Journal Pre-proofs

A simplified two-stage method to estimate the settlement and bending mo-

ment of upper tunnel considering the interaction of undercrossing twin tunnels

Zhiyong Liu, Jianfeng Xue, Jianzhong Ye, Jiangu Qian

PII: S2214-3912(21)00048-9
DOI: https://doi.org/10.1016/j.trgeo.2021.100558
Reference: TRGEO 100558

To appear in: Transportation Geotechnics

Received Date: 3 November 2020

Revised Date: 15 January 2021
Accepted Date: 31 March 2021

Please cite this article as: Z. Liu, J. Xue, J. Ye, J. Qian, A simplified two-stage method to estimate the settlement
and bending moment of upper tunnel considering the interaction of undercrossing twin tunnels, Transportation
Geotechnics (2021), doi: https://doi.org/10.1016/j.trgeo.2021.100558

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2021 Published by Elsevier Ltd.

Title: A simplified two-stage method to estimate the settlement and bending moment

of upper tunnel considering the interaction of undercrossing twin tunnels

Authors: Zhiyong Liu 1,2, Jianfeng Xue1, Jianzhong Ye3, Jiangu Qian 4*

Affiliations: 1, School of Engineering and Information Technology, University of New South Wales,
Campbell, ACT, Australia
2, State Key Laboratory of Hydraulics and Mountain River Engineering, Sichuan University,
Chengdu, 610065, China
3, PSM Engineering Consultants, North Ryde, Sydney, NSW, Australia
4, Department of Geotechnical Engineering, Tongji University, Shanghai, China

Zhiyong Liu PhD candidate, School of Engineering and Information Technology, University of New South
Wales, Campbell, ACT, Australia
MSc graduate, State Key Laboratory of Hydraulics and Mountain River Engineering, Sichuan
University, Chengdu, 610065, China
Jianfeng Xue PhD, Senior Lecturer, School of Engineering and Information Technology, University of New
South Wales, Campbell, Australia

Jianzhong Ye PhD, Senior Engineer, PSM Engineering Consultants, North Ryde, NSW, Australia

Jiangu Qian PhD, Professor, Department of Geotechnical Engineering, Tongji University, Shanghai, China

* Contact of the corresponding author:

Jiangu Qian Department of Geotechnical Engineering, Tongji University

1239 Siping Road, Shanghai, 200092 China
Email: [email protected]

1
Abstract:

With the expansion of existing networks of underground system, many metro tunnels are

excavated in a pair undercrossing existing tunnels. The new excavations will affect the stress

distribution and settlement behavior of the existing tunnels, particularly when the clearance is

small. Generally, the construction of the second undercrossing tunnel induces more settlement in

the upper tunnels compared to that caused by excavating the first one, due to the interaction

between the new excavations. The degree of interaction makes it difficult to predict the settlement

of existing tunnels. In this paper, a simplified two-stage analysis method considering the

interaction between the undercrossing twin tunnels is proposed to estimate the induced response

of existing upper tunnel. In the first stage, the greenfield settlement caused by the first excavation

is estimated using Peck’s empirical function, and that caused by the second excavation is modified

with a modification factor to consider the influence of the first tunnel. In the second stage, the

additional stress calculated using the greenfield displacement is imposed on the existing tunnel

resting on Pasternak foundation model. The calculated settlement profile and bending moment of

the existing tunnel using the two-stage analysis method is compared with those obtained from

finite element analysis and historical field monitored data. The effects of the characteristic

parameter of Pasternak foundation model, the spacing between the new twin tunnels, and the

modification factor of greenfield settlement, on the settlement profile and bending moment of the

existing tunnel are also investigated.

Keywords: Undercrossing twin tunnels; Settlement and bending moment of existing tunnel;

Pasternak foundation model; Interaction;

2
1. Introduction

The rapid growing of traffic in big cities has led to traffic congestion on road. Developing

underground transportation systems is an effective approach to address this problem. Due to space

limitations, more and more tunnels are excavated in a pair undercrossing existing tunnels. New

excavation will affect the stress distribution and settlement behavior of existing tunnels,

particularly when those tunnels are too close. It is critical to accurately estimate the settlement and

bending moment of existing upper tunnels due to the excavation of undercrossing twin tunnels for

risk assessment purpose.

Unlike excavating one tunnel, twin tunnels interact significantly when excavated in a pair,

especially when they are excavated in sequence. Compared with that induced by excavating the

first tunnel, more settlement is anticipated in greenfield or upper tunnels due to the excavation of

the second tunnel. This phenomenon has been confirmed by numerical simulation [8,15], model

tests [6,10] and field monitoring data [7,9,17-19,28]. Therefore, while estimating the settlement of

existing tunnels, the interaction between undercrossing twin tunnels should be considered,

otherwise the settlement may be significantly underestimated. The settlement profile of existing

tunnels also differs to the greenfield subsurface settlement profile which can be described with a

Gaussian distribution curve [26]. If the existing tunnel is stiffer than the surrounding soil, the

settlement of existing tunnels is generally less and the settlement trough is wider than that of

subsurface settlement in greenfield at the same depth [16,21,27,33].

Much research has been done to study settlement profile of surface and subsurface due to the

excavation of twin tunnels, and the interaction between twin tunnels. Fargnoli et al. [13] used the

empirical curves proposed by Peck [30] to predict greenfield settlement due to twin tunnels. The

3
final settlement was calculated using the superposition of the settlement induced by each tunnel.

