Soil constitutive model
A large number of on-site measured data and numerical simulations (Burland 1989; Jardin et al.1986; Mair 1993) show that when soil mass is disturbed in geotechnical engineering, it will undergo small deformations and will have small strain, with a certain relationship between the soil stiffness and the strain. The common strain level of geotechnical engineering is shown in Fig. 6. The engineering strain of the foundation pit is 0.01~1%, which belongs to the small strain range. The soil small strain model (HSS) (Benz 2006) can well reflect the nonlinear and stress-related characteristics of the soil small strain, so HSS is selected as the constitutive model in this study.
Table 1 shows the mechanical parameters of the soil. Where Gref 0 is the initial dynamic shear modulus of the soil, γ0.7 is the shear strain at which the shear modulus decays to 70% of the initial shear modulus, Eref oed is the secant modulus measured by the triaxial consolidation drained shear test, Eref ur is the loading and unloading modulus measured by the triaxial consolidation drained shear test, and Eref oed is the tangent modulus under the reference stress determined by the standard consolidation test. The value of γ0.7 is 0.00011 according to Brinkgreve and Broere (2006), and Gref 0 is 2.5~4.9 Eref ur. According to the PLAXIS 3D software manual and existing engineering experience, the values of Eref 50 and Eref oed are basically the same, Eref ur is generally 2~4 times larger than the loading modulus (the model in this study takes 3 times), and c andφ of the soil are taken from the geotechnical survey report of this project.
Table 1 Physical and mechanical properties of soil and the parameters for numerical simulation
Soil type
|
h(m)
|
ρ
(kN/m3)
|
c
(kPa)
|
φ
(°)
|
Numerical simulation parameters
|
Gref 0(Mpa)
|
γ0.7(10-3)
|
Eref 50(Mpa)
|
Eref oed(Mpa)
|
Eref ur(Mpa)
|
①2Fill
|
5.2
|
19.5
|
15
|
10
|
50
|
0.11
|
6
|
6
|
18
|
②3-2Silt clay
|
4.9
|
19.4
|
25.0
|
14.0
|
80
|
0.11
|
8
|
8
|
24
|
③Silt
|
1.2
|
20
|
11.0
|
15.0
|
80
|
0.11
|
12
|
12
|
36
|
⑤1-1Gravel
|
12.5
|
19.6
|
0
|
34
|
400
|
0.11
|
50
|
50
|
150
|
⑦1-2Sandrock
|
2.7
|
21
|
70
|
22
|
400
|
0.11
|
48
|
48
|
144
|
⑦1-3Sandrock
|
33.5
|
21.1
|
90.0
|
25.0
|
500
|
0.11
|
52.5
|
52.5
|
157.5
|
Numerical model and material parameters
A deep foundation pit of Nanning Metro Line 5 in Guangxi is taken as the case study. The PLAXIS software is used to build a three-dimensional finite element calculation model. The range of soil settlement behind the wall caused by the excavation of the foundation pit is four times the depth of the excavation of the deep foundation pit. Given the influence of the boundary effect, this model conducted full-scale modeling. The size of the model is 280 m (length) × 240 m (width) × 60 m (height). The isolation pile of the foundation pit is simplified to the plate element using the principle of stiffness equivalence (Atkinson et al. 1990), and the simplified formula is as follows:

where, D is the diameter of the isolated piles, t is the distance between the isolated piles, and h is the thickness of the equivalent wall thickness. The value of t is 0.623 m and is calculated by Eq. 1. The concrete internal support is simulated by beam elements, and the contact relationship between ground connecting walls and isolation piles and soil is simulated by interface elements. The three-dimensional numerical model is constrained and fixed at the bottom, and the other four sides are set with normal constraints that allow only vertical displacement of the model. The mesh is refined in places where the structural stress is concentrated. The bottom of the three-dimensional numerical model is restrained and fixed, and normal constraints are placed on the other four sides, so only the vertical displacement of the model occurs. Fig. 7 shows the overall mesh division of the model.
