Professional Documents
Culture Documents
PLAXIS Finite Element Tutorial PDF
PLAXIS Finite Element Tutorial PDF
Plaxis Bulletin
issue 23 / March 2008
Colophon
Editorial
New Developments
Plaxis Practice
Comparison of computed
vs. measured lateral
load/deflection response of
ACIP piles
The Plaxis Bulletin is the combined magazine of Plaxis B.V. and the Plaxis Users
Association (NL). The Bulletin focuses on the use of the finite element method in geotechnical engineering practise and includes articles on the practical application of the Plaxis
programs, case studies and backgrounds on the models implemented in Plaxis.
The Bulletin offers a platform where users of Plaxis can share ideas and experiences with
each other. The editors welcome submission of papers for the Plaxis Bulletin that fall in
any of these categories.
The manuscript should preferably be submitted in an electronic format, formatted as
plain text without formatting. It should include the title of the paper, the name(s) of the
authors and contact information (preferably email) for the corresponding author(s). The
main body of the article should be divided into appropriate sections and, if necessary,
subsections. If any references are used, they should be listed at the end of the article.
The author should ensure that the article is written clearly for ease of reading.
Plaxis Practice 10
In case figures are used in the text, it should be indicated where they should be placed
approximately in the text. The figures themselves have to be supplied separately from the
text in a common graphics format (e.g. tif, gif, png, jpg, wmf, cdr or eps formats are all
acceptable). If bitmaps or scanned figures are used the author should ensure that they
have a resolution of at least 300 dpi at the size they will be printed. The use of colour in
figures is encouraged, as the Plaxis Bulletin is printed in full-colour.
Plaxis Practice 14
Recent Activities 19
Activities 2008 20
Plaxis Bulletin
c/o Erwin Beernink
PO Box 572
2600 AN Delft
The Netherlands
The Plaxis Bulletin has a total circulation of 13.000 copies and is distributed worldwide.
Editorial Board:
Wout Broere
Ronald Brinkgreve
Erwin Beernink
Arny Lengkeek
Editorial
Ronald Brinkgreve
The year 2008 will be a year of many new issues for Plaxis. We started in a new
office in Delft to facilitate a further expansion of activities; we employed new staff
(Delft, Asia); we appointed new agents (China); we introduced a new service (Special
Projects) and last but not least, we are working on new features and will release
new products. Some of these issues are described in more details in this Bulletin or
will come back in the next Bulletin.
In addition to the standard columns, it is a pleasure to notice that more and more Plaxis
users are willing to share their modelling experience with other users, and submit interesting articles for the Plaxis Bulletin. This Bulletin contains three articles about the backgrounds and use of Plaxis 2D and 3D products in practical geotechnical applications.
The first article shows a comparison between the results of a lateral loading test on
cast-in-place piles and a numerical simulation using Plaxis 3D Foundation. The Hardening Soil model with small-strain stiffness was used to model the behaviour of the stiff
over-consolidated clay. The results seem to match the test data reasonably well, but the
specific behaviour of the piles requires more research and a better modelling of the pile
behaviour.
The second article is again an application of the 3D Foundation program. It describes how
3D FEM can be used for piled raft foundations. The embedded piles in the Foundation program prove to be very efficient in the modelling of such complex foundations. Nevertheless, there is a need for improved modelling facilities, like horizontal interface elements.
The third article describes the results of a research project on the possibilities and limitations of the Dynamic module with respect to seismic site response analysis. The researchers have varied numerical parameters to evaluate their influence on the results. They
propose a procedure to calibrate the numerical model and conclude that the standard
settings in Plaxis do not always give the best results.
In all of the articles, the authors give suggestions on how Plaxis could be improved to
model various geotechnical applications in a better way. We are very thankful to the authors for these suggestions and will consider them seriously as a part of future Plaxis
developments.
We wish you an interesting reading experience and look forward to receive new articles
for future Bulletins.
The Editors
New Developments
New Developments
Ronald Brinkgreve
The finite element method is well established in the current geotechnical engineering
practice. Although most calculations are still 2D, there is a tendency to model complicated situations in more detail using 3D models.
