Computers and Geotechnics 64 (2015) 72–82

Contents lists available at ScienceDirect

Computers and Geotechnics

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

Technical Communication

Assessment of tunnel portal stability at jointed rock mass:

A comparative case study
Özgür Satıcı a,⇑, Bahtiyar Ünver b
a
General Directorate of Highways (GDH), Research & Development Department, Division of Geological Surveying, Chief of Geological Surveys, Ankara, Turkey
b
Hacettepe University, Department of Mining Engineering, Ankara, Turkey

a r t i c l e i n f o a b s t r a c t

Article history: This paper presents results of a comparative numerical model study at a highway tunnel in Turkey. The
Received 4 July 2014 tunnel was opened in a sandstone–siltstone rock mass having clear systematic discontinuity sets. In prac-
Received in revised form 15 October 2014 tice, effects of discontinuities of rock mass on tunnel stability are usually disregarded in finite element
Accepted 3 November 2014
(FE) models, which are widely used for design of underground structures. To compensate this deficiency,
Available online 22 November 2014
rock masses’ geotechnical properties are reduced to simulate effect of discontinuities in FE models. Main
objective of this study is to show the effect of systematic and persistent discontinuity sets to FE numerical
Keywords:
modeling when discontinuities exist. On this account, two different conditions were modeled to analyze
Jointed rock mass
NATM
stability conditions in a comparative manner. FE numerical analysis software ‘‘Phase2 8.0’’ was used for
Numerical modeling modeling. Results have shown that presence of discontinuities have a significant effect on tunnel stability
Tunnel portal and model results when FEM software is used for discontinuum media. As a conclusion, in case of using
FE numerical models without taking the discontinuities into account would definitively lead to erroneous
results in assessing tunnel stability.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction on the discontinuity orientation mechanical behavior will be

In nature, most of the rock masses contain discontinuities.
Especially low stress environments may contain more discontinu-
ities than high overburden areas; such as slopes and near surface
excavations [1]. In these areas, discontinuity properties of rock
masses are more important than intact rock properties [2]. There-
fore, significant effects of discontinuities and lateral pressure coef-
ficient on the failure mode and displacement character of numerous
tunnels have been proven by previous studies [1–7]. Discontinuities
generally exhibit brittle or strain softening behavior and their resid-
ual strength is much lower than their peak strength. So, post-failure
behavior of discontinuities leads to development of progressive
failure mechanisms. As a result, stresses are redistributed to adja-
cent areas and joint segments. As a consequence of redistribution
of stresses, far tunnel materials and joint segments may fail [8].
Therefore, deformations having an anisotropic character may take
place mainly in the form of sliding along discontinuities. Depending
⇑ Corresponding author at: Karayolları Genel Müdürlüğü, ARGE Dairesi
Basßkanlığı, Vecdi Diker Eğitim Tesisleri, Ankara, Turkey. Tel.: +90 312 415 8518.
E-mail address: osatici@gmail.com (Ö. Satıcı).
anisotropic elastic or perfectly plastic isotropic. In such cases,
elasto-plastic nonisotropic models or discontinuum models would
be more appropriate [9]. There are methods used for realistic mod-
eling of discontinuum media such as ‘‘Discrete Element Methods’’
(DEM) [10] and ‘‘Discontinuous Deformation Analysis’’ method
(DDA) [11]. These techniques have been available for a long time
and their capabilities are well proven. However their use certainly
require expertise. Therefore it is not possible to use discontinuum
techniques at every project site. On the other hand finite element
modeling (FEM) technique is more widely used due to simple mod-
eling procedure and availability of softwares based on continuum
modeling. The use of continuum based modeling technique may
result in erroneous design if rock mass contains clear systematic
discontinuity sets. The main objective of this study is to demon-
strate how FEM technique should be applied to obtain realistic
results in case of tunnels opened in jointed rock mass. Nevertheless,
authors of this study suggest the use of DEM and DDA methods in
jointed rock mass modeling. Conversely, most of the field engineers
are not familiar with DEM or DDA methods and Phase2 software
based on FEM has been widely used in underground construction
industry in Turkey. Therefore, most of the road tunnel projects in
Turkey are prepared by using Phase2 software without considering