The floor slabs of Buildings A and B are simulated by linear elastic model plate elements, the pillars are beam elements, and the pile elements are simulated by embedded beam elements. The depth of the pile foundation is 14 m. Table 2 shows the specific parameters of the building.
Table 2 Building material parameters
Material
|
γ(kN/m3)
|
E(Mpa)
|
d(m)
|
Floor
|
25
|
30
|
0.1
|
Foundation slab
|
25
|
30
|
1.5
|
Pillar
|
25
|
30
|
/
|
Foundation side wall
|
25
|
30
|
0.2
|
Raft
|
25
|
30
|
0.3
|
Pile foundation
|
5.14
|
30
|
0.6
|
Simulation scheme
The simulated construction steps are as follows. Step 1 generates an initial stress field; Step 2 activates the surrounding buildings; Step 3 activates the diaphragm wall and the isolation piles; Step 4 excavates the first layer of soil (-1.67 m); Step 5 activates the first support; Step 6 excavates the second layer of soil (-6.67 m); Step 7 activates the second support; Step 8 excavates the third layer of soil (-10.67 m); Step 9 activates the third support; and Step 10 completes the excavation of the foundation pit (-15.3 m).
Model validation
Fig. 8 shows the results of the simulation and on-site monitoring of the lateral displacement of the diaphragm wall and the ground settlement value. From Fig. 8(a) it can be seen that after the completion of the excavation of the foundation pit, the measured maximum lateral displacement of the diaphragm wall does not differ much from the simulated values, which are 0.06% H and 0.07% H, respectively. The deformation trends are inward and convex, while the maximum lateral displacement position of the ground connecting wall is slightly higher than the bottom of the pit. After the completion of the excavation of the foundation pit, the measured maximum value of the D45-3 profile does not differ much from the calculated value of the simulated value, which are 0.05% H and 0.04% H, respectively. The deformation trends have surface groove shapes. Combined with the on-site measured data, these results show that the model is reliable.
Hsieh and Ou (1998) provided a method for predicting the form of a surface subsidence groove, proposing a main influence area and a secondary influence area. The main affected area is 0~2 He (He is the excavation depth of the foundation pit), the secondary affected area is 2~4 He, and the maximum settlement occurs at a distance of 0.5 He from the foundation pit edge. Ou (2002) revised the Hsieh experience curve through numerical calculation and actual measurements. This study showed that the scope of the main affected area is not directly related to the excavation depth of the foundation pit, but is determined by the excavation depth, the excavation width, and the hard soil layer depth. As shown in Fig. 9, d is the distance from the edge of the foundation pit, PIZ is the main affected area of surface settlement, δv is the value of surface settlement, and δvm is the maximum value of surface settlement. After correction, the settlement value at the boundary point of the main influence area and the secondary influence area is 1/6 of the maximum settlement value, while the distance between the maximum settlement point of the groove shape and the wall of the enclosure is 1/3 of the width of the main affected zone. The surface settlement curve after the foundation pit excavation is in line with the revised curve of Ou's experience curve, and the maximum surface settlement is located on 1/3 of the main affected area, which further verifies the model reliability.
Building deformation characteristics
To investigate the different influences of two support methods of the diaphragm wall and the combined support method of isolation pile and diaphragm wall on the deformation of different types of building foundations, the following two cases were considered in the numerical simulation. Case 1 is when only the diaphragm wall is used as the retaining structure. Case 2 is the diaphragm wall + isolation pile as the retaining structure. Unless otherwise specified, the deformation values of the buildings are all from the excavation to the bottom of the pit. The support efficiency is defined (Zheng et al. 2018) to quantitatively analyze the deformation of the two schemes:

where, i1 is the evaluation index of building safety, such as the maximum settlement value and the maximum horizontal inclination value of the building for Case 1, and i2 is the building safety evaluation index, such as the maximum settlement value and the maximum horizontal displacement value for Case 2. Therefore, when η (0≤η≤1) is larger, it indicates that the combined support method of isolation pile and diaphragm wall is better.