Since the new millennium, Plaxis offers 3D models that are relatively easy to create. After
the success of the 3D Tunnel program for simplified 3D situations, the recent 3D Foundation program version 2 allows for a realistic modelling of complicated foundation and
excavation projects, including multiple piles or ground anchors. Since the introduction
of this version halfway 2007, more than 500 licences are being used. For the general
modelling of geotechnical applications, Plaxis will offer two independent solutions with
full geometrical flexibility, depending on the users preference:
1. A new-generation geotechnical-oriented input program based on familiar concepts
from 3D Foundation, with arbitrary volume object creation and import facilities, available from mid 2009 for advanced to expert users.
2. A general CAD-like input program, available from mid 2008 for top users with 3D CAD
experience.
Both programs enable the creation of arbitrary 3D finite element models composed of
10-node (quadratic) tetrahedron elements, which are calculated with the Plaxis 3D calculation kernel. In both programs the existing Plaxis soil models are available. The difference is in the modelling approach, either geotechnical-oriented or CAD-like.
The decision to develop this two-leg strategy is based on different work flows in different
companies or different projects: in most companies or projects the geotechnical engineer
has to create a finite element model him/herself based on 1D or 2D (geotechnical) information whereas in some companies or projects a 3D model is created by CAD experts
before a geotechnical finite element analysis is considered. The latter group can soon be
served. The former group has to wait another year before a dedicated product is available,
but in the remainder of this article I will already elaborate some of its details.
The creation of a geotechnical 3D model starts with the composition of the sub-soil. For
this, the borehole feature of the Foundation program may be used. Soil layer boundaries
may be imported as triangulated surfaces and assigned to the soil layer boundary in the
borehole. In addition to the sub-soil, structures and loads are defined independent from
the sub-soil. Arbitrary excavation volumes or other volumes may be created using the
excavation designer, or by importing 3D objects from CAD packages. Structures and loads
may be defined in a similar way as in the Foundation program, but with more flexibility in
vertical direction. During the modelling phase the user is confronted with a 3D view of the
model in which he can directly select the visible objects. Alternatively, all objects appear
in a tree view, which can also be used to create objects or assign properties. Programs
like Google Sketch up have shown that 3D drawing can be almost as easy as 2D drawing.
The new Plaxis program includes similar 3D drawing facilities and allows import of such
models.
New Developments
When proceeding to the definition of calculation phases, all objects are crossed with each
other using a CSG (Constructive Solid Geometry) algorithm and divided into sub-volumes
and sub-structures. Different construction stages may be defined in which (sub-)volumes,
(sub-)structures and loads can be activated or de-activated. The 3D meshing, including
global and local refinement, is considered just before the start of the calculation. Since
all calculation settings are based on geometric objects, it is not necessary to redefine
calculation phases after mesh regeneration. Even minor geometric changes can be made
without the need to redefine the calculation phases. This stimulates improved modelling
and meshing after promising preliminary results have been obtained.
Calculations are performed with the existing Plaxis 3D calculation kernel. To allow for
unstructured 3D meshes, 10-node tetrahedron elements have been implemented. Moreover, a 64-bit version of the calculation kernel will be available to enable hundreds of
thousands elements. Also the general Output program (post-processor), as available with
3D Foundation, has been extended with 10-node tetrahedrons and all existing output
facilities have been adapted for this type of element.
A beta version of the new 3D program is expected by the end of 2008. Users interested in
beta-testing of this program, may contact Plaxis bv. We will keep you informed about the
progress of development. Meanwhile, for those who are interested in CAD-like 3D modelling, we can soon provide general modelling facilities to address the reliable Plaxis 3D
calculation kernel.
Plaxis Practice
Introduction
Five auger cast-in-place piles (ACIP) were installed at the National Geotechnical Experimentation Site located at the University of Houston campus in Houston, Texas in December 1996. Lateral load tests were performed on four of the piles in January 1997. The
purpose of these tests was to evaluate the feasibility for design and construction of ACIP
piles to support concrete screen walls subjected to large wind forces along highways in
urban areas.
Subsoil Profile
The site is located on a Pleistocene age deltaic deposit known locally as the Beaumont
formation. The Beaumont formation is about 8 meters deep at the test site, and it is underlain by an older geologic formation known as the Montgomery formation. The subsoils
in both formations are primarily clay with occasional interbedded seams and layers of
sand and silt. The consistency of the clays is generally stiff to very stiff, and they have
been overconsolidated due to desiccation.