http://dx.doi.org/10.1016/j.compgeo.2014.11.002
0266-352X/Ó 2014 Elsevier Ltd. All rights reserved.
 Ö. Satıcı, B. Ünver / Computers and Geotechnics 64 (2015) 72–82
discontinuities in rock mass. This will certainly lead misleading
results. However if it is not possible to use discontinuum tech-
niques and FEM has to be used, the methodology presented in this
paper should be applied to realistically include the effect of discon-
tinuous media in FEM. Therefore it is suggested that the method
presented in this study should be followed to obtain realistic results
for FEM.
Discontinuities do not usually used in numerical model studies
and instead discontinuity effect is illustrated by simply reducing
the rock mass’ geotechnical properties. This situation can be
explained as follows; if there are continuous discontinuities in
the medium, medium is recognized as continuum by FEM design-
ers. Conversely, decreasing geotechnical properties of rock mass do
not possibly represent the effects of discontinuity on numerical
models in a realistic way because of the fact that existences of dis-
continuities significantly change stress distribution around the
rock mass. Moreover it is rather difficult to predict strength and
deformation characteristics of rock masses having discontinuities
[2]. For this reason to emphasize the effect of discontinuities at
tunnel stability, tunnel entrance portal section was selected for
modeling. Rock mass examined around the portal area consists of
sandstone–siltstone intercalation, which has two clear discontinu-
ity sets. On this account, two different conditions ‘‘discontinuity
and no discontinuity (massive) cases’’ were modeled to show the
effect of discontinuities at tunnel excavation in the vicinity of por-
tal area.
As the main objective of this study is to determine the effect of
discontinuities or deformation around the tunnel, supporting
elements are not applied in our numerical modeling. However,
results of this study were compared with actual field deforma-
tions data.
2. Location of the tunnel
The Küçükbelvar Highway Tunnel is located in the North East-
ern part of Turkey on the route of Kavak–Merzifon road
(40°530 28.5100 N and 35°350 13.0000 E). It connects Samsun to Amasya
(Fig. 1). Besides, this route reaches out also to Ankara which is the
capital city of Turkey. The tunnel consists of two 345 m long and
nearly 11 m wide tubes. The horizontal distance between tubes
is 35 m and overburden thickness reaches up to 90 m. In this study
tunnel section at km: 88+430 (entrance portal) was modeled,
which has only 22 m overburden thickness.
Fig. 1. Location of the tunnel area and its satellite image.
73
3. Site investigation and engineering geology of the tunnel
entrance
Characteristics of the tunnel route in terms of engineering geol-
ogy were determined by means of field and office studies. These
studies include; literature surveys for geological past of the region,
determination of geotechnical properties of the rock mass, rock
mass classifications studies, discontinuity measurements and geo-
technical drillings. At the end of these studies four different geo-
logical formations on the tunnel axis have identified; Armutlu
Formation, Ağılönü Formation, Quaternary Deposits and Colluvium
respectively. Tunnel excavation has been carried out mainly in the
Armutlu Formation, consisting of thinly to thickly bedded sand-
stone–claystone and sandstone–siltstone intercalation, whereas
some excavation parts were located in the Ağılönü Formation hav-
ing brecciaous limestone [12].
Three geotechnical boreholes were drilled along the tunnel axis
for site investigation purpose during design stage and tunnel axis
was divided into three main sections. At the tunnel entrance;
slightly weathered, moderate to high in strength sandstones have
been encountered. Intercalation of sandstone and claystone form
the middle section, which is mainly controlled by sandstone. A
fault has combined the Armutlu and Ağılönü Formations at the
tunnel exit. In this section external side of the limestone is highly
fractured and brecciaus, but it is mainly represented by conglom-
erates and interbedded sandstone (Armutlu Formation). Limestone
with solution opening is also located at the tunnel exit. This
limestone was observed as form of breccia in it is faulty contacts
with sandstone, and named as brecciaus limestone. The Brecciaus
limestone was moderately weathered and moderate to high in
strength [12].
Rock mass around the entrance portal contains two clear
discontinuity sets and eventually shallow overburden thickness.
Entrance portal was intentionally selected to demonstrate effects
of discontinuities. The selected section for analyses is represented
with Armutlu Formation (Figs. 2 and 3) and consists of mainly
sandstone, siltstone, and partly conglomerates. Conglomerates
and sandstones are very strong and thickly bedded. Sandstone
siltstone intercalation is moderate to high in strength, medium
weathered and fractured. Sandstones are thick in this section,
spacing reaches up to 1 m, and persistence reaches up to 12 m.
However, in some sections discontinuity spacing is 20–60 cm
and descends up to 5–10 cm at crushed siltstone sections.
74 Ö. Satıcı, B. Ünver / Computers and Geotechnics 64 (2015) 72–82

Sandstone

Siltstone

Conglomerates

Fig. 2. Sandstone–siltstone and conglomerates in Armutlu formation.