For the settlement of existing tunnels, Fang et al. [12] used Gaussian distribution function and

superposition method to fit the settlement profile of existing tunnels induced by the construction

of undercrossing twin tunnels, using different values of volume loss and trough width for the

undercrossing twin tunnels. In the authors’ work, volume loss and trough width were determined

using monitored settlement of the existing tunnels. Lai et al. [21] classified the influence zones

caused by the construction of twin tunnels undercrossing an existing tunnel into strong, weak and

no influence zones. Based on the classification, the settlement of an existing tunnel can be

predicted using modified Gaussian distribution function. Ocak [28] introduced a coefficient

(1+d/D) to modify the settlement of greenfield induced by excavating the second tunnel to consider

the influence of the first tunnel.

When using empirical models to predict the settlement of existing tunnels, one keystone is to

determine the trough width. As discussed earlier, the settlement profile of existing tunnels is wider

than subsurface settlement of greenfield. Analytical solutions have also been proposed to predict

the settlement of existing tunnels due to underneath excavation. Liu et al. [25] investigated the

responses of an existing tunnel due to the excavation of a new tunnel by removing springs in

Winkler model. This method is limited to study the interactions between upper and lower tunnels

with zero clearance. The authors found that the extension of springs being deleted is the most

significant factor influencing the responses of existing tunnels. Vorster et al. [32] used an

equivalent linear approach with a closed form solution to analyze subsoil displacement, and the

bending moment of existing pipeline caused by single undercrossing tunnel. The method gives an

upper approximation of bending moment for pipes. Zhang et al. [39] proposed an infinitely long

beam model on Winkler foundation to determine the settlement and internal forces of existing

4
tunnel caused by single undercrossing tunnel. Among those studies, the subsoil is normally

simulated with Winkler foundation model. According to Tanahashi [31], however, Winkler

foundation model ignores the continuous deformation of the foundation and the accuracy of results

is scarified. To improve the accuracy, Zhang [37] used Pasternak foundation model, which

considers the interactions between the springs used in Winkler model, to predict the settlement and

bending moment of existing upper tunnels due to excavating twin new tunnels, but the interaction

between the new tunnels was not considered.

This paper investigates the settlement and bending moment of existing upper tunnel due to new

undercrossing twin tunnels using a two-stage analysis method. The interaction of twin tunnels is

considered in calculating the settlement profile of greenfield subsurface, and the settlement of

subsoil is treated as loading acting on the existing tunnel. The settlement profile of existing tunnel

is calculated using Pasternak foundation model. The tunnel settlement and bending moment

calculated using the two-stage analysis method are compared with those obtained from Finite

Element Method (FEM) and field monitored data from historical projects. Parametric study is

performed to compare the results obtained from Pasternak and Winkler foundation model. The

dependence of the settlement and bending moment of existing tunnel on the shear modulus of the

shear layer in Pasternak model, the distance between the new twin tunnels (d'), and the

modification factor of greenfield settlement (M) is also investigated.

2. The two-stage analysis method

2.1 The final subsurface settlement of greenfield

According to Hunt [15] and Divall et al. [11], due to the disturbance of surrounding soils during

the excavation of the first tunnel, the greenfield deformation and the reaction of upper tunnel
5
caused by the construction of the second tunnel would be different to those caused by the first

tunnel. This also has been confirmed with field observations, for example, the settlements of

Bayswater Road in London [34], Metro line 4 in Shenzhen [18], Metro line 2 in Changsha [7], and

Piccadilly Line tunnels in London [9] due to the excavation of new undercrossing twin tunnels, as

illustrated in Figure 1 and Table 1. In the table, Vl is the volume loss and i is the horizontal distance

from the centerline to the inflexion point of the settlement profile. The cases show that, as a

common phenomenon, the settlements of greenfield or the upper tunnels caused by the second

undercrossing tunnel differ to those caused by the first tunnel in three aspects: the second tunnel

induces more settlements than the first one; the settlement trough caused by the second tunnel may

not be symmetrical; and the maximum settlement caused by the second tunnel may not be along

the centerline of the second tunnel. Therefore, to predict the settlement profile caused by twin

tunnels, the interaction between the new twin tunnels should be considered, otherwise the

settlement of greenfield and existing tunnels will be underestimated.

To obtain the subsurface settlement, a global coordinate system is defined in Figure 2 where the

centerlines of tunnel 1 and tunnel 2 are located at x1 and x2 respectively. The subsurface settlement

profile s1(x) of greenfield due to excavating tunnel l can be described using a Gaussian distribution

function recommended by Peck [30] and Mair et al. [26]:

AVl   x  x1 2  (1)
s1 ( x)   exp   
i 2  2i 2

 

where A is the cross-section area of tunnel 1. According to Mair et al. [26], i can be written as:

i  0.5 z0  0.325 z (2)

where z0 is the depth of the tunnel, and z is the depth of subsurface as illustrated in Figure 2.

