International Journal of Non-Linear Mechanics 46 (2011) 3546

Contents lists available at ScienceDirect

International Journal of Non-Linear Mechanics

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

The effect of concrete slabrockll interface behavior on the earthquake

performance of a CFR dam

Alemdar Bayraktar a, Murat Emre Kartal b,n, Suleyman

Adanur a
a
b

Karadeniz Technical University, Department of Civil Engineering, 61080 Trabzon, Turkey

Zonguldak Karaelmas University, Department of Civil Engineering, 67100 Zonguldak, Turkey

a r t i c l e in fo

abstract

Article history:
Received 14 June 2009
Received in revised form
23 June 2010
Accepted 6 July 2010

Earthquake response of the concrete slab is mostly depended upon its conjunction with rockll. This
study aims to reveal the effect of concrete slabrockll interface behavior on the earthquake
performance of a concrete-faced rockll dam considering friction contact and welded contact. Friction
contact is provided by using interface elements with ve numbers of shear stiffness values. 2D nite
element model of Torul concrete-faced rockll dam is used for this purpose. Linear and materially
non-linear time-history analyses considering damreservoir interaction are performed using ANSYS.
Reservoir water is modeled using uid nite elements by the Lagrangian approach. The DruckerPrager
model is preferred for concrete slab and rockll in non-linear analyses. Horizontal component of 1992
Erzincan earthquake with peak ground acceleration of 0.515g is used in analyses. The maximum and
minimum displacements and principal stresses are shown by the height of the concrete slab and
earthquake performance of the dam is investigated considering different joint conditions for empty
and full reservoir cases. In addition, potential damage situations of concrete slab are evaluated.
& 2010 Elsevier Ltd. All rights reserved.

Keywords:
Concrete-faced rockll dam
Damreservoir interaction
DruckerPrager model
Friction contact
Interface element
The Lagrangian approach

1. Introduction
Concrete-faced rockll (CFR) dams are considered to be safe
under seismic excitations because of two following origins [1].
First, porewater development and strength descent do not occur
because the entire CFR dam embankment is waterless during an
earthquake. Second, CFR dams provide more stability with their
whole rockll mass than earth core rockll (ECR) dams, since CFR
dams do not permit water to penetrate inside the dam on the
other hand only downstream rockll mass of the ECR dams may
resist for stability under seismic excitations.
CFR dams involve uidstructure interaction problems. Hydrodynamic pressures resulted from earthquakes considerably affect
dynamic response of dams. The hydrodynamic pressure effects on
dynamic response of dams have been started to be researched in
the 1930s [24]. Dynamic response of damreservoir systems
using the Eulerian and the Lagrangian approaches has been
investigated by many researchers [514]. In the last years,
Bayraktar et al. [1315] paid attention on hydrodynamic
pressures on concrete slab of CFR dams.
Earthquake analysis of CFR dams subjected to strong ground
motion was carried out and published in the literature by various
researchers [1,1322]. In addition, a new approach based on

Corresponding author. Tel.: +90 372 257 4010; fax: + 90 372 257 4023.
E-mail address: murat_emre_kartal@hotmail.com (M.E. Kartal).

0020-7462/$ - see front matter & 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijnonlinmec.2010.07.001

scaled boundary-nite element method was used to obtain

scattered motion along a prismatic canyon with trapezoidal cross
section [23]. The authors performed three-dimensional dynamic
analysis of a typical CFR dam including dam-face slababutments
interaction using scaled boundary-nite element method.
Ghannad [24] performed numerical (nite element method) and
analytical analyses of a CFR dam, which is located in a high
seismicity region of Iran, and compared the results. The effect of
non-linearity and time-dependent deformation on the separation
of the concrete slab from the cushion layer was examined using
contact analysis method [25]. Beyond these studies, there is
limited research related to earthquake performance of CFR dams.
Particularly, performance analysis of a CFR dam including
damreservoir interaction and slippageseparation in concrete
slabrockll interface is rarely seen in the literature.
Interface elements have a wide range of use to describe the
interaction between different media [2631]. Various researchers
investigated discrete joints in non-linear analyses [3238]. Interface elements were used to determine the effect of discontinuities
on the response of circular tunnels established in layered
geological media by Lee and Zaman [39]. The seismic response
of rigid highway bridge abutments, retaining and founded on dry
sand was examined considering sliding and debonding/recontact
between the wall and the soil [40]. Toki et al. [41,42] used joint
elements for dynamic analysis of soilstructure interaction
systems to simulate time-dependent sliding and separation along
the soilstructure interface. The interface behavior in reinforced

36

A. Bayraktar et al. / International Journal of Non-Linear Mechanics 46 (2011) 3546

embankments on soft grounds was researched considering slip

between soil and reinforcement according to MohrCoulomb
strength criterion with interface elements [43]. Nam et al. [44]
used elasto-plastic interface element to predict static and
dynamic behaviors of underground RC structures. Proposed
interface model was quiet well in agreement to describe the
interaction between the underground RC structure and the
surrounding soil media. Uddin [45,46] performed dynamic
analyses using interface elements for the potential sliding interface in embankment of an ECR dam and also for the interaction of
concrete slabrockll in a CFR dam.
This study investigates the effect of interface behavior
between concrete slab and rockll on the earthquake response
and performance of a CFR dam including hydrodynamic effects.
For this purpose, two-dimensional dam and damreservoir nite
element models are used. Hydrodynamic pressure is taken into
consideration by the Lagrangian approach using two-dimensional
uid nite elements. Both material and connection non-linearity
are considered in nite element analyses. DruckerPrager model
is used for concrete slab and rockll in materially non-linear
analyses. Welded and friction contact is considered in concrete
slabrockll interface. Friction is considered with interface
elements. Earthquake response and performance of Torul CFR
Dam are investigated considering different joint conditions
between concrete and rockll. All numerical analyses are
performed using ANSYS [47].