Tea: Armutlu Formation

N
Km: 88+430
Tunnel entrance

SW
Fault;

NE

Geotechnical drilling;

Tunnel entrance
Portal area

Tea: Armutlu Formation

Longitudinal section of Km: 88+430

Fig. 3. Engineering geology map and profile of entrance portal (not to scale).
 Ö. Satıcı, B. Ünver / Computers and Geotechnics 64 (2015) 72–82 75

Discontinuity surfaces are mostly smooth, but rough surfaces can disturbs the rock mass condition, it also affects these constants
be observed at some parts. Average discontinuity separation is and rock mass parameters. So, residual GSI and Hoek–Brown
1–3 mm, with calcite and clay fillings (Table 1). Stereographic parameters were used in design as given in Table 2 [18–20].
projection of discontinuities observed at the entrance section This phenomenon has been first discussed by Cai et al. [18]. If it
was given in Fig. 4. is need to describe; rock mass around excavation will be disturbed
as excavation progress. Owing to the excavation method
4. Geotechnical properties of rock masses and intact rock at the (mechanical or blasting), geotechnical parameters of rock mass
entrance portal (km: 88+430) will not remain same around tunnel perimeter after excavation.
To demonstrate effect of excavation, and establish a realistic
Geotechnical parameter determination of intact rock and rock numerical model, residual values of mi, mb, s, a and GSI should
masses are an inevitable step for realistic numerical modeling. be used. For the sake of clarity, merely the calculation of residual
Therefore, 26 core samples were obtained from 3 geotechnical dril- GSI value was given here in detail (Eq. (1));
lings and tested at General Directorate of Turkish Highways rock
GSIr ¼ GSI  e0:0134GSI ð1Þ
mechanics laboratory in comply with suggested methods of ISRM.
In addition to laboratory tests; RMR [13], Q [14], and GSI [15,16] Peak GSI value was determined as 55 from field studies.
values were also determined by the help of field studies. After that Then GSIr can be obtained as 26.32 from Eq. (1).
some well-known empirical equations were used to obtain neces-
sary rock mass geotechnical parameters for FEM and all results are ‘‘Roclab’’ can also calculate the modulus of elasticity value by
presented in Table 2. using empirical approach of Hoek and Diederichs [21]. Although
‘‘RocLab’’, software and some previous empirical studies were there are many other studies for determining rock mass’ elastic
used to obtain some necessary rock mass parameters. ‘‘mi, mb, s, modulus value [22–25], Hoek and Diederichs [21] approach was
and a’’ are well-known rock mass constants, which were obtained preferred to provide consistency with other rock mass parameters,
from Hoek–Brown failure criterion [17]. But, as the excavation obtained from ‘‘Roclab’’. According to this approach, modulus of
elasticity value of rock mass is obtained from Eq. (2);
Table 1
Engineering geology properties of entrance portal (Armutlu Formation encountered at
 
1  D=2
km: 88+430). Erm ¼ Ei 0:02 þ ð2Þ
1 þ eðð60þ15DGSIÞ=11Þ
Lithology Sandstone, siltstone
Color Grey, yellowish brown  
Strength High – moderate
1  0:7=2
Erm ¼ 19 0:02 þ
Weathering Slightly – medium 1þ eðð60þ150:755Þ=11Þ
Roughness Mostly smooth, some rough surfaces
¼ 2:805 GPa was obtained
Bedding 20–60 cm
Geotechnical drilling SMK 54
RQD% 70 and residual Erm was found as 0.633 GPa from Eq. (2) by using
UCS (Ei) 65 MPa residual GSI (GSIr = 26.32) value obtained from Eq. (1). Where Ei is
Discontinuity sets 2 set main 09/359, 84/017 and 2 set random, the intact rock modulus of elasticity value, Erm is the rock mass
minor importance modulus of elasticity value, D is the disturbance factor, GSI is the
Discontinuity Length 1–12 m
(Persistence)
Geological Strength Index.
Separation (Aperture) 1–3 mm Although cohesion, UCS, and modulus of elasticity values
Infilling Calcite – clay obtained by using Roclab, internal friction angle of the rock mass
was found to be rather high (59°). As it was difficult to determine

Tunnel axis

Fig. 4. Discontinuity sets and stereographic representation of entrance portal.

