See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/332880105

Geotechnical studies and primary support design for a highway tunnel: a case
study in Turkey

Article in Bulletin of Engineering Geology and the Environment · May 2019

DOI: 10.1007/s10064-019-01529-8

CITATIONS READS

14 149

2 authors, including:

Ayberk Kaya
Recep Tayyip Erdoğan Üniversitesi
36 PUBLICATIONS 519 CITATIONS

SEE PROFILE

All content following this page was uploaded by Ayberk Kaya on 29 September 2023.

The user has requested enhancement of the downloaded file.

Geotechnical studies and primary support
design for a highway tunnel: a case study in
Turkey

Ayberk Kaya & Fikri Bulut

Bulletin of Engineering Geology and

the Environment
The official journal of the IAEG

ISSN 1435-9529

Bull Eng Geol Environ

DOI 10.1007/s10064-019-01529-8

1 23
Your article is protected by copyright and
all rights are held exclusively by Springer-
Verlag GmbH Germany, part of Springer
Nature. This e-offprint is for personal use only
and shall not be self-archived in electronic
repositories. If you wish to self-archive your
article, please use the accepted manuscript
version for posting on your own website. You
may further deposit the accepted manuscript
version in any repository, provided it is only
made publicly available 12 months after
official publication or later and provided
acknowledgement is given to the original
source of publication and a link is inserted
to the published article on Springer's
website. The link must be accompanied by
the following text: "The final publication is
available at link.springer.com”.

1 23
 Author's personal copy
Bulletin of Engineering Geology and the Environment
https://doi.org/10.1007/s10064-019-01529-8

ORIGINAL PAPER

Geotechnical studies and primary support design for a highway

tunnel: a case study in Turkey
Ayberk Kaya 1 & Fikri Bulut 2

Received: 7 November 2018 / Accepted: 11 April 2019

# Springer-Verlag GmbH Germany, part of Springer Nature 2019

Abstract
The main aim of this study is to appraise the geotechnical characteristic of the rock masses and to propose the proper support
design for the Cankurtaran Tunnel project situated in NE Turkey. The exhaustive engineering geological investigations were
done to determine the characteristic of rock masses that primarily consist of volcanic and sedimentary rocks. The tunnel route was
divided into 15 segments according to their lithological and structural properties. The rock mass rating (RMR) and Rock Mass
Quality Index (Q) systems were used to determine the quality of rock masses and final tunnel lining support. To check the
capacity of the suggested support units analytically, the convergence-confinement (CC) technique was applied. The efficiency of
the support design, dimensions of the plastic zones and deformations were determined using the 2D and 3D numerical finite
element method (FEM) modeling. The empirical support system suggested in this study reduced the total displacement and
dimension of the plastic zone.

Keywords Tunneling . Convergence-confinement method . Rock mass classification . FEM . Tunnel lining design

Introduction systems by many geotechnical engineers. In spite of the

fact that these techniques are advantageous instruments in
Selection of the most economical and suitable support preliminary support design, they do not provide essential
design contributes positively to an underground excava- information in strain and stress estimations. Therefore,
tion project in design and construction processes. numerical methods gain popularity in determination of
Empirical, analytical and numerical approaches are gener- stress distributions and deformations, and checking the
ally preferred methods when tunnels are projected by de- lining design for the best remedy. Recent research has
signers. In order to design tunnel supports, several re- showed that if empirical techniques are integrated into
searchers have suggested rock mass classification systems numerical methods, more realistic outcomes in lining de-
which have gained universal acceptance (i.e. Terzaghi sign will be acquired (i.e. Basarir 2006; Gurocak et al.
1946; Lauffer 1958; Rabcewicz 1964; Wickham et al. 2007; Sopaci and Akgun 2008; Kaya et al. 2011; Aydin
1972; Barton et al. 1974; Bieniawski 1989; Palmström et al. 2014; Yalcin et al. 2015; Lin et al. 2017; Kanik and
1995; Marinos and Hoek 2000; Aydan et al. 2014). Due Gurocak 2018).
to their practicality, the Rock Mass Quality Index (Q; The Turkish General Directorate of Highways (TGDH)
Barton et al. 1974) and rock mass rating (RMR; planned a project for reclamation of the Artvin–Hopa mo-
Bieniawski 1989) classifications are widely utilized torway (km: 6 + 500 – 13 + 787) owing to increasing
traffic load and congestion. This project includes con-
struction of the double-tubed Cankurtaran Tunnel situated
between km: 7 + 980 – 13 + 208. The total length of the
* Ayberk Kaya
[email protected] modified horseshoe-shaped tunnel is 5.228 km, and the
tunnel is constructed using mechanical excavation and
1 conventional drilling-blasting techniques subjected to
Faculty of Engineering, Department of Civil Engineering, Recep
Tayyip Erdogan University, 53100 Rize, Turkey existing rock mass behaviors and conditions. The
2 planned, unsupported span and height are 12 and 9 m,
Faculty of Engineering, Department of Geological Engineering,
Karadeniz Technical University, 61080 Trabzon, Turkey respectively. The height of surcharge ranges between 80
 Author's personal copy
A. Kaya, F. Bulut

and 410 m above the tunnel. The map showing the tunnel In the present study, the Cankurtaran Tunnel was selected
location is presented in Fig. 1, and the field view of the as an implementation site for the empirical, analytical and
entrance and exit sections are given in Fig. 2a–b. numerical tunnel support process. A geotechnical

Fig. 1 Location map of the study area

 Author's personal copy
Geotechnical studies and primary support design for a highway tunnel: a case study in Turkey

Fig. 2 (a) Pre-excavation view of

a b
the entrance, (b) exit sections and
field view of the (c) A1 and A2,
(d) B1, (e) B2, (f, g) C and (h) D
geotechnical units

Entrance section

Exit section

c d

UNIT B1

UNIT A1

UNIT A2

e f
UNIT B2

UNIT C

UNIT B1

g h
UNIT C UNIT D

investigation followed by rock mass classification and numer- Study area

ical modeling was carried out. The performance of empirical
support design was also controlled via finite element method Based on the structural and lithological arguments, the Eastern
(FEM) modeling. In this framework, the 2D FEM results were Pontide Tectonic Assembly is divided into northern and south-
compared against 3D analysis results and assessed. ern zones (Ketin 1966; Guven 1993). The ages of geological
 Author's personal copy
A. Kaya, F. Bulut

units in the Eastern Pontide vary between Paleozoic and Because of the dense vegetation and steep morphology,
Cenozoic. The tunnel is located in the northern zone, and four seven drillings (total length of 607 m) have been opened
formations crop out along the tunnel route (Capkinoglu 1981). under the control of the TGDH (Figs. 3 and 4) to monitor
The Subasi Ridge Formation consists of andesitic py- the ground properties at the tunnel depth, to characterize
roclastic rocks and intercalations of sandstone, limestone, the joint features and groundwater level, core sampling,
marl and tuff (Fig. 2c). The Late Cretaceous-aged unit is and water pressure tests (Lugeon test). Three of the dril-
generally located at the entrance section of the tunnel. The lings (BH 1–3) were placed at the inlet section, one bore-
Late Cretaceous-Paleocene-aged Cankurtaran Formation hole (BH 7) at the outlet part and another three (BH 4–6)
cropped out in the inner section is defined by limestone towards the center of the tunnel line.
and marl, and conformably overlies the Subasi Ridge The water pressure tests suggested by Lugeon (1933) were
Formation. While the lower and upper parts of the forma- performed to identify the permeability properties of the tunnel
tion are thin-bedded (Fig. 2d), the middle part is thick- ground and, therefore, to collect data on the water inflow
bedded (Fig. 2e). In the inner part of the tunnel, the potency into the tunnel. Along the tunnel alignment, the
Subasi and Cankurtaran Ridge Formations are juxtaposed groundwater depth varied between 3.4 and 17.6 m. The
along the NE–SW-trending inactive vertical faults F1 and Lugeon tests were carried out below the groundwater level
F2. The Paleocene-aged Senkaya Ridge Formation is a in all boreholes at 5-m intervals. The Lugeon experiment re-
sedimentary package including thin- to medium-layered sults showed that the rock mass permeability varied from very
marl and limestone with interbedded claystone (Fig. 2f low to moderate for Bunit A1 [0.92–5.74 Lugeon units (Lu)]^
and g). This unit is generally located towards the end of and Bunit A2 (0.38–6.72 Lu)^ considering the classification
the inner part of the tunnel. The Eocene-aged Kabakoy suggested by Quinones-Rozo (2010). On the other hand, the
Formation which is usually cropped out at the exit part permeability of rock mass varied from very low to low for
of the tunnel consists of basalt-andesite and pyroclastic Bunit B1 (0.00–1.08 Lu)^ and Bunit B2 (0.32–4.17 Lu).^
rocks (Fig. 2h), and unconformably overlies the Senkaya However, the permeability of rock mass is very low (0.00–
Ridge Formation. Near the village of Ciftekopru, the 0.32 Lu) for Bunit C^ and low (1.02–5.00 Lu) for Bunit D.^
Senkaya Ridge and Kabakoy Formations are juxtaposed In other words, based on the average Lugeon values of the
along the NE–SW-trending inactive vertical fault F3. rock masses presented in Table 1, unit A1, unit A2, unit B2
In this study, for convenience, the andesitic pyroclastites in and unit D exhibit low permeability, and unit B1 and unit C
the Subasi Ridge Formation were named as Bunit A1^ and as are impermeable considering the classification of Lugeon
Bunit A2^ for sedimentary intercalations (Fig. 2c). The thin- (1933). A general decrease in permeability with depth was
bedded and thick-bedded levels of the Cankurtaran Formation observed due to the reduction in fracture aperture and filling
were designated as Bunit B1^ and Bunit B2^, respectively of joints with clay particles during weathering. These data
(Fig. 2d and e). Furthermore, the Senkaya Ridge Formation indicate that the geotechnical units are generally dry, and no
was symbolized as Bunit C^ and as Bunit D^ for the Kabakoy sudden water flows were expected during the tunnel excava-
Formation (Fig. 2f-h). Figure 3 shows the generalized geolog- tion. On the other hand, because the fault zone materials lo-
ical map of the study area. The rock masses in the portal cated along the tunnel consist of finer grain size and angular
sections and fault zones (F1, F2 and F3) were also delineated shape of fractured grains, these zones were assumed to be
as separate segments in this study (Fig. 4). more permeable than non-faulted rock masses. Therefore, dif-
The geological cross section was prepared by the help of ficulties due to the risk of sudden water flow are possible
the drilling, seismic survey and field studies data, and the during the excavation of fault zones.
tunnel route was divided into 15 segments based on the struc- Laboratory experiments were performed on the core sam-
tural, lithological and geotechnical properties. The tunnel seg- ples compiled from the drillings and rock blocks based on the
ments are shown in Fig. 4. techniques suggested by ISRM (2007) to define the physical,
mechanical and elastic characteristics of the intact rocks, in-
cluding unit weight (γ), uniaxial compressive strength (σci),
Method point load strength index (IS(50)), and Young’s modulus (Ei). In
addition to this, the rock quality designation (RQD) was de-
Geotechnical studies fined from drillings and scan-line surveys using the methods
proposed by Deere (1964) and Priest and Hudson (1976). The
Geotechnical characteristic of the rock units located along the average RQD values and rock quantities for the geotechnical
Cankurtaran Tunnel alignment were determined performing unit A1, unit A2, unit B1, unit B2, unit C and unit D were
surface, subsurface and laboratory studies. The field work found to be 84% good, 60% fair, 62% fair, 88% good, 3%
consisted of mapping, drilling, scan-line survey, water pres- very poor and 73% fair, respectively. Table 1 demonstrates the
sure test, seismic investigation and geotechnical description. results of laboratory test and RQD values.
 Author's personal copy
Geotechnical studies and primary support design for a highway tunnel: a case study in Turkey