6
Because of the interaction between the two tunnels and the disturbance of soil due to the

construction of the first tunnel, the settlement profile of subsoil induced by excavating tunnel 2 is

different from that suggested in Eq. (1). Hunt [15] proposed a method to consider this effect. The

principle of the method shown in Figure 3 is to introduce a modification factor (M) to modify the

settlement due to excavating tunnel 2. M is the largest at the centerline of tunnel 1 where the

settlement due to excavating tunnel 1 has the largest value, and decreases to zero at the boundary

of the settlement profile due to tunnel 1. Therefore, the interaction has a linear relationship with

the settlement values caused by excavating tunnel 1. Beyond the settlement profile of tunnel 1, the

interaction is eliminated and the settlement profile due to excavating tunnel 2 is not influenced. As

recommended by Hunt [15], the settlement profile s2(x) due to excavating tunnel 2 can be

expressed as:

 x  x1  AVl   x  x2 2  (3)
s2 ( x)   1  M  M  exp   
 ni  i 2  2i 2

 

where n is a multiplier to modify i with typically values ranging from 2.5 to 3.0. Take the value of
|𝑥 ― 𝑥1|
𝑛𝑖 as one if it is greater than one. Hunt [15] performed a series of numerical simulations and

found that M varies from 0.6 to 1.5 for London clay. The settlement profiles of greenfield due to

excavating tunnel 1 and tunnel 2 are revealed in Figure 4 with different M values. The figure shows

that, Eq. (3) well captures the characteristics of settlement profile due to excavating tunnel 2 in the

three aspects mentioned above: the increase of maximum settlement, the asymmetry of settlement

profile and the deviation of the location of the maximum settlement from the second tunnel’s

centerline.

The ground settlement due to excavating twin tunnels can be expressed in Eq. (4) using

superposition method:

7
 S ( x)  s1 ( x)  s2 ( x) (4)

2.2 The settlement of existing tunnel

Equations (1) and (4) are used to calculate the settlement of subsurface soils in greenfield without

considering the effect of existing tunnels shown in Figure 2, therefore they cannot be used directly

to estimate the settlement of existing tunnels. The theory of beam resting on Winkler foundation

model has been widely used to estimate the settlement of existing tunnels as revealed in Figure

5(a) [20, 22, 25, 37, 39]. However, according to Tanahashi [31], Winkler foundation model

assumes that each spring behaves independently, which may not be the case in soils. To better

simulate the behavior of ground, Pasternak [29] introduced a shear layer to consider the interaction

between each spring, which is known as Pasternak foundation model. In this work, the existing

circular tunnel is simplified as a continuous Euler-Bernoulli beam resting on Pasternak foundation

model, as shown in Figure 5(b).

The governing equation for an infinite long beam resting on Pasternak foundation is:

d 4W ( x) d 2W ( x) (5)
EI  kBW ( x )  G p B  q( x) B
dx 4 dx 2

where EI is the bending stiffness of the beam; B is the cross-sectional width of the beam, or the

outer diameter of the existing tunnel; k is the coefficient of subgrade reaction; W(x) is the

deformation of the existing tunnel and q(x) is the loading acting on the tunnel; Gp is the shear

modulus of the shear layer. Note if Gp is set to zero, Eq. (5) becomes the governing equation of

Winkler foundation model. Eq. (5) is a fourth-order differential equation which can be solved using

either finite difference method [23, 40-41] or analytical method [38].

8
Klar et al. [20] found that the coefficient of subgrade reaction, k, depends on the shape of the

greenfield settlement trough and can be determined using:

6 Es (6)
k
i

Tanahashi [31] recommended that the value of Gp can be estimated using:

Es H
Gp  (7)
6 1   

where, Es and υ are the elastic modulus and Poisson’s ratio of soil respectively; and H is the depth

of the elastic layer. According to Xu [35], for pipelines with a large length-width ratio, beyond the

depth of 6B the additional stress due to excavation becomes negligible. Therefore, in this paper H

is assumed to be 6B, as recommended by Zhang and Zhang [36].

Tanahashi [31] introduced a parameter  to describe the relationship between the shear modulus

of the shear element, the bending stiffness of the beam and the coefficient of subgrade reaction:

Gp (8)
=
2 kEI B

Zhang et al. [38] suggested that for tunnels,  is less than 1.

Zhang and Zhang [36] introduced an analytical solution to Eq. (5) with a distributed load q(δ)

acting over an infinitesimal element dδ at an arbitrary position δ, as illustrated in Figure 6:

q ( ) Bd 
dW ( x) 
4 EI    
2 2
e
 x 
  cos  x     sin  x    (9)

where,

9
  kB 4EI  Gp B 4EI ,   kB 4EI  Gp B 4EI (10)

Hence, the settlement of the existing tunnel under external loading can be expressed as:

q ( ) B
  cos  x     sin  x   d

W ( x)  
 x 
e (11)

4 EI    
2 2

The solution of an infinite tunnel resting on Winkler foundation can be obtained if Gp is set to zero,

which leads to:

    4 kB 4EI (12)

and

 q( )
 cos  x    sin  x   d 

W ( x)  
  x 
e (13)
 2k

The induced bending moment (Mb) can be expressed as:

d 2W  x 
M b  x    EI (14)
dx 2

According to Zhang et al. [40], the additional stress exerted on twin tunnels resting on Pasternak

foundations due to ground settlement can be calculated using:

d 2 S ( x)  d 2 s1 ( x) d 2 s2 ( x)  (15)
q( x)  kS ( x)  Gp  ks1 ( x)  ks2 ( x)  Gp   
dx2  dx
2
dx2 

Using equations 13 to 15, the deformation and bending moment of the upper tunnels due to

excavating undercrossing tunnels can be calculated.