2. Formulation of damreservoir interaction by the

Lagrangian approach
The formulation of the uid system based on the Lagrangian
approach is presented as following [48,49]. In this approach, uid
is assumed to be linearly compressible, inviscid and irrotational.
For a general three-dimensional uid, pressurevolumetric strain
relationships can be written in matrix form as follows:
38
8 9 2
C11 0
0
0
P >
ev 9
>
>
>
>
>
>
>
>
>
>
>
6
7
< Px = 6 0
0 7< wx =
C22 0
7
6
1
>
0
0
C33 0 7
P > 6
w >
>
>
5>
> 4
>
> y>
> y>
;
;
:
:
0
0
0
C44
Pz
wz
where P, C11, and ev are the pressures which are equal to mean
stresses, the bulk modulus and the volumetric strains of the uid,
respectively. Since irrotationality of the uid is considered like
penalty methods [50,51], rotations and constraint parameters are
included in the pressurevolumetric strain equation (Eq. (1)) of
the uid. In this equation, Px, Py and Pz, are the rotational stresses;
C22, C33 and C44 are the constraint parameters and wx, wy and wz
are the rotations about the cartesian axis x, y and z, respectively.
In this study, the equations of motion of the uid system are
obtained using energy principles. Using the nite element
approximation, the total strain energy of the uid system may
be written as
1
2

pe UTf Kf Uf

where Uf and Kf are the nodal displacement vector and the

stiffness matrix of the uid system, respectively. Kf is obtained by
the sum of the stiffness matrices of the uid elements as follows:
X e
Kf
Kf
Z
e
e
Kef BeT
3
f Cf Bf dV
V

where Cf is the elasticity matrix consisting of diagonal terms in

Eq. (1). Bef is the straindisplacement matrix of the uid element.

An important behavior of uid systems is the ability to

displace without a change in volume. For reservoir and storage
tanks, this movement is known as sloshing waves in which the
displacement is in the vertical direction. The increase in
the potential energy of the system because of the free surface
motion can be written as
1
2

ps UTsf Sf Usf

where Usf and Sf are the vertical nodal displacement vector and
the stiffness matrix of the free surface of the uid system,
respectively. Sf is obtained by the sum of the stiffness matrices of
the free surface uid elements as follows:
9
P e
=
Sf
Sf
R
5
e
T
e
Sf rf g A hs hs dA ;
where hs is the vector consisting of interpolation functions of the
free surface uid element. rf and g are the mass density of the
uid and the acceleration due to gravity, respectively. Besides,
kinetic energy of the system can be written as
T

1 _T
_
U MU
2 f f f

_ f and Mf are the nodal velocity vector and the mass

where U
matrix of the uid system, respectively. Mf is also obtained by the
sum of the mass matrices of the uid elements as follows:
9
P e
Mf
Mf
=
R
7
Mfe rf V HT H dV e ;
where H is the matrix consisting of interpolation functions of the
uid element. If (Eqs. (2), (4) and (6)) are combined using the
Lagranges equation [52]; the following set of equations is
obtained:
f K* Uf Rf
Mf U
f

f, Uf and Rf are the system stiffness matrix including

where K*f , U
the free surface stiffness, the nodal acceleration and displacement
vectors and time-varying nodal force vector for the uid system,
respectively. In the formation of the uid element matrices,
reduced integration orders are used [48].
The equations of motion of the uid system (Eq. (8)), have a
similar form with those of the structure system. To obtain the
coupled equations of the uidstructure system, the determination of the interface condition is required. Since the uid is
assumed to be inviscid, only the displacement in the normal
direction to the interface is continuous at the interface of
the system. Assuming that the structure has the positive face
and the uid has the negative face, the boundary condition at the
uidstructure interface is
Un Un

where Un is the normal component of the interface displacement

[53]. Using the interface condition, the equation of motion of the
coupled system to ground motion including damping effects are
given by
c Cc U
_ c Kc Uc Rc
Mc U

10

in which Mc, Cc, and Kc are the mass, damping and stiffness
_ c, U
c and Rc are
matrices for the coupled system, respectively. Uc, U
the vectors of the displacements, velocities, accelerations and
external loads of the coupled system, respectively.

A. Bayraktar et al. / International Journal of Non-Linear Mechanics 46 (2011) 3546

3. DruckerPrager model

37

There are many criteria for determination of yield surface or

yield function of materials. DruckerPrager criterion is widely
used for frictional materials such as rock and concrete. Drucker
and Prager [54] obtained a convenient yield function to determine
elasto-plastic behavior of concrete smoothing MohrCoulomb
criterion. This function is dened as
p
f aI1 J2 k
11

v2

v1

u2

x
v4

v3

u4

u3

2 sin f

a p

33sin f
6c cos f
k p
33sin f

12

in Eq. (11), I1 is the rst invariant of stress tensor (sij)

I1 s11 s22 s33

l
13

and J2 is the second invariant of deviatoric stress tensor (sij)

1
J2 sij sij
2

u
14

where sij is the deviatoric stresses as given below

i,j 1,2,3

I1
s11 s22 s33 sii

3
3
3

16

If the terms in Eq. (15) are obtained by the Eq. (16) and replaced in
Eq. (14), the second invariant of the deviatoric stress tensor can be
obtained as follows:
i
1h
s11 s22 2 s22 s33 2 s33 s11 2 s212 s213 s223
J2
6
17
It is observed from Fig. 1 that a smooth surface is obtained
removing Coulomb corner spots [55].

4. Interface element formulation

The formulation of the stiffness matrix of two-dimensional
interface element is presented in this section. The geometry of the
interface element is shown in Fig. 2(a). Since the interface
element represents the interaction characteristics between two

15

In Eq. (15), dij is the Kronecker delta, which is equal to 1 for

ij; 0 for i aj, and sm is the mean stress and obtained as follows:

Fig. 2. (a) Interface nite element, (b) normal strain and (c) shear strain.

different materials and is not a material itself, there exist only a

normal stress and shear stress [56].
Displacements in the upper and lower faces are independently
interpolated as follows:
uupp N1 u^ 1 N2 u^ 2 ,
ulow N3 u^ 3 N4 u^ 4 ,

vupp N1 v^ 1 N2 v^ 2
vupp N3 v^ 3 N3 v^ 4

18

1
1 7 xi 1 7 Zi
19
4
With reference to Fig. 2, strains are computed from Eq. (20) as
shown in Fig. 2(b)
(
) (
)
gyx
uupp ulow 
feg
h
20
ey vupp vlow
Ni

in which ey and gyx represents the normal and tangential (shear)

strains as shown in Fig. 2(c).
Stresses are obtained from strains and constitutive law as
(
)
fsg

sm

where a and k are constants which depend on cohesion (c) and

angle of internal friction (f) of the material given by

sij sij dij sm

u1

tyx
sy

Dfeg

21

Failure Surface of Drucker-Prager

Hydrostatic Axis
(11 = 22 = 33)

in which [D] is the elastic constitutive matrix given by

"
#
d11 0
D
0
d22

22

From Eqs. (18) and (20), strains are written as

t
Co

Failure Surface of Coulomb

3

2
Fig. 1. Failure criteria for Coulomb, DruckerPrager and von Mises [55].

^
feg Bfdg

23

^ and [B] are the nodal displacement vector and the

in which fdg
straindisplacement matrix given by
n
o
^ T u^ 1 v^ 1 u^ 2 v^ 2 u^ 3 v^ 3 u^ 4 v^ 4
fdg
24

B B1 I

B2 I

B3 I

B4 I

with [I] being the identity matrix of order two.