Fig. 3 Simplified geological map of close vicinity of the study area (modified from Capkinoglu 1981)

Discontinuity readings were taken from outcrops to identi- by ISRM (2007). The quantitative descriptions of the discon-
fy the main discontinuity sets. The discontinuity orientations tinuities are shown in Table 3.
were identified by using the Dips v7.0 (Rocscience Inc. In the present study, the most commonly preferred
2016a) software according to stereographic projection classifications such as the RMR and Q system were
(Fig. 5 and Table 2). Additionally, the identification of the employed to describe the ground throughout the tunnel
discontinuities in geotechnical units such as orientation, aper- route and to determine empirical lining design. The data
ture, spacing, roughness, persistence, weathering degree and acquired from field works, scan-line surveys, boreholes
infilling were obtained by examining the core samples and and laboratory studies were utilized in the classification
scan-line surveys by the help of the technique recommended systems.

Fig. 4 Geological cross section showing the borehole locations, segments, chainages and geotechnical units along the tunnel alignment
 Author's personal copy
A. Kaya, F. Bulut

Table 1 Physico-mechanical and elastic properties of rock materials, rock quality designation (RQD) and Lugeon values

Properties Subasi Ridge Formation Cankurtaran Formation Senkaya Ridge Kabakoy

Formation Formation
Unit A1 Unit A2 Unit B1 Unit B2 Unit C Unit D

Unit weight (γ, kN/m3) (24.67–26.73) (20.80–24.67) (23.55–25.89) (18.95–24.73) (23.16–26.04)

a a a a a
25.78 22.97 24.62 22.73 24.73
Uniaxial compressive strength (σci, MPa) (3.10–59.92) (27.59–253.95) (12.46–94.64) (5.37–56.31) (25.83–238.26)
a a a a a
30.73 109.71 55.05 26.19 60.98
Point load strength index (IS(50), MPa) (0.15–4.28) (1.38–12.70) (0.62–4.73) (0.27–2.82) (1.29–11.91)
a a a a a
1.54 5.49 2.75 1.31 3.05
Young’s modulus (Ei, GPa) (13.27–29.50) (11.50–48.72) (15.00–34.75) (8.53–18.05) (17.23–55.08)
a a a a a
18.45 34.24 27.25 12.20 27.53
Rock quality designation (RQD, %) (59–98) (34–90) (36–94) (74–96) (0–8) (25–96)
a a a a a a
84 60 62 88 3* 73
Lugeon (Lu, l/min/m) (0.92–5.74) (0.38–6.72) (0.00–1.08) (0.32–.4.17) (0.00–0.32) (1.02–5.00)
a a a a a a
2.86 3.75 0.32 2.05 0.12 3.29
a
Average value
*The RQD value was selected as 10% in rock mass calculations according to Barton et al. (1974)

The RMR classification system was suggested by these systems was originated from the lack of a stress param-
Bieniawski (1974) and updated by Bieniawski (1989). This eter in the RMR system.
classification system utilizes the following five rock mass pa- The rock mass parameters are essential data for the numer-
rameters: strength of rock material, RQD, spacing of discon- ical modeling. To define the rock mass constants (mb, s, a), the
tinuities, condition of discontinuities and groundwater. In Hoek–Brown failure criterion was utilized. The RMR rating
order to determine the ratings for each five parameters, the can be used to predict the Geological Strength Index (GSI) for
charts for the RQD, intact rock strength and discontinuity RMR > 23. In this case, the modified Q value (Q´) should be
spacing and rating tables suggested by Bieniawski (1989) used. Therefore, the GSI values of the geotechnical units and
were used. The basic RMR values were determined by the fault zones were obtained using Eqs. (2) and (3) suggested by
addition of these ratings. In the next stage, adjustments for Hoek et al. (1995).
the discontinuity orientations considering the tunnel excava- 0
tion direction were applied to the basic RMR ratings. Based GSI ¼ 9lnQ þ 44 ð2Þ



on the both basic and adjusted RMR values, the character of 0 RQD Jr
ground throughout the Cankurtaran Tunnel route varies be- Q ¼ ð3Þ
Jn Ja
tween Bvery poor^ and Bfair^ (Table 4).
The Q classification system was proposed by Barton et al. Variation of the RQD, RMR, Q and GSI values along the
(1974) and it is determined by the following Eq. (1). tunnel route are depicted in Fig. 6.



 Hoek–Brown constants were determined using Eqs.
RQD Jr Jw
Q¼ ð1Þ (4)–(6) proposed by Hoek et al. (2002).
Jn Ja SRF
mb ¼ mi eð 28−14D Þ
GSI−100
ð4Þ
where Jn is joint set number, Jr is joint roughness number, Ja is
s ¼ eð Þ
GSI−100
joint alteration number, Jw is joint-water reduction factor and 9−3D ð5Þ
1 1  −GSI=15 −20=3 
SRF is stress reduction factor.
In terms of Q values, the characteristics of the rocks located a¼ þ e −e ð6Þ
2 6
throughout the tunnel alignment vary from Bextremely poor^
to Bvery poor^ (Table 4). where mi is constant of the rock material and D is disturbance
To be on the safe side, the rock mass quality values of the factor.
fault zones F1-F2 and fault zone F3 were assumed as half of Factor of disturbance (D) was taken into account to be
the rock mass quality values of the unit B2 and unit C, respec- 0.5 for poor quality blasting and zero for the mechanical
tively. According to rock mass classifications, the results ob- excavation. The mi values of the rock materials were iden-
tained from the Q system exhibited more conservative esti- tified using the RocData v5.0 (Rocscience Inc. 2016b)
mates than basic and adjusted RMR values owing to different software and were taken as 13, 7, 8, 8, 7 and 25 for unit
input parameters. The main difference between the results of A1, unit A2, unit B1, unit B2, unit C and unit D,
 Author's personal copy
Geotechnical studies and primary support design for a highway tunnel: a case study in Turkey

Unit A1 Unit A2

J1 J1
J2 J3 J2 J3
N77E N77E

B
J1 : 22/045
J1 : 22/045 J2 : 46/338
J2 : 46/338 J3 : 43/016
J3 : 43/016 B : 47/165
Equal Angle Equal Angle
Lower Hemisphere Lower Hemisphere
715 Poles 772 Poles

Unit B1 + Unit B2

F2
F1

J1
J2

Tunnel Direction
WE

J1 : 41/356
J2 : 44/030
B : 50/173
Equal Angle
Lower Hemisphere
597 Poles

Unit C Unit D

J2 J3

J1
J3
F3
N79W J1 J2 N79W
F3
B
J1 : 60/015
J2 : 57/289 J1 : 68/200
J3 : 37/020 J2 : 33/109
B : 47/170 J3 : 30/323
Equal Angle Equal Angle
Lower Hemisphere Lower Hemisphere
451 Poles 919 Poles

Fig. 5 Stereographic projection of the discontinuity sets in geotechnical units

 Author's personal copy
A. Kaya, F. Bulut

Table 2 Orientations of the

discontinuity sets Formation Geotechnical unit Discontinuity type Dip Dip direction