76 Ö. Satıcı, B. Ünver / Computers and Geotechnics 64 (2015) 72–82

Table 2
Geotechnical properties of entrance portal (km: 88+430).

Rock mass classification according to field studies

RMRbasic: 60 RMRadj: 50 RMR = 9lnQ + 44 = 51 Q: 2,33 GSIpeak: 55 RQD: % 70
Intact rock properties (laboratory results)
3
UCS: 65 MPa Ei: 19 GPa c: 25 (kN/m ) Ci: 10–20 (MPa) /i: 40–51° t: 0.2
Discontinuity surface properties (literature)
Cj1,2: 0.003 MPa /j1,2: 30°
Rock mass properties (empirical studies)
mi: 17 (peak) mb: 1.434 (peak) s: 0.001471 (peak) a: 0.504 (peak) Df: 0.7
mi: 13 (resid.) mb: 0.925 (resid.a) s: 0.0002686 (resid.a) a: 0.529 (resid.a) Crm: 0.321 MPa
Erm: 2.805 GPa Erm: 0.633 GPa (resid.a) UCSrm: 2.428 MPa GSIresid.: 26 /rm: 59°
a
Residual; rock mass constants, GSI and Erm values are obtained from Cai et al. [18] study.

these values precisely, the exact values of cohesion (0.321 MPa)
and internal friction angle (59°) for rock mass were not preferred
to use in this study. Therefore, the internal friction angle and cohe-
sion of the rock mass was estimated by using Hoek and Brown [15]
approach. Anyone who wants to obtain this value can also use
some previous studies like Afrouz [26], Hoek and Bray [27] and
Stillborg [28].
The other important parameter used in FEM is field stress
(stress ratio). There are various approaches used for estimation of
in-situ stress or field stresses ratio ‘‘k’’ [29–31]. Generally high hor-
izontal stresses are caused by different factors such as; erosion,
tectonics, rock anisotropy and rock discontinuities. In case of hav-
ing complex topographic conditions like hills and valleys and mod-
eled area is close to surface, rock mass can be assumed to be under
gravity loads alone [32]. In our study, this phenomenon has also
proven by field observations and numerical analysis. Therefore,
vertical stress (rv) are generated due to gravitational loading (Eq.
(3)), and horizontal stress (rh) was determined from Yokoyama
et al. [31] approach (Eq. (4)).
rv ¼ c  h
rv ¼ 2:5  22 ¼ 55 kPa
For sedimentary rock;
rhmax ¼ 23:5 þ 0:0340h
rhmax ¼ 23:5 þ 0:0340  22 ¼ 24:248 kPa
By the help of these two stresses ‘‘k = rh/rv’’ (stress ratio) was
determined as 0.44. Where rv and rh are the vertical and horizon-
tal earth pressures respectively, c is the average unit weight and h
is the overburden thickness at the analyzed section.
The last geotechnical parameter for accurate design is disconti-
nuities’ cohesion and internal friction angle. Unfortunately, geo-
technical properties of discontinuities were not determined by
laboratory or in situ experiments in this study. Instead, approxi-
mate cohesion and internal friction angle values given in the liter-
ature [8,26–28,32–34] and Table 1 were used for determination of
geotechnical properties of discontinuities. As a result, internal fric-
tion angle is taken as 30° (presence of mostly smooth discontinuity
surfaces), and due to infilling and low overburden thickness 3 kPa
was selected for peak cohesion value.
5. Numerical modeling
Independent from applied excavation method, there will be a
time gap between excavation, hauling and supporting in practice.
Consequently, rock mass around the tunnel excavation loses its
stiffness till installation of the first supporting element and within
this time gap vast amount of deformations have already occurred
[35]. Hence, rock mass’ geotechnical parameters may lessen to
ð3Þ their residual values depending on magnitude of in-situ stress
and post-failure behavior of the rock mass [18–20]. A realistic
numerical modeling application must incorporate with this phe-
nomenon. This philosophy was explained in detail previously by
some researchers [36–38].
ð4Þ
5.1. Modeling procedure
To demonstrate time gap between excavation, hauling and sup-
porting into the model, top heading excavation was completed in 3
steps and this was applied for left and right tube respectively. In
practice to provide accessibility for top heading excavation section

