International Journal of Rock Mechanics & Mining Sciences 76 (2015) 200–208

Contents lists available at ScienceDirect

International Journal of
Rock Mechanics & Mining Sciences
journal homepage: www.elsevier.com/locate/ijrmms

Ground reaction curve for tunnels with jet grouting umbrellas

considering jet grouting hardening
Mahdi Heidari n, Fulvio Tonon
Department of Civil Engineering, University of Texas at Austin, Austin, TX, USA

art ic l e i nf o a b s t r a c t

Article history: In the calculation of the reaction curve for grounds reinforced with jet grouting umbrellas in tunnels, the jet
Received 21 October 2014 grouting progressive hardening is ignored by assuming mean values for the jet grouting properties. In this
Received in revised form paper, we present a new procedure that calculates the reaction curve for these grounds, considering jet
9 March 2015
grouting hardening. Tunnel excavation is represented in the ground plane-strain analysis by increasing the
Accepted 15 March 2015
tunnel convergence at a rate obtained from the tunnel longitudinal deformation profile. This novelty allows
us to perform the analysis at an accurate rate, which is critical for jet grouted tunnels. Using this new
Keywords: approach, one can estimate a lower bound for convergence of supported tunnels and hence an upper bound
Ground reaction curve for pressure on the tunnel supports. Using a finite element model, we calculate the ground reaction curves
Convergence–confinement method
for a case study tunnel with and without considering jet grouting hardening. The resulting curves illustrate
Subhorizontal jet grouting
that ignoring the jet grouting hardening can lead to an unsafe design of these tunnels. We use the ground
Tunnel plane strain analysis
Jet grouting hardening reaction curve to also estimate the savings in tunnel supports due to jet grouting and the efficiency of jet
Tunnel deformation grouting at different tunnel advance rates.
& 2015 Elsevier Ltd. All rights reserved.

1. Introduction
Subhorizontal jet grouting is a pre-confinement method that
reinforces the ground ahead of the tunnel face with jet grouted
umbrellas [1,2] (Fig. 1a). This method is increasingly used in poor
ground conditions, where the primary concern is to control the tunnel
deformation at the walls (convergence) and at the face (extrusion)
(Fig. 1b). Both types of deformation are supplied by the tunnel
convergence ahead of the face (pre-convergence) [3]. By controlling
the tunnel pre-convergence with jet grouted umbrellas, this method
controls both the convergence and the extrusion in a tunnel.
Jet grouted tunnels have been analyzed in different ways in the
past two decades [4–11]. The majority of these analyses use simple
models that assume constant properties – stiffness and strength – for
the jet grouting [4–9]. These properties, however, vary significantly as
a result of the jet grouting progressive hardening over time. This
variation has a considerable impact on the results of the tunnel
analysis because it takes place through the loading period of the jet
grouting. The deformation at a tunnel cross section takes place while
the tunnel face moves for approximately three tunnel diameters
[12,13]—from one tunnel diameter before the cross section to two
n
Correspondence to: Department of Civil Engineering, The University of Texas at
Austin, 1 University Station C1792, Austin, TX 78712-0280, USA.
Tel.: þ 1 512 584 0299.
E-mail address: mahdiheidari@utexas.edu (M. Heidari).
tunnel diameters past it. Given the tunnel diameters and the advance
rates in practice (0.6–1.5 m/day), the deformation at a cross section of
a jet grouted tunnel occurs in a few weeks. Over this loading period,
the jet grouting properties change considerably. The impact of these
properties on the tunnel response is more significant in poor ground
conditions [14]. Few studies that take into account the jet grouting
hardening, on the other hand, use complex models that are unafford-
able to the practice of tunnel design and hard to gain insight into the
impacts of jet grouting on the ground response [10,11].
In this paper, we use the convergence–confinement method for
the tunnel analysis. This method uses two-dimensional models that
are affordable and insightful for the preliminary analysis of tunnels
[15,16]. In this method, the ground (herein, the ground with jet
grouting umbrellas) and the support are analyzed independently and
their resulting responses are represented by the ground reaction
curve (GRC) and the support reaction curve (SRC), respectively. These
curves are used along with the longitudinal deformation profile (LDP)
to calculate the deformation and the stresses in the tunnel. We
present a method to calculate the GRC for grounds with jet grouting
umbrellas that takes into account the jet grouting hardening.
The cement hardening has been considered in the convergence–
confinement method to calculate the SRC for tunnels supported with
shotcrete [17,18]. These considerations assume a longitudinal variation
for the pressure at the tunnel. In this paper, we show that this
variation cannot be derived for jet grouted tunnels; hence, a new
method is suggested that relies on the LDP of the tunnel. We examine

http://dx.doi.org/10.1016/j.ijrmms.2015.03.021
1365-1609/& 2015 Elsevier Ltd. All rights reserved.
 M. Heidari, F. Tonon / International Journal of Rock Mechanics & Mining Sciences 76 (2015) 200–208
Fig. 1. Reinforcement of ground in a tunnel with subhorizontal jet grouting
umbrellas. (a) Implementation of the umbrellas around the tunnel cross section
ahead of the face by creating overlapped jet grouted columns (modified after [11]).
(b) Location of the umbrellas vs. different types of deformation in a jet grouted
tunnel. u0 and uul-sup are the initial and the ultimate convergence of the tunnel with
supports.
a well-established relation that generates LDP’s for unreinforced
tunnels to assess its applicability to jet grouted tunnels. Finally, we
calculate the GRC for a jet grouted tunnel using a finite element model
and discuss the results with special reference to the information that
the GRC of the tunnel can provide for the tunnel preliminary design.