25

A. Bayraktar et al. / International Journal of Non-Linear Mechanics 46 (2011) 3546

On minimizing the potential energy of the element, we obtain

the stiffness matrix, [K], of the interface element in the local
coordinates. Thus
ZZ
K BT DB dx dy
26
This area integral can be easily computed if a change in the
variables is carried out by writing
dx dy DetJ dx dZ

27

in which [J] is the Jacobian matrix. The element thickness is often

assumed to be zero [57].

Cumulative Duration (s)

38

0.36
0.33
0.3
0.27
0.24
0.21
0.18
0.15
0.12
0.09
0.06
0.03
0

May require nonlinear analysis

to estimate damage

Acceptable damage
based on linear analysis
1

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

Demand-Capacity Ratio

5. Structural performance and damage criteria for dams

Fig. 3. Accepted performance curve for CFR dams [59].

Linear time-history analysis is used to formulate a systematic

and rational methodology for qualitative estimate of the level of
damage. In linear time-history analysis, where acceleration
time-histories are the seismic input, deformations, stresses and
section forces are computed in accordance with elastic stiffness
characteristics of various components in time domain. A systematic
evaluation of these results in terms of the demandcapacity ratios
(D/C), cumulative inelastic duration, spatial extent of overstressed
regions, and consideration of possible failure modes comprise the
basis for approximation and appraisal of probable level of damage.
The damage for structural performance amounts to cracking of
the concrete, opening of construction joints, and yielding of the
reinforcing steel. If the estimated level of damage falls below the
acceptance curve for a particular type of structure, the damage is
considered to be low and linear time-history analysis will be
sufcient. Otherwise the damage is considered to be severe in which
case non-linear time-history analysis would be required to estimate
damage more accurately [58].
5.1. Performance criteria for linear and non-linear analysis
The dam response to the maximum design earthquake is
considered to be within the linear elastic range of behavior with
little or no possibility of damage if computed demandcapacity
ratios are less than or equal to 1.0. The stage of non-linear
response or opening and cracking of joints is considered acceptable if demandcapacity ratio is less than 2, overstressed region is
less than 15% of the dam surface area, and the cumulative
inelastic duration falls below the performance curve given in
Fig. 3. Cumulative duration has not been dened for the concrete
slab of CFR dams till now; therefore the performance curve for
concrete gravity dams is used in this study [59].
5.2. Demandcapacity ratios
The demandcapacity ratios for CFR dams can be dened as
the ratio of the computed principal tensile stresses to tensile
strength of the concrete. As discussed previously demand
capacity ratio is limited to 2.0, thus permitting stresses up to
twice the static or at the level of dynamic apparent tensile
strength of the concrete, as long as the overstressed region is less
than 15% of the dam surface area. The cumulative duration
beyond a certain level of demandcapacity ratio is obtained by
multiplying number of stress values exceeding that level of tensile
strength by the time-step of the time-history analysis. The
cumulative duration in Fig. 3 refers to the total duration of all
stress excursions beyond a certain level of demandcapacity
ratio. Although tensile strength of concrete is affected by the
rate of seismic loading, the acceptance criteria employ stable
tensile strength in computation of the demandcapacity ratios.

Fig. 4. Torul Dam [60]: (a) empty reservoir case and (b) full reservoir case.

The reason for this is to account for the lower strength of the lift
lines and provide some level of conservatism in estimation of
damage using the results of linear elastic analysis.

6. Numerical model of Torul CFR dam

6.1. Torul dam
Torul CFR Dam (Fig. 4) is sited on Harsit River and
approximately 14 km northwest of Torul, Gumushane, Turkey.
The dam construction was completed in 2007 by the General
Directorate of State Hydraulic Works [60]. The main goal of the

A. Bayraktar et al. / International Journal of Non-Linear Mechanics 46 (2011) 3546

reservoir is power generation. The volume of the dam body is

4.6 hm3 and the lake area of the dam at the normal water level
is 3.62 km2. The annual total power generation capacity is
322.28 GW. The length of the dam crest and the wide of the
dam crest are 320 and 12 m, respectively. Besides, the maximum
height and base width of the dam are 142 and 420 m, respectively. The thickness of the concrete slab is 0.3 m at the crest level
and 0.7 m at the foundation level. The concrete slab has high
seepage resistance. The two-dimensional largest cross section and
the dimensions of the dam are shown in Fig. 5.
6.2. Material properties
The Torul Dam body consists of concrete face slab and ve
rockll zones: 2A, 3A, 3B, 3C, 3D, respectively, from upstream to
downstream. Rockll zones were arranged from thin granules to
thick particles in upstreamdownstream direction. Table 1 shows
the material properties of the dam and reservoir water used in
linear and non-linear analyses. Performed materially non-linear
analysis procedure is based on the DruckerPrager model. The
cohesion and the angle of internal friction of the dam body are
assumed as 1.225 MPa and 451, respectively. The concrete slab
has tensile strength of 1.6 MPa and compression strength of
20 MPa [61]. The bulk modulus of reservoir water and density are
assumed as 2.07  103 MPa and 1000 kg/m3.

