Deng Et Al 2021 - Pipe Contact Models in Rock

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

Tunnelling and Underground Space Technology 115 (2021) 104063

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology


incorporating Trenchless Technology Research
journal homepage: www.elsevier.com/locate/tust

Analysis and application of friction calculation model for long-distance rock


pipe jacking engineering
Zhiyun Deng a, b, Ninghui Liang b, *, Xinrong Liu b, *, Albert de la Fuente c, *, Peng Lin a,
Haoyang Peng a
a
State Key Laboratory of Hydroscience and Engineering, Tsinghua University, Beijing 100084, China
b
School of Civil Engineering, Chongqing University, Chongqing 400045, China
c
Civil and Environmental Engineering of UPC BarcelonaTECH, Jordi Girona 1-3, 08034 Barcelona, Spain

A R T I C L E I N F O A B S T R A C T

Keywords: For long-distance rock pipe jacking, the coupled actions of mud float and sediment generate pipe-rock in­
Long-distance teractions, which are challenging from the analysis point of view and the general methods available for calcu­
Rock pipe jacking lating the friction resistance of pipe-soil can lead to unrepresentative results. These friction forces must be
Friction resistance
quantified properly, as the magnitude of those determines the technical and economical construction re­
Influence factor
quirements. In this regard, a novel approach that combines experimental and numerical stages is proposed in this
Calculation model
Engineering application research paper for dealing with the assessment of these friction forces. The results derived from the proposed
approach were compared and validated with others obtained from a practical case, with satisfactory outcomes.
The approach proposed herein is found to be a valuable guide for assessing the magnitude of the frictional forces
that occur during the pipe jacketing and for identifying the parameters that govern the magnitude of these forces.
Eventually, design and construction optimization and productivity-oriented measures can be derived from the
application of the method.

1. Introduction tunnel. However, the contact phenomena existing between the pipe and
the surrounding stratum become more complex due to the injection of
Pipe jacking (PJ) is a trenchless construction technique that has bentonite (Ye et al., 2019).
proven to be efficient, safe, environmentally friendly and to lead to The pipe-soil interaction can be modeled as full-contact (Choo and
minor impacts on traffic (Deng et al., 2021; Ji et al., 2019a; Yang et al., Ong, 2020; Ji et al., 2019b; Ong and Choo, 2018; Pellet-Beaucour and
2020). PJ engineering is moving towards the ultra-long distance, large Kastner, 2002; Sofianos et al., 2004; Zhang et al., 2016a), partial-contact
cross-section while adapting the technology to complex stratum. In PJ, (Khazaei et al., 2004; Milligan and Norris, 1999; Zhang et al., 2018) and
the jacking force (JF) is a paramount factor governing both the struc­ pipe-slurry contact (Wang et al., 2014; Ye et al., 2015). The pipe-soil
tural design and the selection of PJ machine (Sheil, 2021; Zhang et al., full-contact model assumes that there is complete contact between the
2016b). An accurate prediction of the JF can guarantee effective control pipe and the surrounding stratum, which generally leads to conservative
of safety and costs. The JF is composed of two parts, namely the head magnitudes of the soil friction resistance. This hypothesis is frequently
resistance of the shield machine and the lateral friction resistance of the adopted in guidelines and specifications, such as Japan’s JMTA (2013),
pipe-soil. For long-distance jacking engineering, the lateral friction UK’s PJA (1995), the Chinese national standard GB 50332 (2002), the
resistance is the greater component, the pipe-soil contact pressure and German standard AVT A - 161 (2010) and the United States standard
the pipe-soil friction coefficient being the main parameters that deter­ ASTM (2011). The quantification of the vertical component of the earth
mine the JF (Zhang et al., 2018). At present, in long-distance PJ projects, pressure is usually assessed by means of the soil column method, the
the friction resistance is generally reduced by injecting lubricating ma­ Promojiyfakonov theory or Terzaghi method (Terzaghi, 1965).
terials (such as bentonite mud) into the overcut between pipe and In case that the excavated front is stable, the pipe is jacked from the

* Corresponding authors.
E-mail addresses: dzhy@mail.tsinghua.edu.cn (Z. Deng), lnh83249@cqu.edu.cn (N. Liang), liuxrong@126.com (X. Liu), albert.de.la.fuente@upc.edu (A. de la
Fuente).

https://doi.org/10.1016/j.tust.2021.104063
Received 8 April 2021; Received in revised form 24 May 2021; Accepted 9 June 2021
Available online 19 June 2021
0886-7798/© 2021 Elsevier Ltd. All rights reserved.
Z. Deng et al. Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 115 (2021) 104063

to provide overestimations of the friction forces.


In addition to these phenomena, rock blocks and other sediments are
likely to flow back through the overcut gap generated during the exca­
vation process. These tend to accumulate at the outer surface at the
bottom of the pipe and alter both distribution and magnitude of the pipe-
rock contact forces. To consider these relevant physical phenomena in
the pre- and design stages, a new engineered-design analytical formu­
lation is proposed herein. The model was satisfactorily validated by
comparing the results derived from the new formulation and those
measured in field in a long-distance jacketed-pipe project.
The model resulted in providing reliable (and safe side) and,
consequently, the authors state that the model might be used for similar
future projects to optimize costs and productivity.

