Modal Behavior of Dam Reservoir

Foundation System

Tahar Berrabah, A.
Civil Engineering Department, Aboubekr Belkaid University, Tlemcen Algeria
[email protected]

Belharizi, M.
Consultant, 37 impasse Armand, 92160 Antony, France
[email protected]

Bekkouche, A.
Civil Engineering Department, Aboubekr Belkaid University, Tlemcen Algeria
[email protected]

ABSTRACT
Gravity dams are fluid – structure – soil interaction problems. It is obvious that the
foundation soil and water reservoir affect the dynamic response of gravity dams during
earthquakes. A parametric study was performed in the present work to view the combined
effect of soil foundation, water reservoir presence and fluid-dam and fluid-soil interfaces
modelling on the modal behaviour of “Brezina” concrete dam situated at Algeria.
Two assumptions were adopted to model the dam-fluid and soil- fluid interfaces: the
contact elements and the coupling equation. Also three approaches are adopted to
investigate the soil-structure interaction phenomenon: the fixed support foundation, the
massless soil foundation and the mass soil foundation. As conclusion it was shown that
modelling the interfaces by contact elements or by coupling equations gives the same results
and that water reservoir presence and the soil foundation modelling produce a decrease of
the modal system frequencies.

KEYWORDS: Soil-structure interaction, modal behaviour, fluid-structure interaction.

- 1593 -
 Vol. 16 [2011], Bund. T 1594

INTRODUCTION
The hydrodynamic pressure acting on dam faces during earthquakes has been recognized as a
major loading in the design of dams. The first rigorous analysis of hydrodynamic forces on dam
faces during earthquakes was reported by westergaard in 1933 (Bang-Fun Chen, 1996).
The dam reservoir interaction problems can be analysed using the three famous approaches:
Westergaard approach: the dynamic effect of the reservoir is modelled as added masses.
Eulerian approach: since in this approach the displacements are the variables in the structure
and the pressures are the variables in the fluid, a special purpose computer program is required
for the solution of the coupled systems.
Lagrangian approach: In this approach the behavior of the fluid and structure is expressed in
terms of displacements. For that reason, compatibility and equilibrium are automatically satisfied
at the nodes along the interfaces between the fluid and structure. This makes a Lagrangian
displacement based fluid finite element very desirable; it can be readily incorporated into a
general purpose computer program for structural analysis, because special interface equations are
not required.
Alemdar Bayraktar and all in 2005, examinated the effect of the base rock characteristics on
the stochastic dynamic response of dam-reservoir-foundation systems subjected to earthquake
forces using the Lagrangian approach. The fluid was assumed to be inviscid in this study. The
normal components of the displacements of the reservoir dam and reservoir foundation interfaces
are to be continuous. This condition is accomplished by using short and axially almost rigid truss
element in the normal direction of the interfaces (Alemdar et al., 2005).
A study of seismic behaviour of a high arch dam with dynamic interaction with reservoir and
foundation based on the wave propagation in a non-uniform and local non-linear medium by
using an explicite finite element method with transmitting boundaries was presented by Chen
Houqun, Du Xiuli and Hou Shunzai in 1998. A finite element mesh automatically generated in
the program and the dam foundation system was modelled by 1364 8-node 3D solid elements. For
simplicity, the dam reservoir interaction was considered through added mass with neglect the
compressibility of the reservoir water (Chen Houqun et al., 1998).
Lot of other research treated the soil -fluid -structure interaction problems are available in the
literature, among these we can cite (M. Akkose et al., 2007) (F.Guan and I.D. Moore 1997)
(Hashem Shariatmadar and Adel Mirhaj 2009; Vahid Lotfi, 2003; Mehmet Akkose et al., 2004;
B. Poursartip and V. Lotfi, 2008).

Fluid –Structure System Formulation

The equations of motion for a fluid system have the similar form to that of the structure when
the Lagrangian approach is used. But requires a different sensitivity to determine interface
condition of the coupled system. At the interface of the fluid-structure system, only the
displacements in the direction normal to the inteface are assumed to be compatible in the
structure as well as the fluid. The fluid is assumed to be lineair-elastic and inviscid.
 Vol. 16 [2011], Bund. T 1595

Constraint Equations and Boundary Conditions

In order to satisfy the continuity conditions between the fluid and solid media at the
boundaries, the nodes at the common lines of the fluid and the plane elements are constrained to
be coupled in the direction normal to the interface, while relative movements are allowed to occur
in the tangential directions, this is implemented by two methods:
First: by the use of contact elements available in the ANSYS software (CONTA 172)
Second: By attaching the coincident nodes at the common lines of the fluid and the plane
elements in the normal direction.

