WP2/TG4

ANALYSIS OF CONCRETE GRAVITY DAM USING THE

FINITE ELEMENT PROGRAM LUSAS

by Dr. A.D. Jefferson and S.C. Hee

Cardiff University
+44-29-2087-5697
e-mail: [email protected]

1
 INTRODUCTION

The document provides details of the analysis of concrete gravity dam using the finite element
program LUSAS (FEA Ltd, 2002), for the purpose of providing a draft template for the
comparison of codes under Task 2.4 of the IALAD contract. The results obtained from the
analysis will be compared with those from other FE programs. The model of the concrete
gravity dam chosen in this analysis was that used by Linsbauer and Bhattacharjee (1999), in
their studies on the effect of uplift pressure on dam safety. The gravity dam has a height of 80m
and the base is 60m wide, as illustrated in Figure 1.

80 m

0m

60 m

Figure 1: Dam cross-section

FINITE ELEMENT ANALYSIS

This section describes the assumptions, elements, loading cases, boundary conditions, and
solution strategy employed in the series of non-linear analyses.

Assumptions

In the present study, rock was assumed to be isotropic elastic and concrete initially isotropic.

2
Description of Finite Element Mesh

The model consists of 781 elements with a total of 4932 degrees of freedom. The gravity dam
was modelled with 8 noded, quadrilateral, isoparametric elements. As for the rock mass, a
combination of 8 noded quadrilateral and 6 noded triangular elements were used in the two
dimensional plane strain analysis.

Figure 2: Mesh configuration

Material Properties

Table 1 shows the material properties of dam and rock used in the analysis.

Material parameters Concrete Rock

Young’s modulus (MPa) 24000 41000
Poisson’s ratio 0.15 0.1
Mass density (kg) 2400 2200
Compressive strength (MPa) 24 40
Tensile strength (MPa) 1.5 2600
Strain at peak comp. strength 0.0022 -
Strain at end of comp. softening curve 0.1 -
Fracture energy (N/m) 100 200

Table 1: Properties of dam and rock materials

3
Boundary Conditions

A compromise between computational efficiency and solution accuracy was obtained by

limiting the total area of rock mesh underlying the gravity dam. The boundaries of the rock
mesh were specified at a distance of 120m from each side of the dam surfaces, and at a depth of
80m from the base of the dam. The degrees of freedom of each node along the edges of the
mesh were constrained in the analysis, i.e. fixed in both horizontal and vertical directions.

Description of Loads

In the analysis, apart from the body force (Loadcase 1) of the overall structure, horizontal
hydraulic pressures (Loadcase 2) were applied on the upstream face of the concrete dam. The
distribution of water pressure varies linearly throughout the depth of the concrete dam, as
depicted in Figure 3.

Water pressure distribution

Figure 3: Hydraulic pressure distribution at upstream of dam

Material Model

A multiple surface plasticity model for concrete was used in the analysis, which is available in
the commercial version of the LUSAS code. A full description of the model is available
elsewhere (Lusas, 2002; Jefferson, 2002). The model uses a local curved Coulomb yield
criterion to simulate directional fracture and isotropic compressive behaviour. A directional-
isotropic transition function was derived from experimental data to distinguish compressive
crushing behaviour from tensile softening behaviour. The local responses from all active plastic
surfaces were coupled in a multi-surface plasticity formulation to provide the interaction
between compressive and tensile behaviour.

4
 NONLINEAR SOLUTION PROCEDURE

This section describes the salient features of the solution process.

Solution Control Parameters

Before running the analysis, important ‘control’ parameters were required as input to define and
adjust the algorithm during the solution process. These parameters were specified to define the
iteration frequency, convergence criteria, initial load factor, and load increment for each of the
analysis cases.

Load Increment and Solution Strategy

The body forces were applied separately from the hydraulic load, with an initial load factor of
1.0. As for the latter load case, an initial load factor of 0.2 with a maximum load increment of
0.1 for subsequent iterations were selected in the analysis, although an automatic load factor
adjustment is applied if the number of iterations to achieve convergence exceeds the target
number. The Newton-Raphson iteration method was used to achieve equilibrium between the
load increment applied and the internal nodal forces produced in the structure.