The first major silt/sand stratum occurs at a depth of 14 m (47 ft) which is below the
depth of influence of the test piles. The water level was at a depth of about 2 m (6.5 ft)
below grade at the time of the load tests.
After deposition, vertical fissures were formed due to shrinkage caused by drying, and the
cracks at the surface were probably more than 5 cm (2 in) wide visually estimated from
presently occurring wet/dry cycles. Soil from the surface was washed down into these
cracks during periods of heavy rain. The soft sediments in the cracks were then compressed when the clays swelled leaving locked-in horizontal stress. This process was repeated throughout Pleistocene to Modern geologic times, and KO values of 3.0 and greater
have been measured in the upper 4 m (13 ft) at this site. The process of desiccation
and subsequent rewetting caused cyclic shearing displacements in the clay mass that
produced polished failure planes referred to as slickensides. The slickensides are widely
variable in size and orientation. The clays are spatially inhomogeneous, and exhibit some
anisotropic properties due to their stress history.
The joints and horizontal locked-in stress affect the strength, deformation, and permeability properties of the clays. The soil parameters have been extensively studied at this site
and they are summarized in Fig. 1. A more detailed summary of the database can be found
on the web site at www.unh.edu/nges. The laboratory and in situ tests in the database indicate a wide range in the strength/deformation properties of the clays due to the effects
of secondary structure, loading stress path, and possible sample disturbance.
Plaxis Practice
Immediately after cleaning the top of the pile heads with a screen, a full length rebar cage
was inserted into the piles. The rebar cages for the 0.46 m diameter piles were comprised
of six number 6 vertical rebars with number 3 ties spaced at 15 cm (6 in) centers. The
cages for the 0.91 m diameter piles were comprised of eight number 10 vertical rebars
with number 4 ties spaced at 23 cm (9 in) centers.
As shown on Photo 1, two ABS tubes were installed in each pile. They were used for sonic
logging to check the integrity of the piles prior to the load tests. Based on results of the
sonic tests, the piles were found to be free of defects such as voids or cracks.
One of the ABS tubes in each pile was later used as a guide to run an inclinometer instrument down the piles during the lateral load tests to measure the rotation during loading.
Plaxis Practice
FE Analysis
The numerical model employed in the FE analysis for the S pile (typical for all piles) is
shown in Fig. 5. The calculations were performed using PLAXIS 3D Foundations V.2 with
about 2,000 elements. The small strain hardening soil model was used to model the stiff
to very stiff clay.
Youngs Modulus
2.5 x 107 MPa
(5.2 x 105 ksf)
Cohesion
7.2 MPa
(150 ksf)
40
-
- Effective stress parameters for the stiff to very stiff clays were used as input, but the
analysis was performed using the undrained mode to simulate the rapid rate of loading.
- Initially, soil parameters were selected from Tand and ONeills article published in the
PLAXIS Bulletin 14 (Sept. 2003). Parametric studies were then performed until good
agreement was obtained with the field load/deflection response of the piles. The final
soil parameters were in good agreement with the prior parameters, but not exact, because there were variations in the subsoil stratigraphy and the stress path was different for the horizontally loaded piles than the vertically loaded underreamed piers. The
soil parameters used in the final FE analysis are summarized in Fig. 6.
- The initial cycle of loading at small loads was not modeled due to the small elastic
deflections that were measured.
- The second cycle of loading was modeled to check the hysteresis cycles. However, only 2
load/unload cycles were computed due to the fact that the load/deflection curves were
almost linearly elastic.
- After the cycling, the pile was loaded to the last measured field load.
Subsoil
Stiff clay
Stiff to very stiff
clay
Very stiff clayey
sand
Very stiff clay
Very stiff sandy
clay
c'
kPa
(ksf)
19.1
(0.4)
19.1
(0.4)
28.7
(0.6)
28.7
(0.6)
28.7
(0.6)
Gref
!'
(deg)
!
