Delft University of Technology

A compact blowout model for shallow tunnelling in soft soils

Vu, Minh Ngan; Broere, Wout

DOI
10.1016/j.tust.2023.105167
Publication date
2023
Document Version
Final published version
Published in
Tunnelling and Underground Space Technology

Citation (APA)
Vu, M. N., & Broere, W. (2023). A compact blowout model for shallow tunnelling in soft soils. Tunnelling and
Underground Space Technology, 138, Article 105167. https://doi.org/10.1016/j.tust.2023.105167

Important note
To cite this publication, please use the final published version (if applicable).
Please check the document version above.

Copyright
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent
of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Takedown policy
Please contact us and provide details if you believe this document breaches copyrights.
We will remove access to the work immediately and investigate your claim.

This work is downloaded from Delft University of Technology.

For technical reasons the number of authors shown on this cover page is limited to a maximum of 10.
 Tunnelling and Underground Space Technology 138 (2023) 105167

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology

journal homepage: www.elsevier.com/locate/tust

A compact blowout model for shallow tunnelling in soft soils

Minh Ngan Vu a , Wout Broere b ,∗
a
Department of Infrastructure Engineering, Hanoi University of Mining and Geology, Viet Nam
b
Geo-Engineering Section, Department of Geosciences and Engineering, Delft University of Technology, Stevinweg 1, 2628 CN Delft, The Netherlands

ARTICLE INFO ABSTRACT

Keywords: Tunnelling in soft soil conditions, especially with a shallow overburden, faces the risk of face instability due
Blow-out to blowout. Although several blowout models have been proposed to estimate the blowout pressure, mostly
Stability based on limit analysis or limit equilibrium, there is a significant gap between the allowable blowout pressures
Shallow bored tunnel
predicted by these models and the values observed in case studies, laboratory experiments and numerical
Support pressure
simulations. This paper proposes a compact blowout model, which is more compact compared to a model
Soft soils
proposed by Balthaus (1991). This new blowout model is able to predict blowout pressures more closely to the
value observed by centrifuge testing, reduced scale experiments and case studies, whilst staying conservative.
Its application on the Hochiminh Metroline No. 1 project in this study resulted in a smaller support pressure
in the boring stage to avoid the occurrence of a blowout.

1. Introduction groundwater flow away from the face, whereas (Perazzelli et al., 2014)
consider seepage towards the face. Other authors used limit analysis to
The increasing population density in cities leads to many social derive upper and lower boundaries of support pressure, including (Leca
problems including traffic congestion and environmental pollution. and Dormieux, 1990) who used failure mechanisms based on conical
As surface space becomes more limited and expensive, use of under- blocks. Mollon et al. (2011, 2013), Soubra (2000a,b) and Subrin and
ground space becomes the preferred solution for new transportation Wong (2002b) developed a 3D logarithmic spiral model and extended
modes. However, tunnel construction in urban areas faces many chal- this to multi-block models. This group of models trends towards a
lenges, including face instability, generating large surface settlements horn-shaped failure mechanism. Senent and Jimenez (2015) extended
and potential damage to surrounding buildings. These risks are es- this approach for layered soils. Recently, kinematic models have also
pecially present when tunnelling in soft soil conditions with shallow been proposed based on limit analysis by Soubra (2000a), Soubra et al.
overburdens. (2008), Mollon et al. (2011) and Qarmout et al. (2019).
In order to limit the effect of tunnelling on surrounding structures A number of laboratory experiments and centrifuge tests have been
and limit soil displacements, the support pressure applied at the exca- carried out in order to investigate the mechanism of the face col-
vation face and at the TBM tail has a vital role in tunnelling design. The
lapse. Chambon and Corte (1994) performed centrifuge tests to identify
support pressure is often derived from face stability models including
the stability of tunnel in dry sandy soil. Takano et al. (2006) carried out
as input factors tunnel depth and geotechnical parameters. The range
1 g experiments to find the 3D shape of the failure in tunnelling. Kirsch
of support pressures should allow a TBM operator to simultaneously
(2010) worked on 1 g experiments to investigate the stability of shallow
satisfy the conditions to prevent collapse (active failure) for minimum
tunnels. Other authors tried to study the stability of the tunnelling
support pressure and blow-out (passive failure) for maximum support
process by applying numerical methods such as finite element methods
pressures. Thus, face stability models have been discussed extensively
(FEM) and discrete element method (DEM) including (Augarde et al.,
in literature.
Several authors use the limit equilibrium method with a wedge 2003; Funatsu et al., 2008; Chen et al., 2011; Lu et al., 2014; Zhang
shaped model based on a relatively simple rectangular soil silo and et al., 2015; Alagha and Chapman, 2019).
triangular wedge, including (Horn, 1961; Jancsecz and Steiner, 1994; Mostly, these stability models focus on the face collapse mechanism
Anagnostou and Kovári, 1994) and (Broere, 2001). Anagnostou and at the lower limit of the support pressure, below which an active
Kovári (1994) consider slurry infiltration and the effect on face sta- collapse takes place. Meanwhile, the maximum allowable pressure
bility during stand-still. Broere (2001) considers slurry infiltration and estimated from the passive failure condition has been studied less.