2. Convergence–confinement method
In the convergence–confinement method, it is assumed that a
tunnel cross section deforms in plane strain conditions as the tunnel
face advances through the cross section. Therefore, the tunnel analysis
is simplified to finding the deformations and pressures between the
ground and the supports in cross section. The analysis is further
simplified to finding a radial deformation and pressure by assuming
axisymmetric conditions for the tunnel, i.e., assuming a circular cross
section for the tunnel and assuming isotropic properties and in situ
stresses for the ground [16].
The ground and the support are analyzed independently in plane
strain conditions. In these analyses, the ground and the support
response are characterized by calculating their radial displacement
(convergence) at different radial pressures, illustrated by two reaction
curves in a pressure-convergence plot (Fig. 2). Fulfilling the kinematic
compatibility and the equilibrium between the ground and the
support, the intersection of the curves signifies the ultimate conver-
gence and pressure at the tunnel, attained far behind the tunnel face.
The support reaction curve is horizontally offset in this plot to take
into account the initial convergence—the ground convergence devel-
oped prior to support installation (Fig. 1b). The initial convergence has
a marked influence on the estimated tunnel deformation and pressure,
particularly in poor ground conditions [19,20].
201
Fig. 2. A schematic convergence–confinement plot showing the ground reaction
curve (GRC) and the support reaction curve (SRC). The GRC intercepts the pressure
axis at the initial ground stress, p0, and the convergence axis at the ultimate
convergence uul (convergence of tunnel with no supports). The SRC intercepts the
convergence axis at the initial convergence, u0. The intersection of the GRC and SRC
defines the ultimate convergence of the tunnel with supports, uul-sup.
The initial convergence is estimated by using longitudinal defor-
mation profiles. These profiles define the variation of the tunnel
convergence along the tunnel axis as a function of the distance to the
tunnel face [15,21]. Among several relations proposed to define the
profile for unsupported tunnels, the one suggested recently by
Vlachopoulos and Diederichs [22] explicitly takes into account the
radius of the ground plastic zone and defines the profile ahead and
behind the tunnel face with separate relations:


x  0:15 Rp
uðxÞ ¼ 13uul exp R for x r 0; ahead of the face
h

 i
R
uðxÞ ¼ uul 1 1 13exp 0:15 Rp expð 1:5 Rxp Þ for x Z 0; behind the face
ð1Þ
where u is the tunnel convergence, and x is the distance to the tunnel
face. R is the radius of the tunnel. Rp is the ultimate radius of the
plastic zone developed in surrounding ground and uul is the ultimate
tunnel convergence, both occurring far from the tunnel face; these
two quantities are calculated by using either analytical solutions [23]
or numerical modeling.
The profiles derived for unsupported tunnels, however, over-
estimate the profile for the tunnels with supports [20,24,25]. This
error leads to an unsafe design of supports as it causes under-
estimation of the pressure on the supports. Among the different
suggestions made to take into account the effect of supports on the
LDP, the one proposed by Nguyen-Minh and Guo [26] is well
established [20]. In this relation, the effect of supports is incorporated
by scaling down the LDP of the unsupported tunnel with a scale
factor φ. By fitting the convergences obtained from axisymmetric
analysis of tunnels with different support stiffnesses, support instal-
lation distances from the face, and ground in situ stresses, this scale
factor was suggested as a function of the ratio between the ultimate
convergence of the tunnel with supports, uulsup , and the one of the
tunnel without supports, uul :
uulsup
uðxÞsup ¼ φðzÞ uðxÞ; φðzÞ ¼ 0:55 þ0:45 z 0:42 ð1 zÞ3 ; z¼ ð2Þ
uul
To estimate the initial convergence, u0 , for a support, Eq. (2) needs
the tunnel ultimate convergence, uul  sup , which itself depends on u0 .
202 M. Heidari, F. Tonon / International Journal of Rock Mechanics & Mining Sciences 76 (2015) 200–208
This interdependence gives rise to an implicit solution for the tunnel
ultimate convergence and pressure.
3. GRC calculation
3.1. A limit proposed for Equation (2)
We first discuss a self-inconsistency in the LDP’s generated by
Eq. (2). Upon investigating the performance of Eq. (2) to construct the
LDP of supported tunnels, we found out that this relation provides
remarkable estimates for the convergence of cross sections near the
face. However, it provides inconsistent convergence estimates for
cross sections far behind the tunnel face: the convergence estimated
for these cross sections exceeds the ultimate convergence supposed
for the tunnel.
To find the reason for the inconsistency, consider an unsupported
tunnel with ultimate convergence uul (Fig. 3a). To construct the LDP
of the tunnel with supports and ultimate convergence unulsup , Eq. (2)
yields scale factor φn , whose magnitude is greater than the ratio
unulsup =uul (Fig. 3b). Consequently, the magnitude of the ultimate
convergence in the profile produced for the tunnel with supports,
being φn unul , will be greater than the supposed ultimate convergence,
unulsup . Accordingly, convergence estimates for the supported tunnel
exceed the supposed ultimate convergence for cross sections far
behind the face (x4xn in Fig. 3a).
Fig. 3. (a) Longitudinal deformation profile of a tunnel with supports (solid curves),
generated by scaling down the profile of the tunnel with no supports (dashed curve
with ultimate convergence, uul) with the scale factor, φ (Eq. (2), and then modifying
it with Eq. (3). Without the modification, the generated profile predicts unrealistic
convergences (greater than the assumed ultimate convergence, unul-sup) for tunnel
cross sections farther than xn from the face. (b) Scale factor φ vs. scale factor uul-sup/
uul. The former is greater than the latter for any magnitude of ultimate convergence
uul-sup.
In order to eliminate the inconsistency yet maintain the remark-
able performance of Eq. (2) for cross sections near the face, we
propose to limit the convergence estimates of Eq. (2) to the supposed
ultimate convergence (Fig. 3a):
uðxÞsup ¼ φ uðxÞ r uulsup ð3Þ
3.2. Implementation of jet grouting hardening
Material models have been developed for cementitious materials
that take into account their property variation due to the hardening
process [10,27,28]. These models represent the material hardening
using a number of intrinsic functions; accordingly, they are sophis-
ticated, require extended experimental data to calibrate their para-
meters, and are not available in finite element packages used in
practice. Here, we suggest a method to implement the jet grouting
hardening in the finite element analysis of tunnels.