Maximum settlement
The curve of the change in the maximum settlement value of different types of building foundation is shown in Fig 10. When the excavation depth of the foundation pit is shallow, excavation of the first and second soil layers, the combined support method of isolation piles and diaphragm walls shows no obvious effect on the maximum settlement value of each building foundation. When excavating the third soil layer, The support efficiency of the maximum settlement value of independent, box, and pile foundation buildings are 41.4%, 38.5%, and 21.0%, respectively. When excavating the last soil layer, the maximum settlement support efficiency for each type of building foundation is 34%, 32%, and 14%, respectively. The support efficiency of the combined support method for the maximum settlement value of each type of building foundation does not increase with increasing the excavation depth of the foundation pit. The excavation depth is 10.67 m (the third layer of soil), which shows the best supporting effect on each building foundation. Comparing the support efficiency of isolation pile diaphragm wall to various foundation types, the combined support method shows the best support effect on the settlement value of independent and box foundations, with small difference in the support efficiency, followed by pile foundation buildings.
Horizontal tilt
The building tilts into the pit due to the excavation of the foundation pit. The higher the floor, the higher the tilt value. The particularly high-rise frame structures are the most sensitive to the value of the tilt. Rankin (1988) proposed 1/500 as the critical slope value that can damage the building. The maximum horizontal displacement in the ux and uy directions of different types of building foundations is taken for analysis (Fig. 11). In the ux direction, the supporting efficiency of the horizontal inclination value of the isolation and diaphragm wall for dependent, box, and pile foundations is 34.7%, 35.8%, and 14.7%, respectively. In the uy direction, these values are 40.5%, 42.9%, and 14.1%, respectively. The combined support of the isolation pile and the diaphragm wall shows a small difference in the support efficiency of the pile foundation. For shallow foundation, the value of horizontal inclination in the uy direction shows a better support, and there is little difference in the horizontal displacement of the pile foundation in two directions. buildings with the pile foundation have a good ability to adjust uneven settlement. Under the action of excavation of the foundation pit, their own horizontal displacement is small, so the control efficiency in the ux and uy directions is not very different.
Differential settlement
Differential settlement of the foundation will cause additional stress on the building structure and will damage it. The frame structure is especially sensitive to differential settlement and deformation (Xin et al. 2011). The slope angle of the building (settlement difference/length of the building) can be calculated through the differential settlement of the building. Bjerrum (1963) found that if the slope angle of a building is greater than 1/500, it will affect the stability of the building. Therefore, it is necessary to study the support efficiency of the combined support method for the slope angle of the building. The settlement values of the short and long sides of the building are shown in Fig. 12. The supporting efficiency of the combined support method of isolation pile and diaphragm wall for a short slope angle of buildings for dependent, box, and pile foundation is 55.5%, 52.3%, and 20.8%, respectively. The supporting efficiency of the long slope angle of each basic building is 64.2%, 40.7%, and 9.3%, respectively. For independent foundations and box foundations, the supporting efficiency is far higher than with pile foundations. Relative to the settlement value and inclination value of the building, the combined supporting method has the best effect on the slope angle of the building. From the analysis of scope angle, the control scope of the long side of the independent foundation is much larger than that of the box and pile foundations.
Angular distortions
After the excavation of the foundation pit, the foundation will tilt and be damaged. Angular distortions can be used as a measure of the degree of damage to a building. The angular distortion is shown in Fig. 13. A, B, C, and D are the positions of the corners before the excavation of the building foundation; E, F, G, and H are the positions after the foundation settlement due to excavation of the foundation pit; the angles between the sides and horizontal plane after settlement are marked as α, β, γ, δ; and the tangent of the four included angles is the foundation angular distortions.