∗ Corresponding author.
E-mail address: [email protected] (W. Broere).

https://doi.org/10.1016/j.tust.2023.105167
Received 30 October 2022; Received in revised form 13 March 2023; Accepted 15 April 2023
Available online 3 May 2023
0886-7798/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
M.N. Vu and W. Broere Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 138 (2023) 105167

Fig. 1. The compact blowout model proposed in this study.

Using limit analysis, the maximum support pressure is often esti- the cover-to-diameter ratio is relatively small, the channel extends to
mated from the upper boundary condition. Leca and Dormieux (1990) the ground surface and forms a channel whose size is basically the
show from a limit analysis for passive failure using elliptic cones same as the diameter of the shield. This is a more compact area than
emanating from the tunnel face, that this approach results in very high predicted by conical limit analysis models. Similarly, Li et al. (2009)
allowable maximum support pressure. Mollon et al. (2011) use a similar argue that for blow out or passive failure a partial face mechanism
approach and find comparable high support pressures for the case of a needs to be considered, as a full face mechanism overestimates the
solid passive failure mechanism, as do Li et al. (2020) and Liu et al. activated resistance of the soil body.
(2021). The theoretical maximum allowable support pressure, or blow- Balthaus (1991) proposed a 3D blow-out limit equilibrium model
out pressure, estimated with the upper boundary condition in limit with the assumption that an obelisk shaped solid mass of soil above the
analysis approaches is often higher than what has been observed in field tunnel face is displaced. This is consistent with a more limited failure
conditions and in experiments. shape than obtained from limit analysis, but still results in relatively
Numerical and analytical simulation has been an important method high allowable support pressure.
to investigate the actual size and shape of the mechanism and the On the other hand, there are simple models based on limit equilib-
resulting scope of the blowout zone. Verruijt and Booker (1998) de- rium which start from different mechanisms and are extremely conser-
veloped a 2D stability analysis including blowout. More recently, using vative, such as models assuming simple fracturing, but these can often
numerical simulation combined with limit analysis models, Wong and be used as a first (overly) conservative estimate. For example, Broere
Subrin (2006) and Mollon et al. (2010, 2013) introduced models in- (2001) introduced a 2D model that assumes that only a rectangular
cluding 3D failure mechanisms and derived a good agreement between soil body above the tunnel lining is pushed upwards by high support
limit analysis models and numerical models, but still these models tend pressure, which is applicable for cases where fracturing is dominant,
to predict higher allowable pressures than actual field observations. Liu but often higher allowable face pressure can be applied before blow
et al. (2022) studies the passive failure and presents one of the few out actually occurs. Vu et al. (2015) introduced a model with simple
model tests available in literature. When a passive failure of the ex- overburden assumptions that seem to be most suitable for the estima-
cavation face occurs in their tests, there is a mud channel formed by tion of the maximum support pressure at the tail of the shield in cases
the slurry splitting the stratum in front of the excavation face. When of shallow overburden.

2
M.N. Vu and W. Broere Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 138 (2023) 105167

Fig. 2. Sketch of centrifuge tests in Bezuijen and Brassinga (2006).

Only a limited number of experimental studies were carried out

studying blowout incidents when tunnelling. Apart from the study
by Liu et al. (2022) mentioned above, Bezuijen and Brassinga (2006)
report on centrifuge tests performed by GeoDelft when studying the
grouting process along the TBM. Berthoz et al. (2012) performed sev-
eral tests to investigate both face collapse and blowout with an original
reduced-scale model of an EPB machine in stratified soils. The blowout
pressure recorded from these test shows a large difference with the
calculated pressures derived from the limit analysis model of Subrin
and Wong (2002a) and Wong and Subrin (2006). As far as case studies
go, only few cases of blowout have been reported in literature for recent
tunnelling projects, even though it is suggested that blowouts occur
more frequently. A clear and documented blowout case occurred during
Second Heinenoord Tunnel project, in the Netherlands, as reported
by Bezuijen and Brassinga (2006) and Broere (2001).
Although many models have been proposed to predict the maximum
support pressure that should be applied in tunnelling to avoid blowouts,
large differences still exist between the predicted blowout pressure
derived from recent models and the actual blowout pressures recorded
in both experimental tests and case studies. Thus, there is a need for
a model that results in predictions closer to observed blow out cases.
This paper proposes a new blowout model, which can be characterized
as a more compact and modified version of the limit equilibrium model
proposed by Balthaus (1991).

2. A new compact blow-out model

As indicated above, relatively simple models are still useful in

practice as they offer the advantage of easy calculation and simple
assumptions, compared to more advanced numerical models. The three-
dimensional limit equilibrium model originally proposed by Balthaus
(1991) is still applied as a useful calculation tool in tunnelling design,
despite the relatively high maximum support pressures it predicts,
which is due to the assumption of an obelisk shaped solid mass of soil
being pushed upward by the blowout pressure. Berthoz et al. (2012)
carried out a study on a 1 g reduced scale model of an earth pressure
balanced (EPB) shield to investigate the failure mechanisms of face
collapse and blowout. The analysis of these experiments shows that
the resulting failure body shape is not an obelisk shaped soil column
as indicated by Balthaus (1991). This might be the cause of too high
maximum support pressures being estimated for blowout conditions in
the case of Balthaus’ model. The observations in Berthoz et al. (2012) Fig. 3. Validations with three centrifuge tests in GeoDelft.
are in good agreement with the illustrations and mechanism derived
from numerical analysis in Soubra (2000b), Li et al. (2009), Zhang et al.
(2015) that all show the scope of the uplift soil volume is limited to shape. The scope of the soil body pushed up during blowout is reduced
above and ahead the tunnelling face. as shown in Fig. 1, to the extent that only the soil mass ahead the
Based on these observations, the obelisk shaped soil mass assumed tunnelling face is taken into account, and the impact of the two side
in Balthaus’s model seems to be inappropriate and should be modified. wedges of the soil body is halved by assuming the failure plane to have
This study proposes a new blow-out model with a more compact failure a steeper angle here.