Ansys Validation
The model is to be developed from the elements available in the ANSYS software, the
validity of the software via the fluid-structure interaction problem is judged by an example about
a rectangular liquid container for the two approaches of fluid-structure interfaces modelling.
The liquid in an open container can oscillate at discrete natural frequencies. The liquid
sloshing, at the lowest of these frequencies is of concern in this study; since the container natural
sloshing frequency is to be tuned to the structure system natural frequency.
From the linear wave theory, the natural frequencies of the first sloshing modes are calculated
as follows (Robert D. Blevins 2001):

Figure 1: Reservoir geometry scheme

1 g i π hi
fi  tanh (1)
2 π a a
where :
 Vol. 16 [2011], Bund. T 1596

g= Acceleration due to gravity ;

h= Depth of liquid in basin ;
a= Basin length ;
i= 0, 1, 2….etc
For the present example : a= 25 m and h=5 m (Figure 1)

1st Mode shape

2 nd Mode shape

3rd Mode shape 4th Mode shape

Figure 2: The four first mode shapes for the rectangular reservoir example

Table 1 (on the next page) shows the consistence of the results between the theoretical and
the finite element results (using the two approaches of fluid-structure modelling).
 Vol. 16 [2011], Bund. T 1597

Table 1: The first four modal frequencies with their corresponding effective masses
for the two approaches of interfaces modelling.

Mode Theoretical Coupling effective Contact effective

coupling contact
Number frequency mass (kg) mass (kg)

1 0,131877274 0,133254 94050,3 0,133254 94050,4

2 0,230453498 0,232988 3,36E-19 0,232988 1,37E-11

3 0,299155572 0,302191 6993,66 0,302191 6993,67

4 0,351190513 0,354286 3,62E-15 0,354286 3,16E-11

MATERIALS AND METHODS

The objective of this work is to study the effects of dam-reservoir-foundation interaction on
modal behaviour of gravity dams. For this reason, Brezina concrete Arch dam situated at the west
of Algeria is chosen as a case study. The dam is 60 m high, its maximum arch length is 78.5 m
and its thickness varies from 5m at the crest to 36.3 m at the foundation level.
It is important to note that this dam is named Arch dam but it have no arch effect in it
behaviour, for this reason a plane strain behaviour is considered for the medium cross section.
Hence, 2-D finite element models are created using ANSYS program. Water is also treated as
compressible fluid. For simplicity, no absorption is considered at reservoir bottom.
Since the extent of the reservoir is large, it is necessary to truncate the reservoir at a
sufficiently large distance from the dam. A length of reservoir equivalent to three times its depth
is appropriate for adequate representation of hydrodynamic effects on the dam, a length of 150 m
on each side in X direction is judged sufficient after doing a sensitivity analysis, and the same
length is chosen as extension of soil foundation. The nodes representing the extreme side of the
reservoir were free to displace in the vertical direction only. The depth of the foundation is taken
as 100 m. Plane strain conditions are taken into account in the calculations. The dam material is
assumed to be linear-elastic, homogeneous and isotropic.
A two dimensional (2D) finite element model with 1521 nodes and 1394 plane elements
(PLANE 42) is used to model Brezina dam with surrounding soil.
A two dimensional (2D) finite fluid element model with 273 nodes and 239 plane elements
(FLUID 79) is used to model the reservoir water.
The soil (foundation) dam interface was modeled by coupling the interface nodes in all
directions. However two methods are used to model the fluid soil / or structure (dam) interface:
First one: using contact elements;
 Vol. 16 [2011], Bund. T 1598