39

The main objective of the couplings is to hold equal the displacements between two reciprocal nodes in normal direction to the
interface. The length of the reservoir is taken as three times of
the dam height to adequately consider reservoir water effects.
6.4. Concrete slabrockll interface
The earthquake response of the concrete slab is mostly
depended upon its conjunction to the rockll. Welded contact
and friction contact models can be used in this joint (Fig. 7).
In fact, concrete slab does not directly contact with the rockll.
According to this observation, the use of interface element in
nite element analysis can procure more realistic results.
Concrete slab may slide over the surface of the rockll by using
this element. This element provides ability for transverse shear
deformation. This study assumes that concrete slab and rockll
dam body are independent deformable bodies by using interface
elements and also dependent deformable bodies considering
welded contact.
The interface element used in this study has four node and
two integration points (Fig. 8). Normal stiffness of the interface
element is considered as 20  103 MPa/m. Five numbers of
transverse shear stiffness values of the interface element are
used as 1.8, 3.6, 18, 180 and 1800 MPa/m in the numerical
analyses.

6.3. Finite element model

7. Earthquake response of Torul CFR Dam
The two-dimensional damreservoir nite element model
used in analyses is shown in Fig. 6. In this model, dam body has
592 solid nite elements, reservoir water has 495 uid nite
elements and 16 interface elements are dened between concrete
slab and rockll. The solid elements used in the analyses have four
nodes and 2  2 integration points and the uid elements have
four nodes and 1  1 integration point. Element matrices are
computed using the Gauss numerical integration technique [48].
The damping ratio of 5% is used in nite element analyses.
Coupling length is chosen as 1 mm at reservoirdam interface
and 15 numbers of couplings are dened in damreservoir model.

This study investigates the earthquake response of Torul CFR

Dam subjected to strong ground motion is. Empty and full
reservoir cases are taken into account in the numerical solutions.
The horizontal component of the 1992 Erzincan earthquake
with peak ground acceleration (pga) 0.515g is utilized in analyses.

Fig. 6. The two-dimensional nite element model including reservoir of

Torul Dam.

Fig. 5. The largest cross section of Torul Dam body [60].

Table 1
Material properties of Torul CFR Dam.
Material

Concrete
2A (sifted rock or alluvium)
3A (selected rock)
3B (lling with quarry rock)
3C (lling with quarry rock)
3D (selected rock)
a

Maximum particle size.

Dmax (mm)

150
300
600
800
1000

Material properties
Modulus of elasticity (kN/m2)

Poissons ratio

Mass density (kg/m3)

3.420E +07
1.400E +07
1.350E +07
1.250E +07
1.150E +07
1.000E +07

0.18
0.26
0.26
0.26
0.26
0.26

2395.5
2905.2
2854.2
2833.8
2803.3
2752.3

40

A. Bayraktar et al. / International Journal of Non-Linear Mechanics 46 (2011) 3546

Concrete Slab

Transition Zones: 2A, 3A

Rockfill Zones: 3B, 3C

Rockfill Zones
Concrete
Slab

2A 3A 3B

Transition
Zones

3D

3A
3C

3B
2A

Interface allowing slippage

Interface

Concrete Slab

Rockfill Zones
2A 3A 3B

Transition
Zones

Concrete
Slab

3D

3A
3B

3C

2A

l
i

i-j and l-k surfaces

are the contact surfaces
x

y
k

Acceleration (m/s2)

Fig. 7. (a) Welded and (b) friction contact in concrete slabrockll interface.

6
5
4
3
2
1
0
-1
-2
-3

t = 2.9s
pga = 5.054 m/s2
t = 3.235s
pga = 0.0 m/s2

t = 2.75s
pga = 0.0 m/s2

j
Fig. 8. The view of two-dimensional interface element in local coordinates [47].

12

15

18

21

24

Time (s)
Fig.10. 1992 Erzincan earthquake acceleration record [62].

This earthquake record is preferred because Torul Dam is close to

Erzincan where severe strong ground motions occurred in last
decades (Fig. 9) and its foundation has similar characteristics with
the place ground motion recorded. Earthquake analyses are
performed during 21.31 s (Fig. 10) and used acceleration record
is available at the PEER Strong Motion Database [62]. The time
interval of the acceleration record is 0.005 s. Displacement and
principal stress components by the height of the concrete slab are
compared.

7.1. Displacements

Fig. 9. The location of Torul Dam [60].

This section presents the horizontal displacements obtained

from linear time-history analyses by the height of the concrete
slab considering different material properties of the interface
element. The analysis results are shown in Figs. 1114 for empty
and full reservoir cases. It is obviously seen from Figs. 1114 that
hydrodynamic pressure of the reservoir water increases the

A. Bayraktar et al. / International Journal of Non-Linear Mechanics 46 (2011) 3546

displacements for all joint conditions. Displacements decrease

with the decrease of shear stiffness of the interface element in
empty reservoir case. However, those increase with the decrease

150

Height (m)

120

1.8 MPa/m

3.6 MPa/m

18 MPa/m

180 MPa/m

1800 MPa/m

Welded Contact

41

of shear stiffness of the interface element if hydrodynamics

pressure effects are included. In addition, displacements obtained
from nite element models including high shear stiffness of the
interface element come close to the ones obtained from the model
including welded contact.
Some deected shapes of Torul Dam during the time interval
of 2.75 and 3.235 s are given in Fig. 15. According to Fig. 15,

90
60
30
0
-40

-35

-30

-25

-20

-15

-10

-5

Displacement (mm)
Fig. 11. The minimum horizontal displacements by the dam height in empty
reservoir case.

150
1.8 MPa/m
180 MPa/m

Height (m)

120

3.6 MPa/m
1800 MPa/m

18 MPa/m
Welded Contact

90
60
30
0
0

15
10
Displacement (mm)

25

20

Fig. 12. The maximum horizontal displacements by the dam height in empty
reservoir case.

150

1.8 MPa/m
180 MPa/m
1800 MPa/m

Height (m)

120

3.6 MPa/m
18 MPa/m
Welded Contact

90
60
30
0
-50

-45

-40

-35

-30

-25

-20

-15

-10

-5

Displacement (mm)
Fig. 13. The minimum horizontal displacements by the dam height in full
reservoir case.

Height (m)

150

1.8 MPa/m
180 MPa/m

120

3.6 MPa/m
1800 MPa/m

18 MPa/m
Welded Contact

90
60
30
0
0

10

15
20
Displacement (mm)

25

30

Fig. 14. The maximum horizontal displacements by the dam height in full
reservoir case.

