Escolar Documentos
Profissional Documentos
Cultura Documentos
by
Victor E. Saouma1 and Yoshihisa Uchita2 and Yoshinori Yagome3
ABSTRACT
Concrete dams, by virtue of the lack of reinforcement, and the potential for well predefined cracks
(along lift joints and the rock/concrete interfaces,
or in-between monoliths, not to mention within the
solid concrete or rock mass) are indeed prime candidates for fracture mechanics. This paper will highlight laboratory tests, numerical models, validation
experiments, and case studies. Numerous laboratory
tests have been conducted to characterize fractured
concrete under varying conditions (such as dynamic
load, effect of water pressure inside the crack, cyclic
loads) and will be reported. Other laboratory tests
were conducted on a concrete dam model mounted
on a shaking table inside a centrifuge to assess the
accuracy of our numerical predictions. Finally, a numerical model (Merlin) for the nonlinear (implicit
and explicit) 3D nonlinear transient analysis of concrete dams has been under continuous development
for the past 15 years. The computational challenges
encountered will be highlighted.
It will also be pointed out that, though nonlinear
fracture mechanics has an important role to play in
the seismic safety assessment of a concrete dam, it
is only one major player amongst many other important ones whose participation is equally essential to
the comprehensive analysis.
1 INTRODUCTION
Japan, which is geographically located on the intersection of three major faults, is one of the most seismically active regions in the world and as such has
taken a leadership role in research and development
to mitigate the destructive forces unleashed by a seismic tremor. Recognizing the computational power
currently available, the advances in various related
1
discontinuity in a dynamic analysis, this model indeed captures the nonlinear response of the joint
through crack opening and normal stress distributions which account for possible (small) cohesive
tensile stresses, Figure 5.
Our focus on discrete crack may have started
with (Saouma 1980), really took roots in modern
plasticity approach through the model developed in
(Cervenka, Chandra and Saouma 1998) which was
extensively used for actual nonlinear dam analysis,
and culminated most recently with the extension of
our previous model to account for reverse cyclic
loads, (Puntel, Bolzon and Saouma 2006).
During an earthquake, joints open and cracks
may form. Currently, there is much controversy as to
whether water can flow into the newly formed crack,
and what the dynamic uplift is. There is some strong
experimental evidence that during crack opening and
closure, we have respectively a decrease or an increase of uplift pressure. Hence during crack propagation, not only is the new crack surface subjected
to an internal uplift pressure, but the magnitude of
the pressure is a function of the rate of crack opening
displacements (as a first approximation), Fig. 6.
Earthquake records (accelerograms) are usually
measured or determined at the surface. Yet the analysis requires modeling the rock foundation (and its
mass), at the base of which accelerations are applied.
Again, in simpler analyses the surface acceleration
is simply applied at the base; however a more rigorous approach would require deconvolution of the
signal under the assumption of a linear elastic rock
foundation. This is done by first applying the (surface) recorded signal at the base, through a preliminary finite element analysis compute the induced acceleration on the surface, transfer those two accelerations from the time to the frequency domain, and
compute the transfer function. In the general three
dimensional case, there will be 9 transfer functions
which constitute the convolution matrix. The inverse
of that matrix times the input accelerations will give
the deconvoluted one. Hence, we the deconvoluted
one is applied at the base of the foundation, the induced acceleration at the surface will be nearly identical to the one recorded, Fig. 7.
Whereas many simpler analyses ignore the foundation mass, this must be accounted for in a rigorous
nonlinear analysis. A major complication becomes
ble (and hence most analyses were linear and performed in the frequency domain) , this is clearly no
longer the case. With modern computers, time history analysis of even small dams is becoming the
norm. Time integration can be either implicit or explicit. Dynamic crack opening and closures can induce numerical problems, and integration schemes
may have to be refined. In Fig. 9 it is apparent that the
both the Newmark and Hughes methods yield
identical results as long as the discrete crack is not
activated, when it is Hughes time integration method
is more suited to dissipate the high frequency content of the response. More recently, we have modified our code to accommodate explicit (which being
conditionally stable requires very small time steps)
method. A major advantage of the explicit time integration is that the global structural stiffness matrix need not be assembled, and hence the code
can be relatively easily parallelized to run on multiple CPUs. This was done through the MPI library,
whereas we used METIS for domain decomposition.