We use the notion of material embedment to represent the
increase in the jet grouting stiffness and strength during its hardening.
This notion allows us to implement the hardening of the jet grouting
with standard elastic-plastic models that are routinely used in finite
element analysis of tunnels—e.g., the Mohr–Coulomb model. In this
technique, when a material is embedded into an existing material in a
model, the displacement of the embedded material is constrained to
that of the host material. In effect, the stiffness and the strength of the
embedded material are added to those of the existing material for
subsequent deformations.
To represent the jet grouting hardening with this concept, we
approximate the continuous growth of the jet grouting properties
(solid graph; Fig. 4) by a sufficient number of discrete property
increments adding up over time (dashed stepped graph; Fig. 4). We
associate each property increment with an artificial jet grouting
material. These materials are embedded successively at their due
time into the jet grouted zone in order to simulate the increase in the
jet grouting properties.
Fig. 4. Continuous (solid graph) and discretized (stepped graph) growth of the jet
grouting unconfined compressive strength (UCS) and Young’s modulus (E) over
time, normalized by their ultimate values, UCSul and Eul, respectively. The
continuous graph is obtained by fitting an exponential curve to the experimental
data from [42]. For the sake of simplicity, a single curve is assumed for both
properties. ΔUCSi and ΔEi denote the increment in the jet grouting UCS and E
developed over Δti.
3.3. Conventional vs. new approach for GRC calculation
The GRC is conventionally calculated by controlling the radial
pressure in the plane strain analysis of the ground cross section
[9,29–35]. The radial pressure on the ground at the tunnel locus is
decreased from the in situ stress to zero, and the tunnel convergence
is acquired during the analysis (Fig. 5a).
In a jet grouted tunnel, the jet grouting hardening results in a
progressive increase in the jet grouting properties over time. This
 M. Heidari, F. Tonon / International Journal of Rock Mechanics & Mining Sciences 76 (2015) 200–208
Fig. 5. Alternate schemes to conduct the plane strain analysis of the ground. The
rings in dark grey represent the jet grouting umbrellas. p0 is the in situ (initial)
ground stress. (a) Conventional scheme, where the analysis is controlled by the
pressure on the ground at the tunnel, p. (b) New scheme, where the analysis is
controlled by the tunnel convergence, u.
time dependent evolution gives rise to a time dependent behavior for
the ground reinforced with jet grouting umbrellas. Consequently, in
the conventional analysis for GRC calculation, the convergences
acquired, and thus the resulting reaction curve, depend on the
temporal variation of the pressure in the analysis. However, deriva-
tion of this variation essentially requires a priori knowledge of the
GRC because this variation, p(t), is obtained by combining the
longitudinal pressure variation, p(x), and the tunnel advance sche-
dule, x(t):
pðxÞ; xðt Þ : pðt Þ ¼ pðxðt ÞÞ
and the longitudinal pressure variation, p(x), is derived by combining
the GRC, p(u), and the LDP, u(x) [20]:
pðuÞ; uðxÞ : pðxÞ ¼ pðuðxÞÞ
Hence, we would essentially require the GRC a priori for its
calculation. In order to avoid this dilemma, we propose an alternative
approach to calculate the GRC.
In the new approach, the roles of the convergence and of the
radial pressure are interchanged in the plane strain analysis of the
ground cross section (Fig. 5b): the convergence controls the analysis,
and the radial pressure is acquired through the analysis. As such, the
tunnel excavation is represented by increasing the tunnel conver-
gence rather than decreasing the pressure at the tunnel. For grounds
with jet grouting umbrellas, which exhibit a time-dependent beha-
vior, the acquired radial pressures depend on the temporal variation
of the convergence, u(t), in the analysis. Nevertheless, the latter is
203
readily estimated from the tunnel LDP, u(x), and the tunnel advance
schedule, x(t):
uðxÞ; xðt Þ : uðt Þ ¼ uðxðt ÞÞ ð6Þ
As such, we eliminate the need for the longitudinal pressure
variation, and thus for the GRC for its calculation.
The use of the LDP is also advantageous to the calculation of the
GRC for jet grouted tunnels. Contrary to the longitudinal pressure
variation, which has been derived essentially for elastic ground
behavior [36], the LDP is provided by analytical, numerical, and
empirical relations for a variety of ground behaviors including
grounds with time-dependent behavior [21,22,37–41]. One can also
take into account the effect of tunnel supports by using implicit
relations such as Eq. (2) for the LDP. Moreover, since the tunnel
convergence can be measured in tunnels as opposed to the (fictitious)
pressure, one can update the GRC during tunnel construction by using
the true LDP measured at the tunnel.
3.4. Proposed procedure for GRC calculation
Based on the approach suggested in Section 3.3, we propose a
procedure to calculate the GRC for grounds reinforced with jet
grouting umbrellas:
1) Select a cross section of the ground with jet grouting umbrellas
and set up a plane strain model where the radial displacements
are constrained at the tunnel locus (Fig. 5b).
2) Calculate the ultimate convergence, uul, and plastic radius, Rp, for
the tunnel in plain ground (no support and no jet grouting) (from
analytical solutions or the model in Step 1).
3) Construct the LDP for the plain ground, given uul and Rp from
Step 2 (for instance, by using Eq. (1)) (dashed curve; Fig. 3a).
4) Assume a magnitude for the ultimate convergence of the
tunnel with supports, unulsup .
5) Modify the profile in Step 3 to accommodate the ultimate
convergence assumed in Step 4 (for instance, by using Eqs.