(deg)
20
20
30
20
30
mPa
Eref
50
ur
(ksf)
mPa
(ksf)
(ksf)
(ksf)
48
(1000)
96
(2000)
144
(3000)
192
(4000)
240
(5000)
9.6
(200)
12.0
(250)
16.8
(350)
14.4
(300)
16.8
(350)
7.2
(150)
9.6
(200)
14.4
(300)
12.0
(250)
14.4
(300)
27.5
(575)
34.5
(720)
69.0
(1440)
41.4
(865)
47.9
(1000)
Eref
oed
mPa
Tension
1.9 MPa
(40 ksf)
with a unit weight of 943 kN/m3 (6,000 pcf) in the initial phase. Thus, the gravity loading induced a preconsolidation pressure of 216 kP (6,000 psf). This layer was then
turned off for the subsequent calculations. The FE computed KO was 2.9 at 1.2 m (4 ft),
and 2.6 at 3.6 m (12 ft). These values correlate reasonably well with those reported in
the U of H database.
Eref
mPa
vur
If
15
15
15
15
15
Plaxis Practice
Results
A graph of the field and FE computed load/deflection relationships for the S pile is shown
in Fig. 7. The N pile was omitted for clarity because both the curves for the field loading
and FE simulation plotted on top of each other. The FE computed load/deflection response
of the pile without the shell is shown for comparison purposes. Also, the load/deflection
response computed assuming linear-elastic parameters for the grout are shown.
A graph of the field and FE computed load/deflection curve for the E pile is shown in Fig.
8. As previously discussed, the W pile appeared to be an anomaly and is not shown. There
were either subsoil variations in front of this pile, or there was an undetected defect in
the pile.
exceeded the tension cut off stress. These observations highlight the advantages of using a proper constitutive model for the grout so that the effective moment of inertia is
automatically reduced during FE loading.
The good correlations are due to parameter studies, not simply selecting the correct input data for the initial computation. FE cannot be expected to model the load/deflection
response within 25 percent on a common basis because the stress/strain behavior of
soil is very complex, and the uncertainty of selecting appropriate strength/deformation
properties of the soil when there is considerable data scatter. Also, the strength/deformation properties of the pile materials must be properly accessed. It is possible that there
might be other combinations of soil and pile properties that could result in correlations as
well as those determined in this study.
Conclusions
The FE computed load/deflection response of the ACIP piles bearing in stiff to very stiff
clay correlated well with results of the full scale lateral load tests. The small strain hardening soil model in PLAXIS 3D Foundations can be used in predicting the lateral load/
deflection response of piles. However, additional research needs to be performed to better
model ACIP piles with internal reinforcing steel if deflections are large enough to cause
tension cracks in the grout.
References:
Fig 8: Comparison of Load/Deflection Curves for E Pile
A graph of the field measured and FE computed horizontal displacements with depth for
the S & E piles is shown in Fig. 9. Note that FE predicts that a plastic hinge formed at
about the same depth as measured in the field loading tests.
The initial FE computed load/deflection response of the piles using the optimized parameters is in excellent agreement with the loading tests. However, FE under predicts the
deflections at the high loads.
The authors speculate that the piles as modeled in the FE analysis are stiffer than the
field piles. The tension stress of the grout had been input so that the grout would crack
and reduce the moment of inertia of the pile during the FE loading phases. However, the
fact that the steel shell modeled in the FE analysis is located at the perimeter of the piles
probably restricted the tension cracks that would be expected form between the rebar and
outer edge of the field piles. Also, PLAXIS has published a notice that the interface elements for round piles have corners in the FE geometry that makes the interface behavior
stiffer than would occur under field conditions (see Plaxis website).
The FE computed load/deflection response of the ACIP piles without the steel shell was
somewhat softer than for the field piles. This occurs because of the reduced moment of
inertia resulting from neglecting the rebar, and the fact that more deflections occur during loading increasing the tensile strains.
The FE computed load/deflection response of the ACIP piles assuming linear elastic properties for the grout was considerably stiffer than for the field piles. This occurs because
the effective moment of inertia was not being reduced when the mobilized tensile stresses
- K.M. Hassan, M.W. ONeill, and C. Vipulanandan, Specifications and Design Criteria for
the Construction of Continuous Flight Auger Piles in the Houston Area, FHA report UH
3921-1, 1998.
- L.J. Mahar and M.W. ONeill, Geotechnical Characterization of Desiccated Clay, Journal
of Geotechnical Engineering, ASCE, January, 1983, pp. 56-71.
- M.W. ONeill, National Geotechnical Experimentation Site University of Houston, Geotechnical Special Publication No. 93, 2000, pp. 72-101.