The angular distortions of different types of building foundations according to two different support methods are shown in Table 3. From Table 3, the combined support method all protects the angular distortions of the four sides of different foundation types, but the degree of impact is different. The supporting effect of the independent and box foundations is more obvious, while the control effect of the pile foundation is not obvious. For the independent foundation, the combined support method most obviously controls the angular distortions of the angle β, and the supporting efficiency reaches 93.5%. For the box foundation, the combined support method most obviously controls the angular distortions of the angle α, and the supporting efficiency is 42.3%. Finally, for the pile foundation, the angular distortions of the angle δ controlled by the combined support method are the most obvious and the support efficiency reaches 31.8%.
Table 3 Results of angular distortions of different foundation types
Foundation type
|
Support method
|
Angular distortion
|
α
|
β
|
γ
|
δ
|
Independent foundation
|
Case 2
|
9.2×10-5
|
8.5×10-5
|
4.5×10-5
|
9.3×10-5
|
Case 1
|
1.6×10-4
|
1.1×10-3
|
4.9×10-5
|
2.6×10-4
|
Box foundation
|
Case 2
|
7.7×10-5
|
5.97×10-5
|
4.8×10-5
|
1×10-4
|
Case 1
|
1.36×10-4
|
7.36×10-5
|
4.96×10-5
|
2.13×10-4
|
Pile foundation
|
Case 2
|
5.76×10-5
|
3.9×10-5
|
1.22×10-5
|
6.07×10-5
|
Case 1
|
7.3×10-5
|
4.2×10-5
|
1.32×10-5
|
8.86×10-5
|
Sensitivity Analysis of Isolation Pile Length
In general, an isolation pile must be inserted under the sliding surface to have a better protection effect. Therefore, the length of the isolation pile and plane position are the main design parameters of the isolation pile. In theory, the closer the isolation pile is to the building, the better the protection effect (Ying et al. 2012). This section mainly discusses the sensitivity of length of the isolated pile to different types of building foundations under the combined support method. To facilitate the practical application of the project, the maximum settlement value, and the maximum value of horizontal displacement of different types of building foundations are selected as sensitivity indicators. As shown in Fig. 14, the maximum settlement and the maximum horizontal displacement of the building decrease with increasing length of the isolation pile. However, there are differences in the amplitude of deformations of each building foundation. The settlement value and horizontal displacement curve of independent and box foundations are steeper than those of the pile foundations, which indicates that buildings with shallow foundations are more sensitive to the length of isolated piles under the combined support method. The change curve of each type of foundation can be roughly divided into diminishing, rapid diminishing, and slow diminishing stages. In addition, the magnitude of the settlement decreases and the horizontal displacement of the building with increasing the pile length is not always the same. For the settlement value and horizontal displacement value of the independent and box foundations, the length of the isolated pile is ≥24 m, ≥27 m, respectively, and the reduction rate becomes slow. For the settlement value of the pile foundation, the isolation pile length is >24 m and the reduction rate is slower. Regarding the horizontal displacement value of the pile foundation, the reduction rate is not much different.
When the calculation results are normalized (Fig. 15), y is the maximum settlement value and the maximum horizontal displacement value of the building when the length of the isolation pile is changed within the range of 15–36 m; ymax is the maximum settlement value and the maximum horizontal displacement value of the building without isolation piles; h0 is the excavation depth of the foundation pit and h is the depth of the isolation pile in the soil.
The supporting efficiency of the combined support method for each type of foundation is shown in Fig. 15. As the depth of the isolation piles increases, the maximum settlement value of the independent and box foundations can be reduced to about 48% of that without isolation piles. For the pile foundations, it is about 60%. The maximum horizontal displacement that can be reduced by the independent and box foundations is about 57% of that without isolation piles. For the pile foundations, it is about 73%. The maximum settlement value of buildings with different foundation types, the horizontal displacement value and the isolation pile depth can be uniformly described as follows:

where, A, B, and C are parameters, and the value size is shown in Fig. 15.
Considering the above research, the combined support method shows a small difference in the shallow foundation support. In the actual construction process, it is necessary to comprehensively assess the safety status and foundation types of the buildings adjacent to the foundation pit.