2. Considerations related to pipe-rock friction resistance in rock


PJ

According to Sterling (2020) and Xiang et al. (2008), factors that


affect the pipe-rock friction resistance in PJ engineering can be divided
Fig. 1. Plane strain model of Persson contact model. into engineering geological (i.e., geological formation properties, depth
of the pipe alignment, underground water pressure, among others) and
bottom of the ring cavity (Milligan and Norris, 1996, 1999; Pellet- construction factors (i.e., overcut, grouting lubrication intermediate
Beaucour and Kastner, 2002) and two different hypotheses can be jacking stations, pipeline-axis alignment deviation, stoppage, and both
used for predicting the JF: 1st friction resistance is only produced by the distance and curvatures of the jacketed-pipeline stretches).
self-weight of the pipe (Pellet-Beaucour and Kastner, 2002) while the As per model derivation, the following assumptions and hypotheses
2nd assumes that this resistance is caused by the pipe-soil contact pres­ were made in relation to the friction resistance:
sure (Milligan and Norris, 1996, 1999), which can be determined by
means of either the Hertz’s or Persson’s contact models (Ciavarella and • Rock strata and overcut annular cavity are constant, so soil arching
Decuzzi, 2001a, 2001b). If the bentonite is effective, the pipe’s buoy­ effect is nonexistent under these theoretical conditions.
ancy can be greater than the pipe’s self-weight and, thus, the pipe is • The geological conditions are introduced by (1) imposing a non-
likely to be in suspension. If so, the suspended pipe is only in contact uniform distribution of pressure on the pipe-rock contact surface;
with slurry, and the JF can be computed with the formulations presented (2) disregarding the influence of tunnel buried depth on friction
in Wang et al. (2014) and Ye et al.(2015), that enables to take into ac­ resistance as rock stratum guarantees the stability of the excavated
count the solid (pipe) – liquid (mud) contact thixotropic properties. cavity.
The suitability of the above-mentioned models has been widely • Pipe-rock interaction due to grouting lubrication is simulated by
validated and verified in shallowly buried PJ. However, for deeply considering the friction coefficient of pipe rock. In this regard, Ji
buried tunnels, the full-contact model based on Terzaghi soil pressure et al. (2019b) and Ye et al. (2015) proposed that the increase of the
theory is unrepresentative because the shear zone does not extend to the frictional resistance (f) caused by axis deviation can be set as con­
surface. Therefore, Zhang et al. (2016a) proposed a modified Terzaghi stant; nevertheless, these authors analyzed short jacking distance
soil pressure model meant to be valid for deeply buried soils. Nonethe­ pipelines. For long-distance PJ, f magnitudes and distribution must
less, this modified model use to lead to negative magnitudes of the be adapted properly to account for the effects of boring machine
component γ-2c/b due to the high cohesion (c) of the rock and, thus, this stoppage and intermediate jacking stations (if necessary), the former
model is still not representative for rock stratum. Yang et al. (2013) being the cause of major friction resistance rise. In addition to this
combined both Terzaghi and Promojiyfakonov theories and proposed a source of friction resistance, those caused by sediment at the bottom
formulation for assessing the vertical component of the earth pressure of the pipe and the buoyancy of slurry mud are also taken into
and applicable to the deep-buried pipe. This formulation was already account.
validated for medium-coarse sand and circular gravel strata, but the role
of bentonite mud was disregarded. In the deep-buried rock stratum, the
surrounding rock usually has good structural integrity and stability, 2.1. Contact angle of pipe-rock
therefore, the magnitude of the exerted pressure is reduced. Likewise,
the lubricating effect of the mud allows reducing the friction coefficient. Zhang et al (2017) pointed out that the traditional Hertz contact
Hence, disregarding the positive effect of the bentonite mud may leads model is valid and representative in case that the contact region size
to an overestimation of the JF magnitude. generated by two elastic bodies is an order of magnitude inferior to the
In this regard, to solve this problem, Zhang et al., 2017; 2018) pro­ size of the latter. This condition is generally nonexistent in JP engi­
posed a model based on the pipe-soil contact characteristic in which the neering since the sizes of both pipe and tunnel are similar, so is the
contact pressure is calculated by using the Persson model. Although the contact region the contact. Under these conditions, (Ciavarella and
model takes into account the pipe-rock contact characteristics and the Decuzzi, 2001a, b) claim that the Persson contact model (Fig. 1) is more
frictional resistance of pipe-mud, the buoyancy effect of bentonite mud appropriate to assess the friction forces.
is disregarded. Previous researchers (Namli and Guler, 2017; Praetorius Zhang et al. (2017) stated that given and specific geometry and di­
and Schößer, 2017; Shou et al., 2010; Ye et al., 2019; Zhou et al., 2009) mensions, the contact characteristics of pipe-soil mainly depend on both
stated that bentonite mud has two action mechanisms: (1) lubrication the elastic modulus and Poisson’s ratio of pipe and soil. The pipe-soil
and (2) filling and supporting effects. The former, if the two latter are friction resistance (f) pipe-soil, per unit of length and accounting for
properly active, can couple with the buoyance effect of bentonite mud the slurry – pipe interaction, can be computed by means of the Eq. (1)
and reduce the pipe-rock contact pressure and the friction forces thereof. (Zhang et al., 2018).
The previously mentioned formulations disregard these effects and tend f = kc Pμs + 2θRp Cs + 2(π − θ)Rp Pm μm (1)

2
Z. Deng et al. Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 115 (2021) 104063

Fig. 2. Diagram of the numerical model. (a) Front view of model (2θ = 90◦ ); (b) Side view of model; (c) Top view of model.

Table 1
Calculation parameters of surrounding rock and pipe.
Material Density (kg⋅m− 3) Elastic modulus (GPa) Poisson’s ratio Cohesion (MPa) Internal friction angle (◦ ) Meshing Size (m)

Rock 2540 1.8 0.33 0.5 27 1.25


Pipe 2000/2500 42.5 0.2 / /

and kc, the tunnel geometry and boundary conditions (Fig. 2) considered
Table 2
in (Li et al., 2019; Yen and Shou, 2015) were implemented with Abaqus.
Table of kc value corresponding to different contact angles.
The Mohr-Coulomb criterion was considered for simulating the rock
Contact Total friction Friction μs Gp Gp × kc while the reinforced concrete pipe was assumed to behave elastically,
angle 2θ resistance (N) resistance per (kN/
see Table 1.
μs
(◦ ) meter (kN/m) m) (kN/
m) A parametric analysis was carried out to assess the friction resistance
for different values of 2θ (see Table 2). For this purpose, a displacement
5 134,716 5.99 0.1 59.42 5.94 1.01
10 134,868 5.99 0.1 59.42 5.94 1.01 control approach (Yen and Shou, 2015) was adopted for applying the
20 136,580 6.07 0.1 59.42 5.94 1.02 jacking load and the penalty function algorithm as a tangential contact
30 136,859 6.08 0.1 59.42 5.94 1.02 setting. Normal contact was set as hard contact (contact materials
45 140,391 6.24 0.1 59.42 5.94 1.05 cannot penetrate the contact surface). The simulation process was the
60 145,293 6.46 0.1 59.42 5.94 1.09
following: (1) applying the initial stress state of equilibrium rock mass;
90 153,472 6.82 0.1 59.42 5.94 1.15
120 163,174 7.25 0.1 59.42 5.94 1.22 (2) determining the contact setting between the rock and the string
180 189,530 8.42 0.1 59.42 5.94 1.42 according to different contact conditions; (3) shifting the string to the
specified position (displacement control hypothesis) to obtain the stress
on the acting surface of the JF; (4) deriving the reaction curve of the
where P is concentrated force acting on the center of a circular pipe; θ is reference point for determining the JF value Fs.
half of the angle corresponding to the contact arc of pipe-soil; Rp is the kc can be cacualated as the ratio between the friction resistance (f)
radius of the pipe; Cs is the adhesion force of the pipe-soil; μm is the pipe- and the product of the friction coefficient (μs) by the pipe self-weight
slurry friction coefficient; Pm is the grouting pressure and kc is the co­ (Gp) according to Eq. (2).
efficient of contact for non-uniform distribution of pipe-soil contact
f
pressure. kc = (2)
G p μs
According to Zhang et al.(2017; 2018), for soft soils, the contact
angle 2θ can be set as 180◦ (hence, the slurry-pipe angle is 180◦ ) and the
The numerical simulation results show that surrounding rock parame­
contact pressure P equal to 1.35 times the self-weight of pipe (kc = 1.35).
ters, buried depth, friction coefficient (μs) and buoyancy have little in­
For hard rock stratum, the pipe-rock contact converges in a point 2θ =
fluence on the calculation results of kc value, which further proved that
0◦ , thus the contact angle of pipe-slurry is 360◦ and kc = 1.00. However,
the numerical simulation result is correct. This content is not detailed
Eq. (1) considers a unique stratum (infrequent in long-distance pipes)
discussed and further details on this aspect can be found elsewhere
and omits the influence of sediments accumulated during the excavation
(Zhang et al. 2017; 2018). The calculation results of kc value under the
process.
condition of friction coefficient of 0.1, the lower limiting value
In this regard, in order to stablish a numerical relation between 2θ

3
Z. Deng et al. Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 115 (2021) 104063

earth pressure for increasing contact angles, the friction resistance


provided by lateral earth pressure should not be ignored (see Fig. 3).
The lateral pressure can be computed as per Eq. (3)

Fa = 2γr hKa × h/2 = Ka γ r R2 (1 − cosθ)2 (3)

where h is the height of sediment; R is the inner radius of the pipe; θ is


half of the contact angle between sediment and pipe; γ r is the unit weight
of the surrounding rock; Ka = tan2(45◦ - φ/2) is the coefficient of active
earth pressure.
When the action of mud is not considered, the frictional resistance
per unit length of the pipe can be assessed by means of Eq. (4).

f = [Ka γ r R2 (1 − cosθ)2 + 1.35P]μs (4)

2.2. Effect of mud buoyancy


Fig. 3. Model diagram of PJ considering lateral earth pressure.
The presence of mud within the annular overcut can lead to pipe
buoyancy and, thus, to modify the total contact force P according to Eq.
Surrounding rock
(6).
H
P = (Gp + Gs − Gup ) (6)