3
M.N. Vu and W. Broere Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 138 (2023) 105167

where 𝜑 the soil friction angle.

Then, the volume of the obelisk shaped soil mass 𝑉𝑜𝑏 can be ex-
pressed:
( ) ( ) ( )
𝜋 𝜑 𝜋 𝜑 4 𝜋 𝜑
𝑉𝑜𝑏 = 𝐴𝐵𝐶 + 𝐴𝐶 2 cot + + 𝐵𝐶 2 cot + + 𝐶 3 cot 2 +
4 2 4 2 3 4 2
(6)

The dimensions 𝐴′ and 𝐵′ of the proposed blowout soil mass as

shown in Fig. 1 can be calculated as:
( )
𝜋 𝜑
𝐴′ = 𝐴 + 𝐶 cot + (7)
4 2
( )
𝜋 𝜑
𝐵 ′ = 𝐵 + 𝐶 cot + (8)
4 2
Then, the volume of the reduced soil mass 𝑉𝑟𝑒 is identified as
follows:
( ) ( ) ( )
1 𝜋 𝜑 1 𝜋 𝜑 𝜋 𝜑
𝑉𝑟𝑒 = 𝐴𝐶 2 cot + + 𝐵𝐶 2 cot + + 𝐶 3 cot 2 + (9)
2 4 2 2 4 2 4 2
Fig. 4. Sectional diagram of EPBS model of ENTPE (Berthoz et al., 2012).
From Eqs. (2), (6), and (9), the volume of the blowout soil body
in Eq. (2) is shown to be equal to:
[ ( ) ( )
Table 1 1 𝜋 𝜑 1 𝜋 𝜑
𝑉 = 𝐶 𝐴𝐵 + 𝐵𝐶 cot + + 𝐴𝐶 cot +
Soil parameters used in centrifuge tests (Bezuijen and Brassinga, 2 4 2 2 4 2
( )]
2006). 1 𝜋 𝜑
+ 𝐶 2 cot 2 + (10)
Soil parameters Speswhite Sand med. 3 4 2
clay dens. and the weight of the soil body in Eq. (1) is:
𝛾𝑤𝑒𝑡 (kN∕m3 ) 17 19.6 [ ( ) ( )
1 𝜋 𝜑 1 𝜋 𝜑
𝐺 = 𝛾𝐶 𝐴𝐵 + 𝐵𝐶 cot + + 𝐴𝐶 cot +
c(kPa) 1 8.3 2 4 2 2 4 2
( )]
1 𝜋 𝜑
Friction angle (deg.) 23 37 + 𝐶 2 cot 2 + (11)
Dilatancy angle (deg.) – 9
3 4 2
Poisson’s ratio(–) 0.45 0.3 The total support force at the top of the tunnelling face can subse-
𝐸50 (MPa) 0.53 0.4 quently be estimated as:
n(–) – 0.394 𝑃 = 𝐴𝐵𝑠 (12)