Subasi Ridge Formation Unit A1 Joint 22 045

Joint 46 338
Joint 43 016
Unit A2 Joint 22 045
Joint 46 338
Joint 43 016
Bedding 47 165
Cankurtaran Formation Unit B1 Joint 41 356
+ Joint 44 030
unit B2 Bedding 50 173
Senkaya Ridge Formation Unit C Joint 60 015
Joint 57 289
Joint 37 020
Bedding 47 170
Kabakoy Formation Unit D Joint 68 200
Joint 33 109
Joint 30 323

respectively (Table 5). It was assumed to be 7, 7 and 8 for where σci is uniaxial compressive strength of intact rock in
intact rocks in fault zones F1, F2 and F3, respectively. MPa and Ei is Young’s modulus in GPa.
The rock mass strength (σcm) and deformation modulus Because it is hard to determine the σci and Ei values of
(Em) were obtained using Eqs. (7) and (8) proposed by Hoek intact rocks in fault zones due to sampling difficulties, the
et al. (2002) and Hoek and Diederichs (2006). σcm and Em values were estimated using Eqs. (9) and (10) that
are the basis for the RMR proposed by Aydan et al. (1997).
ðmb þ 4s−aðmb −8sÞÞðmb =4 þ sÞa−1 These equations are more suitable to identify the elastic prop-
σcm ¼ σci : ðMPaÞ ð7Þ
2ð 1 þ a Þ ð 2 þ a Þ erties of weak rocks such as phyllite, mudstone, siltstone, salt,

 potash or weathered and/or sheared/faulted rocks at relatively
1−ðD=2Þ
Em ¼ Ei ðGPaÞ ð8Þ shallow depths (<400 m).
1 þ eð60þ15D−GSI Þ=11

Table 3 Quantitative description of the discontinuities in geotechnical units

Unit A1 Unit A2 Unit B1 Unit B2 Unit C Unit D

Joint Bedding Joint Bedding Joint Bedding Joint Bedding Joint Joint

Spacing (0.5–240) (1–110) (0.5–240) (1–45) (1–84) (1–64) (1–84) (0.5–13) (0.5–33) (0.5–119)
a a a a a a a a a a
(cm) 13.31 8.94 13.31 6.70 13.53 7.84 27.06 2.12 4.28 12.24
Persistence (0.2–35) (4–50) (0.2–35) (12–40) (1–34) (12–40) (1–34) (4–50) (0.2–35) (1–22)
a a a a a a a a a a
(m) 9.10 24.45 9.10 21.20 7.50 21.20 7.50 27.11 9.66 5.18
Aperture (0–24) (0–6.5) (0–24) (5–35) (0–5) (0–1) (0–5) (0–5) (0–4) (0–22)
a a a a a a a a a a
(mm) 1.92 0.99 1.92 12.39 0.89 0.19 0.89 1.05 1.1 2.67
Infilling <5 mm soft <5 mm <5 mm soft >5 mm soft <5 mm soft <5 mm soft <5 mm soft <5 mm soft <5 mm <5 mm soft
clay and calcite clay and clay and clay and clay and clay and clay soft clay clay
calcite calcite calcite calcite calcite calcite
Roughness Rough, Smooth, Rough, Smooth, Rough, Smooth, Rough, Rough, Rough, Rough,
undulating undulating undulating undulating undulating undulating undulating planar undulating undulating
Weathering Slightly Fresh Slightly Slightly Slightly Slightly Slightly Slightly Slightly Slightly
weathered weathered weathered weathered weathered weathered weathered weathered weathered
Groundwater Dripping Dripping Dripping Dripping Dripping Dripping Dripping Dripping Dripping Dripping
condition
a
Average
 Author's personal copy
Geotechnical studies and primary support design for a highway tunnel: a case study in Turkey

Table 4 RMR and Q classifications of the rock masses along the Cankurtaran Tunnel

Segment Chainage (km) Rock mass Basic RMR Adj. RMR Q´ Q

1 7 + 980–8 + 125 Entrance portal (unit A1) 46.4 41.7/Fair 2.6 0.26/Very poor
2 8 + 125–8 + 310 Unit A1 46.4 41.7/Fair 5.3 0.70/Very poor
3 8 + 310–9 + 225 Unit A2 41.0 33.8/Poor 2.0 0.27/Very poor
4 9 + 225–9 + 910 Unit A1 46.4 41.7/Fair 5.3 0.70/Very poor
5 9 + 910–10 + 660 Unit B1 34.1 27.4/Poor 1.0 0.14/Very poor
6 10 + 660–11 + 255 Unit B2 45.2 37.8/Poor 3.7 0.49/Very poor
7 11 + 245–11 + 255 F1 fault zone 22.6 15.8/Very poor 1.8 0.24/Very poor
8 11 + 255–11 + 645 Unit A1 46.4 38.9/Poor 5.3 0.70/Very poor
9 11 + 645–11 + 655 F2 fault zone 22.6 15.8/Very poor 1.8 0.24/Very poor
10 11 + 655–12 + 090 Unit B2 45.2 37.8/Poor 3.7 0.49/Very poor
11 12 + 090–12 + 320 Unit B1 34.1 27.4/Poor 1.0 0.14/Very poor
12 12 + 320–12 + 645 Unit C 31.7 26.7/Poor 0.5 0.05/Extremely poor
13 12 + 645–12 + 655 F3 fault zone 15.9 11.1/Very poor 0.3 0.025/Extremely poor
14 12 + 655–13 + 170 Unit D 46.9 42.2/Fair 4.6 0.61/Very poor
15 13 + 170–13 + 208 Exit portal (unit D) 46.9 42.2/Fair 2.3 0.23/Very poor

RMR Rock mass rating

σcm ¼ 0:0016RM R2:5 ðMPaÞ ð9Þ GSI r ¼ GSIe−0:0134GSI ð11Þ

E m ¼ 0:0097RMR =1000ðGPaÞ: ð10Þ mbr ¼ mi eð Þ
3:54 GSIr−100
28 ð12Þ

s r ¼ eð Þ
GSIr−100
The post-peak attitude of a rock mass is an essential input 9 ð13Þ
1 
in the numerical modeling of a tunnel, because it has an im-
1
portant influence on the excavation stability. Residual param- ar ¼ þ e−GSIr=15 −e−20=3 ð14Þ
2 6
eters of the rock mass are needed to design underground struc-
tures properly. Much investigation has been focused on the In order to determine the Poisson’s ratio (νm) of geo-
specification of peak parameters, and limited attempts have technical units dynamically, geophysical studies were
been made to predict the residual parameters of rocks. In this carried out using the seismic refraction method in six
study, because no field testing was performed to investigate lines (Fig. 2). With the help of the seismic studies,
the post-peak behavior of geotechnical units upon excavation, primary-wave (Vp) and shear-wave (Vs) velocities were
it was estimated by the help of the method suggested by Cai defined for each geotechnical unit cropped out along the
et al. (2007). The residual GSI (GSIr) value and residual tunnel alignment. Considering the elastic theory, the dy-
Hoek–Brown constants (mbr, sr, ar) were calculated using namic Poisson’s ratio was determined using the seismic
Eqs. (11)–(14). velocity values (Eq. 15).

100 Overburden 100.0

RQD
Adj. RMR
Adj. RMR, GSI, RQD

80 GSI
Q
10.0
60
Q

40
1.0

20

0 0.1
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
DISTANCE (m)

Fig. 6 Rock mass characterization along the Cankurtaran Tunnel

Table 5 Calculated rock mass parameters of the geotechnical units and fault zones for numerical analyses

Entrance portal Unit A1 Unit A2 Unit A1 Unit B1 Unit B2 F1 fault Unit A1 F2 fault Unit B2 Unit B1 Unit C F3 fault Unit D Exit portal
(unit A1) zone zone zone (unit D)

Segment 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Elastic type Isotropic Isotropic Isotropic Isotropic Isotropic Isotropic Isotropic Isotropic Isotropic Isotropic Isotropic Isotropic Isotropic Isotropic Isotropic
Rock mass strength, 4.80 5.70 14.11 5.70 5.37 7.43 3.89 5.70 3.89 7.43 5.37 1.97 1.60 14.93 12.68
σcm (MPa)
Deformation modulus, Em (GPa) 2.89 4.39 4.49 4.39 2.27 5.26 0.60 4.39 0.60 5.26 2.27 0.60 0.17 6.05 3.94
Poisson’s ratio, υm 0.23 0.23 0.4 0.23 0.46 0.26 0.34 0.23 0.38 0.26 0.46 0.31 0.42 0.39 0.39
Material type Plastic Plastic Plastic Plastic Plastic Plastic Plastic Plastic Plastic Plastic Plastic Plastic Plastic Plastic Plastic
mi constant 13 13 7 13 8 8 8 13 8 8 8 7 7 25 25
GSI 53 59 50 59 44 56 49 59 49 56 44 38 32 58 51
mb constant 1.37 1.84 0.94 1.84 0.56 0.97 1.32 1.84 1.32 0.97 0.56 0.36 0.61 3.33 2.47
s constant 0.0018 0.0042 0.0013 0.0042 0.0006 0.0027 0.00364 0.0042 0.0036 0.0027 0.0006 0.0002 0.000496 0.0035 0.0015
a constant 0.505 0.503 0.506 0.503 0.508 0.504 0.506 0.503 0.506 0.504 0.508 0.513 0.520 0.503 0.505
GSIr 26 27 26 27 24 26 25 27 25 26 24 23 21 27 26
mbr residual constant 0.93 0.95 0.70 0.95 0.54 0.58 0.56 0.95 0.56 0.58 0.54 0.44 0.41 1.82 1.77
sr residual constant 0.00027 0.00029 0.00026 0.00029 0.00023 0.00028 0.00025 0.00029 0.00025 0.00028 0.00023 0.00019 0.000148 0.00029 0.00026
ar residual constant 0.529 0.528 0.530 0.528 0.532 0.528 0.530 0.528 0.530 0.528 0.532 0.536 0.542 0.528 0.530
Author's personal copy