Pipe where Gp is the pipe self-weight; Gs is the counterweight (weight of


R+Rcos θ various pipelines) and Gup is the buoyancy of the pipe. For the analysis
Mud
dθ carried out herein, the unit weight of mud and Gs were set to be 11 kN/
Rθ m3 and 0.55 kN/m, respectively (Praetorius and Schößer, 2017).
Likewise, the accumulation of sediments at the bottom of the pipe in
Sediment Mud pressure long-distance rock PJ tunnels affects the pipe-slurry contact area, which
is reduced or even eliminated and consequently, the buoyancy is not
effectively produced. For the sake of simplification, and on the safe side,
Fig. 4. Schematic diagram of buoyancy calculation under different sediment
angle conditions.
it is assumed that the small amount of mud within the range of sediment
is disabled to transfer buoyancy effectively.
Therefore, it is necessary to calculate the buoyancy of the pipe under
different sediment angle conditions. For this purpose, the model pre­
sented in Fig. 4 is assumed as valid.
In the model Fig. 4, the height of the free liquid surface measured
from the pipe crown is H; the outer radius of the pipe is R; the sediment
angle at the bottom of the pipe is 2θ; the length of the pipe is taken as
unit length, and the unit weight of mud is γm.
Gup can be obtained by integration (see Eq. (7))
∫π
Gup = γm ⋅ 2(H + R + Rcosθ)Rdθ.cosθ
θ
[ ] (7)
R2
= γm ⋅ R2 (π − θ) − 2(H + R)Rsinθ − sin2θ
2

Fig. 5 presents the relations of P and Gup (positive upward) with 2θ for R
= 1.585 m, Gc = 59.4 kN/m and H = 0.0 m.
According to the results presented in Fig. 5, for small sediment
Fig. 5. Relationships for pipe-rock contact and buoyancy forces with the con­
tact angle 2θ. accumulation (2θ < 30◦ ), under the action of buoyancy (Gp greater than
0, upward), the crown of the pipe is in contact with the surrounding
rocks and P greater than 0 concentrated in a contact point. Thereafter,
suggested by Stein et al. (1989) for pipe-soil friction coefficient under
Gup decreases with the increase of 2θ, and the contact pressure (P < 0) is
the action of mud lubrication, and pipe unit weight of 25 kN/m3 are
distributed at the bottom of the pipe. The absolute value of buoyancy
summarized as shown in Table 2.
increases continuously as 2θ increases in the range of 120◦ ~180◦ and
Based on the results presented in Table 2, kc = 1.35 for 2θ ≤ 180 ◦ C,
then decreases slightly as 2θ increases in the range of 180◦ ~360◦ .
as proposed by Zhang et al. (2017), might lead to conservative values of
In case that the sediment at the pipe bottom is not removed, the
Fs. For 2θ = 180 ◦ C, kc = 1.35 differs only in 5% respect to the numerical
frictional resistance increases dramatically. The pipe buoyancy force
value (kc = 1.42), and this can also be accepted for designing purposes.
reaches a minimum of − 11.9 kN /m when 2θ is 180◦ , and then the
This lower values of kc resulting from the numerical model are the
contact force of pipe-rock reaches a maximum of − 71.8 kN /m, this may
consequence of the contribution of the lateral earth pressure to the
cause the maximum friction force caused by the self-weight of pipe.
friction mechanism, which is neglected by the analytical formulation
After that, as the buoyancy force of the pipe decreases, the contact force
(Eq. (1)). However, when the contact angle further increases, the uneven
of the pipe-rock gradually decreases to the self-weight of the pipe.
distribution of the gravitational forces is mainly concentrated in the
range of contact angle 180◦ below the pipe springline. Thus kc value can
be set as the theoretical value 1.35. Given the gradual increase of lateral

4
Z. Deng et al. Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 115 (2021) 104063

2.3. Influence of the pipe axis deviation

Due to geometric tolerances (pipe geometry and construction),


geological variations, as well as the deformation of the cutter head, the
actual pipeline suffers from deviations respect to the designed pipeline,
this leading to increase of the required JF. In this regard, Milligan and
Norris (1999), based on field monitoring, concluded that the jacking
trajectory can be assumed S-shaped. Posteriorly, Ji et al. (2019b) pro­
posed a pipe-soil interaction model (Fig. 6) applicable to pipe-jacketing
engineering.
As shown in Fig. 6, in the length range of S/2, the deviation angle of
the first pipe is 0◦ , the second pipe presents a relative angle deviation to
the first pipe of λ, and the angle deviation of the nth pipe is the accu­
mulated angle deviation (n-1)⋅λ. It is assumed that the nth pipe reaches
the maximum accumulated angle deviation at the end of the S/2 curve,
which is corrected by applying the proper rectifying measures.
The total friction resistance (FJ) of the n pipes of length Lo contained
within the distance S/2 can be computed with the Eq. (8)
1 1 1
FJ = L0 f [ + +⋯+ ] (8)
cos0◦ cosλ cos(n − 1)λ

where f is the friction resistance per meter of pipe. If the pipe axis de­
viation is 0◦ (straight jacking), FJ0 is expressed by means of Eq. (9)
Fig. 6. Diagram of pipe-soil interaction during the jacking process (Ji et al.,
2019b). (a) S-shaped pipeline layout; (b) Force balance of the pipe. FJ0 = L0 fn (9)

Therefore, the coefficient of axis deviation fk0 can be expressed ac­


Table 3 cording to Eq. (10)
Table of increase coefficient of axis deviation fk0 under different deflection
angles. FJ 1 1 1
fk0 = =( + + ... + )/n (10)
(n-1)⋅λ ( )

λ( )

n fk0 fk0,m FJ0 cos0◦ cosλ cos(n − 1)λ

6 0.2 31 1.0019 1.0020 According to previous researchers (Broere et al., 2007; Cui et al.,
0.5 13 1.0019 2015; Milligan and Norris, 1999; Norris, 1992; Zhang et al., 2016b),
1.0 7 1.0020
even in curved PJ engineering, λ ≤ 2◦ , and the maximum accumulated
2.0 4 1.0021
10 0.2 51 1.0052 1.0054 angle deviation is inferior to 30◦ ; otherwise, the pipes are prone to suffer
0.5 21 1.0052 from damaging. As per this analysis, the accumulated angle deviations
1.0 11 1.0054 considered were 6◦ , 10◦ , 20◦ and 30◦ , and the angle deviations between
2.0 6 1.0056 different pipes were 0.2◦ , 0.5◦ , 1.0◦ and 2.0◦ , respectively. The co­
20 0.2 101 1.0211 1.0215
efficients of axis deviation fk0 are gathered in Table 3. (See Table 4.)
0.5 41 1.0212
1.0 21 1.0215 As it can be noticed from the results of Table 3, on the one hand, fk0 is
2.0 11 1.0221 hardly sensitive to variations of λ from 0.2◦ to 2.0◦ . On the other hand,
30 0.2 151 1.0493 1.0500 fk0 varies a 5% for accumulative angle deviations from 10◦ to 30◦ . It
0.5 61 1.0496
must be remarked, however, that the jacking distance (L) was neglected
1 31 1.0500
2 16 1.0510
in the above table, so the value had almost no effect on the fk0 value. The
λ affected the JF by affecting the length of the S-shaped jacking curve.
Considering that the increase coefficient fk0 only corresponds to the
increase coefficient of axis deviation within the range of the semi-curve
Table 4
length (S/2), in the long-distance PJ project, considering the jacking
Allowable deviation of PJ (mm) (GB50268-08, 2008).
distance L and the length of single pipe L0, the Eq. (11) can be re-written.
Inspection items Allowable deviation (mm)
FJ 1 1 1 1
(12)
[ L ]
Horizontal direction L < 300 m 50 fk = = [( + + +⋯+ )/n] n.L0
300 ≤ L ≤ 1000 m 100 FJ0 cos0◦ cosλ cos2λ cos(n − 1)λ
L ≥ 1000 m L/10
Vertical direction L < 300 m +40, − 50 When the jacking distance L is S/2, L/(nL0) = 1, then fk = fk0, and when
300 ≤ L ≤ 1000 m +60, − 80 the jacking distance L is S, L/(nL0) = 2, then fk = fk02.
In practice, Norris (1992) found that λ was mostly between 0 ~ 0.30◦
and occasionally up to 0.5◦ when the construction process was properly
controlled, and λ might exceed 1◦ when the construction control was
poor or the geological conditions were greatly different through actual
monitoring. In general, under the condition of suitable construction

Table 5
The suggested value table for the axis deviation increase coefficient fk0 in linear jacking engineering.
Jacking distance (m) 100 200 300 400 500 600 700 800 900 1000

fk0 1.001 1.002 1.003 1.004 1.005 1.006 1.007 1.009 1.010 1.011

Note: L0 = 2.5 m, λ = 0.5 , (n-1)⋅λ = 2.5 .