where 𝑠 is the support pressure. Although defined in Eq. (11) only

Table 2
as a function of the friction angle of the soil, the effective support
Parameters of soil used in reduced scale experiments in Berthoz et al.
(2012). pressure length at the tunnel roof 𝐵 in reality depends on many factors
Test Weight unit Cohesion Friction angle Elastic modulus including the ratio of depth to diameter of the tunnel 𝐶∕𝐷, the soil
(kN∕m3 ) (kPa) (𝑜 ) (MPa) types encountered, the type of support fluids used and the penetration
MC1 12.20 3.2 36 6 mechanism of the support fluid. This actual length of 𝐵 is difficult to
MC3 13.20 2.5 36 10 determine a priori as it depends on the interaction of the soil conditions
MC5 13.05 1.5 36 10 and support fluid applied on the tunnelling face. Therefore, Balthaus
(1991) suggested the weight of the soil body might be presented as a
function of the effective support pressure length at the tunnel roof 𝐵
as:
The weight of the soil body 𝐺 can be calculated as follows:
𝐺 = 𝛼 + 𝛽𝐵 (13)
𝐺 = 𝛾𝑉 (1)
where the indexes 𝛼 and 𝛽 are given as:
where 𝑉 the volume of the blowout soil body, 𝛾 the soil weight unit.
( )[ ( )]
The volume of the blowout soil body 𝑉 can be estimated as follows: 𝜋 𝜑 𝐴 𝐶 𝜋 𝜑
𝛼 = 𝛾𝐶 2 cot + + cot + (14)
4 2 2 3 4 2
[ ( )]
1 𝜋 𝜑
𝑉 = 𝑉𝑜𝑏 − 𝑉𝑟𝑒 (2) 𝛽 = 𝛾𝐶 𝐴 + 𝐶 cot + (15)
2 4 2
where 𝑉𝑜𝑏 the volume of the obelisk shaped soil mass indicated in The maximum allowable support pressure can then be derived from
Balthaus (1991) and 𝑉𝑟𝑒 the volume by which is it reduced as proposed the equilibrium of the total support force 𝑃 and the weight of soil body
in this study. 𝐺. When including the safety factor, in line with (Balthaus, 1991), the
The volume of the obelisk shaped soil mass 𝑉𝑜𝑏 is calculated as: maximum support pressure is estimated as:
𝐶[ ]
𝐺 𝛽 𝛼
𝑉𝑜𝑏 = 𝐴𝐵 + 𝐴𝑜𝑏 𝐵𝑜𝑏 + (𝐴 + 𝐴𝑜𝑏 )(𝐵 + 𝐵𝑜𝑏 ) (3) 𝜂= = + (16)
6 𝑃 𝐴𝑠 𝐴𝐵𝑠
where 𝐴 the width of equivalent rectangle, 𝐵 the effective support where 𝜂 is the safety factor.
pressure length at the tunnel roof and 𝐴𝑜𝑏 , 𝐵𝑜𝑏 are dimensions of the Since Eq. (16) involves the effective support pressure length at the
obelisk shaped soil mass as shown in Fig. 1, and 𝐶 the overburden depth tunnel roof 𝐵, which is difficult to determine, two other safety indexes
of the tunnel.
have been proposed by Balthaus (1991) as:
The obelisk dimensions 𝐴𝑜𝑏 and 𝐵𝑜𝑏 can be calculated as:
( ) 𝛽 𝛼 𝛽 𝛾𝐶
𝜋 𝜑 𝜂= + > 𝜂1 = > 𝜂2 = (17)
𝐴𝑜𝑏 = 𝐴 + 2𝐶 cot + (4) 𝐴𝑠 𝐴𝐵𝑠 𝐴𝑠 𝑠
4 2
( ) where 𝜂1 , 𝜂2 are safety factors. The safety factor 𝜂2 is often used for the
𝜋 𝜑
𝐵𝑜𝑏 = 𝐵 + 2𝐶 cot + (5) case of very deep tunnels (Balthaus, 1991) and safety factor 𝜂1 is used
4 2

4
M.N. Vu and W. Broere Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 138 (2023) 105167

in case of shallow and moderate tunnels. These safety factors might

be different based on the national construction code applied. Balthaus
(1991) proposed a value of safety factor 𝜂1 = 1.1 for blowout analysis.
In this paper, the blowout in the case of shallow tunnelling in soft soils
is studied, thus the blowout pressures can be estimated from Eq. (17)
with the factor of safety 𝜂1 as:
𝛽
𝑠= (18)
𝐴𝜂1

3. Validations with experimental data

3.1. Validation with GeoDelft centrifuge tests

Validation of the new model is first performed against three cen-

trifuge tests reported by Bezuijen and Brassinga (2006). These exper-
iments were carried out in the GeoDelft geotechnical centrifuge when
investigating the grouting process around the tunnel after a blowout
incident during construction of the Second Heinenoord Tunnel in the
Netherlands. In the centrifuge tests, a tube with an outer diameter of
130 mm and an inner diameter of 125 mm is used for representing
the tunnel. The 25 mm tail void in this model was directly filled by
a bentonite slurry as can be seen in Fig. 2.
The parameters of the soil used in the three centrifuge tests, as
reported by Bezuijen and Brassinga (2006), are shown in Table 1. The
tunnel tube in the first centrifuge experiment is covered by 200 mm
saturated sand. The second and the third tests have similar conditions
in that the tube is covered by sand and clay layers including a 77.5 mm
thick sand layer directly of above the tunnel, a 170 mm clay layer above
the sand layer and a 5 mm sand layer on top, with the water level at
the top of the 5 mm sand layer.
The first centrifuge test was performed at 150 g representing a
large diameter tunnel of 18.75 m and an excess bentonite pressure was
recorded at 620 kPa when the blow-out occurred. Meanwhile, the sec-
ond and the third centrifuge tests were carried out at 40 g representing a
tunnel with a smaller diameter of 5 m. The blowout pressures recorded
in these tests are 190 kPa and 215 kPa, correspondingly. Bezuijen and
Brassinga (2006) also used a numerical analysis by the Plaxis FEM
programme with a hardening soil model in the case of the second and
the third centrifuge tests. The blow-out pressure found in the numerical
analysis was equal to the results from the third centrifuge case, as the
total of the pore pressure and 2.5 times the vertical effective stress.
In order to compare the accuracy of the newly proposed model, the
blowout pressure is compared to experimental data and overburden
models proposed by Broere (2001) and Vu et al. (2015), the limit equi-
librium model by Balthaus (1991), the limit analysis for passive failure
as indicated in Leca and Dormieux (1990) and numerical analysis
results.
Fig. 3 shows the validation against blowout centrifuge tests from
GeoDelft. The blowout pressures predicted by the limit analysis model
by Leca and Dormieux (1990) are too high compared to the other
results, so these pressures are off the chart and are not shown. Fig. 3(a)
shows results for conditions identical to the first centrifuge test. This
shows that the blow-out pressure predicted by the new model is close
to the blow-out pressure recorded in the centrifuge test. The allowable
pressure predicted by Vu et al. (2015) is rather similar. Meanwhile, the
pressure predicted by Balthaus’ model is about 1.5 times larger than
the experimental values and the pressure predicted by Broere’s model
seems rather conservative.
For the second and the third centrifuge tests, as shown in Fig. 3(b),
the blowout pressures predicted by the new model are nearly equal to
the centrifuge test data and numerical analysis results. In these cases,
the blowout pressures predicted by Broere (2001) and Vu et al. (2015)
are still conservative, and the blowout pressures predicted by Balthaus
(1991) are about two times higher than numerical analysis data and
experimental data.
This validation shows the new model can predict blowout pres- Fig. 5. Validations with blowout cases in reduced scale experiments in Berthoz et al.
sure more accurately than Balthaus’ model and closer to the observed (2012).
pressures in the tests.