Convergence Criterion

The convergence criteria used in the iteration process was based on the sum of the squares of all
the residual forces, normalised by the sum of the squares of all the external forces. In this
analysis, a load criterion set at 1.0% produced an acceptable balance between solution
performance and numerical accuracy.

5
 RESULTS

Deformed Mesh

The deformed configuration of the dam and rock is shown in Figure 4.

Figure 4: Mesh deformation at water level of 80m (Magnification factor = 500)

Crest Displacement

Figure 5 shows the crest displacement responses with increasing water level.

85

80
Water level (m)

75

70

65

60

55
4 6 8 10 12 14 16 18
Crest displacement (mm)

Figure 5: Variation of crest displacement with water level

6
Stress Distribution

Figure 6 and 7 show the maximum and minimum principal stress distributions at a water level
of 60m and 80m respectively.
RESULTS FILE = 1
STRESS
2
CONTOURS OF SMax (N/m )

-201.298E3
-100.649E3
0
100.649E3
201.298E3
301.947E3
402.596E3
503.245E3
603.894E3
704.543E3
805.192E3
905.842E3
1.00649E6
1.10714E6
1.20779E6
1.30844E6
Max 0.1378E+07 at Node 125
Min -0.2320E+06 at Node 3683

(a)
RESULTS FILE = 1
STRESS
2
CONTOURS OF SMin (N/m )

-1.20486E6
-1.09533E6
-985.793E3
-876.26E3
-766.728E3
-657.195E3
-547.663E3
-438.13E3
-328.598E3
-219.065E3
-109.533E3
0
109.533E3
219.065E3
328.598E3
438.13E3
Max 0.4516E+06 at Node 1247
Min -0.1301E+07 at Node 2

(b)

Figure 6: (a) Maximum and (b) minimum principal stress contour at a water level of 60m

7
 STRESS
2
CONTOURS OF SMax (N/m )

-508.129E3
-381.097E3
-254.064E3
-127.032E3
0
127.032E3
254.064E3
381.097E3
508.129E3
635.161E3
762.193E3
889.225E3
1.01626E6
1.14329E6
1.27032E6
1.39735E6
Max 0.1500E+07 at Node 125
Min -0.5325E+06 at Node 3683

(a)

RESULTS FILE = 1
STRESS
2
CONTOURS OF SMin (N/m )

-2.91038E6
-2.66785E6
-2.42532E6
-2.18278E6
-1.94025E6
-1.69772E6
-1.45519E6
-1.21266E6
-970.127E3
-727.595E3
-485.063E3
-242.532E3
0
242.532E3
485.063E3
727.595E3
Max 0.7960E+06 at Node 3739
Min -0.3085E+07 at Node 2

(b)

Figure 7: (a) Maximum and (b) minimum principal stress contour at a water level of 80m

8
Crack Propagation

Figure 8 depicts the formation of crack at the vicinity of the upstream dam toe.

(a)

(b)

(c)

Figure 8: Formation of crack at the upstream dam toe area at water level of (a) 60m, (b) 70m,
and (c) 80m

9
 COMMENTS ON THE ANALYSIS

In the present analysis, the interaction between concrete and rock at the base of the dam was
idealised as a fully connected interface, but the accuracy of the results could be improved by
assigning joint or contact elements to the interface between the two materials. Despite this, the
program correctly predicts the position at which crack is most likely to form and the direction of
crack propagation.

REFERENCES

FEA Ltd. (2002), LUSAS User and theory manuals version 13.4.

Jefferson A. D. (2002), Local plastic surfaces for cracking and crushing in concrete, Proc. Instn.
Mech. Engrs. Materials Design and application, 216, p257-266.

Linsbauer H.N. and Bhattacharjee S. (1999), Dam safety assessment due to uplift pressure
action in a dam-foundation interface crack, Fifth Benchmark Workshop on Numerical Analysis
of Dams, Denver, Colorado USA.

10