◦ ◦

5
Z. Deng et al. Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 115 (2021) 104063

Fig. 7. Schematic diagram of specimen size design and experiment process.

2.5◦ . Therefore, for the analyzed PJ project, fk should be dependent on


the total pipe jacketed length (see Table 5).
As shown in Table 5, if the PJ axis deviation is well controlled, the
increase of fk is limited and its influence on the frictional resistance in
long-distance linear PJ engineering can be disregarded. However, it is
worth noting that the values of fk reported in Table 5 fail to consider the
additional compressive stresses between the pipe and surrounding rock
caused by axis deviation. However, in practical engineering, this phe­
nomenon only leads to a sudden and local increase of the JF, and this
drop down to the normal level after the jacking deviation is rectified.
Therefore, the fk values of Table 5 can be considered as a suitable
reference.

2.4. Influence of pipe-rock friction coefficient

Fig. 8. Direct shear test results of pipe-rock interface with bentonite concen­ Previous research was mainly focus on the pipe-soil friction coeffi­
tration of 4.5%. cient of the soil layer. Stein et al. (1989) found that this friction coeffi­
cient for soils ranges 0.1 ~ 0.3 under the action of mud lubrication.
However, there are few studies on the pipe-rock friction coefficient.
Table 6 Referencing to Li et al. (2019), who established friction coefficients
Summary of friction test results. through the screening of concrete specimens and rock surfaces and
characterizing pipe-rock friction mechanism by means of the shear test,
Contact surface No. of tests Mean Reduction
condition Value percentage (%) experimental process of the shear test for obtaining pipe-rock friction
Test 1 Test 2 Test 3 coefficients is established (Fig. 7). Combined with the possible pipe-rock
Slurry 0.341 0.331 0.324 0.332 / contact conditions in the field, this paper considers the pipe-rock shear
Slurry: sediment 0.412 0.409 0.406 0.409 23.1 friction tests under different contact conditions, including pipe-rock
= 1.5: 1
friction test under 4.5% lubricating mud to the sediment of different
Slurry: sediment 0.432 0.449 0.437 0.439 32.2
= 1: 1 mass ratio. The lubricating slurry concentration is defined as the mass
Slurry: sediment 0.506 0.505 0.493 0.501 50.9 ratio of bentonite to water. This paper presents the pipe-rock shear test
= 0.5: 1 results under the condition that the lubricating slurry concentration is
Note: lubricating slurry concentration is 4.5%, and the funnel viscosity of the 4.5% as an example, as shown in Fig. 8.
bentonite slurry is 64 s. As shown in Fig. 8, after the friction coefficient quickly reaches the
stable stage in the early stage, it can keep floating up and down at a
control, λ = 0.5◦ leads to safe results. For linear PJ engineering, the certain value in the later stage. The mean value of the stable section of
allowable deviation of the pipe during jacking should meet the re­ the friction coefficient was taken as the results of the single friction test,
quirements of GB50268-08 (2008). and the mean value of the results of the three tests was taken as the pipe-
The allowable horizontal deviation within the range of 1000 m for a rock friction coefficients and the pipe-rock friction coefficient of this
single pipe is taken as the limit. Considering the most unfavorable sit­ group of tests was 0.332. Further, the test results of each group are
uation λ = (n-1)⋅λ; (n-1)⋅λ corresponding to the axis deviation of 100 mm summarized in Table 6.
for a 2.5 m pipe is arctan(100/2500) = 2.3◦ . Considering adverse con­ As shown in Table 6, the pipe-rock friction coefficient under the
struction factors, (n-1)⋅λ in a straight jacking project can be taken as slurry condition is about 0.332. In addition, it should be pointed out
that, since the mass ratio of lubricating mud and sediment is in a

6
Z. Deng et al. Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 115 (2021) 104063

P
Surrounding rock Pm Pm
Lubrication mud
Pipe
Sediment
FJ Jacking force
FF Resistance of jacking Model I: Pipe-slurry contact model Model II: Top point contact model
machine head

Working
Bentonite mud
well

FF
FJ
Pipe

Jacking machine Sediment

Pm
Pm

P (θ ) P (θ )

Model III: Bottom-filled contact model Model IV: Upper-filled contact model

Fig. 9. Four contact states of pipe-rock during jacking.

of the long-distance rock PJ project.


Table 7
Table of application condition for the combined friction calculation model.
3. Combined calculation model of frictional resistance in rock
Model Model description Application condition PJ
I Pipe-slurry contact The grouting effect is good, the buoyancy of the
model pipe is greater than gravity, and the pipe is under Four pipe-rock contact components during jacking are identified in
the traction action of the head machine Fig. 9:
II Top point contact The grouting effect is good, the pipe-rock angle 2θ is
model less than 30◦ , and the buoyancy of the pipe is
greater than its self-weight (1) Model I: Pipe-slurry contact model
III Bottom-filled The pipe-rock angle 2θ is greater than 30◦ , the
contact model buoyancy of the pipe is less than its self-weight, and Under the action of gravity, the machine head and the pipes con­
the lateral friction due to lateral pressure
taining the working trolley just behind the machine head are in contact
disregardable
IV Upper-filled When the pipe-rock angle 2θ is greater than 180◦ , with the bottom of the surrounding rock. In the case of a good lubri­
contact model the lateral friction due to lateral pressure cating effect, the subsequent pipes are in suspension under the traction
considered action of the head machine and the buoyancy action of bentonite slurry.
The friction resistance of the pipe under this model is taken asfI (Eq.
(13)) the adhesion force between mud and pipe being disregarded.
dynamic change in the actual project, the mixtures of lubricating mud
and sediment with three mass ratios of lubricating mud to sediment were fI = πDp Pm μm fk (13)
considered. The mass ratio of lubricating mud to the sediment of 1.5:1,
1:1 and 0.5:1 indicating the case of excellent, general and poor lubri­
(2) Model II: Top point contact model
cating effect of lubricating mud, respectively, according to the reference
(Ji et al., 2019b).
When the pipe is far away from the machine head, the traction force
At this stage, although there is sediment between the pipe and sur­
is small. If the buoyancy force applied to the pipe is greater than the self-
rounding rock, the friction coefficient between the pipe and sediment is
weight, then the pipe is in contact with the surrounding rock at the top.
also called the pipe-rock friction coefficient. When there exist sediments
According to the research results of the reference (Zhang et al., 2017;
at the bottom of the pipe, pipe-rock friction coefficient is the lowest
2018), it can be considered that the pipe and surrounding rock are
(0.409) under the condition of sufficient mud and efficient lubricating
almost in a point contact state. The friction resistance is mainly
effect. Contrarily, when the amount of slurry is minor and the lubrica­
composed of the friction resistance caused by pipe-rock contact pressure
tion effect is insufficient, the pipe-to-rock friction coefficient increases to
and pipe-mud friction resistance. The friction resistance of the pipe can
0.501, an increase of 22.5% (50.9%) compared to that of 0.409 (0.332).
be calculated by Eq. (1). The friction resistance of the pipe can be
Generally, it can be seen that when the pipe is in contact with the sur­
computed as fII by substituting Eq. (6) into Eq. (1), this resulting in Eq.
rounding rock, if the pipe is only in partial contact with the complete
(14).
surrounding rock, the pipe-rock friction coefficient is relatively small.
( ⃒ ⃒ )
Under certain contact pressure, the friction resistance of the pipe can be fII = πDp Pm μm + ⃒Gp + Gs − Gup ⃒μs fk (14)
ensured to be in a small state, which is conducive to the jacking control

7
Z. Deng et al. Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 115 (2021) 104063

Fig. 10. Geologic profile of each NPT. (a) No. 4 NPT; (b) No. 7 NPT; (c) No. 8 NPT.