5
M.N. Vu and W. Broere Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 138 (2023) 105167

Fig. 6. The blow-out case in the Second Heinenoord Tunnel project (Bezuijen and Brassinga, 2006).

Table 3
Blowout pressures recorded in reduced scale experiments in Berthoz et al. (2012) and estimated from blowout models.
Test Experimental data Kinematic approach Balthaus (1991) Broere (2001) Vu et al. (2015) The new compact
Berthoz et al. (2012) Subrin and Wong (2002a) model model model blowout model
(kPa) (kPa) (kPa) (kPa) (kPa) (kPa)
MC1-B1 34 612 15.50 15.68 27.37 9.61
MC1-B2 10 612 15.50 15.68 27.37 9.61
MC3-B2 21 612 17.85 16.51 29.04 10.40
MC5-B1 47 515 17.65 12.02 22.43 10.29
MC5-B2 21 515 17.65 12.02 22.43 10.29

3.2. Validation with the reduced scale experiments in Berthoz et al. (2012) to Subrin and Wong (2002b) are too large, as are the results from Leca
and Dormieux (1990), so that they are off the graph in Fig. 5.
Berthoz et al. (2012) carried out a number of 1 g Earth Pressure Fig. 5 shows the validation against the reduced scale experiments
Balance (EPB) shield reduced scale experiments to investigate the by Berthoz et al. (2012). In Fig. 5(a), in the case of test MC1-B1, all
mechanism of face collapse and face blowout. The dimensions of the predicted blowout pressures are smaller than the pressure recorded in
experimental container are 2 × 1.3 × 1.3 m. The EPB equipment involves the test, whereas in the case of test MC1-B2, only the new model can
a 55 cm cutter head, a conical working chamber, a screw conveyor, a predict the blowout pressure with good agreement, while other models
horizontal screw conveyor, a cylindrical shield tail, a cutter wheel and predict larger values. Fig. 5(b) shows the validation with the test MC3-
four thrust jacks as can be seen in Fig. 4. The soil used in the experiment B2. It can be seen that the blowout pressure predicted by Balthaus
was Houston 𝑆28 sand with soil parameters shown in Table 2. In this (1991) is slightly lower than the observed value in the test. The simple
study, only the experiments leading to blowout are considered. models of Broere (2001) and Vu et al. (2015) show a relatively good
Table 3 compares predicted blowout pressures calculated by the agreement for this case, while the new model delivers a conservative
kinematic approach as indicated in Berthoz et al. (2012), by simple prediction. Fig. 5(c) presents the validation with the tests MC5-B1 and
models proposed by Broere (2001) and Vu et al. (2015), by the limit MC5-B2. For MC5-B1, all models predict a lower blowout pressure
equilibrium model proposed by Balthaus (1991) and the new model. than the observed pressure in the test. However, in the case of test
The blowout pressures predicted by the kinematic approach according MC5-B2, the (Balthaus, 1991) model shows a lower predicted blowout

6
M.N. Vu and W. Broere Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 138 (2023) 105167

Table 4
The geotechnical conditions of the blow-out location in Hochiminh Metroline No. 1 project.
Layer Thickness Unit weight Cohesion Friction angle
(m) (kN∕m3 ) (kPa) (𝑜 )
Fill - F 1.6 19.0 10 28
Alluvium soft clayed silt - Ac2 2.65 16.5 10 24
Alluvium sand - As1 8.01 20.5 0 31
Alluvium sand - As2 7.51 20.5 0 31
Alluvium sand - As3 15.21 20.5 0 31
Dilluvium hard clayed silt - Dc 14.07 21.0 22 0
Dilluvium dense silty sand - Ds – 21.0 0 34

The simple overburden model proposed by Vu et al. (2015) gives the

blowout pressure equal to the blowout pressure recorded at the top
of the tunnel, but higher than the recorded value if safety factors
are applied. Meanwhile, the blowout pressure derived from the new
model is about 0.86 times the recorded blowout pressure at the top
of the tunnel and slightly larger than the blowout pressure predicted
by Broere (2001).
The validation against the Second Heneinoord Tunnel case shows
that the new model results in a close but still conservative prediction of
blowout pressure for tunnelling in soft soil, although all simple models
result yield similar blowout pressures.