Fig. 15. The deected shapes of Torul Dam between 2.75 and 3.235 s: (a) the
deected shape on second 2.750 (acceleration is equal to zero); (b) the deected
shape on second 2.900 (pga is equal to 0.515g); (c) the deected shape on
second 2.975 (minimum displacement at the crest); (d) the deected shape on
second 3.200 (excessive deformations in downstream side) and (e) the deected
shape on second 3.235 (acceleration is equal to zero).

42

A. Bayraktar et al. / International Journal of Non-Linear Mechanics 46 (2011) 3546

the minimum displacement of the crest does not occur on second

2.9 in which the maximum ground acceleration of the earthquake
exists. The deected shape, where the minimum horizontal
displacement occurs, shown in Fig. 15c cannot be adequately
distinguished because of relatively high vertical displacements of
reservoir water. Relatively excessive deformations at downstream
side near foundation and besides separation and transverse shear
deformation of the concrete slab appear on second 3.20 as shown
in Fig. 15d.

7.2. Stresses
This section presents the principal tensile and compression
stresses occurred in the concrete slab by the dam height. Figs. 16
and 17 refer that maximum and minimum principal stresses
decrease with the decrease of the shear stiffness of the interface
element in empty reservoir case. Besides, those increase with the
decrease of the shear stiffness of the interface element in full
reservoir case as shown in Figs. 18 and 19. If the analysis ignores

150

1.8 MPa/m

3.6 MPa/m

18 MPa/m

180 MPa/m

1800 MPa/m

Welded Contact

Height (m)

120
90
60
30
0
-7000

-6000

-5000

-4000

-3000

-2000

-1000

Stress (kPa)
Fig. 16. The principal compression stresses by the dam height in empty
reservoir case.

150

1.8 MPa/m

3.6 MPa/m

18 MPa/m

180 MPa/m

1800 MPa/m

Welded Contact

Height (m)

120
90
60
30
0
0

500

1000 1500 2000 2500 3000 3500 4000 4500

Stress (kPa)

Fig. 17. The principal tensile stresses by the dam height in empty reservoir case.

reservoir water effects and considers friction in concrete

slabrockll interface, concrete slab may behave itself easily
and avoid unnecessary stress intensity resulting from rockll.
However, in full reservoir case, hydrodynamic pressure causes
additional stress density in the concrete slab.

8. Performance analysis of Torul CFR Dam

This part of the study presents earthquake performance
analysis of Torul CFR Dam. The main objective of this study is to
reveal the effect of concrete slabrockll interface on the earthquake performance of a CFR dam. Therefore, this study considers
ve shear stiffness values of the interface element for empty and
full reservoir cases. Time-history analyses are performed according to northsouth component of 1992 Erzincan earthquake
record shown in Fig. 10.
The demandcapacity ratios, which are evaluated between
1 and 2, are considered for the principle tensile stresses occurred in
the concrete slab. The principal tensile stress cycles obtained from
linear time-history analyses are given for different shear stiffness
values of the interface element in Fig. 20. As seen from Fig. 20,
principal tensile stresses exceed the tensile strength of the concrete
numerous times in full reservoir case even if D/C is equal to 2.
Besides, the tensile stresses exceed the tensile strength of the
concrete several times in empty reservoir case as well. Fig. 20 refers
that reservoir water increases the principal tensile stresses. But,
this increase is more evident in the case that shear stiffness of the
interface element is lower. For the higher shear stiffness values of
the interface element, hydrodynamic pressure effects on the
concrete slab are relatively low. In addition to this, numerical
results of the concrete slab for the maximum shear stiffness of the
interface element are fairly close to ones of the dam including
welded contact in concrete slabrockll interface.
The performance curves are drawn to determine the earthquake performance of the concrete slab of Torul CFR Dam
according to linear time-history analyses. Those frequently
exceed the acceptable level in empty reservoir case and
completely exceed it by the effect of the reservoir water. The
earthquake performance of the dam involving hydrodynamic
pressure effects is lower for the shear stiffness of 1.8 and
3.6 MPa/m than the other ones, which are relatively close to
earthquake performance of the dam modeled with welded contact
in concreterockll interface. Nevertheless, these shear stiffness
values have the increase effect on earthquake performance when
hydrodynamic pressure effects are ignored. In this conditions,
earthquake performance curve fall below the acceptable level into
1.51.8D/C interval for the shear stiffness of 1.8 MPa/m, and
1.71.8D/C interval for 3.6 MPa/m. The performance curve for
18 MPa/m is a bit lower than the curves drawn for 1.8 and
3.6 MPa/m. Figs. 21 and 22 indicate that the use of higher shear

150
3.6 MPa/m
1800 MPa/m

150

18 MPa/m
Welded Contact

90
60
30
0
-9000 -8000 -7000 -6000 -5000 -4000 -3000 -2000 -1000
Stress (kPa)

1.8 MPa/m

3.6 MPa/m

180 MPa/m

1800 MPa/m

18 MPa/m
Welded Contact

120
Height (m)

Height (m)

120

1.8 MPa/m
180 MPa/m

90
60
30
0

Fig. 18. The principal compression stresses by the dam height in full reservoir
case.

50

00

10

00

15

00

20

00 000 500 000 500 000 500

3
25
3
5
5
4
4
Stress (kPa)

Fig. 19. The principal tensile stresses by the dam height in full reservoir case.

A. Bayraktar et al. / International Journal of Non-Linear Mechanics 46 (2011) 3546

5600

5600
Stress

4800

D/C=1

D/C=2

4800

4000

4000

3200

3200

2400

2400

1600

1600

800

800

Stress

D/C=1

D/C=2

0
0

12
15
Time (s)

18

21

24

12
15
Time (s)

18

21

24

5600

5600
Stress

4800

D/C=1

D/C=2

4800

4000

4000

3200

3200

2400

2400

1600

1600

800

800

43

12
15
Time (s)

18

21

24

Stress

D/C=1

12
15
Time (s)

18

D/C=2

21

24

5600

5600
Stress

4800

D/C=1

D/C=2

Stress

4800

4000

4000

3200

3200

2400

2400

1600

1600

800

800

D/C=1

D/C=2

0
0

12
15
Time (s)

18

21

24

12
15
Time (s)

18

21

24

5600

5600
Stress

4800

D/C=1

D/C=2

Stress

4800

4000

4000

3200

3200

2400

2400

1600

1600

800

800

D/C=1

D/C=2

0
0

12
15
Time (s)

18

21

24

5600

12
15
Time (s)