Contact elements generated by ANSYS software were applied at the interface soil-concrete
dam considering the soil and the dam as target areas (TARGE 169) and water as contact area
(CONTA 172).
Second: using coupling equations; by coupling the interface nodes in the normal direction
only and let it free in all other directions.
Three assumptions of soil foundation models are suggested using the direct method; the fixed
support foundation, the mass soil foundation to take into account the inertial effect of the
foundation and the massless soil foundation to neglect it.
Figure 3 shows the dam with fixed support without water and the dam with fixed support
with full reservoir water. However, Figure 4 shows dam with soil without water and dam with
soil and full reservoir water.
To investigate the modal behaviour of the dam, nine different cases are taken as followings:
Dam with fixed support; Empty reservoir (case 1)
Dam with fixed support; Full reservoir; fluid dam interface modeled by contact elements
(case 2)
Dam with fixed support; Full reservoir; fluid dam interface modeled by coupling equations
(case 3)
Dam with massless soil foundation; Empty reservoir (case 4)
Dam with massless soil foundation; Full reservoir; fluid dam and fluid soil interfaces
modeled by contact elements (case 5)
Dam with massless soil foundation; Full reservoir; fluid dam fluid soil interface modeled by
coupling equations (case 6)
Dam with mass soil foundation; Empty reservoir (case 7)
Dam with mass soil foundation; Full reservoir; fluid dam fluid soil interface modeled by
contact elements (case 8)
Dam with mass soil foundation; Full reservoir; fluid dam fluid soil interface modeled by
coupling equations (case 9)

Figure 3: Finite element modeling of dam-reservoir interaction (Soil modeled as fixed

support)
 Vol. 16 [2011], Bund. T 1599

Figure 4: Finite element modeling of dam reservoir foundation interaction (foundation

modeled as mass soil/massless soil)

The material properties for the concrete arch dam, foundation soil and reservoir water are
presented by Table 2, Table 3 and Table 4 respectively.

Table 2: Material properties of Brezina dam

Young Modulus
Poisson’s Ratio Density (kg/m3)
(N/m2)

28.5E+09 0.2 2500

Table 3: Soil properties of Brezina dam foundation

Young Modulus
Poisson’s Ratio Density (kg/m3)
(N/m2)

14.5E+09 0.1 2100

Table 4: Reservoir water properties

Compressibility Modulus
Poisson’s Ratio Density (kg/m3)
(N/m2)

2.068E+09 0.49 1000

 Vol. 16 [2011], Bund. T 1600

RESULTS AND DISCUSSION

Effect of water reservoir and interface modeling
Case fixed support foundation
Frequency Value (Hz)

Case 1
Case 2
Case 3

Mode Number

Figure 5: Combined effect of reservoir water and fluid dam/or soil modeling
on the modal frequencies of the soil-dam-fluid system for the case of
fixed support foundation model

The Figure 5 represents for the case of fixed support foundation the effect of the reservoir
water presence and the interface modelling on modal frequencies of the dam object of this study.
As mentioned above, the case 1, case 2 and case 3 represent respectively the system without
water effect, the system with full reservoir and interfaces modelled by contact element and finally
the case of the system with full reservoir and interfaces modelled by coupling equations.

It is clear that the presence of water leads to a decrease of the modal frequencies of the dam
and that modelling the interface fluid structure by coupling equations or using contact elements
give the same results.
 Vol. 16 [2011], Bund. T 1601

Massless soil foundation

Frequency Value (Hz)

case 4
case 5
case 6

Mode number

Figure 6: Combined effect of reservoir water and fluid dam/or soil modelling on the
modal frequencies of the soil-dam-fluid system for the case of massless soil foundation
model

Soil foundation with mass

Frequency Value (Hz)

case 7
case 8
case 9

Mode number

Figure 7: Combined effect of reservoir water and fluid dam/or soil modelling on the
modal frequencies of the soil-dam-fluid system for the case of mass soil foundation
model
 Vol. 16 [2011], Bund. T 1602

Figures 5, 6 and 7 show the same results but for the case of massless soil foundation and mass
soil foundation respectively.

Effect of soil structure interaction modelling on the modal

frequencies of soil -fluid-structure system
Empty reservoir
Frequency value (Hz)

Case 1
Case 4
Case 7

Mode number

Figure 8: Effect of soil structure interaction modelling on the modal frequencies of

the dam-soil system.

Figure 8 represents the modal frequencies for case1, case 4 and case 7 which represent
respectively the case of dam with fixed support with empty reservoir, the case of dam with
massless soil foundation with empty reservoir and the case of dam with mass soil foundation with
empty reservoir.