10
3 Application of the elastic two-stage analysis method

3.1 Comparison with FEM

A three-dimensional finite element analysis was performed to investigate the response of an upper

tunnel to the excavation of undercrossing twin tunnels using Plaxis 3D [5]. As illustrated in Figure

7, the existing tunnel with the diameter of 6 m is at the depth of 12 m. The new twin tunnels are at

the depth of 23 m, with the diameter of 6 m. To better consider soil and tunnel interaction,

Hardening soil model with small-strain stiffness (HSS) is employed to simulate the behavior of

soil [14]. The existing tunnel is simulated using continuous elastic solid body with the elastic

modulus of 2.825 GPa, as used by Liang et al [23]. This gives a bending stiffness of 180 GPa.m4

for the existing tunnel. The tunnel boring machine (TBM) and the lining of the new tunnels are

simulated with 0.35 m thick plate elements. The input parameters for the finite element analysis

are summarized in Table 2.

The excavation is simulated with step-by-step method [2,42]. In the procedure as illustrated in

Figure 8, soils and the existing tunnel are activated to reach the initial equilibrium. The first

calculation phase is to simulate the installation of the 12 m long TBM. After that, in each

consequential excavation, the soil in front of TBM is deactivated, the TBM segment and lining

segment are activated with assigned TBM material and lining material respectively, and both face

support and grouting pressure are activated. The face support and grouting pressure at the crown

of new tunnel are 200 kPa, and the value increases with depth at 14kPa/m. A contraction factor of

1.2% at the TBM facing with an increase of -0.1%/m along the alignment of TBM is employed to

simulate over-excavation. The TBM advances 2 m in each calculation phase and a total of 125

11
phases were performed. The construction of Tunnel 1 and tunnel 2 were simulated in the 2nd to the

62th phases and the 63th to the 124th phases respectively.

After the installation of tunnel 1, the volume loss of 1.95% is achieved. The value of i is 8.2m

which is slightly larger than that determined with Eq. (2). Figure 9 illustrates the greenfield

settlement profiles due to new twin tunnel excavation. It is found that the greenfield settlement

due to excavating tunnel 2 is significantly larger than that due to excavating tunnel 1, as expected.

For M = 0.5, Eq. (4) well catches the behavior of greenfield settlement while compared with FEM.

Figure 10 compares the calculated results using the two-stage method and the results from FEM.

The input parameters of Pasternak foundation model are summarized in Table 3 as obtained from

finite element analysis. The k and Gp are calculated using Eqs. (6) and (7), where Es and υ are 10
ref
MPa and 0.2 respectively, which are the same with E50 and υur as shown in Table 2. It is found that

the calculated settlement and bending moment of existing tunnel agree well with those from FEM.

In addition, the comparison shows that the Pasternak foundation model and Winkler foundation

model produce almost similar predictions, which is consistent with the findings in Liang et al [24].

This is because the bending stiffness of existing tunnel is large but the elastic modulus of soil is

small, which leads to a small characteristic parameter ρ (0.05 for this case) as determined in

Equation 8. The effects of ρ values on the settlement and bending moment of existing tunnel will

be detailed later.

3.2 Comparison with field monitored data

In this section, the predicted settlement profiles of existing tunnels using the proposed method are

compared with the field monitored data of Shenzhen metro line 4 [18] and Piccadilly Line tunnels

in London [9].

12
As revealed in Figure 11, the centerline of the Shenzhen Metro line 4 tunnels is at 12 m below

ground surface. The internal and external diameters of the tunnels are 5.4 m and 6.0 m respectively.

The centerline of the new tunnels is at the depth of 21 m, with the diameter of approximately 6.0

m. The tunnels are in gravelly clay and sandy soil. Piccadilly Line tunnels are in London clay. The

internal and external diameters of the existing Piccadilly Line tunnels are 3.8 m and 4.1 m

respectively at the depth of approximately 13 m. The diameter of the new tunnels is approximately

9.1 m at the depth of about 26 m. The tunnel linings are concrete for all the cases. The values of

M for the projects are estimated to be 1.5 based on the simulation results from Hunt [15]. The

typical cross-sections of the two projects and soil properties used in the analysis are summarized

in Figure 11. The elastic moduli of the soils in Shenzhen Metro Line 4 are estimated using Es=1.4N

MPa as recommended by Architectural Institute of Japan [1], where N is the SPT values shown in

Figure 11(a). Since the analytic solution assumes that the tunnel is resting on Pasternak foundation

which is semi-infinite, for multi-layered soil, an equivalent modulus may be obtained using weight

average method as proposed by Baguelin et al. [3]. For simplicity, the elastic modulus of the stratum

underlain the tunnel is used in Shenzhen Metro Line 4 as this layer is relatively deep. The elastic

modulus of London clay is from the work of Burland and Kalra [4] and shown in Figure 11(b).

The Poisson’s ratio of the soils is assumed to be 0.3 for both cases. The parameters used for

settlement analysis are summarized in Table 4.

The settlement profiles determined with the proposed method are compared in Figure 12 with the

field monitored values. The comparison shows that the proposed method can well predict the total

settlement of the upper tunnels considering the interaction of the two undercrossing tunnels. The

comparison also shows that, though the solution is proposed for tunnels undercrossing existing

tunnels at 90 degrees, the method can well predict the settlement of tunnels that are not

13
undercrossing at 90 degrees. For example, in the case of Shenzhen Metro Line 4, the tunnels

intersect at 83 degrees, and for Piccadilly line tunnels in London, the tunnels intersect at about 69

degrees.

4 Parametric analysis

In this section, the dependence of settlement profile of the existing tunnel on input parameters of

the two-stage method, including the characteristic parameter of Pasternak foundation model (ρ),

the spacing between the new twin tunnels (d'), and the modification factor of greenfield settlement

(M), is analyzed. The parameters listed in Table 3 are used as the baseline.