4.2. Validation with a case study of Hochiminh Metroline no. 1, Vietnam

A second case study used to validate the new model is Hochiminh

Metroline No1 project, where a blowout occurred. Hochiminh Metro-
line No. 1 – a pilot metro project in Vietnam – launched in the year
2012 and completed the underground construction works in the year
2020. This metro line has a total length of 19.7 km from Ben Thanh
Fig. 7. The validation with the blowout case study at the Second Heneinoord Tunnel. station to Suoi Tien park with fourteen stations. The underground part
of this metroline is about 2.6 km with two circular tunnels under
a dense urban area with many important buildings such as Ba Son
pressure than the value recorded in the test. The other models show a shipyard and the Saigon Municipal Opera House. Two circular tunnels
good agreement, deriving lower blowout pressures than the test value, were excavated at depths varying from 15 to 20 m in the case of
with (Broere, 2001) model and the new model obtaining conservative The West Tunnel and varying from 20 to 30 m in the case of the
results in comparison to Vu et al. (2015). East tunnel. The outer diameter of the tunnel is 6.65 m and the inner
The validation against reduced scale experiments carried out by diameter is 6.05 m. The tunnel is located in Mekong deltaic soils with
Berthoz et al. (2012) shows that the proposed model can predict predominantly soft clay layers and silty sand layers.
blowout pressure more closely to the observed pressures in the ex- During tunnel construction, a blowout occurred on the 23th April
periments than previous models, even though these predictions are 2018, at chainage Km1 + 154.4 of the West Line as can be seen in Fig. 8.
sometimes still conservative. The tunnel here is located at a depth of −11.627 m corresponding to
a 𝐶∕𝐷 ratio of 1.25. The support pressure at the site when blowout
4. Validations with case studies occurred is recorded as 335 kPa. Detailed soil condition at this location
are presented in Table 4.
4.1. Validation with a case study of Second Heinenoord Tunnel The validation with the blowout case at Hochiminh Metroline No. 1
is shown in Fig. 9. As the 𝐶∕𝐷 ratio is approximately 1.25, the differ-
The next validation is carried out with the case study of the ob- ence between predicted blowout pressures for the various models can
served blowout in Second Heinenoord Tunnel project in the Nether- be seen clearer. The limited equilibrium model proposed by Balthaus
lands (see Fig. 6). The tunnel was constructed under the Oude Maas (1991) derives a higher blowout pressure than the value observed.
river, in the period from 1996 to 1999. The tunnel has an outer Meanwhile, the simple overburden model proposed by Vu et al. (2015)
diameter of 8.3 m. As shown in Fig. 6(a), the blowout occurred when gives a blowout pressure close to the blowout pressure recorded at the
the tunnel passes the Oude Maas river. At that location, the tunnel is top of the tunnel. Broere (2001) delivers the most conservative value,
covered by 4 m of Pleistocence sand with a friction angle of 36.5◦ whilst about 0.6 times of recorded site blowout pressure. It can be seen that
the total cover of 8.6 m consist of this sand layer and very soft Holocene the new model derives a more suitable, but still conservative, blowout
layers and recent river deposits, as well as 11 m of water column above pressure of about 0.7 times the recorded site blowout pressure. The
the soil (Bezuijen and Brassinga, 2006). blowout pressure predicted by Leca and Dormieux (1990) is again too
Fig. 6(b) shows the support pressure recorded at the tunnelling face high and off the graph.
when the blow-out occurred, which was 450 kPa at the centre of the
tunnel or 405 kPa at the crown of the face. 5. Discussion and application
Fig. 7 shows that the results from blowout models proposed by
Balthaus (1991), Broere (2001), Vu et al. (2015) and the new model Validations with laboratory results and two case studies as indicated
are all close to the case study observation. The blowout pressure above show that the new blowout model can obtain an accurate and
derived (Leca and Dormieux, 1990) by limit analysis for passive failure safe estimate of blowout support pressures. For the GeoDelft centrifuge
is again too high and off the graph. It can be seen that Balthaus’ tests, the (Balthaus, 1991) limit equilibrium model predicts a high
model predicts a pressure slightly higher than the recorded site value. blowout pressure in comparison to other models and experimental

7
M.N. Vu and W. Broere Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 138 (2023) 105167

Fig. 8. Blowout case in Hochiminh Metroline No. 1 at the chainage of Km1 + 154.4.