Stage 2: Left tube top Stage 3: Left tube top Stage 4: Left tube top
Stage 1: Initial heading excavation heading excavation heading excavation
element loading step1 step2 step3

Stage 5: Right tube top Stage 6: Right tube top Stage 7: Right tube top
heading excavation heading excavation heading excavation
step1 step2 step3

Stage 10: Left tube bench

Stage 8: Left tube bench Stage 9: Right tube
excavation step2
excavation step1 bench excavation step1 (excavation completed)

Stage 11: Right tube bench

excavation step2
(excavation completed)

Fig. 5. Stages of numerical modeling for a realistically representation of the tunneling operation.
 Ö. Satıcı, B. Ünver / Computers and Geotechnics 64 (2015) 72–82 77

for mounting of support system and for the next excavation step,
bench excavation do not finalized at once. Therefore, bench exca-
vation was modeled in 2 stages for left tube and right tube respec-
tively. Stages of the model can be explained better in the form of a
flow chart in Fig. 5.
Boundary conditions were taken as 2.5 times of tunnel diameter
for side walls and 2 times of tunnel diameter from the bottom,
respectively (Figs. 6 and 7). Whereas the excavation model bound-
ary was fixed from both sides and bottom, top of the model was set
free because of low overburden thickness. Mesh and discretization
was generated automatically by software and three-noded
triangles were chosen as mesh element type. In both cases; rock
mass’ geotechnical properties (residual and peak) presented in
Table 2 were used as input parameters. Material type was chosen
as plastic. If the material type would have been chosen as elastic,
the input rock mass parameters would only have been used for
calculation and plotting of strength factor. In this case, any failure
would not be observed in the model. Conversely, by selecting the
material type as plastic, residual strength parameters can be used
in case of failure of the material. As our aim is to show how much
deformation occurs in case of discontinuities taking into account
for the same modeling conditions, plastic strength parameters will

Top heading excavation steps

Bench excavation steps

Fig. 6. Numerical model section for homogenous condition together with geotechnical parameters.

Top heading excavation steps

Bench excavation steps

Discontinuity set 09/359

Discontinuity set 84/017

Fig. 7. Numerical model section for discontinuity added model together with geotechnical parameters.
78 Ö. Satıcı, B. Ünver / Computers and Geotechnics 64 (2015) 72–82

1 2

3 4

5 6

7 8

9 10

11

Fig. 8. 11-staged excavation model results, without any discontinuity set.

 Ö. Satıcı, B. Ünver / Computers and Geotechnics 64 (2015) 72–82 79

1 2

3 4

5 6

7 8

9 10

11

Fig. 9. 11-staged excavation model results, with 2 discontinuity sets.