4.1 The characteristic parameter (ρ)

The shear modulus (Gp) of shear layer strongly affects the value of characteristic parameter (ρ)

which distinguishes the Pasternak model with the Winkler model. The following shear moduli are

considered: 0 (Winkler foundation model), Gp, 4Gp, 7Gp, 10Gp (Gp is determined with Eq. (7) and

is 46.2 MPam from the finite element analysis), which gives ρ values of 0, 0.05, 0.2, 0.35 and 0.5

respectively.

The settlement profiles of the existing tunnel due to excavating tunnel 1 are given in Figure 13(a).

It is found that the deviation of Pasternak foundation model from Winkler foundation model (ρ

=0) increases with ρ as expected. Settlement profile becomes narrower and deeper as ρ increases.

The maximum settlement in Pasternak foundation model is approximately 2%, 6% and 15 %

greater than that in Winkler foundation model when ρ is set as 0.05, 0.2 and 0.5 respectively.

Figure 13(b) shows the bending moment of the existing tunnel with different ρ values. Similar with

settlement, the maximum bending moment obtained with Pasternak foundation model increases

14
with ρ, which is about 6%, 25% and 50 % greater than that from Winkler foundation model when

ρ is set be to 0.05, 0.2 and 0.5 respectively. Such phenomena suggest that only when ρ is small

would Pasternak foundation model be replaced by Winkler foundation model.

4.2 The spacing between the new twin tunnels (d')

Five spacings are considered: d'= 2D, 3D, 4D, 5D, 6D, to study the sensitivity of upper tunnel

settlement to the spacing between the undercrossing tunnels. The typical settlement profiles with

different d' values are given in Figure 14(a) which shows that with the increase of spacing, the

settlement curve of the upper tunnel changes from ‘V’ shape to ‘U’, then to ‘W’ shape. The

maximum settlements under those five scenarios are summarized in Figure 14(b). The result shows

that the maximum settlement of existing tunnel decreases with the spacing between the

undercrossing twin tunnels. When the spacing is greater than 6D, the interaction effect diminishes

and the settlement due to excavating tunnel 2 is not affected by that induced by excavating tunnel

1.

Figure 15(a) shows that the maximum bending moment in the upper tunnel does not always

increase as the new twin tunnels get closer. This can be better seen in Figure 15(b) where both the

positive and negative maximum bending moment are the largest when d' is 2D. The negative

bending moment almost halves when d'=4D compared with that at d'=2D, and the value increases

slightly as d' increases to 6D. This is similar to the positive bending moment which is the lowest

at d'=4~5D. The reason is explained in Figure 15(c), when d'=4D, the positive bending moment

induced by excavating tunnel 1 compensates the negative bending moment induced by excavating

tunnel 2. As a result, the final maximum bending moment due to excavating two tunnels is less

than that induced by single tunnel. This may suggest that d'=4D may be the optimum distance in

terms of reducing the induced bending moment in the upper tunnel.

15
4.3 The modification factor (M)

The modification factor (M) determines the degree of interaction between the undercrossing twin

tunnels. Four different modification factors are considered: 0, 0.5, 1.0, 1.5. The settlement profiles

of the existing tunnel under different M values are illustrated in Figure 16(a). The maximum

settlement of each case is also summarized in Figure 16(b). These figures show that M greatly

affects the settlement profile of the existing tunnel, and the maximum settlement increases linearly

with M. When M increases from 0.5 to 1, the maximum settlement in the upper tunnel increases

from 37 mm to 47 mm as shown in the figure, which is almost 30% increment. In addition, as M

gets larger, the location of the maximum settlement moves towards the centerline of tunnel 2.

Figure 17(a) and (b) illustrate the final bending moment and maximum bending moment under

different M values. Both the maximum positive and negative bending moment values increase

linearly with M.

5 Conclusion

A simplified two-stage analysis method is proposed to predict the settlement and bending moment

of existing upper tunnels due to excavation of undercrossing twin tunnels. The additional stresses

acting on existing tunnels are determined using superposition method considering the interaction

of the undercrossing twin tunnels. The settlement profile and bending moment of existing tunnels

are determined using Pasternak foundation model. The calculated results using the two-stage

analysis method are comparable with the results from finite element analysis and historical field

monitored data. In the method, the variation of water table is not considered, which needs to be

further investigated. The two-stage analysis method can be used to estimate the settlement and

16
bending moment of existing tunnels due to the excavation of undercrossing twin tunnels by

considering their interaction.

Parametric study shows that the settlement and bending moment of existing tunnels are affected

by the characteristic parameter of Pasternak foundation model (ρ), the distance of new twin tunnels

(d'), and the modification factors of greenfield settlement (M). Based on the limited number of

modelling, it was found that, the difference between the results obtained using Pasternak

foundation model and Winkler foundation model increases with ρ. Only when ρ is small would

Pasternak foundation model be replaced with Winkler foundation model. The interaction of the

twin tunnels decreases with the increase of d'. The shape of the final settlement profile of an

existing tunnel can be “V”, “U” or “W” shape, depending on d'. However, the maximum positive

and negative bending moments do not decrease monotonously with the increment of d', and the

values are the lowest when d'=4D based on the analyses performed. The modification factor has

great effect on the settlement profile and bending moment of existing tunnel. It was found from

one case that, when modification factor doubles (from 0.5 to 1), the maximum settlement of and

the maximum bending moment in the existing tunnel increase by approximate 30%. Yet more

research is required to determine the value for M in practice.