18

21

24

5600
Stress

4800

D/C=1

D/C=2

Stress

4800

4000

4000

3200

3200

2400

2400

1600

1600

800

800

D/C=1

D/C=2

0
0

12
15
Time (s)

18

21

24

5600

12
15
Time (s)

18

21

24

5600
Stress

4800

D/C=1

D/C=2

Stress

4800

4000

4000

3200

3200

2400

2400

1600

1600

800

800

D/C=1

D/C=2

0
0

12
15
Time (s)

18

21

24

12
15
Time (s)

18

21

24

Fig. 20. The principal tensile stress cycles according to linear analyses: (a) shear stiffness is 1.8 MPa/m in empty reservoir case; (b) shear stiffness is 1.8 MPa/m in full
reservoir case; (c) shear stiffness is 3.6 MPa/m in empty reservoir case; (d) shear stiffness is 3.6 MPa/m in full reservoir case; (e) shear stiffness is 18 MPa/m in empty
reservoir case; (f) shear stiffness is 18 MPa/m in full reservoir case; (g) shear stiffness is 180 MPa/m in empty reservoir case; (h) shear stiffness is 180 MPa/m in full
reservoir case; (i) shear stiffness is 1800 MPa/m in empty reservoir case; (j) shear stiffness is 1800 MPa/m in full reservoir case; (k) welded contact in empty reservoir case
and (l) welded contact in full reservoir case.

44

A. Bayraktar et al. / International Journal of Non-Linear Mechanics 46 (2011) 3546

Cumulative Duration (s)

stiffness causes the earthquake performance closer to the one

of the dam including welded contact. The decrease of the shear
stiffness of the interface element increases the earthquake
performance in empty reservoir case and reduces it in full

1.35
1.2
1.05
0.9
0.75
0.6
0.45
0.3
0.15
0
1.1

1.2

1.8 kPa

3.6 kPa

1800 kPa

Welded Contact

18 kPa

180 kPa

Acceptance Curve

1.3 1.4 1.5 1.6 1.7

Demand-Capacity Ratio

1.8

1.9

Cumulative Duration (s)

Fig. 21. Performance assessment of the dam in empty reservoir case.

1.35
1.2
1.05
0.9
0.75
0.6
0.45
0.3
0.15
0

reservoir case. According to linear analyses, performance curves

are usually over the acceptance curve in both cases so damage in
concrete appears inevitable.
The estimation of the earthquake performance of the dam
implies non-linear analysis to predict realistic earthquake
performance of the dam. Hence, non-linear analyses are performed to estimate the essential performance of the dam for only
welded contact and shear stiffness of 1.8 MPa/m of the interface
element because the most critical tensile stresses are obtained in
these contact situations.
The maximum principle tensile stresses obtained from nonlinear analyses are entirely small from the maximum tensile
strength of the concrete in empty and full reservoir cases as
shown in Fig. 23. Non-linear analyses point out that principle
tensile stresses occurred in concrete slab are less than tensile
strength of the concrete during earthquake, so performance curve
is not required to draw. According to non-linear analysis results,
crack formation does not appear in the concrete slab; therefore
damage does not occur in concrete.

9. Conclusions

1.1

1.2

1.8 MPa

3.6 MPa

18 MPa

1800 MPa

Welded Contact

Acceptance Curve

1.3 1.4 1.5 1.6 1.7

Demand-Capacity Ratio

1.8

180 MPa

1.9

This paper presents the effect of the interface element, which

represents the friction contact in concrete slabrockll interface,
on the earthquake response and earthquake performance of Torul
CFR Dam considering damreservoir interaction. The reservoir
water is modeled using two-dimensional uid nite elements by
the Lagrangian approach. The DruckerPrager model is used in
non-linear time-history analyses.
The reservoir water has an obvious effect on the earthquake
response of the dam. According to linear analyses hydrodynamic
pressure increases the displacements and principle stresses and
this increase is more evident for low shear stiffness of the

Fig. 22. Performance assessment of the dam in full reservoir case.

5600

5600
Stress

4000

Stress

2400

D/C=2

Stress

4800
Stress (kN/m2)

(kN/m2)

4800

D/C=1

3200

1600

4000
3200
2400
1600

0
0

12

15

18

21

24

5600
D/C=1

D/C=2

Stress

4800
Stress (kN/m2)

Stress

15

18

21

24

5600
Stress

4800
(kN/m2)

12
Time (s)

Time (s)

2400

D/C=2

800

800

4000

D/C=1

3200

1600
800

D/C=1

D/C=2

4000
3200
2400
1600
800

0
0

12
Time (s)

15

18

21

24

12

15

18

21

24

Time (s)

Fig. 23. The principal tensile stress cycles according to non-linear analyses: (a) shear stiffness is 1.8 MPa/m in empty reservoir case; (b) Shear stiffness is 1.8 MPa/m in full
reservoir case; (c) welded contact in empty reservoir case and (d) welded contact in full reservoir case.

A. Bayraktar et al. / International Journal of Non-Linear Mechanics 46 (2011) 3546

interface element. The numerical results for higher shear stiffness

values are close to ones of the model including welded contact.
Earthquake performance assessment of Torul CFR Dam
indicates that signicant damages will occur in the concrete slab
according to linear time-history analyses for each reservoir case.
The hydrodynamic pressure has also considerable inuence on
the earthquake performance of the dam especially for low shear
stiffness of the interface element. The linear analysis results
indicate that hydrodynamic pressure increases the damage level.
So materially non-linear time-history analyses are performed and
analysis results refer that damage formation will not appear in
both reservoir cases.
As a consequence of this study, some suggestions may be
arranged as follows:

 The more realistic CFR dam models may be achieved

considering friction contact in the joints.

 In earthquake performance assessment of a CFR dam, interface

elements should be used in concrete slabrockll interface.

 The hydrodynamic pressure should be considered in earthquake performance analyses to obtain more critical results.

 The materially non-linear analyses should be performed to

evaluate reliable earthquake performance of a CFR dam.

Acknowledgement
The authors would like express heartfelt thanks to Dr. Yasemin
Bayram working at General Directorate of State Hydraulic
Works, 22, Regional Directorate, Trabzon, for her contributions
to this study.