Table 8
The geological table along with the PJ of each tunnel.
Tunnel number Jacking distance (m) Length (m) Sandstone Mudstone Ave. buried depth (m) Max. buried depth (m)

4 0 ~ 510 510 √ 31 53
510 ~ 680 170 √
680 ~ 720 40 √
7 0 ~ 30 30 √ 33 52
30 ~ 85 55 √
85 ~ 130 45 √
130 ~ 215 85 √
215 ~ 355 140 √
355 ~ 380 25 √
380 ~ 530 150 √
530 ~ 890 360 √
890 ~ 910 20 √
8 0 ~ 125 125 √ 51 88
125 ~ 255 130 √
255–1413 1158 √

the pipe. With the increase of the jacking distance, the contact area
Table 9
between the debris and the pipe increases continuously. When the
Table of physical and mechanical parameters of surrounding rock on site.
contact angle 2θ is lower than 180◦ , the frictional resistance is mainly
Lithology Density Elastic Poisson’s Internal Cohesive composed of the frictional resistance caused by contact pressure and the
(g/cm3) modulus ratio friction force c
frictional resistance of lubricating mud. The frictional resistancefIII can
angle φ (◦ ) (MPa)
be computed by coupling Eqs. (1), (6), and (8) for deriving Eq. (15).
Sandstone 2.55 1.6–2.0 0.32 27.5–30.5 0.40–0.60
⃒ [ ]⃒
Mudstone 2.55 1.5–2.0 0.35 16.7–26.5 0.25–0.40 ⃒ R2 ⃒
fIII = (Kc ⃒⃒Gp + Gs − γ m R2 (π − θ) − 2(H + R)Rsinθ − sin2θ ⃒⃒μs
2
+ (π − θ)Dp Pm μm )fk (15)
(3) Model III: Bottom-filled contact model
When the angle of bottom sediment is 0, fIII equals fII.
If the rock debris produced by the machine head cutting cannot be
effectively cleaned, the debris accumulates gradually at the bottom of (4) Model IV: Upper-filled contact model

8
Z. Deng et al. Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 115 (2021) 104063

Jacking force The jacking machine [ ]


12000 R2
Friction resis tance reached the exposed part fIV = (1.35|Gp + Gs − γm R2 (π − θ) − 2(H + R)Rsinθ −
Jacking force and friction resis tance (kN)
sin2θ |μs
2 (16)
10000 Vertical rise in JF due ◦

to pipe mis alignment +(π − θ)Dp Pm μm + γr R2 (1 − cosθ)2 tan2 (45 − φ/2)μs )fk
8000 For Eqs. (13) to (16), fk can be obtained by means of applying Eq. (12).
FF Thus, the influence of sediment at the bottom of the pipe, mud buoyancy
6000 k=92.76
effect, pipe-soil friction coefficient and the influence of axis deviation
influence coefficient fk are considered in the combined calculation
4000 k=32.71 k=10.75
model.
JF decreas ed by 46.8% The conditions under which each model is valid and representative
2000 The JF continued to rise sharply, but it are summarized Table 7.
fell down after sediment removal and
0 the huge injection of bentonite mud
k=2.94 4. Verification and application of the combined calculation
0 100 200 300 400 500 600 700 800 model
Jacking distance (m)
4.1. Verification of the combined calculation model
(a)
IJS was then Guanjingkou Water Conservancy Project in Chongqing is located in
Jacking force
12000 the section from Dongwenquan Town to Tea Garden, with a total length
Jacking force and friction resis tance (kN)

Friction resis tance activated


of 21.6 km, and a total of eight pressure-free tunnels and two pressure-
10000 free tunnels. The total length of ten tunnels is 15.6 km, among which,
Vertical rise in JF due
to pipe mis alignment No. 3 pressure-free tunnel is the longest with a distance of 3,224 m. The
8000 FF inner diameter of the pipe is 2.65 m, the outer diameter is 3.17 m, and
the wall thickness is 0.26 m. The longest one-way jacking distance is
6000 2,000 m. There are a total of nine drives with a one-way jacking length
greater than 1,000 m, and three tunnels are constructed by jacking in
4000 k=12.48 two directions. This project is selected for validation purposes since it
deals with long-distance pipe jacking and because this is probably one of
2000
the most challenging projects in China of this typology. The diameter of
k=2.06 the excavated tunnel is 3236 mm, and the overcut clearance is 33 mm.
0
The uniaxial compressive strength is an important parameter for rock
0 100 200 300 400 500 600 700 800
engineering designs (Sheil et al., 2016; Sun et al., 2021). The main strata
Jacking distance (m)
along the water transmission line are mudstone, sandstone and lime­
(b) stone, and the maximum uniaxial saturated compressive strength is 20
MPa, 70 MPa and 90 MPa, respectively. The jacking distance of pipes
Jacking force IJS was then jacked by the main jacking station alone in No. 4, No. 7 and No. 8–1
12000
Jacking force and friction resis tance (kN)

Friction resis tance activated


no-pressure tunnels (NPTs) are 836 m, 707 m and 690 m, respectively.
Vertical rise in JF due The geological conditions of them are shown in Fig. 10(a), (b) and (c),
10000
to pipe mis alignment FF respectively.
8000 In Fig. 10, Q4eld is the diluvial layer consisting of Quaternary strata,
k=16.17 J3sn is the Suining Formation of the Upper Jurassic, J2s is the Shaximiao
6000 k=31.68 Formation of the Medium Jurassic, Ss is sandstone, Cr is mudstone. The
JF decreas ed by 15.7% geological conditions along the jacking distance by the main jacking
4000 station alone were summarized as shown in Table 8, in which No. 4 NPT
The JF continued to rise sharply,
has only counted the geological conditions within the range of 0–720 m
but it fell down after sediment
2000
removal and the huge injection because the surface was exposed within the range of 720–809 m.
FF As shown in Table 8, mudstone and sandstone are distributed along
of bentonite mud
0 k=2.84 the pipeline mainly. Specifically, sandstone is mainly along the line of
0 100 200 300 400 500 600 700 800 No.4 NPT, with 170 m of mudstone in part. Mudstone and sandstone are
Jacking distance (m) staggered along the line of No.7 NPT, while mudstone is mainly along
the line of 8–1 NPT, with 130 m of sandstone in part. The physical and
(c) mechanical parameters of the field surrounding rocks are shown in
Fig. 11. Measured JF and friction resistance for NPTs: (a) No. 4; (b) No. 7 and Table 9.
(c) No. 8. The field measurement and analysis of the JF of No. 4, No. 7, and No.
8–1 tunnels were carried out in detail. The JF and the jacking machine
When the sediment is enough and the upper overcut area is filled, the head resistance is real-time measured. The friction resistance is calcu­
lateral friction force caused by the lateral pressure should be considered. lated as the difference between JF and the jacking machine head resis­
The friction force is composed of pipe-mud friction, friction caused by tance. The measured JF and calculated friction resistance curves of No.
pipe-rock contact pressure and side friction caused by lateral pressure. 4, No. 7, and No. 8–1 NPTs are shown in Fig. 11(a), (b) and (c),
This component of the frictional resistance (fIV) can be obtained by respectively.
combing Eqs. (1), (5), (6), (8) to obtain Eq. (16). As shown in Fig. 11(a), for No. 4 NPT, the unit length frictional
resistance k is 2.94 kN/m (disregarding the data fluctuation in the early
stage) for distances 40–100 m, this increasing to 10.75 kN/m for
100–720 m of jacking distance. For No. 7 NPT (Fig. 11(b)), the k is 2.06
kN/m for 40–120 m and 12.48 kN/m for 120–708 m. Finally, for No. 8–1