Disturbance factor (D) 0.5 0.5 0.5 0.5 0.5 0.5 0 0.5 0 0.5 0.5 0.5 0 0.5 0.5
Dilation parameter 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Overburden (m) 40 55 200 360 490 420 250 300 260 200 150 70 85 125 45
Vertical stress, 1.03 1.42 4.59 9.28 12.06 10.34 5.00* 7.73 5.20* 4.92 3.69 1.59 1.70* 3.09 1.11
σv (MPa)
Horizontal stress, σh (MPa) 0.31 0.42 3.06 2.77 10.28 3.64 5.00 2.31 5.20 1.73 3.15 0.72 1.70 1.98 0.71

*Unit weight was selected arbitrarily as 20 kN/m3 for fault zone material
GSI Geological Strength Index
A. Kaya, F. Bulut
 Author's personal copy
Geotechnical studies and primary support design for a highway tunnel: a case study in Turkey

V P 2 −2V S 2 systems (final lining) recommended by Q classification

νm ¼  2  ð15Þ
2 V P −V S 2 (Barton, 2002) were implemented for the Cankurtaran
Tunnel. Based on the Q system values, four support categories
where Vp is primary-wave velocity in m/s and Vs is were determined for 15 segments (Table 6). The planned con-
shear-wave velocity in m/s. struction properties, including round length, construction
Because the seismic studies were not feasible in the fault phase, support time and stand-up time, determined consider-
zones due to the steep morphology, the Poisson’s ratio values ing the adjusted RMR ratings are also given in Table 6.
of the fault zones were estimated using the empirical equation To clutch the process of tunnel lining design, it is important
(Eq. 16) that is the basis for the rock mass strength suggested to investigate basic notions of how a rock mass around an
by Aydan et al. (1993). The given equation gives the best opening deforms and how the supports check this deforma-
solution for weak, weathered and/or faulted rocks at shallow tion. Therefore, the convergence–confinement (CC) method
depths (<400 m). was utilized to determine the interaction among the surround-
  ing rock mass and lining support. This methodology, which
ν m ¼ 0:25 1 þ e−σcm =4 ð16Þ satisfies the Hoek–Brown criterion, has been delineated by
Carranza-Torres and Fairhurst (1999) for rock masses. The
where σcm is rock mass strength in MPa. CC method authorizes the load charged on a lining applied
The rock mass parameters of the geotechnical units and behind the tunnel face to be predicted. In this method, uniform
fault zones are presented in Table 5. These parameters were internal pressure pi and far-field stress σo are taken into ac-
later used in the FEM analyses as input parameters. count. The Pi and So are described as (Carranza-Torres and
The vertical stress (σv) acting at a point below the ground Fairhurst 2000);
surface is due to the weight of overburden. The loading con-
pi s
dition for vertical stress was calculated by: Pi ¼ þ 2 ðMPaÞ ð19Þ
mb σci mb
σv ¼ γH ðMPaÞ ð17Þ σo s
So ¼ þ 2 ðMPaÞ ð20Þ
mb σci mb
where γ is rock mass unit weight in MN/m3 and H is overbur-
den depth in meters. where mb and s are the Hoek–Brown constants, σci is the intact
The horizontal stresses (σh) acting on a rock element at a rock strength in MPa, pi is the support pressure (uniform in-
depth below the surface are harder to predict than the vertical ternal) in MPa, σo is the far-field stress (σo = (σv + σh)/2) in
stresses. The horizontal stress in fractured rock mass was ob- MPa, Pi is scaled internal pressure in MPa and So is scaled far-
tained from the following Eq. (18) proposed by Sheorey et al. field stress in MPa.
(2001). On the other hand, the horizontal stress in the fault The scaled critical internal pressure Picr representing the
zones was assumed to be equal to the vertical stress. The rock elastic limit and the actual critical internal pressure picr are
mass parameters considered in the analysis models are given defined as (Carranza-Torres and Fairhurst 2000);
in Table 5.
1 h pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffii2
Pi cr ¼ 1− 1 þ 16S o ðMPaÞ ð21Þ
νm βE m G 16
σh ¼ σv þ ðH þ 1000ÞðMPaÞ ð18Þ
1−ν m 1−ν m s
pi cr ¼ Pi cr − 2 mb σci ðMPaÞ ð22Þ
where 훽= 8 × 10−6/oC (coefficient of linear thermal expan- mb
sion), G = 0.024 °C/m (geothermal gradient), υm is the
If the pi is less than the picr, failure is expected to occur. On
Poisson’s ratio of rock mass and Em is deformation modulus
the other hand, if the pi is greater than this picr, no failure will
of rock mass in GPa. occur.
In this study, the CC analyses were carried out in two
Tunnel lining design stages. In the first stage, the pi was supposed to be 0 for the
unsupported state. The essential rock characteristics were re-
Bieniawski (1989) published a set of guidelines for the selec- placed due to the rock units and they were utilized as data. The
tion of support in tunnels in rock for which the RMR value has computed values of each segment are presented in Table 7.
been defined. The support chart is applied only for a tunnel When pi and picr values are compared for unsupported tunnel
with a 10-m horseshoe-shaped span, excavated using drill- segments, apart from segments 2 and 15, the stability problem
blast technique and in situ stress <25 MPa. However, in the is expected, and the deformation type is plastic.
Q system, there are no specific limitations such as span of In the second stage, the psmax (maximum pressure) and the
underground openings and excavation shape. Therefore, be- Ks (elastic stiffness) of the proposed lining supports were de-
cause of restrictions of the RMR system, the empirical support fined for each segment. In calculations, the methods suggested
 Author's personal copy
A. Kaya, F. Bulut

Table 6 Empirical tunnel support categories (final lining) for the rock masses along the Cankurtaran Tunnel

Support type 0.4 < Q < 1 /Poor rock 0.1 < Q < 0.4 /Poor rock 0.04 < Q < 0.1 /Extremely poor 0.01 < Q < 0.04 /Extremely poor

Segment 2, 4, 6, 8, 10, 14 1, 3, 5, 7, 9, 11, 15 12 13

Construction phase Top heading and bench Top heading and bench Top heading, bench and invert Top heading, bench and invert
Excavation method Drill and blast Drill and blast Mechanical excavation Mechanical excavation
Round length 1.5–3.0-m advance in top 1.0–1.5-m advance in top 0.5–1.0-m advance in top 0.5–1.0-m advance in top
heading and bench heading and bench heading, bench and invert heading, bench and invert
Stand-up time Immediate collapse Immediate collapse Immediate collapse Immediate collapse
Support time Install support concurrently Install support concurrently Install support concurrently with Install support concurrently with
with excavation with excavation excavation excavation
Support requirements
Roof Systematic rock bolts 4 m Systematic rock bolts 4 m Systematic rock bolts 4 m long, Systematic rock bolts 4 m long,
long, spaced 1.5–1.7 m long, spaced 1.3–1.5 m spaced 1.2–1.3 m and spaced 1.0–1.2 m and
and 90–120-mm-thick and 120–150-mm-thick 150–250-mm-thick steel 150–250-mm-thick steel
steel fiber/wire steel fiber/wire fiber/wire mesh-reinforced fiber/wire mesh-reinforced
mesh-reinforced mesh-reinforced shotcrete (Sfr + B) and steel shotcrete (Sfr + B) and steel
shotcrete (Sfr + B) shotcrete (Sfr + B) ribs (RRS) ribs (RRS)
Wall Systematic rock bolts 3.5 m Systematic rock bolts 3.5 m Systematic rock bolts 3.5 m Systematic rock bolts 3.5 m
long, spaced 1.7–2.1 m long, spaced 1.5–1.7 m long, spaced 1.2–1.3 m and long, spaced 1.0–1.2 m and
and 50–90-mm-thick and 90–120-mm-thick 150–250-mm-thick steel 150–250-mm-thick steel
steel fiber/wire steel fiber/wire fiber/wire mesh-reinforced fiber/wire mesh-reinforced
mesh-reinforced mesh-reinforced shotcrete (Sfr + B) and steel shotcrete (Sfr + B) and steel
shotcrete (Sfr + B) shotcrete (Sfr + B) ribs (RRS) ribs (RRS)
Invert – – 150–250-mm-thick steel 150–250-mm-thick steel
fiber/wire mesh-reinforced fiber/wire mesh-reinforced
shotcrete (Sfr + B) and steel shotcrete (Sfr + B) and steel
ribs (RRS) ribs (RRS)