(2) and (3)) (solid curves; Fig. 3a).
6) Derive the variation of the cross section convergence in time.
Given the tunnel advance schedule and the profile in Step 5, the
convergence time history can be constructed by using Eq. (6).
The time history starts (t¼0) when the convergence of the cross
section initiates (x0; Fig. 3a) and ends (t¼tn) when the conver-
gence of the cross section stabilizes (xn; Fig. 3a), as indicated by
the LDP of the tunnel (solid curve; Fig. 3a).
7) Calculate the time when the cross section is jet grouted, t0, given
the tunnel advance schedule, x(t), and the value of x0 from Step
6, x(t0)¼x0.
8) Discretize the time period [0, tn] into time intervals defined by
times [0,…,t0,…,ti,…,tn].
ð4Þ
9) Calculate convergence increments Δui for every time interval
Δti ¼ [ti, ti þ 1] from the convergence history derived in Step 6.
10) Calculate increments of the jet grouting properties, ΔPi,
associated with the time intervals Δti for tZt0 (Fig. 4) (see
ð5Þ
Section 3.2).
11) Conduct the plane strain analysis. For each time interval Δti:
 For t¼ t0, replace the stressed ground material at the jet
grouting ring around the tunnel with the stress-free jet
grouting material with properties ΔP0.
 For t4 t0, increase the properties of the jet grouting material
by ΔPi (for instance, as suggested in Section 3.2).
 Displace the radial constraints at the tunnel locus by Δui
towards the tunnel center.
12) Calculate the radial pressure on the radial constraints at the end
(t¼tn) of the analysis in Step 11. This ultimate pressure together
with the ultimate convergence supposed in Step 4 defines a
point on the GRC.
204 M. Heidari, F. Tonon / International Journal of Rock Mechanics & Mining Sciences 76 (2015) 200–208

Perform Steps 4–12 for an adequate number of magnitudes
assumed for the ultimate convergence, unulsup , starting from zero
and increasing until the corresponding pressure calculated in Step 12
vanishes (or becomes negative). Note that one needs to construct the
curve point by point because different points are associated with
different LDP’s (Step 5) and thus different convergence histories
(Step 6).
4. Procedure application
For our case study, we use the jet grouted tunnel, Cassia 2,
constructed in Rome, Italy, according to the ADECO method [3]. A
comprehensive description of the tunnel construction and geotech-
nical conditions can be found in [12]. The tunnel underpasses critical
roads, ancient buildings, and two existing tunnels under minimal
cover (5–30 m) in a weak clayey ground; therefore, the tunnel
deformation was the main concern in the tunnel design. Conse-
quently, the tunnel was constructed with subhorizontal jet grouting
together with full face excavation. This method of excavation allows
for the implementation of the invert close to the tunnel face,
resulting in a considerable increase in the ground confinement near
the face.
The tunnel is 14 m in diameter. The jet grouting umbrellas are
0.5 m thick (0.6 m in the actual tunnel) and 16 m long, composed of
eighty jet grouted columns. Jet grouting was carried out subhorizon-
tally after every 8 m tunnel excavation, resulting in an overlap of 8 m
for the umbrellas along the tunnel (see Fig. 1a). In the tunnel cross
section, we represent these overlapped umbrellas by two concentric
jet grouted rings created at different times.
The tunnel is in a ground mainly composed of stiff silty clay. The
ground water table is located below the tunnel. The numerical
analyses are performed in finite element code Abaqus assuming
undrained conditions for ground. In these conditions, the ground
behavior can be characterized by the Tresca model (Mohr–Coulomb
model with zero friction and dilation angle) with a cohesion equal
to the ground undrained strength, cu. The value of the parameters of
the ground model is given in Table 1. These values are assumed
within the ranges provided by two substantial geotechnical inves-
tigations carried out for the tunnel. We assume an isotropic in situ
stress of 1 MPa for the ground. The tunnel face is assumed to
advance at a mean constant rate of 1 m/day.
The pre-confinement of the Cassia 2 tunnel also included the use
of fiber glass dowels into the tunnel face and steel pipes into the jet
grouting umbrellas. For the sake of simplicity, these reinforcements
are not considered herein. One should also take into account the
effect of the proximity of the ground surface and the twin tunnel,
which are not considered in the convergence–confinement method
and thus in the following results.
4.1. Implementation of jet grouting hardening
We use the elastic-plastic Mohr–Coulomb model to represent
the jet grouting material. Young’s modulus (E) and the cohesion
coefficient (c) are the jet grouting properties that vary over time—
P¼[E, c]; other properties are constant with values given in Table 1.
The temporal variation of the cohesion coefficient is obtained from
the temporal variation of the unconfined compressive strength
(UCS) by invoking the relation between these two quantities in
the Mohr–Coulomb model. The UCS temporal variation is obtained
by fitting the experimental data from [42] with an exponential
curve of the form:
UCSðt Þ ¼ UCSul ð1  expð  k t ÞÞ ð7Þ
UCSul is the ultimate UCS of the jet grouting, and k is a constant
representing the rate of UCS growth with time. The value of these
parameters is given in Table 1. For the sake of simplicity, the same
curve is used for the temporal variation of the jet grouting Young’s
modulus, E.
In order to reduce the computational effort, we take into
consideration only the hardening that takes place within the jet
grouting loading period (Fig. 4) because subsequent hardening will
make no difference in the results. We estimate the loading period
from the tunnel advance rate and assuming a loading zone of three
tunnel diameters along the tunnel—one diameter ahead of and two
diameters behind the cross section.
4.2. Applicability of Eqs. (2) and (3) to the jet grouted tunnel
The longitudinal deformation profile is central to our procedure.
