and stability increase, and the axial force tended to be stable. e excavation of the second basement has little effect on it.

Building Settlement.
e settlement of apartment, bank, community, and Beijing road caused by ground deformation during construction is shown in Figures 9(a)-9(d), respectively. e deformation of surrounding buildings at the initial stage of construction shows a fluctuating state. With the casting of the slab, the supporting structure's weight increases, and with the completion of the stage 5 construction, the measured settlement of buildings tends to increase with time. Figure 10 plots the relationship between the building settlements due to ground settlement, δ v , and distance, d, behind the diaphragm wall at the final excavation depth (stage 5), H e , for this study, in which H e normalized both δ v and d.
e ground settlement in this study showed a typical concave profile, and the maximum ground settlements did not occur right behind the diaphragm wall but in the zones around 2.7H e away. e ground settlement range is mostly located in Zone I of Peck [10], and the maximum value is located in Zone II of Peck [10]. As shown in Figure 10, the normalized δv/He in this study were same as the subway foundation pit which supported by rock-socketed pile [26], but smaller than the subway foundation pit which is supported by rocksocketed diaphragm wall [27]. e pit's size, the ratio of length to width, supporting structure form, and surrounding environment are the main reasons for these differences. Figure 11 presents the relationships between the normalized building settlements caused by ground deformation, (δ v /δ vm ) and distance ratio (d/H e ), in which δ v is the building settlement at a survey point along the a critical section, and δ vm is the maximum building settlement at the same section. However, the settlement data of this study had much wider influence zones than those defined by Clough and O'Rourke [12] and Hsieh and Ou [11] and the subway foundation pit which was supported by rock-socketed diaphragm wall (2019) [27] and rock-socketed pile (2015) [29]. e settlement influence zone extended beyond 3.8H e , distant from the pit, but exhibited an apparent tendency of decreasing with distance. e significantly wider distribution of (δ v /δ vm ) in this study should mainly derive from its super-large size, which incurred a much bigger displacement of wall deformation and ground settlement and incurred a much wider stress field during excavation.  Table Level. Figure 12 presents the measured groundwater table levels during construction at SW3, SW4, SW10, and SW11 outside the pit. During the construction, the measured groundwater table levels at SW3, SW4, SW10, and SW11 fluctuated slightly between 0 and 0.5 m according to seasonal changes, which were within the Kunming areas' long-term groundwater table levels. is proved that the diaphragm walls successfully cut off the water flow during excavation.

Influence on Subway Tunnel
In this section, the three-dimensional geotechnical finite element MIDAS/GTS software is used to simulate and    analyze the foundation's construction without considering groundwater's influence. According to the simulation results, the foundation pit excavation on the surrounding subway tunnel's displacement and internal force is analyzed.

ree-Dimensional Finite Element Calculation Model.
According to the actual situation and the scope of the pit's excavation, the model size is 330 m × 266 m × 60 m (as shown in Figure 13). e model simplifies the surrounding environment and soil layers. e distribution of each soil layer in the model is shown in Figure 14. e soil adopts the Mohr-Coulomb criterion, and the concrete structure adopts the elastic criterion. e beam element is used to simulate the column pile, and the slab element is used to simulate the diaphragm wall, the slab, and the tunnel lining. e parameters used for the soil in the calculation and analysis are shown in Table 2, and the physical and mechanical parameters of the foundation pit supporting structure and tunnel lining are shown in Table 3.

Deformation of Subway Tunnel Caused by Foundation Pit
Excavation. During excavation, with the diaphragm wall's deformation and the uplift of the soil at the bottom of the pit, the subway tunnel adjacent to the pit will produce horizontal and vertical displacement. e maximum deformation is located at the center of the long side of the pit (as shown in Figures 15(a) and 15(b)). Figure 16 is the maximum value of the tunnel deformation computed and measured during  construction. Overall, the displacements of subway tunnels caused by the construction of pit are not very significant. e measured maximum horizontal displacement is 5.2 mm, the vertical deformation is 2 mm, and the computed results are 6.5 mm and 1.8 mm, respectively. Due to stress release, the soil at the bottom of the pit rises uplift, the retaining structure of the foundation pit deforms laterally, and the ground around the pit subsides, which leads to the deformation of the subway tunnel outside the pit during excavation [30] (as shown in Figure 17).

Changes in Internal Forces of Subway Tunnel Caused by
Foundation Pit Excavation. Figures 18 and 19 show the axial force and bending moment change at the tunnel's maximum deformation point during excavation. Consistent with the displacement change, the axial force and bending moment are minor changes during the construction stages 1 to 3. e axial force fluctuates during the current construction stage. However, compared with the initial construction period, the axial force on the subway tunnel increased at stage 5, and the horizontal axial force

Conclusions
Based on the analyses of the field data, the following significant findings were obtained: (1) e relationship between the measured maximum wall deflection, δ hm , and excavation depth, H, in this study is δ hm � 0.06%H ∼ 0.27%H, which is quite different compared with the relationship of soft-soil foundation pit δ hm � 0.02%H ∼ 1.2%H, and compared with the normalized curve of the rock-socketed supported foundation pit, δ hm � 0.01%H ∼ 0.45%H (rock-socketed pile), and δ hm � 0.031%H ∼ 0.129%H (rock-socketed diaphragm wall) is closer.
(2) e relationship between the maximum settlement of column (δ p ) and excavation depth (H) is δ p � −0.09% ∼ 0.04%H. e maximum distortion between the diaphragm wall and the column is less than 1/500 of the limit range proposed by Bjerrum.
(3) Due to the area, deformation of supporting structure, and the surrounding environment, the building settlements in this study showed a typical concave profile. e maximum ground settlements did not occur right behind the diaphragm wall, but in the zones around 2.7H e , and the influence zone extended beyond 3.8H e . (4) e foundation pit's excavation leads to the maximum vertical settlement of 2 mm and maximum horizontal displacement of 5.2 mm in the subway tunnel. e maximum change of axial force and bending moment is 8.8 kN (the vertical direction) and 6.4 kN·m/m (the horizontal direction).

Data Availability
e data used to support the findings of this study are included with in the article.

Conflicts of Interest
e authors declare that they have no conflicts of interest. 14 Advances in Civil Engineering