by Hoek (2007) and Carranza-Torres and Engen (2017) were FEM software RS2 v9.0 and 3D FEM software RS3 v1.0
used. According to the acquired results, the psmax values were (Rocscience Inc. 2016c, 2017) were used in the numerical
higher than the picr values for segments 1, 2, 3, 10, 12, 14, and analyses. Unlike the RS3, RS2 uses a plane strain analysis
15. However, the psmax values were smaller than the picr values where two principal in situ stresses are in the plane of the
for other segments. As a result, considering the CC analysis, excavation, and the third principal stress is out of plane. In
the empirical lining design was not satisfactory for these this model, the 3D stress tensor is divided into three orthogo-
segments. nal stresses which are aligned with the 2D model of excava-
Numerical analyses have been progressively used today tion (2 in plane and 1 out of plane), and they decompose. An
compared to the past in tunnel designs to check the accuracy automatic mesh around the excavation was formed, and con-
of results acquired from empirical methods owing to the quick sidering the elasto-plastic analysis, deformations and stresses
developments in computer systems. Because numerical tech- were calculated using these software programs. A very simple
niques consider the geometry of tunnels, field stresses and model was used to analyze the tunnel stability and to identify
elastic/strength parameters, they are valuable design instru- the concept of rock support interaction. The six-noded trian-
ments in tunneling. The finite element method (FEM) has gular elements were applied to optimize the meshes, and sen-
become one of the most chosen methods in underground ap- sitive zoning was used around the excavation. The boundary
plications by many researchers (i.e. Ozsan and Karpuz 2001; of excavation was generated based on its height and span in
Javadi and Snee 2002; Kockar and Akgun 2003; Ozsan and three steps as (i) top heading, (ii) bench and (iii) invert for the
Basarir 2003; Park 2004; Sari and Pasamehmetoglu 2004; Bunit C^ and fault zone F3 (segments 12 and 13), and (i) top
Basarir et al. 2005; Basarir 2006; Genis et al. 2007; Gurocak heading and (ii) bench for the other segments (Fig. 7). The
et al. 2007; Sopaci and Akgun 2008; Kaya et al. 2011; Aydin external boundary of design was applied as ten times the ex-
et al. 2014; Kanik et al. 2015; Yalcin et al. 2015; Kaya and cavation width. The Hoek–Brown failure criterion was ap-
Sayin 2017; Lin et al. 2017; Ozdogan et al. 2018; Kanik and plied to define the yielded elements and plastic zones in close
Gurocak 2018). vicinity to the excavation. In general, deformation of the rock
In the present study, to define the displacements and plastic mass reaches its maximum rate between one and two circular
zones around the opening and to check the validity of the tunnel radii behind the face (Hoek 2007). However, noncircu-
suggested tunnel lining design given in Table 6, the 2D lar tunnel cross sections are widespread in application.
Table 7 Results of the convergence–confinement analyses for the Cankurtaran Tunnel

Entrance Unit Unit Unit Unit Unit F1 fault Unit F2 fault Unit Unit Unit F3 fault Unit Exit
portal A1 A2 A1 B1 B2 zone A1 zone B2 B1 C zone D portal
(unit A1) (unit D)

Segment 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Uniform internal pressure, pi (MPa) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Far-field stress, σo (MPa) 0.67 0.92 3.83 6.03 11.17 6.99 5.00 5.02 5.20 3.33 3.42 1.15 1.70 2.53 0.91
Scaled far-field stress, So (MPa) 0.02 0.02 0.04 0.11 0.36 0.13 0.14 0.09 0.15 0.07 0.11 0.12 0.22 0.01 0.01
Equivalent diameter, D (m) 10.76 10.76 10.76 10.76 10.76 10.76 10.76 10.76 10.76 10.76 10.76 10.76 11.22 11.22 10.76
Equivalent radius, R (m) 5.38 5.38 5.38 5.38 5.38 5.38 5.38 5.38 5.38 5.38 5.38 5.38 5.61 5.61 5.38
Scaled critical internal pressure, Picr (MPa) 0.00101 0.00108 0.00467 0.02651 0.16113 0.03725 0.04008 0.01979 0.04256 0.01153 0.02816 0.03296 0.07678 0.00060 0.00015
Actual critical internal pressure, picr (MPa) 0.002 0.0 0.325 1.428 4.942 1.835 1.375 1.048 1.465 0.461 0.816 0.294 0.599 0.056 0.0
Maximum support pressure provided by 0.55 0.44 0.55 0.44 0.55 0.44 0.55 0.44 0.55 0.44 0.55 0.87 0.87 0.44 0.55
shotcrete, pscmax (MPa)
Elastic stiffness of shotcrete, Ksc (MPa/m) 0.11 0.09 0.11 0.09 0.11 0.09 0.11 0.09 0.11 0.09 0.11 0.17 0.17 0.09 0.11
Maximum support pressure provided by rock 0.148 0.111 0.148 0.111 0.148 0.111 0.148 0.111 0.148 0.111 0.148 0.250 0.250 0.111 0.148
bolt, psbmax (MPa)
Geotechnical studies and primary support design for a highway tunnel: a case study in Turkey

Elastic stiffness of rock bolt, Ksbmax (MPa/m) 0.02 0.01 0.02 0.01 0.02 0.01 0.02 0.01 0.02 0.01 0.02 0.03 0.03 0.01 0.02
Maximum support pressure provided by steel – – – – – – – – – – – 0.39 0.39 – –
Author's personal copy

set, pssmax (MPa)

Elastic stiffness of steel set, Kss (MPa/m) – – – – – – – – – – – 34.43 34.43 – –
Maximum support pressure provided by 0.13 0.10 0.13 0.10 0.13 0.10 0.13 0.10 0.13 0.10 0.13 0.59 0.59 0.10 0.13
combined support, psmax (MPa)
Elastic stiffness of combined support, Ks 4.93 4.93 4.93 4.93 4.93 4.93 4.93 4.93 4.93 4.93 4.93 0.01 0.01 4.93 4.93
(MPa/m)
Maximum elastic deformation of the combined 0.64 0.51 0.64 0.51 0.64 0.51 0.64 0.51 0.64 0.51 0.64 0.01 0.01 0.51 0.64
support, urmax (m)
 Author's personal copy
A. Kaya, F. Bulut

Fig. 7 Tunnel excavation

sections for segments (a) 1–11 a
and 14–15, and (b) 12–13
Concrete
lining
Ex
bo cava
un tio
da n
ry

tcr ced
te
sho nfor
i
Re
Roof bolt

TOP HEADING
6m

12 m
Wall bolt

BENCH
3m

INVERT
1.5 m

Invert fi
lling lt
rt bo
Inve

Concrete
lining
Ex
bo cava
un tio
da n
ry
tcr ced
te
shonfor
i
Re

Roof bolt

TOP HEADING
6m

12 m
Wall bolt

BENCH
3m

Therefore, based on the method proposed by Curran et al. (discontinuum or continuum models). The fundamental sig-
(2003) for noncircular cross sections, the slice thickness was nificance in discontinuum models is the acting of discontinu-
taken as 10 m in the 3D models. ity behavior. However, joints in continuum models are shown
All rock masses contain joints. Representation of these implicitly, with the intention that the behavior of the continu-
joints in numerical models differs based on the model type um design is substantially equivalent to the real jointed rock
 Author's personal copy
Geotechnical studies and primary support design for a highway tunnel: a case study in Turkey

mass being represented. Continuum models assume material segments 5, 6 and 8 due to the great field stresses. However,
is continuous throughout the body. Joints are treated as special considering the 3D models, few stability problems are expect-
cases by introducing interfaces among continuum bodies ed in all segments. Thus, the lining application (rock bolts and
(Wyllie and Mah 2004). Because the discontinuity spacing shotcrete; sometimes steel ribs) is recommended for the secu-
values of the geotechnical units varied between close and very rity of rock masses whose quality ranges between extremely
close (Table 3), the rock masses surrounding the excavation poor and very poor along the Cankurtaran Tunnel route.
are well-suited for use in the continuum model. Therefore, in On the other hand, it should be remembered that RS2
numerical analyses, continuum models were used to represent and RS3 are small-strain software programs and, thus, they
rock mass behavior in the 15 segments. In this study, numer- cannot provide the very large strains. Therefore, it is more
ical modeling consisted of three stages. appropriate to take into account the size of the failure area
In the first and second stages, deformations (Figs. 9, 10, 11, rather than the deformation size. It can be seen from
12 and 13) and plastic zones (Figs. 14, 15, 16, 17 and 18) that Figs. 14, 15, 16, 17 and 18 that the extent of the plastic
occurred around the openings were examined using top head- zones demonstrates that there would be a serious matter in
ing followed by excavating the entire tunnel (bench/invert). In terms of stability, especially in segments 4–13, if they are
the final stage, the effectiveness of the lining support (i.e., not supported. When Figs. 15 and 16 and Table 8 are
shotcrete, rock bolting and steel set) was analyzed. checked, the most problematical sections along the tunnel
According to the 2D models for unsupported cases, the route are segments 4, 5, 6 and 8 driven in the units A1, B1,
maximum value of total displacement for all segments ranges B2 and A1, respectively. A larger plastic zone and maxi-
between 0.23 and 17.71 cm (Table 8, Fig. 8). On the other mum total displacement were developed in these segments.
hand, the maximum total displacements are very small, and However, the extents of plastic zones in the fault zones
range between 0.19 and 3.42 cm when considering the 3D (F1, F2 and F3) are nearly half of these segments and are
analyses (Table 8, Fig. 8). It was seen that higher deformations 6.29, 6.14 and 6.43 m, respectively (Table 8). In general,
occurred in the 2D models of segments 5, 8 and 12 compared the maximum plastic zone concentrations occurred at the
to the 3D models (Fig. 8). Moreover, there is a discrepancy tunnel roof due to the raised field stresses. Considering the
among the convergence system of the 2D and 3D analysis 2D designs, the dimensions of the plastic zone for seg-
results for segment 4. An exaggerated displacement has de- ments varied between 1.22 and 14.21 m. On the other
veloped at the external walls considering the 2D model of hand, since there is no a ruler option in RS3 v1.0, the failed
segment 4. However, no wall deformations were detected in areas in the 3D modeling could not be measured. In con-
the 3D models of segment 4 (Fig. 10). This is a quirky finding trast, when the plastic zones occurred around the opening,
and an unforeseen outcome of the 2D model since small de- given in Figs. 14, 15, 16, 17 and 18, are compared, it can
formations at the tunnel walls possibly occur under low field be understood that the dimension of the failure areas in the
stresses. In general, the deformation concentrations in the 2D 3D designs are partially larger than that of the 2D models,
designs are larger than those of the 3D design. The justifica- particularly in segments 5, 12, 13 and 15. This result is
tion in these variations may be that in the 2D design, the tunnel primarily owing to the lost arching activity in the 2D de-
opening upright to the plane is infinitive. In addition to this, in signs. It is believed that in numerical modelling, these dis-
the analogous direction, the excavation round is merely 10 m tinctions are reasoned by a finite discretization.
in the 3D designs. Figure 8 shows the maximum total dis- In the last step, the efficiency of the suggested lining sup-
placement (Ut), normalized vertical displacement (Ux/span) port was examined utilizing the same 2D and 3D analysis
and normalized horizontal displacement (Uy/tunnel height) models. The lining designs were the same as those suggested
as a function of vertical distance. Each of the 15 plots corre- in Table 6, and their properties implemented in FEM models
sponds to a different value of displacement. The plots clearly are given in Table 9. Variances in the deformations and dimen-
showed that an increase in field stress leads to higher displace- sions of the plastic zones following the support installations
ment in segments 5, 8 and 12. were analyzed, and outcomes were cross-checked with the
Strain is described as the ratio of tunnel closure to span. unsupported states. The load split option in RS2 v9.0 software
Hoek and Marinos (2000) proposed that for rock masses with authorizes the user to Bsplit^ the field stress-induced load,
strains less than 1%, few matters in terms of stability can between any stages of the model, rather than applying the
possibly occur. On the other hand, minor squeezing problems entire field stress load in the first stage. Therefore, the load
are expected with strain values between 1 and 2.5%. In the splitting option is used to liken the retarded support installa-
present study, according to the 2D models, the strain values of tion. However, RS3 v1.0 software does not have a load split
15 segments ranged between 0.05 and 2.22% (Table 8). option. Therefore, in order to compare the 2D and 3D analyses
However, the strain values obtained from 3D models varied results realistically, it was implicitly assumed that the lining is
between 0.03 and 0.57%, as shown in Table 8. According to installed instantly after excavation, and that no deformation
2D models, minor squeezing problems are expected in occurs before the support installation.
Table 8 Results of the 2D and 3D FEM analyses for the Cankurtaran Tunnel