We use Eq. (2) with the limit defined in Eq. (3) to estimate the profile
for the jet grouted tunnel because it takes into account the effect of
tunnel supports on the profile. However, since this relation has been
derived for tunnels in unreinforced grounds, we first investigate its

Table 1
Ground and jet grouting mechanical parameters.

Parameter Value

Ground
Young’s modulus 200 MPa
Poisson’s ratio 0.49
Internal friction angle 01
Dilation angle 01
Cohesion (cu) 0.2 MPa
Jet grouting
Ultimate Young’s modulus (Eul) 13,500 MPa
Ultimate unconfined compressive strength (UCSul) 8 MPa Fig. 6. (a) Axisymmetric modeling of jet grouted tunnel. Shown in dark gray is one
Poisson’s ratio 0.2 jet grouting umbrella. (b) Longitudinal deformation profile of the jet grouted
Internal friction angle 201 tunnel, generated with Eq. (2) (dotted curve) and modified by Eq. (3) (dashed
Constant of hardening rate (k) 0.054 1/day curve), vs. the one obtained from the tunnel axisymmetric analysis (solid curve).
The profiles are normalized by the ultimate convergence of the jet grouted tunnel.
 M. Heidari, F. Tonon / International Journal of Rock Mechanics & Mining Sciences 76 (2015) 200–208
applicability to the jet grouted tunnel. For this purpose, we compare
the LDP provided by this relation to the one obtained from axisym-
metric analysis of the tunnel (Fig. 6a).
In the axisymmetric analysis, the tunnel is excavated in 1 m
advances, and the jet grouting umbrellas are sequentially installed
along the tunnel. The jet grouting hardening is implemented with the
material embedment method explained in Section 4.1. For Eq. (2), on
the other hand, we employ Eq. (1) to construct the LDP of the plain
ground (with no jet grouting umbrellas), where the ultimate con-
vergence uul and plastic radius Rp is calculated from analytical
solutions [23].
It is seen that Eq. (2), together with Eq. (3), can reproduce the
profile obtained from the axisymmetric analysis with a remarkable
agreement (Fig. 6b). The simple limit suggested by Eq. (3) proves
efficient to improve convergence estimations of Eq. (2) for cross
sections far from the face. It should be noted that further analyses
with various sets of jet grouting and support properties are necessary
to prove the performance of these equations for jet grouted tunnels
with different jet grouting and support scenarios.
4.3. The GRC of the tunnel
We now calculate the reaction curve of the ground with jet
grouting umbrellas (Fig. 7a) by following the procedure outlined in
Section 3.4. We use Eqs. (2) and (3) to construct the LDP of the tunnel
associated with ultimate convergences assumed for the tunnel with
supports (Fig. 7b). We also use the method explained in Section 4.1 to
implement the hardening of the internal and the external jet grouted
rings at the tunnel cross section (Figs. 6a, 8), considering their
different installation times.
To examine the significance of jet grouting hardening to the GRC
of the tunnel, we compare the calculated GRC (Fig. 7a) to a GRC
calculated by assuming mean values for the jet grouting properties
(Fig. 9). These values are calculated as the mean value of the jet
grouting properties over the jet grouting loading period, which is 42
Fig. 7. (a) GRC for the tunnel with and without the jet grouting umbrellas (plain
ground). The support reaction curve (SRC) is schematic. (b) Longitudinal deforma-
tion profiles associated with different magnitudes of uul-sup. The horizontal
coordinate of the points shown by ★ signifies the installation distance of a rigid
support with respect to the tunnel face in order to attain the corresponding
magnitude of uul-sup.
205
Fig. 8. Simultaneous evolution of hoop stress (solid curves) and unconfined
compressive strength (UCS) (dashed curves) in the internal and the external jet
grouting layers as the tunnel face progresses past the studied cross section. The
stress value given by Eq. (8) is shown by ▲, representing the mean value of the
ultimate hoop stress in the jet grouting layers.
days for the studied tunnel assuming a loading zone of three tunnel
diameters. It is seen that ignoring the jet grouting hardening leads to
an overestimation of the jet grouting capability and thus to an unsafe
design of the tunnel for most support pressures. This error arises
from the nonlinearity of the jet grouting hardening vs. time, the LDP,
and the ground behavior.
5. Discussion
5.1. A limit for tunnel convergence and support pressure
Following the proposed procedure (Steps 3 and 4), we assumed a
number of magnitudes for uul-sup and then constructed the associated
LDP’s (Fig. 7b). In light of these profiles, we determine a lower bound
for the tunnel convergence. In fact, we show that, with even a rigid
tunnel support installed at the face, we could not attain tunnel
convergences smaller than a minimum. Given the fact that the tunnel
convergence does not change behind a rigid support, the point on
each LDP where the profile flattens, indicated by asterisks in Fig. 7b,
represents the installation point of the rigid support along the tunnel.
It is observed that the support would have to be installed ahead of
the face in order to attain small tunnel convergences such as uul-
sup ¼1 cm. The minimum tunnel convergence attainable with sup-
ports may thus be defined as uul-sup associated with the profile
flattening at the face, which is nearly 3 cm for the studied tunnel.
The pressure associated with this minimum convergence on the
GRC, on the other hand, defines an upper bound for the pressure on
the tunnel supports. For the studied tunnel, this maximum pressure
is approximately one third the ground initial stress. This magnitude
agrees with measurements and other analyses reading that support-
ing the tunnel could maintain only a minor fraction of the ground
initial stress [43].
The minimum convergence also defines a practical range for the
GRC; the curve is practically meaningful only for the range of tunnel
convergences higher than this minimum. By determining the prac-
tical range of the reaction curves (before their calculation), data
points will be calculated only for this range, resulting in a reduction
in the calculation efforts.