Plaxis Practice
Introduction
The quick growth of cities in the last two decades all over the world led to a rapid increase
in the number and height of high rise buildings even in unfavourable subground conditions. Since the 80's, a new foundation technique, the so-called piled rafts, has been
developed and used extensively in order to reduce the maximum as well as the differential
settlements and the associated tilting of the buildings. The analysis of piled raft is a
very interesting example of the soil-structure interaction that requires the co-operation
between the geotechnical and structural engineers to reach the most economic foundation system. Enhanced numerical analyses play a decisive role for the analyses of such
complex foundation system. The piled raft foundation has shown its validity as a very
economic geotechnical foundation type, where the structural loads are carried partly by
the piles and partly by the raft contact stresses. This foundation system was successfully
applied in stiff as well as soft subsoil. An innovative application of the piled raft is its
special adjustment to cases of foundations with large load eccentricities or very different
loaded parts of buildings to avoid the need of complex settlement joints especially below
ground water table.
10
Calculation procedures to model the behavior of such complex three-dimensional problems have been developed since the 1970s (e.g., by Butterfield and Banerjee 1971, Poulos
and Davis 1980 and Randolph 1993). But some important requirements concerning the
raft stiffness, the nonlinear behavior of the pile support and the slip developing along the
pile shafts even under working loads were not sufficiently considered in these analyses.
For these reasons improved numerical models based on three dimensional finite element
method are applied taking into account all above mentioned effects (El-Mossallamy
1996).
A traditional 3D finite element technique with the appropriate soil constitutive laws presents a powerful tool to model this complex soil-structure interaction problem. Nevertheless, the main disadvantage applying the 3D FE analyses is the need of a huge number
of volume elements which can exceed the available computer capacities. To cover this
problem, a new technique combined the so called embedded pile model with the 3D finite
element model was developed by Plaxis B.V. under the name Plaxis 3D Foundation version
2. The following sections present an example demonstrating the ability of this program
to deal with a complex piled rafts. A case history in Frankfurt will be resolved applying
this program.
Plaxis Practice
General information
Height (m)
Foundation area (m2)
Raft thickness (m)
Foundation depth (m)
Groundwater
Slenderness ratio
No of piles
Pile length (m)
Pile diameter (m)
114
1930
3.5 - 1.0
-15.75
- 6.0
3.5
25
22
1.3
Geometry
The foundation of the building has a total area of about 1930 m. Only 25 large diameter
bored piles were constructed beneath the raft as a piled raft foundation. The pile arrangements are shown in Figure 4. The rafts are 3.5 meters thick in the middle and 1.0 m at
the edges. The raft base lies at a depth of 15.75 meters below the soil surface. The piles
where designed with a diameter of 1.3 m and a length of 22 m. The total working loads
reach about 900 MN.
11
Plaxis Practice
Figure 5: 3D FE-Model
Numerical model
Soil Parameters
The soil stress-strain relationship was modelled applying the Hardening soil model. The
main advantage of this constitutive law is its ability to consider the stress path and its
effect on the soil stiffness and its behavior. For the concrete piles and raft, a linear elastic
material set was applied using the concrete weight and its stiffness. The ultimate skin
friction of the pile is assumed to start with 60 kPa at the pile head and increased with
depth to reach 120 kPa at the pile tip. The ultimate pile base resistance was taken equal
to 2.0 MPa.
3D Finite element model
Work-planes are defined. The Work-planes are needed at each level where a discontinuity
in the geometry or the loading occurs in the initial situation or in the construction process. Figure 5 shows the applied three dimensional finite element mesh. The main model
geometries are given in figure 6.
12
Inspect output
The initial conditions should be generated using the K0-procedure. A value of K0 = 0.8 is
applied to consider the effect of overconsolidation. The aim of the calculation is to determine the average settlement of the rafts under working load (serviceability limit state).
Figure 7 demonstrates the raft settlements under working loads.
Settlement of about 4 cm is calculated at the raft center. This value agrees well with the
measured value and approves the ability of the three dimensional analyses to predict the
settlement of the piled raft as a main part of the foundation design. Figure 8 shows the
load distribution among the individual piles within the pile group. It can be recognized
that the contribution of the edge piles by carrying the loads is very small. This is due to
the presence of the outer wall that works also as shoring system, which is modelled as
fully connected with the foundation raft. The effect of the outer walls can be investigated
by applying a new model in which the outer walls are not modelled.