Entrance portal Unit Unit Unit Unit Unit F1 fault Unit F2 fault Unit Unit Unit F3 fault Unit Exit portal
(unit A1) A1 A2 A1 B1 B2 zone A1 zone B2 B1 C zone D (unit D)

Segments 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
2D-strain (ɛ, %) 0.05 0.06 0.21 0.97 2.95 2.17 0.24 1.72 0.26 0.53 0.38 0.76 0.16 0.08 0.04
2D unsupported total displacement (Ut, cm) 0.31 0.38 1.28 5.80 17.71 12.99 1.46 10.30 1.54 3.19 2.28 4.55 0.94 0.50 0.23
2D unsupported vertical displacement (Ux, cm) 0.30 0.37 1.25 4.66 14.58 12.84 1.46 9.69 1.54 3.17 2.27 4.40 0.84 0.61 0.22
2D unsupported horizontal displacement (Uy, cm) 0.31 0.35 1.08 4.88 15.43 6.62 1.28 6.13 1.36 1.75 2.07 3.81 0.89 0.50 0.21
2D-supported total displacement (Ut, cm) 0.30 0.36 1.08 4.10 13.30 8.81 1.28 8.55 1.37 3.11 1.94 1.00 0.32 0.50 0.22
2D-supported vertical displacement (Ux, cm) 0.29 0.36 1.07 3.97 13.15 8.71 1.28 8.46 1.36 3.11 1.94 0.36 0.32 0.61 0.22
2D supported horizontal displacement (Uy, cm) 0.28 0.26 0.76 3.12 11.95 4.01 0.57 3.43 0.64 1.08 1.36 0.99 0.23 0.50 0.19
2D unsupported plastic zone (Rpl, m) 1.83 1.87 2.78 13.83 14.21 11.75 6.29 13.97 6.14 5.69 6.08 7.05 6.43 2.57 1.22
2D supported plastic zone (Rpl, m) 1.17 0.81 0.58 4.90 6.61 4.79 3.09 3.21 3.41 2.70 1.97 0.00 1.39 0.58 0.49
3D strain (ɛ, %) 0.04 0.05 0.11 0.57 0.53 0.49 0.10 0.44 0.10 0.21 0.20 0.26 0.05 0.06 0.03
3D unsupported total displacement (Ut, cm) 0.24 0.30 0.68 3.42 3.20 2.94 0.57 2.61 0.59 1.25 1.17 1.55 0.30 0.36 0.19
0.17 0.20 0.53 2.38 2.73 2.16 0.42 1.89 0.42 0.80 0.92 1.39 0.27 0.28 0.15
Author's personal copy

3D unsupported vertical displacement (Ux, cm)

3D unsupported horizontal displacement (Uy, cm) 0.23 0.29 0.68 3.39 3.15 2.81 0.57 2.59 0.59 1.19 1.06 1.55 0.28 0.36 0.19
3D supported total displacement (Ut, cm) 0.14 0.15 0.40 1.84 3.03 1.76 0.30 1.37 0.29 0.56 0.62 0.56 0.15 0.29 0.10
3D supported vertical displacement (Ux, cm) 0.14 0.15 0.36 1.32 2.03 1.18 0.21 1.05 0.21 0.47 0.46 0.55 0.12 0.20 0.10
3D supported horizontal displacement (Uy, cm) 0.11 0.13 0.40 1.84 3.01 1.76 0.30 1.37 0.29 0.56 0.62 0.28 0.15 0.29 0.10
A. Kaya, F. Bulut
 Author's personal copy
Geotechnical studies and primary support design for a highway tunnel: a case study in Turkey

20.00
17.71
2D unsupported case
3D unsupported case
15.00
13.30 12.99 2D supported case
3D supported case
Ut (cm)

10.30
8.81
10.00
8.55

5.80
5.00 2.28 4.55
1.28 4.10 1.46 3.19 1.94
1.08 3.20 1.54 3.11 1.17 1.55 0.50
0.38 2.94 1.28 2.61 1.37 0.94
0.31 0.36 0.68 3.42 0.57 1.25 0.62 1.00 0.32 0.50 0.23
0.30 3.03 0.59 0.56 0.56 0.30 0.36 0.22
0.30 0.40
0.24 0.15 1.84 1.76 0.30 0.29 0.15 0.29 0.19
0.14 1.37 0.10

Segment 1 Segment 2 Segment 3 Segment 4 Segment 5 Segment 6 Segment 7 Segment 8 Segment 9 Segment 10Segment 11Segment 12Segment 13Segment 14Segment 15

0.012
Ux / Span

0.008

0.004

Segment 1 Segment 2 Segment 3 Segment 4 Segment 5 Segment 6 Segment 7 Segment 8 Segment 9 Segment 10Segment 11Segment 12Segment 13Segment 14Segment 15
Uy / Tunnel Height

0.016

0.012

0.008

0.004

0
Segment 1 Segment 2 Segment 3 Segment 4 Segment 5 Segment 6 Segment 7 Segment 8 Segment 9 Segment 10 Segment 11 Segment 12 Segment 13 Segment 14 Segment 15

Fig. 8 Graphs showing the maximum total displacement (Ut), vertical displacement (Ux)/span and horizontal displacement (Uy)/tunnel height variations
along the tunnel line for unsupported and supported cases according to 2D and 3D numerical analyses

The maximum total displacements for all supported seg- segments 6–10 compared to the 3D models (Figs. 10, 11 and
ments range between 0.22 and 13.30 cm, and 0.10 and 12). However, in the 2D models of segments 1, 4, 12 and 15,
1.84 cm with respect to 2D and 3D analyses, respectively. an overdone deformation condensation demonstrating a fail-
The size of the displacements was indistinctly decreased after ure formed at the external walls. In contrast to the 2D FEM
support installation (Fig. 8, Table 8). It was concluded that no models, the reasonable deformations in the 3D analysis results
crown and wall displacements occurred in the 2D models of of all segments occurred inside the reinforced area, and these

Table 9 Characteristics of the support units used in the numerical analyses (TGDH 2013)

Properties Shotcrete Wire mesh Rock bolt Steel set

Young’s modulus (E, GPa) 20 200 200 200

Poisson’s ratio (υ) 0.2 0.35 – 0.35
Peak uniaxial compressive strength (σcp, MPa) 20 400 – 400
Residual uniaxial compressive strength (σcr, MPa) 3.5 – – –
Peak tensile strength (σtp, MPa) 3.1 500 – 500
Residual tensile strength (σtr, MPa) 0 – – –
Peak load (MN) – – 0.25 –
Residual load (MN) – – 0.025 –
Type – Ø6.5 mm, Ø28 mm I-beam
150 × 150 mm fully bonded 160 cm × 17.9 kg/m
 Author's personal copy
A. Kaya, F. Bulut

SEGMENT 1 SEGMENT 2 SEGMENT 3

2D MODEL (UNIT A1 / ENTRANCE PORTAL) (UNIT A1) (UNIT A2)
UNSUPPORTED CASE

3D MODEL
2D MODEL
SUPPORTED CASE

3D MODEL

Fig. 9 2D and 3D numerical analyses showing the maximum total displacement developed in segments 1, 2 and 3 for the unsupported and supported
cases

results verify the stability of the tunnel (Figs. 9, 10, 11, 12 and models shown in Figs. 14, 15, 16, 17 and 18, the plastic zones
13). hardly ever formed. The empirical lining applications have
Furthermore, checking against an unsupported state, the been efficient in decreasing the plastic zone around the most
size of the plastic zone has been decreased remarkably by problematic segments 7, 9 and 13 driven in fault zones
implementation of the rock bolts and shotcrete for all tunnel (Figs. 16 and 18). The extremely poor rock mass quality of
segments (Figs. 14, 15, 16, 17 and 18). Following the appli- segments 12 and 13 was supported by the heaviest lining,
cation of lining, the size of the plastic zone determined in the including steel ribs. According to 3D analysis results of seg-
2D designs of all segments reduced from 1.22–14.21 m to ments 4, 6 and 8, there are still plastic zones at the crown level,
0.00–6.61 m. However, as it may be observed from the 3D as in the 2D model (Figs. 15 and 16). Considering the 3D