9
Z. Deng et al. Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 115 (2021) 104063

Table 10
Comparison between the measured (Fs,M) and the computed (Fs,N) friction force, and relative errors.
Calculation method Fs,N (kN/ (Fs,N - Fs,M)/ Fs,M 2θ (◦ )
m)
Fs,M of No. 4 (kN/m) Fs,M of No. 7 (kN/m) Fs,M of No. 8–1 (kN/m)

40–100 m: 100–720 m: 40–120 m: 120–708 m: 30–140 m: 140–690 m:


2.94 10.75 2.06 12.48 2.84 16.17

Promojiyfakonov theory 358 120.77 32.30 172.79 27.69 125.06 21.14 /


JMTA (2013), Terzaghi (1965) 85.4 28.05 6.94 40.46 5.84 29.07 4.28 /
PJA (1995), Ong and Choo 28.5 8.69 1.65 12.83 1.28 9.04 0.76 /
(2018)
AVT (2010) 42.8 13.56 2.98 19.78 2.43 14.07 1.65 /
ASTM (2011) 1296.7 440.05 119.62 628.47 102.90 455.58 79.19 /
GB 50332 (2002) 835.6 283.22 76.73 404.63 65.96 293.23 50.68 /
Zhang et al (2018) 21.9 6.45 1.04 9.63 0.75 6.71 0.35 /
Model I 2.99 0.02 − 0.72 0.45 − 0.76 0.05 − 0.82 /
Model II 11.89 3.04 0.11 4.77 ¡0.05 3.19 − 0.26 /
Model III 10.95 2.73 0.02 4.32 − 0.12 2.86 − 0.32 3◦
10.32 2.51 − 0.04 4.01 − 0.17 2.63 − 0.36 5◦
8.67 1.95 − 0.19 3.21 − 0.31 2.05 − 0.46 10◦
5.43 0.85 − 0.49 1.64 − 0.56 0.91 − 0.66 20◦
3.28 0.12 − 0.69 0.59 − 0.74 0.16 − 0.80 30◦
7.82 1.66 − 0.27 2.80 − 0.37 1.75 − 0.52 45◦
12.32 3.19 0.15 4.98 ¡0.01 3.34 ¡0.24 60◦
17.4 4.92 0.62 7.45 0.39 5.13 0.08 75◦
20.44 5.95 0.90 8.92 0.64 6.20 0.26 90◦
27.07 8.21 1.52 12.14 1.17 8.53 0.67 120◦
32.23 9.96 2.00 14.65 1.58 10.35 0.99 150◦
35.31 11.01 2.28 16.14 1.83 11.43 1.18 180◦
Model IV 47.01 14.99 3.37 21.82 2.77 15.55 1.91 240◦
52.7 16.93 3.90 24.58 3.22 17.56 2.26 300◦

Note: rock density: 2.55 g /cm3; rock internal friction angle: 30◦ ; cohesive force: 0.5 MPa, and Pmum: 0.3 kN/m2; the friction coefficient: 0.332 (experimentally
obtained); height of the slurry head at the top of the pipe: 0.0 m

floats. This means that the pipe crown can be considered, as per calcu­
lation purposes, in contact with the top of the tunnel or gradually sink
inside the sediment to finally be in contact with the tunnel bottom with a
certain contact angle. Under this condition, the pipe is a dynamic state of
rising and sinking. Compared with the existing analytical formulations
(Table 10), the errors between the theoretical and the measured values
resulting from assuming the top point contact model (model II) for No.4,
No.7 and No. 8–1 NPTs are 11%, − 5% and − 26%, respectively; these
errors being lower than those obtained with the other existing formulas.
Particularly, model II can better reflect the frictional resistance in the
later stage of long-distance PJ, especially the theoretical friction resis­
tance is almost consistent with the measured frictional resistance in the
later jacking stage of No. 4 and No. 7 NPTs.
For the bottom-filled contact model (model III), the calculation re­
Fig. 12. Layout of grouting holes. sults are closely associated with the pipe-rock contact angle (2θ), and the
theoretical pipe-rock friction force first decreases and then increases as
NPT (Fig. 11(c)) k is 2.84 kN/m for 30–140 m and 16.17 kN/m for 2θ increases. The theoretical pipe-rock friction force reached the mini­
140–690 m. It is worth noting that the JF increases sharply due to the mum when 2θ is about 30◦ , with the buoyancy of pipe and gravity of
axis deviation locally and turns to normal magnitude during the recti­ itself is in balance, thus the friction force of pipe is mainly made of the
fying operation. Bentonite grouting failure and sediment accumulation pipe-mud friction resistance. When 2θ is 60◦ , the errors between the
lead to the growth of JF. Removing the bottom sediment and injecting a theoretical value and the measured values in No.4, No.7 and No. 8–1
large amount of bentonite allows reducing the JF (see Fig. 11(a), (c)) by NPTs are minor, with the error of 15% in No.4 NPT during the jacking
46.8% and 15.7%, respectively. distance of 100–720 m and error of 1% in No.7 NPT during the jacking
For validation purposes, the JFs were assessed by means of the distance of 120–708 m. Through the geological field record, it was found
different existing models and those presented herein, see Table 10. that the surrounding rock in No.8–1 NPT is more broken than that in
The results gathered in Table 10 shows that in long-distance PJ en­ No.4 and No.7 NPTs. Thus the sediment at the bottom of No.8–1 NPT is
gineering, pipes are in a state of suspension due to the good grouting more than that in No.4 and No.7 NPTs, hence 2θ of No.8–1 NPT during
effect and the traction action of the jacking machine in the initial jacking the jacking distance of 140–690 m will be greater than 60◦ . It shows that
stage. At this stage, the unit friction force (k) is small (2.06–2.94 kN/m) when 2θ is 75◦ , friction force calculated by model III can well reflect the
and the pipe-slurry contact model (model I) can well predict the measured friction force in No. 8–1 NPT during the jacking distance of
measured friction resistance, with the errors for No.4, No.7 and No. 8–1 140–690 m, with the error of just 8%. In general, model III with 2θ of
NPTs being 2%, 45% and 5%, respectively. It can thus be stated that the about 60◦ can better reflect the growth of JF in the later jacking stage of
model I can better predict the magnitude of the frictional resistance of No. 4 and No.7 NPTs, while model III with 2θ of about 75◦ can better
pipe in the early jacking stages. reflect the growth of JF in the later jacking station of No. 8–1 NPT. In
In the later jacking stage, since the buoyancy of the pipe is superior to practical engineering, 2θ is in a dynamic process, and the theoretical
the pipe self-weight, the pipe far away from the jacking machine head calculation of k also fluctuates in a range. Overall, the pipe-rock friction

10
Z. Deng et al. Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 115 (2021) 104063

Fig. 13. Sediment removal procedure: (a) Flushing sediment through opening holes on the pipe; (b) Sediment discharge at the intermediate jacking station;
(c) Sediment.