It is our strong expectation that complex 3D nonlinear analysis of dam-foundation-reservoir system can
most effectively be analyzed in parallel on a network
of computers connected by 1 GB Ethernet network.
A special purpose 2D-3D finite element mesh generator, Kumo-no-su, for concrete dams with the capability of supporting all the above features was developed, (Saouma 2007).
4 ANALYSIS
numerical values on the top), through a graphical user interface. Furthermore, the user can
select two accelerograms and seek the evaluation of the transfer function between the two
signals (FFT and transfer function all being internally computed), or plot accelerations and
FFTs, Fig. 11.
7 CONCLUSIONS
Historically, Dam Engineering has constituted some
of the most complex and challenging engineering
problems. Hence, solutions initially developed for
dams were subsequently extended to other Engineering disciplines. Furthermore, given what is at stake,
5
processor; users manual, Technical report, Rewe no longer can satisfy ourselves with simple engiport Submitted by the University of Colorado
neering approaches, instead we must take advantage
to the Tokyo Electric Power Service Company.
of the latest developments and lead the way in Civil
http://civil.colorado.edu/~saouma/Spider.
Engineering research and development.
Nonlinear fracture mechanics plays a crucial role
in the seismic safety assessment of concrete dams, Miura, F. and Okinaka, H.: 1989, Dynamic analysis method for 3d soil-structure interaction sysit is by far the largest source of nonlinearity. Joints
tems with the viscous boundary based on the
are present between monoliths, between the rock and
principle of virtual work, Japanese Journal of
the concrete and potentially in the rock foundation as
Civil Engineering pp. 395404.
joints or in the dam as cracks. Nevertheless, it should
be made clear that as important as fracture mechan- Puntel, E., Bolzon, G. and Saouma, V.: 2006, A fracics is, it is only one player amongst many others
ture mechanics based model for joints under
which should also be accounted for.
cyclic loading, ASCE J. of Engineering Mechanics 132(11), 11511159.
8 ACKNOWLEDGMENTS
The research reported in this paper could not have Saouma, V.: 1980, Finite Element Analysis of Reinforced Concrete; a Fracture Mechanics Apbeen possible without the precious assistance of nuproach, PhD thesis, Cornell University, Departmerous collaborators: Prof. Bruhwiler, Prof. Slowik,
ment of Structural Engineering.
Dr. Puntel, and particularly Dr. Cervenka. Their contributions is gratefully acknowledged. The extensive
Saouma, V.: 2007, KumoNoSu, a 3d interacsoftware development was made possible through
tive graphics mesh generator for merlin;
the invaluable assistance of Mr. Takashi Shimpo and
users manual, Technical report, Report
Mr. Yoshinori Yagome. Financial support for the reSubmitted by the University of Colorado to
search was made possible orginally by the Electric
the Tokyo Electric Power Service Company.
Power Research Institute (EPRI), and for the past
http://civil.colorado.edu/~saouma/Kumo.
seven years by the Tokyo Electric Power Company
(TEPCO) through the Tokyo Electric Power Service Saouma, V., Cervenka,
J. and Reich, R.: 2006,
Company (TEPSCO).
Merlin finite element users manual, Technical report, Electric Power Research In9 REFERENCES
stitute (EPRI), Palo Alto; Tokyo Electric
Power Service Company (TEPSCO), Tokyo.
References
http://civil.colorado.edu/~saouma/Merlin.
Bruhwiler, E., Broz, J. and Saouma, V.: 1991, Fracture model evaluation of dam concrete, ASCE, Slowik, V. and Saouma, V.: 2000, Water pressure in
Journal of Civil Engineering Materials 4, 235
propagating cracks, ASCE J. of Structural En251.
gineering 126(2), 235242.