SEGMENT 4 SEGMENT 5 SEGMENT 6

(UNIT A1) (UNIT B1) (UNIT B2)
UNSUPPORTED CASE

2D MODEL
3D MODEL
2D MODEL
SUPPORTED CASE

3D MODEL

Fig. 10 2D and 3D numerical analyses showing the maximum total displacement developed in segments 4, 5 and 6 for the unsupported and supported
cases
 Author's personal copy
Geotechnical studies and primary support design for a highway tunnel: a case study in Turkey

SEGMENT 7 SEGMENT 8 SEGMENT 9

(F1 FAULT ZONE) (UNIT A1) (F2 FAULT ZONE)
2D MODEL
UNSUPPORTED CASE

3D MODEL
2D MODEL
SUPPORTED CASE

3D MODEL

Fig. 11 2D and 3D numerical analyses showing the maximum total displacement developed in segments 7, 8 and 9 for the unsupported and supported
cases

models, it was seen that the failure areas are in the zone of the elasto-plastic tunnel analysis underestimates the deformations.
assembled bolts for all tunnel parts except these segments. On the other hand, it was determined that the 3D elasto-plastic
Therefore, it was suggested to use rock bolts longer than analysis yields results which compare reasonably well with
4 m at the level of the crown in segments 4, 6 and 8. the in situ measurements. Ucer (2006) noted that the deforma-
The 2D and 3D numerical analyses have become the most tion analyses results of the 3D analyses in tunneling are in
chosen ones in underground applications by many good agreement with the site data compared to the 2D analy-
researchers. Dhawan et al. (2002) revealed that the 2D ses. Trinh et al. (2010) indicated that the 3D model is a much

SEGMENT 10 SEGMENT 11 SEGMENT 12

(UNIT B2) (UNIT B1) (UNIT C)
2D MODEL
UNSUPPORTED CASE

3D MODEL
2D MODEL
SUPPORTED CASE

3D MODEL

Fig. 12 2D and 3D numerical analyses showing the maximum total displacement developed in segments 10, 11 and 12 for the unsupported and
supported cases
 Author's personal copy
A. Kaya, F. Bulut

SEGMENT 13 SEGMENT 14 SEGMENT 15

(F3 FAULT ZONE) (UNIT D) (UNIT D / EXIT PORTAL)
2D MODEL
UNSUPPORTED CASE

3D MODEL
2D MODEL
SUPPORTED CASE

3D MODEL

Fig. 13 2D and 3D numerical analyses showing the maximum total displacement developed in segments 13, 14 and 15 for the unsupported and
supported cases

more powerful tool than 2D model in stress and strain practical problems. Kaya and Sayin (2017) reported that the
estimations. Xu et al. (2015) noted that the precision of the 3D FEM model presents the best remedy in lining pattern
3D design is larger than that of the 2D prediction models. compared to 2D models. It is clearly seen that the findings
Furthermore, it was indicated that the 3D estimation models of the present study exactly coincide with the results of re-
are required for absolute prediction owing to their higher search mentioned above. According to results of the 2D
accuracy and applicability in a wider range of complex models under great in situ rock stresses, the over-estimated

SEGMENT 1 SEGMENT 2 SEGMENT 3

(UNIT A1 / ENTRANCE PORTAL) (UNIT A1) (UNIT A2)
2D MODEL

Plastic zone
UNSUPPORTED CASE

Plastic zone
3D MODEL
2D MODEL
SUPPORTED CASE

3D MODEL

Fig. 14 2D and 3D numerical analyses showing the plastic zones developed in segments 1, 2 and 3 for the unsupported and supported cases
 Author's personal copy
Geotechnical studies and primary support design for a highway tunnel: a case study in Turkey

SEGMENT 4 SEGMENT 5 SEGMENT 6

(UNIT A1) (UNIT B1) (UNIT B2)
Plastic zone
2D MODEL
UNSUPPORTED CASE

Plastic zone
3D MODEL
2D MODEL
SUPPORTED CASE

3D MODEL

Fig. 15 2D and 3D numerical analyses showing the plastic zones developed in segments 4, 5 and 6 for the unsupported and supported cases

SEGMENT 7 SEGMENT 8 SEGMENT 9

(F1 FAULT ZONE) (UNIT A1) (F2 FAULT ZONE)
2D MODEL

Plastic zone
UNSUPPORTED CASE

Plastic zone
3D MODEL
2D MODEL
SUPPORTED CASE

3D MODEL

Fig. 16 2D and 3D numerical analyses showing the plastic zones developed in segments 7, 8 and 9 for the unsupported and supported cases
 Author's personal copy
A. Kaya, F. Bulut

SEGMENT 10 SEGMENT 11 SEGMENT 12

(UNIT B2) (UNIT B1) (UNIT C)
2D MODEL

Plastic zone
UNSUPPORTED CASE

Plastic zone
3D MODEL
2D MODEL
SUPPORTED CASE

3D MODEL

Fig. 17 2D and 3D numerical analyses showing the plastic zones developed in segments 10, 11 and 12 for the unsupported and supported cases

displacements and plastic zones were obtained compared to Consequently, the outcomes of the CC method in all mat-
the 3D models. Therefore, the utilization of 3D modeling in ters depending on the deformational behavior and perfor-
design of the tunneling seems to be safe. mance of the empirical lining were not in good agreement

SEGMENT 13 SEGMENT 14 SEGMENT 15

(F3 FAULT ZONE) (UNIT D) (UNIT D / EXIT PORTAL)
2D MODEL
UNSUPPORTED CASE

Plastic zone
3D MODEL

Plastic zone
2D MODEL
SUPPORTED CASE

3D MODEL

Fig. 18 2D and 3D numerical analyses showing the plastic zones developed in segments 13, 14 and 15 for the unsupported and supported case
 Author's personal copy
Geotechnical studies and primary support design for a highway tunnel: a case study in Turkey

with the related outputs of the FEM-based 2D and 3D models. Acknowledgements The authors would like to acknowledge to the editor
and reviewers for their valuable contribution. Also, thanks to MSc.
However, as checked against to the 2D method, the 3D model
Geology Engineer Aytuna Sayin from the Turkish General Directorate
may be very advantageous for tunnel engineers to define the of Highways (TGDH) for the office work associated with this study.
safety in a more accurate way.

References
Conclusions
Aydan O, Akagi T, Kawamoto T (1993) The squeezing potential of rocks
In the present study, the stability evaluation and the lining around tunnels; theory and prediction. Rock Mech Rock Eng 26(2):
137–163
design of the Cankurtaran Tunnel, which is built in Artvin
Aydan O, Ulusay R, Kawamoto T (1997) Assessment of rock mass
City (Turkey), was researched. Considering the data compiled strength for underground excavations, proceedings of the 36th US
from field and laboratory investigations, the quality of the rock mechanics symposium, New York, pp 777–786
tunnel ground segments were defined by the help of the Aydan O, Ulusay R, Tokashiki N (2014) A new rock mass quality rating
RMR and Q systems as ranging from extremely poor to fair. system: rock mass quality rating (RMQR) and its application to the
estimation of geomechanical characteristics of rock masses. Rock
After the application of the rock mass classifications, the four Mech and Rock Eng 47(4):1255–1276
empirical lining design classes and the essential engineering Aydin A, Ozbek A, Acar A (2014) Geomechanical characterization, 3-D
geological characteristic properties forming the tunnel ground optical monitoring and numerical modeling in Kirkgecit-1 tunnel,
were obtained. Turkey. Eng Geol 181:38–47
Barton NR (2002) Some new Q-value correlations to assist in site char-
The applicability of the empirically determined support
acterization and tunnel design. Int J Rock Mech Min Sci 39:185–
element was firstly controlled by the help of the 216
convergence-confinement (CC) technique. Interaction be- Barton NR, Lien R, Lunde J (1974) Engineering classification of rock
tween rock masses and tunnel support was also investigated masses for the design of tunnel support. Rock Mech (4):189–239
analytically. According to the acquired results, the empirical Basarir H (2006) Engineering geological studies and tunnel support de-
sign at Sulakyurt dam site, Turkey. Eng Geol 86:225–237
lining design was only satisfactory for segments 1, 2, 3, 10, Basarir H, Ozsan A, Karakus M (2005) Analysis of support requirements
12, 14 and 15. Except for these segments, it was concluded for a shallow diversion tunnel at Guledar dam site, Turkey. Eng Geol
that the proposed lining designs are not successful to prevent 81(2):131–145
the stability problem owing to the great field stresses. Bieniawski ZT (1974) Geomechanics classification of rock masses and its
application in tunneling, Proceedings of the Third International
In addition to the CC analysis, the FEM-based 2D and 3D
Congress on Rock Mechanics, Vol. 1A. International Society of
software programs were utilized to check the efficiency of the Rock Mechanics, Denver, 27–32
empirical lining, and to numerically define the plastic zones Bieniawski ZT (1989) Engineering rock mass classifications. Wiley, New
and convergences that occurred around the ground surround- York, p 251
ing the tunnel. Following application of the lining support, the Cai M, Kaiser PK, Tasaka Y, Minami M (2007) Determination of residual
strength parameters of jointed rock masses using the GSI system. Int
size of the plastic zones around the opening was ordinarily J Rock Mech Min Sci 4(2):247–265
decreased. Thus, it was determined that the empirical lining Capkinoglu S (1981) Geology of the district between Borcka and Cavuslu
suggestions given for the Cankurtaran Tunnel project were (Hopa), MSc. thesis, Karadeniz Technical University, Trabzon,
generally acceptable. Support systems have been accom- Turkey
plished in suppressing the thickness of the plastic zone occur- Carranza-Torres C, Engen M (2017) The support characteristic curve for
blocked steel sets in the convergence-confinement method of tunnel
ring around the tunnel excavation. However, it was deter- support design. Tun Und Space Tech 69:233–244
mined that there are still plastic zones at the level of the tunnel Carranza-Torres C, Fairhurst C (1999) The elasto-plastic response of
roof in segments 4, 6 and 8. Application of rock bolts longer underground excavations in rock masses that satisfy the Hoek–
than 4 m was suggested for these zones. The results of the Brown failure criterion. Int J Rock Mech Min Sci 36(6):777–809
Carranza-Torres C, Fairhurst C (2000) Application of the convergence-
numerical analyses indicated that the most unsuitable defor- confinement method of tunnel design to rock-masses that satisfy the
mation and the load scenario on lining elements are within Hoek–Brown failure criterion. Tun. Und. Space Tech. 15(2):187–
admissible limits, and that the empirical support designs are 213
considered to be reliable enough to support the excavation. Curran JH, Hammah RE, Thamer EY (2003) A two dimensional ap-
proach for designing tunnel support in weak rock, Proc. 56th
Finally, it was concluded that the 3D FEM modelling presents
Canadian Geotech. Conference, Winnebeg, Monibota
the best answer in lining design compared to the CC method Deere DU (1964) Technical description of rock cores for engineering
and 2D FEM analysis. purposed. Rock Mech Rock Eng 1:17–22
As an important addition of future studies on this topic, it is Dhawan KR, Singh DN, Gupta ID (2002) 2D and3D finite element anal-
recommended that the in situ monitoring data should be taken ysis of underground openings in an inhomogeneous rock mass. Int J
Rock Mech Min Sci 39:217–227
throughout the construction stage of the excavation for adjust- Genis M, Basarır H, Ozarslan A, Bilir E, Balaban E (2007) Engineering
ment of the FEM designs and for checking the accuracy of the geological appraisal of the rockmasses and preliminary support de-
suggested lining support. sign, Dorukhan tunnel, Zonguldak, Turkey. Eng Geol 92:14–26
 Author's personal copy
A. Kaya, F. Bulut