force calculated by model III can well reflect the field measured friction some stretches (see Fig. 11a). For instance, k of 36.7 kN/m was reached
force in the later jacking stage. between distances 100–150 m in of the No. 4 NPT and sediment was
For the upper-filled contact model (model IV), it reflects the pipe- unremoved owe to the safety margin assumed. However, this decision
rock contact state when 2θ is greater than 180◦ . For the No.4 NPT in led k to increase sharply up to 92.76 kN/m in the stretch 150–171 m
the range of 150–171 m, k is 92.76 kN/m, which is far bigger than that at and, thus, to unacceptably trigger the probability of pipe-stuck. To avoid
any other jacking distance range. Through the site investigation, it was this, the sediment was removed through the existing holes and dis­
found that for part of the pipes, there was no slurry flowing out from charged at the intermediate jacking station (Fig. 13).
grouting holes on both sides of themselves, with the slurry flowing out After this sediment removal maneuver, k gradually decreased and
from grouting holes on top of it showing the state of dripping. Thus, it the jacketing operations were carried out smoothly. Sediment accumu­
can be deduced that 2θ is almost 360◦ . k calculated by model IV is 55.35 lation was also detected in the No. 8–1 NPT (270–320 m) with k
kN/m, with an error of 67% compared to the measured friction resis­ reaching 31.7 kN/m > klim, and the same removal operation was satis­
tance. However, when μs is selected as 0.501 under the condition of poor factorily performed with the subsequent positive expected results. It is,
lubricating effect (see Table 6), k is 83.53 kN/m and is just of 9.9% error hence, evident that timely sediment removal is an effective measure to
when compared to the measured friction resistance. Thus, it shows that reduce the magnitude of the frictional resistance.
Model IV can reflect the frictional resistance of the pipe when the upper
part of the overcut is filled with sediment. 5. Conclusions

An analytical design-oriented model for assessing the surrounding


4.2. Application of the combined calculation model soil – pipe forces existing during pipe-jacketing operations is proposed
herein. The model was validated by means of both numerical and
In long-distance rock PJ, sediment accumulation is a phenomenon experimental results. Based on the outcomes derived from this research,
that affects both the distribution and magnitude of pipe-rock contact the following conclusions can be drawn:
pressure and buoyancy of pipe. Sediment removal can effectively
smoothen the pipe-rock contact and, thus, reduce the friction resistance • The magnitude of the coefficient of uneven contact pressure (kc)
and the likelihood of the pipe from getting stuck in long-distance rock obtained numerically proved to be aligned with that theoretical
pipe-jacketing. Frequent sediment removal delays the construction (error inferior to 5%) when a pipe-rock contact angle ranging from
period whilst its untimely removal leads to blockages in the bentonite 0◦ to 180◦ is considered. Thus, the numerical model was found to be
grouting holes, and slurry erosion increases both pipe-rock contact a suitable approach for assessing kc for other boundary conditions
pressure and pipe-rock friction coefficient. not reproducible by means of theoretical models.
In practical PJ engineering, sediment accumulation can be accepted • Both numerical and analytical results led to an upwards resultant of
even if the friction resistance increases to a certain extent provided this the pipe buoyancy force provided by the bentonite mud for pipe-rock
does not affect the normal jacking process. If the friction resistance af­ contact angles ranging from 0◦ to 120◦ , this causing a net reduction
fects the normal process, timely sediment removal is necessary. In this of the pipe-rock contact pressure. For angles superior to 120◦ ,
project, the sediment cleaning angle of pipe-rock was set according to opposite results were obtained for both the resultant buoyancy force
the layout of the grouting holes (Fig. 12). (downwards) and the pipe-rock contact pressure (increasing). These
The limit value of k (klim) was imposed for those cases in which the results further reinforce the effectiveness of the prompt removal of
pipe-rock contact angle was 120◦ in order to prevent the lateral grouting the sediments.
holes (B and C, Fig. 12) from sediment clogging. To this purpose, model • The experimental results confirmed that in the early stage of the PJ
III was assumed to calculate klim, this resulting in 27.1 kN /m. Taking construction, the grouting effect proved to be efficient and the
klim as reference, sediment removal operations were applied when k > accumulated sediment of the minor amount and, therefore, the pipe
klim. It must be highlighted that klim was computed assuming the was in suspension. In this case, the pipe-slurry contact model (model
effectiveness of the grouting and that the pipe-sediment friction coeffi­ I) proposed can be used to assess the friction resistance per unit
cient (0.332) was lower than that theoretical (0.409 ~ 0.531) so that a length (k). In later jacking stages, k can be estimated by considering
safety margin (23.2% ~ 59.9%) could be guaranteed. either a single pipe-tunnel point (at the outer section of the pipe
During the jacketing operations, local values of k exceeded klim in