5.2. Performance of tunnel jet grouting
In order to assess the performance of jet grouting the tunnel, we
compare the reaction curves calculated for the ground with jet
grouting umbrellas and for the plain ground (without jet grouting
umbrellas) (Fig. 7a). The vertical difference between the reaction
206 M. Heidari, F. Tonon / International Journal of Rock Mechanics & Mining Sciences 76 (2015) 200–208
Fig. 9. GRC of the jet grouted tunnel for advance rates of 1 and 2 m/day vs. the GRC
of the jet grouted tunnel calculated assuming mean values for the jet grouting
properties (E¼ 6 GPa and c ¼1.1 MPa) and the GRC of the tunnel with no jet
grouting umbrellas (plain ground).
curves represents the amount of saving that the tunnel jet grouting
results in the pressure on the tunnel supports, and thus in their
quantity, for a given tunnel convergence. The relative amount of the
saving is significant within the practical tunnel convergences.
The horizontal difference between the curves, on the other hand,
represents the amount of reduction in the tunnel convergence for a
given support, resulting from the tunnel jet grouting. The relative
amount of the reduction is significant within the practical support
pressures, amounting to nearly 90% when no supports are used
(from 57 cm to 6 cm).
These comparisons indicate the high performance of jet grouting,
particularly in poor ground conditions (cu ¼200 kPa). Nonetheless, it
should be noted that our analysis assumes continuity in the plane of
jet grouting umbrellas and thus does not take into account defects
that impair this continuity, resulting from poor columns due to grout
loss or poor overlapping of the columns. Further, our analysis
assumes closure of jet grouting umbrellas, which is not the case at
least ahead of the face. The adverse effect of these factors is
important and thus should also be considered in the assessment of
the performance of these umbrellas.
5.3. The stress in the jet grouting umbrellas
We can estimate the hoop stress in the jet grouting umbrellas
from the reaction curves calculated for the ground with and without
jet grouting umbrellas (plain ground) (Fig. 7a). Considering the jet
grouted zone around the tunnel as a tube, the pressure denoted by
pi in Fig. 7a represents the radial pressure between the supports and
the intrados of this tube, and the pressure denoted by pe in Fig. 7a
represents the pressure between the ground and the extrados of this
tube. We can estimate the mean hoop stress in the tube, σt,
according to Mariotte’s formula as
  R
σ t ¼ pe  pi  ð8Þ
s
where s stands for the thickness of the jet grouting tube (twice the
thickness of an umbrella), and R is the tunnel radius. Eq. (8) indicates
that the hoop stress in the jet grouting tube is proportional to the
vertical difference between the reaction curves for the ground with
and without jet grouting umbrellas. This stress thus maximizes
when no support is installed at the tunnel, amounting to 3.8 MPa for
the studied tunnel. Eq. (8) provides us with the ultimate value of the
mean hoop stress in the jet grouting tube.
5.4. The behavior of the jet grouting
The strength of the jet grouting umbrellas varies during their
transient loading period. Therefore, to determine the jet grouting
behavior during this period, we need to compare the jet grouting
stress to the concurrent strength over this period.
We obtain the jet grouting stress evolution from the plane strain
analyses conducted to calculate the GRC and compare it to the jet
grouting strength evolution (Fig. 4). The jet grouting stress is
obtained by adding up the stresses of the jet grouting (artificial)
materials embedded in the jet grouted area through the analysis.
The external ring is created first at the cross section as it is installed
in a jet grouting operation prior to the internal ring (Fig. 6a); as a
result, the stress is higher in the external umbrella at any given time
(Fig. 8). Constant yielding of the jet grouting umbrellas during their
transient loading near the face (Fig. 8) is indicated by the equality of
the jet grouting stress and yield strength. The resulting inelastic
deformation is not unacceptable for the jet grouting umbrellas as a
primary reinforcement.
Before the tunnel face reaches the cross section (negative face
distances), the hoop stress can exceed the unconfined jet grouting
strength because the ground ahead of the tunnel face (core) provides
confinement and thus extra strength for the jet grouting. For the
external layer, this extra strength persists even after the removal of
the core because of the confinement provided by the internal layer.
The hoop stress calculated by using Eq. (8) is in agreement with the
average of the ultimate stresses in the jet grouting umbrellas (Fig. 8).
5.5. The shape of the GRC
It is noteworthy that the jet grouting hardening causes an atypical
trend for the tunnel GRC at large convergences (low pressures); GRC’s
typically end up with a positive curvature while the GRC of the
ground with jet grouting umbrellas ends up with a negative curvature
(Figs. 7a, 9). Note that the slope of the GRC represents the stiffness of
the ground response, which, in the case of grounds with jet grouting
umbrellas, is the resultant of the ground’s response and the umbrel-
las’ response. The ground’s response softens progressively as the
tunnel convergence increases – the curve slope decreases with the
tunnel convergence – because of the progressive expansion of the
plastic zone in the ground. The umbrellas’ response, on the other
hand, hardens progressively, which, in the case of the studied tunnel,
outweighs the progressive softening of the ground’s response.
5.6. Jet grouting efficiency vs. tunnel advance rate
Jet grouting hardening over time results in a rate dependent
response for the ground with jet grouting umbrellas. Accordingly, the
GRC of the tunnel will depend on the tunnel advance rate. To assess
this dependency, we calculate the GRC for two advance rates of 1 and
2 m/day (Fig. 9). The rate dependent response of the ground with jet
grouting umbrellas is reflected by the marked difference between
these curves (Fig. 9).
Increasing the tunnel advance rate decreases the efficiency of the
ground reinforcement with jet grouting umbrellas as these umbrellas
are loaded in a shorter time period, and thus with less grown strength
and stiffness. At excessively high advance rates, the jet grouting would
perform even more weakly than the substituted ground, and thus
would impair the ground strength rather than reinforce it [10]. One
can assess the reduction in the efficiency of the jet grouting umbrellas
due to increasing the advance rate by comparing the associated GRC’s
(Fig. 9). The vertical difference illustrates the increase in the amount
of tunnel supports, and the horizontal difference represents the
increase in the tunnel convergence. The convergence of the studied
tunnel with no supports (the curve horizontal intercept) almost
doubles (11.3 vs. 5.7 cm) when the tunnel advance rate increases
from 1 to 2 m/day.