Gurocak Z, Solanki P, Zaman MM (2007) Empirical and numerical anal- Ozsan A, Karpuz C (2001) Preliminary support design for Ankara sub-
yses of support requirements for a diversion tunnel at the Boztepe way extension tunnel. Eng Geol 59(1–2):161–172
dam site, eastern Turkey. Eng Geol 91:194–208 Palmström A (1995) RMi-a rock mass characterization system for rock
Guven IH (1993) 1:250000-scaled geology and compilation of the engineering purposes, Ph.D. thesis, University of Oslo, Norway, 400
Eastern Pontide. General Directorate of Mineral Research and Park KH (2004) Elastic solution for tunneling-induced ground move-
Exploration (MTA) of Turkey, Ankara (unpublished) ments in clays. Int. J. Geomech. 4:310–318
Hoek E (2007) Practical Rock Engineering, Evert Hoek Consulting Priest SD, Hudson JA (1976) Discontinuity spacing in rock. Int J Rock
Engineer Inc., Vancouver, Canada (Available for download at), Mech Min Sci Geo Abs 13:135–148
https://www.rocscience.com/learning/hoek-s-corner/books Quinones-Rozo C (2010) Lugeon test interpretation, revisited, In:
Hoek E, Diederichs MS (2006) Empirical estimation of rock mass mod- Collaborative Management of Integrated Watersheds, 30rd Annual
ulus. Int J Rock Mech Min Sci 43:203–215 USSD (United States Society on Dams) Conference, US Society on
Hoek E, Marinos P (2000) Predicting tunnel squeezing, Tunnels and Dams, Denver, CO, USA, 405–414
Tunneling International, Part 1 – November 2000, Part 2– Rabcewicz L (1964) The new Austrian Tunnelling method. Water Power
December 2000 16:453–457
Hoek E, Kaiser PK, Bawden WF (1995) Support of underground exca- Rocscience Inc (2016a) Dips v7.0 graphical and statistical analysis of
vations in hard rock. AA Balkema, Rotterdam orientation data, Toronto, Ontario, Canada, www.rocscience.com
Hoek E, Carranza-Torres C, Corkum B (2002) Hoek-Brown failure Rocscience Inc (2016b) RocData v5.0 rock, soil and discontinuity
criterion-2002 edition. proceedings of NARMS-TAC2002, mining strength analysis, Toronto, Ontario, Canada, www.rocscience.com
innovation and technology. Toronto, Canada, pp 267–273 Rocscience Inc (2016c) RS3 v1.0 3D finite element analysis for rock and
ISRM (2007) The complete ISRM suggested methods for rock character- soil, Toronto, Ontario, Canada, www.rocscience.com
ization, testing and monitoring: 1974–2006, International Society of
Rocscience Inc. (2017) RS2 v9.0 finite element analysis for excavations
Rock Mechanics Turkish National Group, Ankara, Turkey, 628
and slopes, Toronto, Ontario, Canada, www.rocscience.com
Javadi AA, Snee CPM (2002) Numerical modeling of air losses in com-
Sari D, Pasamehmetoglu AG (2004) Proposed support design, Kaletepe
pressed air tunneling. Int J Geomech 2(4):399–417
tunnel, Turkey. Eng Geol 72:201–216
Kanik M, Gurocak Z (2018) Importance of numerical analyses for deter-
Sheorey PR, Murali MG, Sinha A (2001) Influence of elastic constants on
mining support systems in tunneling: a comparative study from the
the horizontal in situ stress. Int J Rock Mech Min Sci 38(1):1211–
Trabzon-Gumushane tunnel, Turkey. J Afr Earth Sci 143:253–265
1216
Kanik M, Gurocak Z, Alemdag S (2015) A comparison of support sys-
tems obtained from the RMR89 and RMR14 by numerical analyses: Sopaci E, Akgun H (2008) Engineering geological investigations and the
Macka tunnel project, NE Turkey. J Afr Ear Sci 109:224–238 preliminary support design for the proposed Ordu peripheral high-
Kaya A, Sayin A (2017) Engineering geological appraisal and prelimi- way tunnel, Ordu, Turkey. Eng Geo 96:43–61
nary support design for the Salarha tunnel, Northeast Turkey. Bul Terzaghi K (1946) Rock defects and loads on tunnel supports, in Proctor,
Eng Geo Env. https://doi.org/10.1007/s10064-017-1177-2 R.V., and White, T.L., eds., Rock tunneling with steel support,
Kaya A, Bulut F, Alemdag S, Sayin A (2011) Analysis of support require- Youngstown, Ohio, Commercial Shearing and Stamping
ments for a tunnel portal in weak rock: a case study in Turkey. Sci Company, 1:17–99
Res Ess 6(31):6566–6583 TGDH (2013) Specification for highway works (in Turkish), Turkish
Ketin I (1966) Tectonic units of Anatolia. J Gen Direc Min Res Exp, Ministry of Public Works, General Directorate of Highways, Ankara
(MTA) 66:23–34 Trinh QN, Broch E, Lu M (2010) 2D versus 3D modeling for tunneling at
Kockar MK, Akgun H (2003) Engineering geological investigations a weakness zone, ISRM Regional Symposium-EUROCK 2009,
along the Ilıksu tunnels, southern Turkey. Eng Geol 68(3–4):141– Croatia
158 Ucer S (2006) Comparison of 2D and 3D finite element models of tunnel
Lauffer H (1958) Classification of in-situ rock in tunnel construction. advance in soft ground: A case study on Bolu tunnels, MSc Thesis,
Geology Bauwesen 24:46–51 Middle East Technical University, Ankara, Turkey
Lin SY, Hung HH, Yang JP, Yang YB (2017) Seismic analysis of twin Wickham GE, Tiedemann HR, Skinner EH (1972) Support determination
tunnels by a finite/infinite element approach. Int. J. Geomech. based on geologic predictions, In: Lane, K.S.A.G., L. A., ed., North
https://doi.org/10.1061/(ASCE)GM.1943-5622.0000677 American Rapid Excavation and Tunneling Conference: Chicago,
Lugeon M (1933) Barrages et geologic methods de recherche´ New York: Society of Mining Engineers of the American Institute of
terrasement et un permeabilisation. Litrairedes Universite, Paris Mining, Metallurgical and Petroleum Engineers, 43–64
Marinos P, Hoek E, (2000) GSI: a geologically friendly tool for rock mass Wyllie DC, Mah CW (2004) Rock slope engineering civil and mining,
strength estimation, In: Proceedings of the GeoEng2000 at the in- Spon Press, Taylor and Francis e-library
ternational conference on geotechnical and geological engineering, Xu Q, Xiao Z, Liu T, Lou T, Song X (2015) Comparison of 2D and 3D
Melbourne, Technomic publishers, Lancaster, 1422–1446 prediction models for environmental vibration induced by under-
Ozdogan MV, Yenice H, Gonen A, Karakus D (2018) Optimal support ground railway with two types of tracks. Comput Geotech 68:
spacing for steel sets: Omerler underground coal mine in western 169–183
Turkey. Int. J. Geomech. https://doi.org/10.1061/(ASCE)GM.1943- Yalcin E, Gurocak Z, Ghabchi R, Zaman M (2015) Numerical analysis
5622.0001069 for a realistic support design: case study of the Komurhan tunnel in
Ozsan A, Basarir H (2003) Support capacity estimation of a diversion eastern Turkey. Int. J. Geomech. https://doi.org/10.1061/(ASCE)
tunnel in weak rock. Eng Geol 68:319–331 GM.1943-5622.0000564

View publication stats