Acknowledgement

The project is funded by the Zhejiang Institute of Transportation, the State Key Laboratory of

Hydraulics and Mountain River Engineering, China (SKHL1704), the National Natural Science

Foundation of China (Grant No.51578413) and the Fundamental Research Funds for the Central

Universities (Grant No. 22120190220). The first author received PhD scholarship from the

University of New South Wales.

17
References

[1] Architectural Institute of Japan. Recommendations for design of building foundations,

Tokyo;2001. [in Japanese].

[2] Avgerinos V, Potts DM, Standing JR. Numerical investigation of the effects of tunnelling

on existing tunnels. Geotechnique 2017;67(9):808-22.

https://doi.org/10.1680/jgeot.SiP17.P.103.

[3] Baguelin F, Jézequel JF, Shields DH. The pressuremeter and foundation engineering, series

on Rock and Soil Mechanics. Trans Tech Publications, Clausthal‐Germany;1978.

[4] Burland JB, Kalra CJ. Queen Elizabeth II Conference Centre, Geotechnical Aspects. Proc

Inst Civ Eng, Part 1 1986;80:1479-503. https://doi.org/10.1680/iicep.1986.527.

[5] Brinkgreve RBJ, Kumarswamy S, Swolfs WM, editors. (2017). Plaxis 3D 2017 tutorial

manual; 2017.

[6] Chapman DN, Ahn SK, Hunt DVL. Investigating ground movements caused by the

construction of multiple tunnels in soft ground using laboratory model tests. Can Geotech

J 2007;44(6):631-43. https://doi.org/10.1139/t07-018.

[7] Chen RP, Lin XT, Kang X, Zhong ZQ, Liu Y, Zhang P, Wu HN. Deformation and stress

characteristics of existing twin tunnels induced by close-distance EPBS under-crossing.

Tunn Undergr Space Technol 2018;82:468-81. https://doi.org/10.1016/j.tust.2018.08.059.

[8] Chen SL, Gui MW, Yang MC. Applicability of the principle of superposition in estimating

ground surface settlement of twin- and quadruple-tube tunnels. Tunn Undergr Space

Technol 2012;28:135-49. https://doi.org/10.1016/j.tust.2011.10.005.

18
[9] Cooper ML, Chapman DN, Rogers CDF, Chan AHC. Movements in the Piccadilly Line

tunnels due to the Heathrow Express construction. Geotechnique 2002;52(4):243-57.

https://doi.org/10.1680/geot.2002.52.4.243.

[10] Divall S, Goodey RJ. Twin-tunnelling-induced ground movements in clay. Proc

Inst Civ Eng - Geotech Eng 2015;168(3):247-56. https://doi.org/10.1680/geng.14.00054.

[11] Divall S, Goodey RJ, Stallebrass SE. Twin-tunnelling-induced changes to clay

stiffness. Geotechnique 2017;67(10):906-13. https://doi.org/10.1680/jgeot.sip17.P.151.

[12] Fang Q, Zhang DL, Li QQ, Wong LNY. Effects of twin tunnels construction

beneath existing shield-driven twin tunnels. Tunn Undergr Space Technol 2015;45:128-

37. https://doi.org/10.1016/j.tust.2014.10.001.

[13] Fargnoli V, Boldini D, Amorosi A. Twin tunnel excavation in coarse grained soils:

Observations and numerical back-predictions under free field conditions and in presence

of a surface structure. Tunn Undergr Space Technol 2015;49:454-69.

https://doi.org/10.1016/j.tust.2015.06.003.

[14] Hejazi Y, Dias D, Kastner R. Impact of constitutive models on the numerical

analysis of underground constructions. Acta Geotech 2008;3(4):251-58.

https://doi.org/10.1007/s11440-008-0056-1.

[15] Hunt DVL. Predicting the ground movements above twin tunnels constructed in

London clay, PhD thesis. University of Birmingham; 2005.

[16] Jin DL, Yuan DJ, Li XG, Zheng HT. Analysis of the settlement of an existing tunnel

induced by shield tunneling underneath. Tunn Undergr Space Technol 2018a;81:209-20.

https://doi.org/10.1016/j.tust.2018.06.035.

19
[17] Jin DL, Yuan DJ, Li XG, Zheng HT. An in-tunnel grouting protection method for

excavating twin tunnels beneath an existing tunnel. Tunn Undergr Space Technol 2018b;

71:27-35. https://doi.org/10.1016/j.tust.2017.08.002.

[18] Jin DL. Research on mechanism and control of existing subway tunnel deformation

induced by multi-crossing of shield tunnel underneath, PhD thesis. Beijing Jiaotong

University; 2018. [in Chinese].

[19] Jin DL, Yuan DJ, Liu SY, Li XG, Luo WP. Performance of Existing Subway

Tunnels Undercrossed by Four Closely Spaced Shield Tunnels. J Perform Constr Facil

2019;33(1):04018099. https://doi.org/10.1061/(ASCE)CF.1943-5509.0001230.

[20] Klar A, Vorster TEB, Soga K, Mair RJ. Soil-pipe interaction due to tunnelling:

comparison between Winkler and elastic continuum solutions. Geotechnique

2005;55(6):461-6. https://doi.org/10.1680/geot.2005.55.6.461.