Cervenka, J., Chandra, J. and Saouma, V.: 1998, Slowik, V., Saouma, V. and Thompson, A.: 1996,
Mixed mode fracture of cementitious bimaterLarge scale direct tension test of concrete, Ceial interfaces; part ii: Numerical simulation, Enment and Concrete Research 26(6), 949954.
gineering Fracture Mechanics 60(1), 95107.
Uchita, Y., Noguchi, H. and Saouma, V.: 2005, Dam
Gillan, C., Saouma, V. and Shimpo, T.: 2004, Censafety research, International Water Power &
trifuge tests of concrete dams, International
Dam Constuction pp. 1622.
Journal of Water Power and Dam Construction
Uchita, Y., Shimpo, T. and Saouma, V.: 2005, Dypp. 3841.
namic centrifuge tests of concrete dams, Earthquake Engineering and Structural Dynamics
Haussman, G. and Saouma, V.: 2006, Spider, a
34, 14671487.
3d interactive graphics finite element post6
180
160
Top Load/CMOD
Bottom Load/CMOD
Total Load/Average CMOD
140
Load [kN]
120
100
80
60
40
20
0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50
CMOD [mm]
Figure 4: Large Scale Testing of Concrete Joints Subjected to Reverse Cyclic Loading
Case2: Constant
Case3: Linear
Case2: Constant
COD (mm)
2
1.5
10
Case3: Linear
Case2: Constant
Case3: Linear
COD (mm)
COD (mm)
4
Case3: Linear
COD (mm)
3
2.5
Case2: Constant
6
4
0.5
5.00
10.00
15.00
20.00
0
0.00
25.00
5.00
10.00
15.00
20.00
0
0.00
25.00
-8
-10
0.00
0
-2
-4
-6
-8
10.00
15.00
Case3: Linear
Case2: Constant
-10
0.00
5.00
20.00
-6
20.00
0
0.00
25.00
5.00
10.00
15.00
20.00
25.00
-4
15.00
-2
10.00
5.00
5.00
10.00
15.00
20.00
0
0.00
-2
-4
-6
-8
Case3: Linear
-10
0.00
25.00
Ligament Length (m)
5.00
10.00
15.00
-4
-6
-8
-10
-12
0.00
Case2: Constant
25.00
0
-2
20.00
25.00
5.00
10.00
15.00
20.00
25.00
Figure 5: Special Interface Elements Simulating Cohesive Crack Joints During Seismic Excitation. Note
Out of Phase Crack Opening Displacements and Cohesive Stresses
U Dyn
U Stat
dCOD
dt
-1
(TFI-O .FFTI)
-1
TFI-O
O(t), FFTO
I(t), FFTI
10
Y
X
Figure 9: Effect of Time integration Scheme on Dynamic Response with Discrete Cracks
11
X Gb Ethernet
Domain Decomposition:
METIS
Profiles_of_crack_001,Uplift
Profiles_of_crack_001,Crack opening
'D:\MERLIN\New_bucharest_analyses\S3D\size_2m\Bucharest-3D-size2.xyzvdat'
'D:\MERLIN\New_bucharest_analyses\S3D\size_2m\Bucharest-3D-size2.xyzvdat'
0.3
0.001
0.0008
Values
0.25
Values
0.2
0.0006
0.15
0.0004
0.001
0.0008
0.0006
0.0004
0.0002
0
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
0.0002
0
25
0
2
20
4
15
6
10
5
12
14
15
10
8
Z
25
20
4
0.1
0.05
10
8
Z
10
5
12
14
Figure 13: 3D Crack Opening Displacement Profile of a Joint, and Corresponding Nonlinear Uplift Pressures
13
Figure 15: Aluminum container showing dam model and accompanying instrumentation
Figure 16: Centrifuge Facility at Obayashi Corporation, and Gravity Dam model
14
Acceleration (m/sec )
1800
1400
1000
600
200
-200
-600
-1000
-1400
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Time (sec)
Figure 17: Crest Acceleration of Dam Model: Yellow Predictive Analysis, Orange: Experiments
15