80 Ö. Satıcı, B. Ünver / Computers and Geotechnics 64 (2015) 72–82

be more suitable for our study. Therefore, behavior of the rock case of continuum modeling approach (without discontinuities)
mass after yielding can only be simulated with plastic deformation. yielding for surrounding rock was not observed, almost any yield-
Residual rock mass parameters were obtained as suggested by Cai ing spots or plastic zone were not developed around the excavation
et al. [18]. boundary at the end of 11th stage, in other words model was
Original Hoek–Brown was chosen for the failure criterion, as behaved elastically (Fig. 8). Conversely in the discontinuity added
because the rock mass is controlled by tightly interlocking angular model, yielding spots have started to develop especially in roof sec-
rock pieces and can be defined as good rock. The original Hoek– tion at the end of 4th stage and maximum total deformation was
Brown failure criterion was found to work well with this kind of obtained as 2.5 cm (0.02492 m) for left tube roof and 1.8 cm
rock masses [17]. Geometry, mesh, discretization and geotechnical (0.01804 m) for right tube roof after completing of excavation
parameters were given in detail in Figs. 6 and 7 for each model. (Fig. 9). That means, residual strength related parameters repre-
Only difference between two modeling cases is discontinuity senting plastic behavior are started to use especially in roof section
sets. Two main discontinuity sets (09/359, 84/017) were measured at the end of 4th stage. While almost no shear and tension spots
in the vicinity of portal area and these discontinuities were added observed in the massive model, many shear and tension spots were
to one of the models in accordance with their dip/dip direction. In observed in the roof section for discontinuity added model, and
massive case (discontinuity sets were not defined) continuum these spots reached up to the model surface. This deformation
modeling approach [2] was represented, while discontinuity sets zone, which starts from the excavation roof through the surface
added model represents discontinuum modeling approach [9]. is similar to Terzaghi’s ground arch concept (Fig. 9 stage 11).
However, as it can be clearly seen from Figs. 8 and 9, any sup-
5.2. Comparison of results porting element is not necessary. Of course, this will not be a real-
istic approach. It should be kept in mind that, time dependent
Stage based results obtained from numerical analyses are deformations could not be modeled by FEM softwares. So, in prac-
presented in Figs. 8 and 9 for continuum and discontinuum tice if the excavated span left unsupported by trusting calculated
models, respectively. Surrounding rock behavior at continuum amount of deformation, results may be catastrophic. 45-day tunnel
and discontinuum models are examined and compared in the form monitoring data (field deformation records), which was given in
of total deformation at roof and right walls of the tunnels. Table 3, Figs. 10 and 11. for supported tunnel section (Küçükbelvar
11-staged finite element analysis results have shown that even if Tunnel) prove this phenomenon. Excavation stability has been
all the geotechnical parameters were the same, obtained total achieved only by application of appropriate supporting system.
deformation results are completely different for both models. In In fact actual deformations after excavation are more than the
measured one. Namely, before installation of first monitoring
device, excavation face have to advance forward to at least 8 to
Table 3 10 m. Otherwise, monitoring device equipment will suffer damage
45-day field monitoring results (supported).
from blasting or excavation equipment. This shortcoming about
Date Displacements (m) installation of monitoring system was explained in some previous
Left tube roof Right tube roof studies. According to these studies 30–80% of immediate deforma-
tion could not be measured by monitoring devices, owing to the
08.06.2006 0 0
09.06.2006 0.0110 0.0097 contractor ability or unwillingness to install these devices
10.06.2006 0.0115 0.0099 [35,39–41]. This makes our discontinuity-added numerical model
11.06.2006 0.0128 0.0101 approach more reliable, if 60% is accepted as an average measure-
12.06.2006 0.0133 0.0106 ment deficiency. In case of actual final deformation measurement
13.06.2006 0.0138 0.0110
14.06.2006 0.0141 0.0115
values will increase up to 60% (0.0156 m  1.06 for the left tube
18.06.2006 0.0148 0.0117 roof and 0.0123 m  1.06 for the right tube roof) then 0.02496 m
20.06.2006 0.0153 0.0119 deformation is obtained for the left tube (0.02492 m was obtained
26.06.2006 0.0155 0.0120 from FEM) and 0.01968 m for the right (0.01804 m was obtained
02.07.2006 0.0156 0.0122
from FEM). These results show the good agreement with
10.07.2006 0.0156 0.0123
20.07.2006 0.0156 0.0123 discontinuity-added FEM deformation results with actual field
measurement data.

0.025
FEM displacement result for the
Total displacements of supported

unsupported left tube roof is nearly 60%

0.02
more of the actual field monitoring
displacement data
left tube roof (m)