References
[1] J.L. Sherard, J.B. Cooke, Concrete-face rockll damI. Assessment, and II.
Design, J. Geotech. Eng. 113 (10) (1987) 10961132.
[2] H.M. Westergaard, Water pressures on dams during earthquakes, Transactions, ASCE 98 (1933) 418433.
[3] C.N. Zangar, R.J. Haefei, Electric analog indicates effects of horizontal
earthquake shock on dams, Civil Eng. (1952) 5455.
[4] O.C. Zienkiewicz, B. Nath, Earthquake hydrodynamic pressures on arch
damsan electric analogue solution, Proc. Int. Civil Eng. Congr. 25 (1963)
165176.
[5] A.K. Chopra, Earthquake behavior of reservoirdam systems, J. Eng. Mech.
Div. 94 (1968) 14751500.
[6] W.D.L. Finn, E. Varoglu, Dynamics of gravity damreservoir systems, Comput.
Struct. 3 (4) (1973) 913924.
[7] S.S. Saini, P. Bettess, O.C. Zienkiewicz, Coupled hydrodynamic response of
concrete gravity dams using nite and innite elements, Earthquake Eng.
Struct. Dyn. 6 (4) (1978) 363374.
[8] A.K. Chopra, P. Chakrabarti, Earthquake analysis of concrete gravity dams
including damwaterfoundation rock interaction, Earthquake Eng. Struct.
Dyn. 9 (1981) 363383.
[9] E.J. Greeves, A.A. Dumanoglu, The Implementation Of An Efcient Computer
Analysis for FluidStructure Interaction using the Eulerian Approach within
SAP-IV, Department of Civil Engineering, University of Bristol, Bristol, 1989.
[10] A.C. Singhal, Comparison of computer codes for seismic analysis of dams,
Comput. Struct. 38 (1) (1991) 107112.
[11] Y. Calayr, A.A. Dumanoglu, A. Bayraktar, Earthquake analysis of gravity dam
reservoir systems using the Eulerian and Lagrangian approaches, Comput.
Struct. 59 (5) (1996) 877890.
[12] A. Bayraktar, A.A. Dumanoglu, Y. Calayr, Asynchronous dynamic analysis of
damreservoirfoundation systems by the Lagrangian approach, Comput.
Struct. 58 (5) (1996) 925935.

[13] A. Bayraktar, A.C. Altuns-k, B. Sevim, M.E. Kartal, T. Turker,

Near-fault ground
motion effects on the nonlinear response of damreservoirfoundation
systems, Struct. Eng. Mech. 28 (4) (2008) 411442.

[14] A. Bayraktar, A.C. Altuns-k, B. Sevim, M.E. Kartal, T. Turker,

Y. Bilici,
Comparison of near- and far-fault ground motion effect on the nonlinear
response of damreservoirfoundation systems, Nonlinear Dyn. 58 (4) (2009)
655673.
[15] A. Bayraktar, K. Haciefendioglu, M. Muvak, Asynchronous seismic analysis of
concrete-faced rockll dams including damreservoir interaction, Can. J. Civil
Eng. 32 (2005) 940947.

45

[16] F.B. Guros, G.R. Thiers, T.R. Wathen, C.E. Buckles, Seismic design of concretefaced rockll dams, in: Proceedings of the Eighth World Conference on
Earthquake Engineering, San Francisco, vol. 3, Prentice-Hall, Englewood Cliffs,
NJ, 1984, pp. 317323.
[17] G. Bureau, R.L. Volpe, W. Roth, T. Udaka, Seismic analysis of concrete face
rockll dams, in: Proceedings of the Symposium on Concrete Face Rockll
DamsDesign, Construction and Performance, Detroit, Michigan, ASCE, New
York, 1985, pp. 479508.
[18] H.B. Seed, R.B. Seed, S.S. Lai, B. Khamenehpour, Seismic design of concrete
faced rockll dams, in: Proceedings of the Symposium on Concrete Face
Rockll DamsDesign, Construction and Performance, ASCE, New York,
1985, pp. 459478.
[19] R. Priscu, A. Popovici, D. Stematiu, C. Stere, Earthquake Engineering for Large
Dams, Editura Academiei, Bucuresti and John Wiley & Sons, New York, NY,
1985.
[20] G. Han, X. Kong, J. Liu, Dynamic experiments and numerical simulations
of model concrete-face rockll dams, in: Proceedings of the Ninth
World Conference on Earthquake Engineering, Tokyo, Japan, vol. 2, 1988,
pp. 331336.
[21] G. Gazetas, P. Dakoulas, Seismic analysis and design of rockll damsstate of
the art, Soil Dyn. Earthquake Eng. 11 (1) (1992) 2761.
[22] N. Uddin, G. Gazetas, Dynamic response of concrete-face rockll dams to
strong seismic excitation, J. Geotech. Eng. ASCE 121 (2) (1995) 185197.
[23] S.M. Haeri, M. Karimi, Three-dimensional response of concrete faced rockll
dams to strong earthquakes considering damfoundation interaction and
spatial variable ground motion, in: First European Conference on Earthquake
Engineering and Seismology (a Joint Event of the 13th ECEE & 30th General
Assembly of the ESC), Geneva, Switzerland, 38 September, 2006, pp. 1406.
[24] Z. Ghannad, S. Malla, Dynamic analysis of concrete face rockll dams using
numerical and analytical methods, in: First European Conference on
Earthquake Engineering and Seismology (a Joint Event of the 13th ECEE &
30th General Assembly of the ESC), Geneva, Switzerland, 38 September,
2006, p. 649.
[25] B. Zhang, J.G. Wang, R. Shi, Time-dependent deformation in high concretefaced rockll dam and separation between concrete face slab and cushion
layer, Comput. Geotech. 31 (7) (2004) 559573.
[26] P.C.F. Ng, I.C. Pyrah, W.F. Anderson, Assessment of three interface elements
and modication of the interface element in CRISP 90, Comput. Geotech. 21
(4) (1997) 315339.
[27] C.Y. Dong, A simple benchmark problem to test frictional contact, Comput.
Meth. Appl. Mech. Eng. 177 (12) (1999) 153162.
[28] D.A. Karabatakis, T.N. Hatzigogos, Analysis of creep behaviour using interface
elements, Comput. Geotech. 29 (4) (2002) 257277.