11
Z. Deng et al. Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 115 (2021) 104063

crown) contact model (model II) or a 60◦ bottom-filled contact model Cui, Q.-L., Xu, Y.-S., Shen, S.-L., Yin, Z.-Y., Horpibulsuk, S., 2015. Field performance of
concrete pipes during jacking in cemented sandy silt. Tunn. Undergr. Space Technol.
(model III). In case of expecting (or allowing) accumulated sediment,
49, 336–344.
the pipe-rock contact angle may exceed 180◦ and the upper-filled Deng, Z., Liu, X., Chen, P., de la Fuente, A., Zhou, X., Liang, N., Han, Y., Du, L., 2021.
contact model (model IV) is more representative. Basalt-polypropylene fiber reinforced concrete for durable and sustainable pipe
• It was noticed that accumulation of sediments resulted in a lower production. Part 1: Experimental Program. Struct. Concr. https://doi.org/10.1002/
suco.202000759.
mud lubrication effect, and μs increased by 50.9% with respect to the GB 50332-02, 2002. Structural design code for pipeline of water supply and waste water
μs observed for stretches in which sufficient slurry content and good engineering, Beijing, pp. 11–12.
lubrication was consistently provided. These results permit to state GB50268-08, 2008. Code for construction and acceptance of water and sewerage
pipeline works. Ministry of Housing and Urban-Rural Development of the People’s
that early-stage sediment removal is an effective solution to reduce Republic of China, Beijing, China (in Chinese).
the contact pressure and its technical–economic consequences. Japan Microtunnelling Association, 2013. Microtunnelling methods serious II, design,
• To prevent the grouting holes from being blocked by sediments, a construction management and rudiments, Tokyo, pp. 69-72.
Ji, X., Ni, P., Barla, M., 2019a. Analysis of jacking forces during pipe jacking in granular
klim of 27.1 kN/m computed by assuming the model III (pipe-rock materials using particle methods. Undergr. Space 4, 277–288.
contact angle of 120◦ ) was imposed as a threshold of k above which Ji, X., Zhao, W., Ni, P., Barla, M., Han, J., Jia, P., Chen, Y., Zhang, C., 2019b. A method to
sediment removal operations were necessary. The application of this estimate the jacking force for pipe jacking in sandy soils. Tunn. Undergr. Space
Technol. 90, 119–130.
criterion proved to be an effective indirect control measure to avoid Khazaei, S., Shimada, H., Matsui, K., 2004. Analysis and prediction of thrust in using
the blockage of the grouting holes and the as well as of the pipe. slurry pipe jacking method. In: Proceedings of the 30th ITA-AITES world tunnel
congress, Singapore, pp. 22–27.
Li, C., Zhong, Z., Liu, X., Tu, Y., He, G., 2019. Numerical simulation for an estimation of
As remarked in several points throughout the manuscript, the pipe-
the jacking force of ultra-long-distance pipe jacking with frictional property testing
rock interaction is a complex phenomenon; nevertheless, the models at the rock mass–pipe interface. Tunn. Undergr. Space Technol. 89, 205–221.
proposed herein for assessing the friction resistance have led to esti­ ASTM, 2011. Standard Guide for Use of Maxi-Horizontal Directional Drilling for
mation with acceptable errors from the engineering point of view when Placement of Polyethylene Pipe or Conduit under Obstacles Including River
Crossings. West Conshohocken, PA, USA.
these are compared with those experimentally measured for a specific Milligan, G.W.E., Norris, P., 1996. Site-based research in pipe jacking—objectives,
linear pipe-jacketing project. The applicability of this model to pipe- procedures and a case history. Tunn. Undergr. Space Technol. 11, 3–24.
jacketing projects with curved axis still remains to be proven. Like­ Milligan, G.W.E., Norris, P., 1999. Pipe–soil interaction during pipe jacking. In:
Proceedings of the Institution of Civil Engineers - Geotechnical Engineering,
wise, the friction coefficient between the pipe and different bodies (ex., pp. 27–44.
single formation of sandstone, mudstone and the resulting composite) is Namli, M., Guler, E., 2017. Effect of Bentonite Slurry Pressure on Interface Friction of
to be further researched. Pipe Jacking. J. Pipeline Syst. Eng. Pract. 8, 04016016.
Norris, P., 1992. The behaviour of jacked concrete pipes during site installation.
Ong, D.E.L., Choo, C.S., 2018. Assessment of non-linear rock strength parameters for the
CRediT authorship contribution statement estimation of pipe-jacking forces. Part 1. Direct shear testing and backanalysis. Eng.
Geol. 244, 159–172.
Pellet-Beaucour, A.-L., Kastner, R., 2002. Experimental and analytical study of friction
Zhiyun Deng: Conceptualization, Methodology, Investigation, forces during microtunneling operations. Tunnell. Undergr. Space Technol. 17,
Writing - original draft. Ninghui Liang: Conceptualization, Methodol­ 83–97.
ogy. Xinrong Liu: Conceptualization, Methodology, Supervision. Al­ PJA, 1995. Guide to Best Practice for the Installation of Pipe Jacks and Microtunnels.
Pipe Jacking Association, London, UK.
bert Fuente: Conceptualization, Supervision, Visualization. Peng Lin: Praetorius, S., Schößer, B., 2017. Bentonite Handbook: Lubrication for Pipe Jacking.
Conceptualization, Supervision, Visualization. Haoyang Peng: Visuali­ John Wiley & Sons.
zation, Writing - review & editing. Sheil, B., 2021. Prediction of microtunnelling jacking forces using a probabilistic
observational approach. Tunn. Undergr. Space Technol. 109, 103749.
Sheil, B.B., Curran, B.G., McCab, B.A., 2016. Experiences of utility microtunnelling in
Irish limestone, mudstone and sandstone rock. Tunn. Undergr. Space Technol. 51,
Declaration of Competing Interest 326–337.
Shou, K., Yen, J., Liu, M., 2010. On the frictional property of lubricants and its impact on
jacking force and soil–pipe interaction of pipe-jacking. Tunn. Undergr. Space
The authors declare that they have no known competing financial
Technol. 25, 469–477.
interests or personal relationships that could have appeared to influence Sofianos, A., Loukas, P., Chantzakos, C., 2004. Pipe jacking a sewer under Athens. Tunn.
the work reported in this paper. Undergr. Space Technol. 19, 193–203.
Stein, D., Möllers, K., Bielecki, R., 1989. Microtunneling: Installation and renewal of
nonman-size supply and sewage lines by the trenchless construction method. Ernst,
Acknowledgments Berlin, Germany.
Sterling, R.L., 2020. Developments and research directions in pipe jacking and
microtunneling. Underground Space 5, 1–19.
This study is supported by the National Key Research and Develop­ Sun, H., Du, W.S., Liu, C., 2021. Uniaxial compressive strength determination of rocks
ment Program of China (Grant No. 2018YFC1504802), Natural Science using X-ray computed tomography and convolutional neural networks. Rock Mech.
Foundation Project of Chongqing (Grant No. cstc2018jscx-mszdX0071). Rock Eng. https://doi.org/10.1007/s00603-021-02503-1.
Terzaghi, K., 1965. Theoretical Soil Mechanics. John Wiley and Sons.
Prof. Albert de la Fuente also wants to express his gratitude to the
Wang S., Xia C., Ge J., 2014. Formulae of lateral friction resistance for pipe-jacking
Spanish Ministry of Science and Innovation for the financial support considering different forms of mud screen. Rock Soil Mech. 35, 159-166+174 (in
received under the scope of the project CREEF (PID2019-108978RB- Chinese). DOI: 10.16285/j.rsm.2014.01.024.
C32). Xiang, A., Zhu, H., Ding, W., 2008. Relationship of jacking force and average friction
with jacking path in pipe-jacking construction. Rock and Soil Mechanics 29,
1005–1009. https://doi.org/10.16285/j.rsm.2008.04.039 in Chinese.
References Yang, X., Zhang, K., Li, Y., Peng, H., Huang, C., 2013. Theoretical and experimental
analyses of jacking force during deep-buried pipe jacking. Rock and Soil Mechanics
34, 757–761. https://doi.org/10.16285/j.rsm.2013.03.027 in Chinese.
ATV-A 161, 2010. Statische Berechnung von Vortriebsrohren - Entwurf, DWA.
Yang, S., Wang, M., Du, J., Guo, Y., Geng, Y., Li, T., 2020. Research of jacking force of
Broere, W., Faassen, T.F., Arends, G., van Tol, A.F., 2007. Modelling the boring of curves
densely arranged pipe jacks process in pipe-roof pre-construction method. Tunn.
in (very) soft soils during microtunnelling. Tunn. Undergr. Space Technol. 22,
Undergr. Space Technol. 97, 103277.
600–609.
Ye, Y., Peng, L., Zhou, Y., Yang, W., Shi, C.-H., Lin, Y., 2019. Prediction of Friction
Choo, C.S., Ong, D.E.L., 2020. Assessment of non-linear rock strength parameters for the
Resistance for Slurry Pipe Jacking. Applied Sciences 10, 207.
estimation of pipe-jacking forces. Part 2. Numerical modeling. Eng. Geol. 265,
Ye, Y., Peng, L., Yang, W., Lei, M., Shi, C., Wang, Z., 2015. Calculation of jacking force for
105405.
pipe-jacking considering mud slurry thixotropy. Chinese. Journal of Geotechnical
Ciavarella, M., Decuzzi, P., 2001a. The state of stress induced by the plane frictionless
Engineering 37, 1653–1659. https://doi.org/10.11779/CJGE201509012 in Chinese.
cylindrical contact. I. The case of elastic similarity. Int. J. Solids Struct. 38,
Yen, J., Shou, K., 2015. Numerical simulation for the estimation the jacking force of pipe
4507–4523.
jacking. Tunn. Undergr. Space Technol. 49, 218–229.
Ciavarella, M., Decuzzi, P., 2001b. The state of stress induced by the plane frictionless
cylindrical contact. II. The general case (elastic dissimilarity). Int. J. Solids Struct.
38, 4525–4533.

12
Z. Deng et al. Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 115 (2021) 104063

Zhang, H., Zhang, P., Zhou, W., Dong, S., Ma, B., 2016a. A new model to predict soil Zhang, P., Ma, B., Zeng, C., Xie, H., Li, X., Wang, D., 2016b. Key techniques for the
pressure acting on deep burial jacked pipes. Tunn. Undergr. Space Technol. 60, largest curved pipe jacking roof to date: A case study of Gongbei tunnel. Tunn.
183–196. Undergr. Space Technol. 59, 134–145.
Zhang, P., Behbahani, S.S., Ma, B., Iseley, T., Tan, L., 2018. A jacking force study of Zhang, P., Ma, B., Zeng, C., Tan, L., 2017. Numerical model for jacking force based on
curved steel pipe roof in Gongbei tunnel: Calculation review and monitoring data pipe-soil contact characteristics. Chinese J. Geotech. Eng. 39, 244–249.
analysis. Tunn. Undergr. Space Technol. 72, 305–322. Zhou, S., Wang, Y., Huang, X., 2009. Experimental study on the effect of injecting slurry
inside a jacking pipe tunnel in silt stratum. Tunn. Undergr. Space Technol. 24,
466–471.

13

You might also like