The jet grouting exhibits other time-dependent behaviors such as
creep and shrinkage in addition to hardening. Contrary to hardening,
increasing the advance rate (loading rate) reduces the adverse effect
 M. Heidari, F. Tonon / International Journal of Rock Mechanics & Mining Sciences 76 (2015) 200–208
of these behaviors on the efficiency of the umbrellas. Therefore, these
behaviors need to be considered for an accurate estimation of the
effect of the tunnel advance rate on the efficiency of the jet grouting
umbrellas.
6. Summary and conclusions
We present a procedure to calculate the reaction curve for
grounds reinforced with jet grouting umbrellas in tunnels, taking
into account the jet grouting progressive hardening over time. The
resulting curve hence provides a simple and accurate representation
of the reinforced ground for tunnel preliminary design.
To take into account the jet grouting hardening, one needs to define
the temporal variation of ground unloading at the tunnel in the ground
plane strain analysis. For this purpose, we propose a new scheme to
conduct the analysis, whereby the ground unloading is represented by
increasing the tunnel convergence rather than decreasing the pressure
at the tunnel. The temporal variation of the convergence increase is
obtained from the longitudinal deformation profile.
The LDP is of elevated importance in our approach. It not only
contributes to the GRC but also is used to estimate the minimum
tunnel convergence that could be attained by supporting the tunnel.
This minimum convergence defines an upper bound for the pressure
on the tunnel supports and also a practical range for the GRC of the
tunnel. We find and eliminate a self-inconsistency in a well-est-
ablished implicit relation that provides the LDP for supported tunnels
without ground reinforcement. We also show that the modified
relation maintains its performance for a jet grouted tunnel. Never-
theless, our approach allows one to update the profile estimated with
such relations with the true one measured during tunnel construc-
tion, resulting in a more realistic GRC.
The study of a jet grouted tunnel in poor ground conditions
indicates that assuming mean values for the jet grouting stiffness and
strength leads to overestimation of the jet grouting capability, and
thus to unsafe design of the tunnel. The study also indicates that the
jet grouting umbrellas undergo inelastic deformation throughout
their transient loading period, yet considerably reduce the tunnel
deformation and the need for tunnel supports. The decent ductility of
the fresh jet grouting prevents failure of the umbrellas due to the
inelastic deformation.
In addition to the deformation and the stresses in the ground and
the supports, we show that the GRC of a jet grouted tunnel can be
used to estimate the stress in the jet grouting umbrellas, the reduction
in the tunnel deformation and the amount of tunnel supports due to
tunnel jet grouting, and the reduction in the efficiency of the
umbrellas due to increasing the tunnel advance rate.
Subhorizontal jet grouting is typically used in poor ground
conditions. In these conditions, the ground is often associated with
other time dependent mechanisms as well, including creep or
consolidation. Jet grouting also exhibits time-dependent behaviors
such as creep and shrinkage in addition to hardening. Our proposed
procedure can be extended to take into account these time depen-
dent mechanisms as well in the GRC calculation.
Acknowledgements
This research was sponsored by the International Tunneling
Consortium (J.F. Shea, Maccaferri, Grace, CSI) established by the
second author at The University of Texas at Austin.
References
[1] Pelizza S, Peila D. Soil and rock reinforcements in tunnelling. Tunnelling
Underground Space Technol 1993;8:357–72.
207
[2] Tonon F. Adeco full-face tunnel excavation of two 260 m2 tubes in clays with
sub-horizontal jet-grouting under minimal urban cover. Tunnelling Under-
ground Space Technol 2011;26:253–66.
[3] Lunardi P. Design and construction of tunnels: analysis of controlled deforma-
tion in rocks and soils (adeco-rs). Berlin: Springer; 2008.
[4] González-Cao J, Varas F, Bastante FG, Alejano LR. Ground reaction curves for
circular excavations in non-homogeneous, axisymmetric strain-softening rock
masses. J Rock Mech Geotech Eng 2013;5:431–42.
[5] Oreste PP. Analysis of structural interaction in tunnels using the convergence–
confinement approach. Tunnelling Underground Space Technol 2003;18:347–63.
[6] Lignola G, Flora A, Manfredi G. Simple method for the design of jet grouted
umbrellas in tunneling. J Geotech Geoenviron Eng. 2008;134:1778–90.
[7] Peila D, Oreste PP, Rabajoli G, Trabucco E. The pretunnel method, a new italian
technology for full-face tunnel excavation: a numerical approach to design.
Tunnelling Underground Space Technol 1995;10:367–74.
[8] Barla M, Bzowka J. Comparing numerical alternatives to model jet grouting in
tunnels. Electron J Geotech Eng 2013;18:2997–3008.
[9] Fang Q, Zhang D, Zhou P, Wong LNY. Ground reaction curves for deep circular
tunnels considering the effect of ground reinforcement. Int J Rock Mech Min
Sci 2013;60:401–12.
[10] Pichler C, Lackner R, Spira Y, Mang H. Thermochemomechanical assessment of
ground improvement by jet grouting in tunneling. J Eng Mech 2003;129:951–62.
[11] Coulter S, Martin CD. Effect of jet-grouting on surface settlements above the
aeschertunnel, switzerland. Tunnelling Underground Space Technol
2006;21:542–53.
[12] Graziani A, Boldini D, Ribacchi R. Practical estimate of deformations and stress
relief factors for deep tunnels supported by shotcrete. Rock Mech Rock Eng
2005;38:345–72.
[13] Sterpi D, Gioda G. Visco-plastic behaviour around advancing tunnels in
squeezing rock. Rock Mech Rock Eng 2009;42:319–39.