by Broere (2001) and Vu et al. (2015) models. Also, as can be seen in

the figure, the recorded blowout pressure is not the highest support
pressure applied in boring stage during the tunnelling process. This
can be explained by the presence of other structures such as nearby
buildings and road pavement that are not taken into account in the
various models due to the complicated calculation it would result in.
Thus, the actual blowout pressure including the effect of surrounding
structures might be higher than the blowout pressure when tunnelling
in ‘‘greenfield’’ conditions, but it is not recommended to include these
effects when estimating safe support pressure ranges, especially when
tunnelling in urban areas.
The support pressure during the erecting stage, when the TBM is at
standstill for erecting tunnel segments inside the TBM, is less than the
maximum support pressures predicted by the new model and higher
than the values predicted by Broere (2001) and Vu et al. (2015), and
much higher than the minimum support pressure. This indicates that
the TBM operators tended to drive the TBM at a relatively high support
pressure in order to reduce overexcavation of soil ahead of the face and
Fig. 9. The validation with the blowout case at the Hochiminh Metroline No. 1.
improve face stability, but at an increased risk of blowout. This analysis
shows that a smaller support pressure should be applied in the boring
stage and less than 300 kPa should have been applied in the West tunnel
data. In the reduced scale experiments by Berthoz et al. (2012), the in Hochiminh Metroline No. 1 project to avoid any risk of instability.
simple overburden model by Vu et al. (2015) shows a higher predicted
blowout pressure than the test data, although this model performs
better in the two case studies. In all validation cases, the simple 6. Conclusions
overburden model by Broere (2001) shows to be very conservative. The
new limit equilibrium model proposed in this study gives a reasonable Tunnelling in soft soil conditions, especially in urban areas with
predicted blowout in all validation cases. However, it is also clear deltaic soils, faces the challenge of face stability. Based on the observed
that the number of laboratory tests focusing on blowout pressures, scope and shape of the failure mechanism of blowout in experiments
as well as the number of documented case studies, is rather small. A and numerical simulations, a new three-dimensional blowout is pro-
more extensive comparison with field cases would be preferred, also
posed in this study. Validations with three centrifuge tests and a
to see what uncertainties still remain in the models and the blowout
number of reduced scaled experiment by Berthoz et al. (2012) show
mechanisms observed in practice.
that the new model can predict a more precise and safe blowout pres-
Fig. 10 shows the support pressure range calculated for Hochiminh
sure than blowout models by Balthaus (1991), Broere (2001) and Vu
Metroline No. 1 chainages from Km1 + 080 to Km1 + 170. The maximum
et al. (2015). Further validation in two blowout case studies of the
support pressure lines in the figure are derived from blowout models
Second Heinenoord Tunnel and the Hochiminh Metroline No. 1 also
proposed by Balthaus (1991), Broere (2001), Vu et al. (2015) and the
new model. The minimum support pressures are obtained from the confirm the good agreement between the site observations and the
wedge model proposed by Jancsecz and Steiner (1994). The operational blowout pressure predictions from the new model. The application of
support pressures recorded at the site for the boring stage and the the new model to the Hochiminh Metroline No. 1 shows that smaller
erecting stage of the tunnelling process are also shown. The blowout support pressures should have been applied in the boring stage when
occurrence as indicated above is also marked on the graph. tunnelling with a low cover to diameter ratio 𝐶∕𝐷 of less than 2, in
It can be seen that the support pressures applied in the boring stage order to avoid the risk of blowout. Overall the new model provides
fluctuated around the maximum support pressures predicted by the new a more reliable but still conservative estimate of allowable support
model, which is lower than the maximum support pressures predicted pressure. This study also highlights the limited number of reported
by Balthaus and higher than the maximum support pressures predicted blowout case studies where detailed data for analysis is available.

8
M.N. Vu and W. Broere Tunnelling and Underground Space Technology incorporating Trenchless Technology Research 138 (2023) 105167

Fig. 10. Predicted and recorded support pressures in Hochiminh metroline No. 1 chainages from Km1 + 080 to Km1 + 170.