[29] H. Karutz, R. Chudoba, W.B. Kratzig,

Automatic adaptive generation of a
coupled nite element/element-free Galerkin discretization, Finite Elem.
Anal. Des. 38 (11) (2002) 10751091.
[30] H. Shakib, A. Fuladgar, Dynamic soilstructure interaction effects on the
seismic response of asymmetric buildings, Soil Dyn. Earthquake Eng. 24 (5)
(2004) 379388.
[31] D.V. Oliveira, P.B. Lourenc-o, Implementation and validation of a constitutive
model for the cyclic behaviour of interface elements, Comput. Struct.
82 (1719) (2004) 14511461.
[32] R.E. Goodman, R.L. Taylor, T.L. Brekke, A model for the mechanics of jointed
rock, J. Soil Mech. Found. Div., ASCE 94 (3) (1968) 637659.
[33] O.C. Zienkiewicz, Analysis of nonlinear problems in rock mechanics with
particular reference to jointed rock systems, in: Proceedings of the Second
Congress International Society for Rock Mechanics, Belgrade, Yugoslavia,
1970.
[34] M.A. Mahtab, R.E. Goodman, Three-dimensional nite element analysis of
jointed rock slopes, in: Proceedings of Second Congress of the International
Society of Rock Mechanics, Belgrade, vol. 3, 1970, pp. 353360.
[35] J. Ghaboussi, E.L. Wilson, J. Isenberg, Finite element for rock joints and
interfaces, J. Soil Mech. Found. Div. ASCE 99 (10) (1973) 833848.
[36] A.D. Tzamtzis, B. Nath, Application of a three-dimensional interface element
to nonlinear static and dynamic nite element analysis of discontinuous
systems, Eng. Syst. Des. Anal. Conf., ASME 1 (1992) 219222.
[37] G. Formica, V. Sansalone, R. Casciaro, A mixed solution strategy for the
nonlinear analysis of brick masonry walls, Comput. Meth. Appl. Mech. Eng.
191 (5152) (2002) 58475876.
[38] J.P.M. Gonc- alves, M.F.S.F. de Moura, P.M.S.T. de Castro, A three-dimensional
nite element model for stress analysis of adhesive joints, Int. J. Adhes.
Adhes. 22 (5) (2002) 357365.
[39] K.S. Lee, M.M. Zaman, Interface behavior of circular subway tunnels in
layered discontinuous geological media, Jurnal Institusi Jurutera Malaysia 38
(1986) 1420.
[40] A.S. Al-Homoud, R.V. Whitman, Seismic analysis and design of rigid bridge
abutments considering rotation and sliding incorporating non-linear soil
behavior, Soil Dyn. Earthquake Eng. 18 (4) (1999) 247277.
[41] K. Toki, T. Sato, F. Miura, Separation and sliding between soil and structure
during strong ground motion, Earthquake Eng. Struct. Dyn. 9 (1981)
263277.
[42] K. Toki, F. Miura, Non-linear seismic response analysis of soilstructure
interaction systems, Earthquake Eng. Struct. Dyn. 11 (1983) 7789.
[43] C.C. Hird, C.M. Kwok, Finite element studies of interface behavior in
reinforced embankments on soft ground, Comput. Geotech. 8 (2) (1989)
111131.

46

A. Bayraktar et al. / International Journal of Non-Linear Mechanics 46 (2011) 3546

[44] S.H. Nam, H.W. Song, K.J. Byun, K. Maekawa, Seismic analysis of underground
reinforced concrete structures considering elasto-plastic interface element
with thickness, Eng. Struct. 28 (8) (2006) 11221131.
[45] N. Uddin, A single-step procedure for estimating seismically-induced displacements in earth structures, Comput. Struct. 64 (56) (1997) 11751182.
[46] N. Uddin, A dynamic analysis procedure for concrete-faced rockll dams
subjected to strong seismic excitation, Comput. Struct. 72 (13) (1999) 409421.
[47] Swanson Analysis Systems Inc., Houston, PA, USA, 2008.
[48] E.L. Wilson, M. Khalvati, Finite elements for the dynamic analysis of
uidsolid systems, Int. J. Numer. Meth. Eng. 19 (11) (1983) 16571668.
[49] Y. Calayr, Dynamic analysis of concrete gravity dams using the Eulerian and
the Lagrangian approaches, Dissertation, Karadeniz Technical University,
1994 (in Turkish).
[50] O.C. Zienkiewicz, R.L. Taylor, The Finite Element Method, McGraw-Hill, 1989.
[51] K.J. Bathe, Finite Element Procedures in Engineering Analysis, Prentice-Hall,
Englewood Cliffs, NJ, 1996.
[52] R.W. Clough, J. Penzien, Dynamics of Structures, second ed., McGraw-Hill,
Singapore, 1993.
[53] N. Akkas, H.U. Akay, C. Ylmaz, Applicability of general-purpose nite element
programs in soliduid interaction problems, Comput. Struct. 10 (5) (1979)
773783.

[54] D.C. Drucker, W. Prager, Soil mechanics and plastic analysis of limit design, Q.
Appl. Math. 10 (2) (1952).
[55] W.F. Chen, E. Mizuno, Nonlinear Analysis in Soil Mechanics, Elsevier,
New York, 1990.
[56] A.L.G.A. Coutinho, M.A.D. Martins, R.M. Sydenstricker, J.L.D. Alves, L. Landau,
Simple zero thickness kinematically consistent interface elements, Comput.
Geotech. 30 (2003) 347374.
[57] C.S. Desai, M.M. Zaman, J.G. Lightner, H.J. Siriwardame, Thin layer element for
interface and joints, Int. J. Numer. Anal. Math. Geomech. 8 (1984) 1943.
[58] Y. Ghanaat, Seismic performance and damage criteria for concrete dams, in:
Proceedings of the Third USJapan Workshop on Advanced Research on
Earthquake Engineering for Dams, San Diego, California, 2002.
[59] USACE, US, Army Corps of Engineers, Time History Dynamic Analysis of
Concrete Hydraulic StructuresEngineering and Design (Engineer Manual),
EM 1110-2-6051, 2003.
[60] DSI, General Directorate of State Hydraulic Works, /http://www.dsi.gov.tr/
english/S, accessed 10 June 2009.
[61] TS 500, Turkish Standard, Requirements for Design and Construction of
Reinforced Concrete Structures (2000).
[62] PEER, Pacic Earthquake Engineering Research Centre, /http://peer.berkeley.
edu/smcat/dataS, accessed 10 June 2009.