The Figure 8 shows that adding soil to dam in the modelling which means taking into account
the soil structure interaction effect leads to a decrease in the modal frequencies values
furthermore for the mass soil foundation model.
 Vol. 16 [2011], Bund. T 1603

Full reservoir (fluid dam/or soil interface modelled by contact element)

Frequency value (Hz) case 2

case 5
case 8

Mode number

Figure 9: Effect of soil structure interaction modelling on the modal frequencies of the
dam-soil-fluid system (interfaces modelled by contact elements).

Figure 9 represents the modal frequencies for case 2, case 5 and case 8 which represent
respectively the case of dam with fixed support, the case of dam with massless soil foundation
and the case of dam with mass soil foundation with fluid dam and fluid soil interface modelled by
contact elements.

Figure 10 represents the modal frequencies for case 3, case 6 and case 9 which represent
respectively the case of dam with fixed support with, the case of dam with massless soil
foundation and the case of dam with mass soil foundation with fluid dam and fluid soil interface
modeled by coupling equations.

From Figure 9 and Figure 10 it’s clear that soil structure interaction modelling has the same
effect on the modal frequencies for both cases of interface modelling; with means that taking into
account the soil structure interaction effect leads to a decrease in the modal frequencies even if
the reservoir water is present and for interface modelled as contact element or by coupling
equations.
 Vol. 16 [2011], Bund. T 1604

Full reservoir (fluid dam/or soil interface modelled by coupling

equations)

Frequency value (Hz)

case 3
case 6
case 9

Mode number

Figure 10: Effect of soil structure interaction modelling on the modal frequencies of
the dam-soil-fluid system (interfaces modelled by coupling equations).

CONCLUSIONS
A parametric study was performed in the present work to investigate the combined effect of
soil foundation, water reservoir presence and fluid dam / fluid soil interfaces modelling on the
modal behaviour of Brezina concrete dam.

The study had shown that reservoir water presence and soil foundation modelling have the
same effect on the modal frequencies results which means on the modal behaviour of the concrete
dam, they decrease the frequencies values.

Two assumptions were adopted in this study in the subject of the interface fluid-dam and
fluid-soil modelling: the contact elements and the coupling equation. As conclusion it was shown
that modelling the interfaces by contact element or by coupling equations give the same results.

REFERENCES
1. Alemdar Bayraktar, Ebru Hancer, Mehmet Akkose (2005) ”Influence of base-rock
characteristics on the stochastic dynamic response of dam-reservoir-foundation systems”. J.
Engineering structures, V. 27, 1498-1508
2. B. Poursartip, V. Lotfi (2008) “Modal analysis of concrete arch dams in time domain
including dam-reservoir interaction”. 14th world conference on earthquake Engineering.
3. Bang- Fun Chen (1996) “Nonlinear hydrodynamic effects on concrete dam”. J. Engineering
structures , V18, 201-212
 Vol. 16 [2011], Bund. T 1605

4. Chen Houqun, Du Xiuli, Hou Shunzai (1998) “Application of Transmitting Boundaries to

Non Linear Dynamic Analysis of an Arch Dam-Foundation-Reservoir System”. J. Elsevier
Science B.V, V. 83, 115-124
5. F.Guan, I.D. Moore (1997) “New techniques for modelling reservoir-dam and foundation-
dam interaction”. J. Soil dynamics and Earthquake Engineering, V.16, 285-293
6. Hashem Shariatmadar, Adel Mirhaj (2009) ”Modal response of dam reservoir foundation
interaction”. 8th International congress on civil Engineering.
7. M. Akkose, A. Bayraktar, A.A. Dumanoglu (2007) “Reservoir water level effects on
nonlinear dynamic response of arch dams”. J. Journal of fluids and structures, V. 27, 418-435
8. Mehmet Akkose, A. Aydin Dumanoglu, M. Emin Tuna (2004) “Investigation of
Hydrodynamic Effects on Linear and Nonlinear Earthquake Responses of Arch Dams by the
Lagrangian Approach”. J. Turkish J. Eng. Env. Sci. V 28, 25-40
9. Robert D. Blevins (2001) “Formulas for natural frequency and mode shape”. Van Nostrand
reinhold Company, New York, Book
10. Vahid Lotfi (2003) “Seismic analysis of concrete gravity dams by decoupled modal approach
in time domain”. J. EJSE International V.3.

© 2011 ejge