Declaration of competing interest Leca, E., Dormieux, L., 1990. Upper and lower bound solutions for the face stability
of shallow circular tunnels in frictional material. Géotechnique 40 (4), 581–606.
Li, Y., Emeriault, F., Kastner, R., Zhang, Z., 2009. Stability analysis of large slurry
The authors declare that they have no known competing finan-
shield-driven tunnel in soft clay. Tunnell. Undergr. Space Technol. 24 (4), 472–481.
cial interests or personal relationships that could have appeared to Li, D., Zhao, L., Cheng, X., Zuo, S., Jiao, K., 2020. Upper-bound limit analysis of passive
influence the work reported in this paper. failure of a 3D shallow tunnel face under the bidirectional inclined ground surfaces.
Comput. Geotech. 118, 103310.
Liu, D., Liu, X., Han, Y., Xiong, F., Liu, R., Lin, C., Deng, Z., Xiao, Y., Luo, W.,
Data availability
2022. Model test on the passive failure of slurry shield tunneling in circular-gravel
stratum. Earth and Space Science 9 (6), e2021EA002199.
Data will be made available on request Liu, W., Wu, B., Shi, P., Cai, G., 2021. Upper bound analysis of working face
passive failure in large-diameter shield tunneling based on a composite mechanism.
Comput. Geotech. 138, 104362.
Acknowledgements
Lu, X., Wang, H., Huang, M., 2014. Upper bound solution for the face stability of shield
tunnel below the water table. Math. Probl. Eng. 2014.
This study was supported by the Ministry of Education and Training Mollon, G., Dias, D., Soubra, A.-H., 2011. Rotational failure mechanisms for the face
of Vietnam, Grant No. B2021-MDA-05. stability analysis of tunnels driven by a pressurized shield. Int. J. Numer. Anal.
Methods Geomech. 35 (12), 1363–1388.
Mollon, G., Dias, D., Soubra, A.-H., 2013. Continuous velocity fields for collapse and
References blowout of a pressurized tunnel face in purely cohesive soil. Int. J. Numer. Anal.
Methods Geomech. 37 (13), 2061–2083.
Alagha, A.S., Chapman, D.N., 2019. Numerical modelling of tunnel face stability Mollon, G., Phoon, K.K., Dias, D., Soubra, A.-H., 2010. Validation of a new 2D failure
in homogeneous and layered soft ground. Tunnell. Undergr. Space Technol. 94, mechanism for the stability analysis of a pressurized tunnel face in a spatially
103096. varying sand. J. Eng. Mech. 137 (1), 8–21.
Anagnostou, G., Kovári, K., 1994. The face stability of slurry-shield-driven tunnels. Perazzelli, P., Leone, T., Anagnostou, G., 2014. Tunnel face stability under seepage
Tunnell. Undergr. Space Technol. 9 (2), 165–174. flow conditions. Tunnell. Undergr. Space Technol. 43, 459–469.
Augarde, C.E., Lyamin, A.V., Sloan, S.W., 2003. Stability of an undrained plane strain Qarmout, M., König, D., Gussmann, P., Thewes, M., Schanz, T., 2019. Tunnel face
heading revisited. Comput. Geotech. 30 (5), 419–430. stability analysis using kinematical element method. Tunnell. Undergr. Space
Balthaus, H., 1991. Tunnel face stability in slurry shield tunnelling. In: Proceeding Technol. 85, 354–367.
12th International Conference on Soil Mechanics and Foundation Engineering, Rio Senent, S., Jimenez, R., 2015. A tunnel face failure mechanism for layered ground,
de Janeiro, 13–18 August 1989 V2, P775–778. 28, (6), Pergamon, p. A391. considering the possibility of partial collapse. Tunnell. Undergr. Space Technol.
Berthoz, N., Branque, D., Subrin, D., Wong, H., Humbert, E., 2012. Face failure in 47, 182–192.
homogeneous and stratified soft ground: Theoretical and experimental approaches Soubra, A.-H., 2000a. Kinematical approach to the face stability analysis of shallow
on 1g EPBS reduced scale model. Tunnell. Undergr. Space Technol. 30, 25–37. circular tunnels. In: 8th International symposium on plasticity.
Bezuijen, A., Brassinga, H.E., 2006. Blow-out pressures measured in a centrifuge model Soubra, A.-H., 2000b. Three-dimensional face stability analysis of shallow circular
and in the field. In: Tunnelling: A Decade of Progress: GeoDelft 1995-2005. Taylor tunnels. In: ISRM International Symposium. OnePetro.
& Francis Group, p. 143. Soubra, A.-H., Dias, D., Emeriault, F., Kastner, R., 2008. Three-dimensional face stability
Broere, W., 2001. Tunnel Face Stability & New CPT Applications (Ph.D. thesis). Delft analysis of circular tunnels by a kinematical approach. In: GeoCongress 2008:
University of Technology. Characterization, Monitoring, and Modeling of GeoSystems. pp. 894–901.
Subrin, D., Wong, H., 2002a. Stabilité du front d’un tunnel en milieu frottant: un
Chambon, P., Corte, J.-F., 1994. Shallow tunnels in cohesionless soil: stability of tunnel
nouveau mécanisme de rupture 3D. C. R. Méc. 330 (7), 513–519.
face. J. Geotech. Eng. 120 (7), 1148–1165.
Subrin, D., Wong, H., 2002b. Tunnel face stability in frictional material: a new 3D
Chen, R., Tang, L., Ling, D., Chen, Y., 2011. Face stability analysis of shallow shield
failure mechanism. C. R. Mec. 330 (7), 513–519.
tunnels in dry sandy ground using the discrete element method. Comput. Geotech.
Takano, D., Otani, J., Nagatani, H., Mukunoki, T., 2006. Application of x-ray CT
38 (2), 187–195.
on boundary value problems in geotechnical engineering: research on tunnel
Funatsu, T., Hoshino, T., Sawae, H., Shimizu, N., 2008. Numerical analysis to better
face failure. In: GeoCongress 2006: Geotechnical Engineering in the Information
understand the mechanism of the effects of ground supports and reinforcements on
Technology Age. pp. 1–6.
the stability of tunnels using the distinct element method. Tunnell. Undergr. Space
Verruijt, A., Booker, J., 1998. Surface settlements due to deformation of a tunnel in
Technol. 23 (5), 561–573.
an elastic half plane. Geotechnique 48 (5), 709–713.
Horn, N., 1961. Horizontal earth pressure on the vertical surfaces of the tunnel tubes. Vu, M.N., Broere, W., Bosch, J., 2015. The impact of shallow cover on stability when
In: National Conference of the Hungarian Civil Engineering Industry, Budapest. pp. tunnelling in soft soils. Tunnell. Undergr. Space Technol. 50, 507–515.
7–16. Wong, H., Subrin, D., 2006. Stabilité frontale d’un tunnel: mécanisme 3D en forme
Jancsecz, S., Steiner, W., 1994. Face support for a large mix-shield in heterogeneous de corne et influence de la profondeur. Revue Européenne de Génie Civil 10 (4),
ground conditions. In: Tunnelling’94. Papers Presented at the Seventh International 429–456.
Symposium’TunnellinG’94’, Held 5-7 JULY 1994, London. Zhang, C., Han, K., Zhang, D., 2015. Face stability analysis of shallow circular tunnels
Kirsch, A., 2010. Experimental investigation of the face stability of shallow tunnels in in cohesive–frictional soils. Tunnell. Undergr. Space Technol. 50, 345–357.
sand. Acta Geotech. 5 (1), 43–62.