[14] Cosciotti L., Lembo-Fazio L., Boldini D., Graziani A. Simplified behavior models
of tunnel faces supported by shotcrete and bolts. In: Proceedings of the
international conference on modern tunneling science and technology. Kyoto,
Japan;2001. p. 407–12.
[15] Panet M. Le calcul des tunnels par la méthode convergence–confinement.
Paris: Presses de l’école nationale des Ponts et chaussées; 1995.
[16] Oreste PP. The convergence–confinement method: roles and limits in modern
geomechanical tunnel design. Am J App. Sci 2009;6:757–71.
[17] Oreste PP. A procedure for determining the reaction curve of shotcrete lining
considering transient conditions. Rock Mech Rock Eng 2003;36:209–36.
[18] Gschwandtner GG, Galler R. Input to the application of the convergence
confinement method with time-dependent material behaviour of the support.
Tunnelling Underground Space Technol 2012;27:13–22.
[19] Cantieni L, Anagnostou G. The effect of the stress path on squeezing behavior
in tunneling. Rock Mech Rock Eng 2009;42:289–318.
[20] AFTES. Aftes recommandations on the convergence–confinement method.
Association Française des Tunnels et de l’espace Souterrain; 2001. p. 79–89.
[21] Alejano LR, Rodríguez-Dono A, Veiga M. Plastic radii and longitudinal
deformation profiles of tunnels excavated in strain-softening rock masses.
Tunnelling Underground Space Technol 2012;30:169–82.
[22] Vlachopoulos N, Diederichs MS. Improved longitudinal displacement profiles
for convergence confinement analysis of deep tunnels. Rock Mech Rock Eng
2009;42:131–46.
[23] Carranza-Torres C. Dimensionless graphical representation of the exact elasto-
plastic solution of a circular tunnel in a Mohr–Coulomb material subject to
uniform far-field stresses. Rock Mech Rock Eng 2003;36:237–53.
[24] Bernaud D, Rousset G. The ‘new implicit method’ for tunnel analysis. Int J
Numer Anal Methods Geomech 1996;20:673–90.
[25] Oke J, Vlachopoulos N, Diederichs MS. Modification of the supported long-
itudinal displacement profile for tunnel face convergence in weak rock.
American Rock Mechanics Association. Paper ARMA 2013-397; 2013 .
[26] Nguyen-Minh D, Guo C. A new approach to convergence confinement method.
Proc Adv Rock Mech Conf 1998:267–79.
[27] Hellmich C, Ulm F, Mang H. Multisurface chemoplasticity. I: Material model
for shotcrete. J Eng Mech 1999;125:692–701.
[28] Ulm F, Coussy O. Modeling of thermochemomechanical couplings of concrete
at early ages. J Eng Mech 1995;121:785–94.
[29] Alejano LR, Alonso E, Rodríguez-Dono A, Fernández-Manín G. Application of the
convergence–confinement method to tunnels in rock masses exhibiting Hoek–
Brown strain-softening behaviour. Int J Rock Mech Min Sci 2010;47:150–60.
[30] Brown E, Bray J, Ladanyi B, Hoek E. Ground response curves for rock tunnels. J
Geotech Eng 1983;109:15–39.
[31] Carranza-Torres C, Fairhurst C. The elasto-plastic response of underground
excavations in rock masses that satisfy the Hoek–Brown failure criterion. Int J
Rock Mech Min Sci 1999;36:777–809.
[32] Wang S, Yin X, Tang H, Ge X. A new approach for analyzing circular tunnel in
strain-softening rock masses. Int J Rock Mech Min Sci 2010;47:170–8.
[33] Chen R, Tonon F. Closed-form solutions for a circular tunnel in elastic-brittle-
plastic ground with the original and generalized Hoek–Brown failure criteria.
Rock Mech Rock Eng 2011;44:169–78.
[34] Serrano A, Olalla C, Reig I. Convergence of circular tunnels in elastoplastic rock
masses with non-linear failure criteria and non-associated flow laws. Int J
Rock Mech Min Sci 2011;48:878–87.
[35] Alonso E, Alejano LR, Varas F, Fdez-Manin G, Carranza-Torres C. Ground
response curves for rock masses exhibiting strain-softening behaviour. Int J
Numer Anal Methods Geomech 2003;27:1153–85.
208 M. Heidari, F. Tonon / International Journal of Rock Mechanics & Mining Sciences 76 (2015) 200–208
[36] Panet M, Guenot A. Analysis of convergence behind the face of a tunnel. In:
Proceedings of international symposium “IST 82”. Brighton; 1982. p. 197–4.
[37] Sulem J, Panet M, Guenot A. An analytical solution for time-dependent
displacements in a circular tunnel. Int J Rock Mech Min Sci 1987;24:155–64.
[38] Chang Y. Tunnel support with shotcrete in weak rock: a rock mechanics study
[PhD thesis]. Stockholm, Sweden: Royal Institute of Technology; 1994.
[39] Unlu T, Gercek H. Effect of poisson's ratio on the normalized radial displace-
ments occurring around the face of a circular tunnel. Tunnelling Underground
Space Technol 2003;18:547–53.
[40] Pilgerstorfer T, Radončić N. Prediction of spatial displacement development.
Vorhersage der räumlichen verschiebungsentwicklung. Geomechanik und
Tunnelbau. 2009;2:250–59.
[41] Basarir H, Genis M, Ozarslan A. The analysis of radial displacements occurring
near the face of a circular opening in weak rock mass. Int J Rock Mech Min Sci
2010;47:771–83.
[42] Coulter S, Martin CD. Single fluid jet-grout strength and deformation proper-
ties. Tunnelling Underground Space Technol 2006;21:690–5.
[43] Mair RJ. Tunnelling and geotechnics: new horizons. Géotechnique 2008;58:695–736.