0.015

0.01

0.005

Monitoring dates

Fig. 10. Left tube roof total displacements results obtained from field monitoring.
 Ö. Satıcı, B. Ünver / Computers and Geotechnics 64 (2015) 72–82
0.02
Total displacements of supported
0.015
right tube roof (m)
0.01
0.005
Fig. 11. Right tube roof total displacements results obtained from field monitoring.
6. Conclusion
In line with the objective of this study, numerical modeling
strikingly demonstrates the paramount effect of discontinuities
on stability when FEM is used. Total displacement at the roof of left
tube reached to a maximum value of 0.25 cm in massive (contin-
uum) case. By adding discontinuities and leaving every other detail
as unchanged, this value was obtained as 2.5 cm which is quite
coherent with actual deformation results. Our study revealed that
discontinuum conditions could not have illustrated properly only
by using reduced geotechnical parameters in FE numerical model-
ing applications. Therefore, the procedure presented in the paper is
believed to be a good application example for FE numerical model-
ing, which includes discontinuities. It is quite common that to form
a homogeneous medium for the FE models because of its simplic-
ity. However, this may lead to catastrophic consequences. Consid-
eration of discontinuity characteristics of rock mass would
certainly make 3D numerical analysis very complex. Therefore,
2D numerical modeling incorporating discontinuities like DEM
and DDA can be a more viable solution for this kind of modeling
application. However, these are not common in commercial appli-
cations and are not user friendly softwares for discontinuum
modeling.
This study also proves the importance of proper sensitive geo-
logical modeling, field studies, (site investigations) and their appli-
cations for numerical models. For the sake of simplicity it is quite
often to construct commercial FE numerical models without con-
sidering discontinuities. This would certainly lead to erroneous
results. Discontinuities should be always taken into account and
represented in FE numerical modeling in order to obtain reliable
and realistic results.
Acknowledgements
This study could not be happened without supports of A. GUL-
SEN and S. KIZIROGLU from General Directorate of Highways and
M.K. AKMAN from Yuksel Proje Uluslararası A.S.
References
[1] Jia P, Tang CA. Numerical study on failure mechanism of tunnel in jointed rock
mass. Tunn Undergr Space Technol 2008;23(5):500–7.
[2] Curran JH, Hammah RE, Yacoub T, Corkum B. The practical modeling of
discontinuous rock masses with finite element analysis. ARMA 2008:08–180.
[3] Bandis SC, Lumsden AC, Barton NR. Fundamentals of rock joint deformation.
Int J Rock Mech Min Sci 1983;20(6):249–68.
81
FEM displacement result for the
unsupported right tube roof is nearly
60% more of the actual field monitoring
displacement data
Monitoring dates
[4] Cundall P, Voegele M, Fairhurst C. Computerized design of rock slopes using
interactive graphics for the input and output of geometrical data. In: 16th
Symposium on Rock Mechanics, Minneapolis, USA; 1975.
[5] Goodman RE, Taylor RL, Brekke TL. A model for the mechanics of jointed rock. J
Soil Mech Found Div 1968:637–59.
[6] Manfredini G, Martinetti S, Ribacchi R. Inadequacy of limiting equilibrium
methods for rock slopes design. In: 16th Symposium on Rock Mechanics,
Minneapolis, USA; 1975. p. 35–43.
[7] Song JJ, Lee CI, Seto M. Stability analysis of rock blocks around a tunnel using a
statistical joint modeling technique. Tunn Undergr Space Technol
2001;16(4):341–51.
[8] Hammah RE et al. The practical modeling of discontinuous rock masses with
finite element analysis. ARMA 2008:08–180.
[9] Carranza-Torres C, Fairhurst C. Application of the convergence-confinement
method of tunnel design to rock masses that satisfy the Hoek–Brown failure
criterion. Tunn Undergr Space Technol 2000;15:187–213.
[10] Jing L, Stephansson O. Fundamentals of discrete element methods for rock
engineering: theory and applications. Develop Geotech Eng 2007;85.
[11] Wu JH, Ohnishi Y, Nishiyama S. Simulation of the mechanical behavior of
inclined jointed rock masses during tunnel construction using discontinuous
deformation analysis. Int J Rock Mech Min Sci 2014;41:731–43.