[21] Lai HP, Zhao X, Kang Z, Chen R. A new method for predicting ground settlement

caused by twin-tunneling under-crossing an existing tunnel. Environ Earth Sci

2017;76:726. https://doi.org/10.1007/s12665-017-7079-6.

[22] Liang RZ, Xia TD, Hong Y, Yu F. Effects of above-crossing tunnelling on the

existing shield tunnels. Tunn Undergr Space Technol 2016;58:159-76.

https://doi.org/10.1016/j.tust.2016.05.002.

[23] Liang RZ, Wu WB, Yu F, Jiang GS, Liu JW. Simplified method for evaluating

shield tunnel deformation due to adjacent excavation. Tunn Undergr Space Technol

2018;71:94-105. https://doi.org/10.1016/j.tust.2017.08.010.

20
[24] Liang RZ, Xia TD, Huang MS, Lin CG. Simplified analytical method for evaluating

the effects of adjacent excavation on shield tunnel considering the shearing effect. Comp

Geotech 2017;81:167-87. https://doi.org/10.1016/j.compgeo.2016.08.017.

[25] Liu X, Fang Q, Zhang DL. Mechanical responses of existing tunnel due to new

tunnelling below without clearance. Tunn Undergr Space Technol 2018;80:44-52.

https://doi.org/10.1016/j.tust.2018.06.001.

[26] Mair RJ, Taylor RN, Bracegirdle A. Subsurface Settlement Profiles above Tunnels

in Clays. Geotechnique 1993;43(2):315-20. https://doi.org/10.1680/geot.1993.43.2.315.

[27] Marshall AM, Klar A, Mair RJ. Tunneling beneath Buried Pipes: View of Soil

Strain and Its Effect on Pipeline Behavior. J Geotech Geoenviron Eng 2010;136(12):1664-

72. https://doi.org/10.1061/(ASCE)GT.1943-5606.0000390.

[28] Ocak I. A new approach for estimating the transverse surface settlement curve for

twin tunnels in shallow and soft soils. Environ Earth Sci 2014;72(7):2357-67.

https://doi.org/10.1007/s12665-014-3145-5.

[29] Pasternak P L. On a new method of an elastic foundation by means of two

foundation constants, Moscow; 1954. [in Russian].

[30] Peck RB. Deep excavations and tunneling in soft ground state of the art report. In:

7th International Conference on Soil Mechanics and Foundation Engineering, Mexico City;

1969.p. 225-90.

[31] Tanahashi H. Formulas for an infinitely long Bernoulli-Euler beam on the

Pasternak model. Soils Found 2004;44(5):109-18. https://doi.org/10.3208/sandf.44.5_109.

21
[32] Vorster TEB, Klar A, Soga K, Mair RJ. Estimating the effects of tunneling on

existing pipelines. J Geotech Geoenviron Eng 2005;131(11):1399-410.

https://doi.org/10.1061/(ASCE)1090-0241(2005)131:11(1399).

[33] Wan JC, Zhang DL, Zhang CP, Fang Q, Su J, Du NX. Deformation characteristics

of existing tunnels induced by excavation of new shallow tunnel in Beijing. Chin J Rock

Mech Eng 2014;33(5):947-56. [in Chinese].

[34] Wan MSP, Standing JR, Potts DM, Burland JB. Measured short-term subsurface

ground displacements from EPBM tunnelling in London Clay. Geotechnique

2017;67(9):748-79. http://dx.doi.org/10.1680/jgeot.SIP17.P.148.

[35] Xu L. Study on the longitudinal settlement of shield tunnel in soft soil, PhD thesis.

Tongji University (China); 2005. [in Chinese].

[36] Zhang H, Zhang ZX. Vertical deflection of existing pipeline due to shield tunneling.

J Tongji Univ 2013;41(8):1172-78. [in Chinese].

[37] Zhang QF. Field observation and theoretical study on an existing tunnel

underpassed by new twin tunnels. Adv Civil Eng 2018:1598672.

https://doi.org/10.1155/2018/1598672.

[38] Zhang XM, Ou XF, Yang JS, Fu JY. Deformation Response of an Existing Tunnel

to Upper Excavation of Foundation Pit and Associated Dewatering. Int J Geomech

2017;17(4): 04016112. https://doi.org/10.1061/(ASCE)GM.1943-5622.0000814

[39] Zhang ZG, Huang MS, Wang WD. Responses of existing tunnels induced by

adjacent excavation in soft soils. Rock Soil Mech 2009;30(5):1373-80. [in Chinese].

22
[40] Zhang ZG, Huang MS, Xu C, Jiang YJ, Wang WD. Simplified solution for tunnel-

soil-pile interaction in Pasternak's foundation model. Tunn Undergr Space Technol

2018;78:146-58. https://doi.org/10.1016/j.tust.2018.04.025.

[41] Zhang ZG, Huang MS, Zhang C P, Jiang KM, Lu MH. Time-domain analyses for

pile deformation induced by adjacent excavation considering influences of viscoelastic

mechanism. Tunn Undergr Space Technol 2019; 85:392-405.

https://doi.org/10.1016/j.tust.2018.12.020.

[42] Zhao CY, Lavasan AA, Barciaga T, Zarev V, Datcheva M, Schanz T. Model

validation and calibration via back analysis for mechanized tunnel simulations - The

Western Scheldt tunnel case. Comp Geotech 2015;69:601-14.

https://doi.org/10.1016/j.compgeo.2015.07.003

23
List of Tables

Table 1 The Gaussian fitting results of four undercrossing tunneling projects.