[12] Genç Sß , et al. Merzifon (Amasya) Dolayının Jeolojisi. MTA Jeo. Etüd Dai. Yay.,
9527; 1991.
[13] Bieniawski ZT. Engineering rock mass classification. New York: John Wiley
Sons; 1989.
[14] Barton N. Some new Q-value correlations to assist in site characterization and
tunnel design. Int J Rock Mech Min Sci 2002;39:185–216.
[15] Hoek E, Brown ET. Practical estimates of rock mass strength. Int J Rock Mech
Min Sci 1997;34(8):1165–86.
[16] Sönmez H, Ulusay R. A discussion on the Hoek–Brown failure criterion and
suggested modification to the criterion verified by slope stability case studies.
Yerbilimleri Dergisi 2002;26:77–9.
[17] Hoek E, Carranza-Torres C, Corkum B. Hoek–Brown failure criterion 2002
edition. In: Proceedings of the NARMS-TAC. Mining Innov and Technol,
Toronto, Canada; 2002. p. 267–73.
[18] Cai M, Kaiser PK, Tasaka Y, Minamic M. Determination of residual strength
parameters of jointed rock masses using the GSI system. Int J Rock Mech Min
Sci 2007;44:247–65.
[19] Kwon S et al. An investigation of the excavation damaged zone at the KAERI
underground research tunnel. Tunn Undergr Space Technol 2009;24:1–13.
[20] Martino JB, Chandler NA. Excavation-induced damage studies at the
underground research laboratory. Int J Rock Mech Min Sci 2004;41:1413–26.
[21] Hoek E, Diederichs MS. Empirical estimation of rock mass modulus. Int J Rock
Mech Min Sci 2006;43(2):203–15.
[22] Sönmez H, Gökçeoğlu C, Ulusay R. Indirect determination of the modulus of
deformation of rock masses based on the GSI system. Int J Rock Mech Min Sci
2004;1:849–57.
[23] Nicholson GA, Bieniawski ZT. A non-linear deformation modulus based on rock
mass classification. Int J Min Geol Eng 1990;8:181–202.
[24] Palmstrom A, Singh R. The deformation modulus of rock masses: comparisons
between in situ tests and indirect estimates. Tunn Undergr Space Technol
2001;16.
[25] Sönmez H, Gökçeoğlu C, Nefeslioğlu HA, Kayabasßı A. Estimation of rock
modulus: for intact rocks with an artificial neural network and for rock masses
with a new empirical equation. Int J Rock Mech Min Sci 2005;43(2):224–35.
[26] Afrouz A. Rock mass classification systems and modes of ground
failure. Florida: CRC Press; 1992.
[27] Hoek E, Bray JW. Rock slope engineering. London: The Inst. of Min. and Metall;
1981.
82 Ö. Satıcı, B. Ünver / Computers and Geotechnics 64 (2015) 72–82
[28] Stillborg B. Professional users handbook for rock bolting. Germany: Trans Tech.
Publ.; 1986.
[29] Hoek E, Brown ET. Underground excavation in rock. London: The Inst. of Min.
and Metall.; 1982.
[30] Sheorey PR. A theory for in situ stresses in isotropic and transversely isotropic
rock. Int J Rock Mech Min Sci 1994;31(4):23–4.
[31] Yokoyama T, Ogawa K, Kanagawa T, Tanak M, Ishida T. Regional in-situ stress
in Japan based on Measurements. In: Rock Stress: Proc. 3rd Int. Symp. on Rock
Stress, Kumamoto, Japan; 2003. p. 335–41.
[32] Zhang L. Engineering properties of rocks. Elseiver Geo-Engineering Book
Series; 2005.
[33] Wyllie CD, Mah CW. Rock slope engineering. New York: Spon Press; 2005.
[34] Pasßamehmetoğlu AG, Özgenoğlu A, Karpuz C. Kaya Sß ev Stabilitesi. TMMOB
Maden Müh. Odası Yayını Ankara; 1999.
[35] Kavvadas MJ. Monitoring ground deformation in tunneling: current practice in
transportation tunnels. Eng Geol 2005;79:93–113.
[36] Curran JH, Hammah RE, Yacoub T. A two-dimensional approach for designing
tunnel support in weak rock. Canada: Rocscience Inc., <http://
www.rocscience.ca/assets/files/uploads/7689.pdf> [26.09.13].
[37] Vlachopoulos N, Diederichs MS. Improved longitudinal displacement profiles
for convergence confinement analysis of deep tunnels. Rock Mech Rock Eng
2009;42(2):131–46.
[38] Millard F, Krishnaswamy S. A theory of stress-softening in incompressible
isotropic materials. J Mech Phys Solids 2000;48(9):1931–65.
[39] Bizjak KF, Petkovsek B. Displacement analysis of tunnel support in soft rock
around a shallow highway tunnel at Golovec. Eng Geol 2004;75:89–106.
[40] Kim BYN, Chung HS. A study on prediction of final displacement of road tunnel
section during excavation in highly weathered rock by NATM. KSCE J Civil Eng
2002;6(4):399–405.
[41] Kontogianni VA, Stiros SC. Predictions and observations of convergence in
shallow tunnels: case histories in Greece. Eng Geol 2002;63:333–45.