Table 2 The input parameters of FEM.

Table 3 The input parameters of Pasternak foundation model.

Table 4 The input parameters of the two projects.

24
 Table 1 The Gaussian fitting results of four undercrossing tunneling projects.
Vl due to i due to Vl due to i due to
Project excavating excavating excavating excavating
tunnel 1 (%) tunnel 1 (m) tunnel 2 (%) tunnel 2 (m)
Twin-tunneling beneath Bayswater
0.46 13.2 1.2 17.2
Road, London [34]
Metro line 4 in Shenzhen [18] 0.38 9.8 0.63 13.1
Metro line 2 in Changsha [7] 0.24 11.3 0.65 10.8
Piccadilly Line tunnels in London
1.2 10.5 1.8 12
[9]

25
Table 2 The input parameters of FEM.

Parameter and unit Soil Existing tunnel Lining of tunnels TBM

Constitutive model HSS Linear elastic Linear elastic Linear elastic
Bulk density (kN/m3) 20
Cohension (kN/m2) 10
Friction angle (˚) 25
Dilation angle (˚) 0
ref
E50 (kN/m2) 10×103
ref
Eoed (kN/m2) 10×103
ref
E ur (kN/m2) 30×103
υur 0.2
m 0.8
G0 (kN/m2) 100×103
γ0.7 2×10-4
Elastic modulus 2.825×106 30×106 200×106
(kN/m2)
Poisson’ ratio 0.1 0.1 0.2

26
Table 3 The input parameters of Pasternak foundation model.

z0 Z D A=πD2/4 Vl i EI d' k Gp
M
(m) (m) (m) (m2) (%) (m) (GPa·m4) (m) (MPa/m) (MP.m)

23 12 6 28.3 1.95 8.2 180 12 0.5 7.3 46.2

Note: D is the outside diameter of new twin tunnels.

27
 Table 4 The input parameters of the two projects.

z0 z D A=πD2/4 Vl EI d'
Projects M
(m) (m) (m) (m2) (%) (GPa.m4) (m)
Shenzhen metro line 4, the left
21 12 6 28.3 0.42 480 15.6 1.5
tunnel [18]
Shenzhen metro line 4, the
21 12 6 28.3 0.38 480 15.6 1.5
right tunnel [18]
Piccadilly Line tunnels in
26 13 9.1 65 1.2 80 23 1.5
London [9]
Note: The elastic modulus of lining is estimated to be 22 GPa. D is the outside diameter of new
twin tunnels.

28
List of figures:
Figure 1 (a): The measured settlement of Bayswater Road, London [33], (b): settlements of Metro
line 4 in Shenzhen [17], (c): Metro line 2 in Changsha [6], and (d): Piccadilly Line tunnels in
London (d) [8] due to excavating of new undercrossing twin tunnels. EX1: excavating tunnel
1, EX2: excavating tunnel 2 (same abbreviations are used in following figures).
Figure 2 The settlement of existing tunnel due to excavating new twin tunnels.
Figure 3 The surface/subsurface modification factor (after Hunt [14]).
Figure 4 The greenfield settlement due to excavating each tunnel with different M (data from Hunt
[14]).
Figure 5 Analytical model of excavating twin undercrossing tunnels using Winkler foundation
model (a) and Pasternak foundation model (b).
Figure 6 Deflection of Pasternak foundation model induced by external load.
Figure 7 Finite element model used to study the interaction between existing upper and
undercrossing tunnels.
Figure 8 The simulation of excavation process in finite element model.
Figure 9 The greenfield settlement caused by the exaction of twin tunnels.
Figure 10 Comparison between the settlement (a) and bending moment (b) of the existing tunnel
obtained using the two-stage method and FEM.
Figure 11 The typical cross-section of Shenzhen Metro Line 4 (a) (data from Jin et al. [17], N is
number of Standard Penetration Test), and Piccadilly line tunnels in London (b) (cross-section
modified after Cooper et al. [8] and the elastic modulus of soil from Burland and Kalra [3]).
Figure 12 The comparison of simulated settlement with field monitored data of, (a): the left tunnel,
and (b) the right tunnel of Shenzhen Metro Line 4 (monitored data from Jin et al. [17]), and
(c): Piccadilly Line tunnels (monitored data from Cooper et al. [8]).
Figure 13 The settlement (a) and bending moment (b) of existing tunnel due to the excavation of
tunnel 1.
Figure 14 The settlement profile (a) and maximum settlement (b) of the existing tunnel with
undercrossing twin tunnels at different spacings.
Figure 15 The bending moment profile (a), maximum bending moment (b) of the existing tunnel
with undercrossing twin tunnels at different spacings, and the bending moment profile due to
each tunnel excavation (c).
Figure 16 The settlement profile (a) and maximum settlement (b) of existing tunnel settlement
with different M values.
Figure 17 The bending moment profile (a) and maximum bending moment (b) of existing tunnel
with different M values.

29
Author Credit Statement

Zhiyong Liu: Conceptualization, Methodology, Software, Writing- original draft preparation.

Jianfeng Xue: Supervision, Methodology, Writing - review & editing, Funding acquisition.

Jianzhong Ye: Investigation, Resources.

Jiangu Qian: Funding acquisition, Visualization, Writing - review & editing.

30
Dear Editor,

The authors declare that we do not have any commercial or associative interest that represents a conflict of

interest in connection with the work submitted

Best regards,

Jiangu Qian

31