Plaxis Practice
The illustrated examples show that understanding of the effects of the interaction between construction and subsoil based on the appropriate theoretical knowledge and on
experienced application of measurement techniques and numerical modelling are the
necessary qualification for a safe and economic design for such complex foundations.
- Burland, J.B., Broms, B.B., and de Mello, V. (1977) Behaviour of foundations and structures, State-of-the- art report, Proc. 9th ICSMFE Tokyo, vol. 2, 495 - 546
The piled raft foundation can be modelled using the embedded piles that are available in
Plaxis 3D foundation. The results should be further compared with cases where the piles
are modelled using volume elements. There is still need of horizontal interface elements
to investigate the raft contact stresses in a direct manner. The embedded piles help to
reduce the required number of elements needed to model the complex three dimensional
feature of piled rafts. The experience with this model type should be gathered with time
and shared among the Plaxis users. The effect of the shoring system on the behavior of
piled raft needs further investigation.
- Butterfield, R., and Banerjee, P.K. (1971) "The problem of pile group-pile cap interaction." Gotechnique, Vol. 21, No. 2, pp. 135-142.
- DIN 1054-100 (2005) Baugrund: Sicherheitsnachweis im Erd- und Grundbau
- Drrwang, R., El-Mossallamy, Y. und Reininger-Behrenroth, M.(2007) Neue Erkenntnisse
zum Verformungsverhalten des Frankfurter Tones, Bautechnik, Vol.3, 190-192
- El-Mossallamy, Y. (1996) Ein Berechnungsmodell zum Tragverhalten der kombinierten
Pfahl-Plattengrndung., Dissertation, Fachbereich Bauingenieur-wesen der Technischen Hochschule Darmstadt
- El-Mossallamy, Y., Lutz, B., and Richter, Th. (2006) Innovative application and design
of piled raft foundation. 10th International Conference on Piling and Deep Foundations,
(31 May - 2 June 2006), Amsterdam, Netherlands
- El-Mossalamy, Y., El-Nahhas, F. and Essawy, A. (2006) Innovative Use of Piled Raft
Foundation to Optimize the Design of High-Rise Buildings. 10th Arab Structural Engineering Conference, 13-15 November 2006, Kuwait
Outer piles
- El-Mossallamy, Y (2007) Piled raft foundation in Frankfurt clay. Validation manual,
Plaxis 3D Foundation, Version 2
- Franke, E.; El-Mossallamy, Y.; and Wittmann, P.(2000) Calculation methods for raft
foundations in Germany. Design applications of raft foundation and ground slabs, Edited by Hemsley, Published by Thomas Telford Ltd, London, 2000, p.p.283-322
- Hanisch, J., Katzenbach, R., und Knig, G. 2002. Kombinierte Pfahl-Plattengrndung,
In Zusammenarbeit mit dem Arbeitskreis Pfhle der Deutschen Gesellschaft fr Geotechnik e.V. (DGGT), Ernst & Sohn.
- Poulos, H.G., and Davis, E.H. (1980) Pile Foundation Analysis and Design. Wiley, New
York.
Middle piles
- Randolph, M.F. and Clancy, P. (1993) Efficient design of piled rafts. Proc. 2nd Int. Seminar, Deep foundation, Ghent, 119-130.
13
Plaxis Practice
Introduction
Dynamic FE analyses can be considered the most complete available instrument for the
prediction of the seismic response of a geotechnical system, since they can give detailed
indication of both the soil stress distribution and deformation. However, they require at
least a proper soil constitutive model, an adequate soil characterization by means of in
situ and laboratory tests, a proper definition of the seismic input.
This article discusses how to calibrate a finite element model in order to obtain a realistic
response of the given system subjected to seismic loading. Plaxis 2D v.8.2 (Brinkgreve,
2002) that includes the dynamic module was used in this research. A series of dynamic
analyses of vertical propagation of S-waves in a homogeneous elastic layer was carried
out. This scheme was chosen because a theoretical solution of the problem is available
in literature and some comparisons can be easily done. The influences on the response of
boundaries conditions, mesh dimensions, input signal filtering and damping parameters
was investigated.
The information obtained in this preliminary calibration process can be used thereafter
for the analysis of any geotechnical system subjected to seismic loadings.
PHYSICAL PROPERTIES
!
RAYLEIGH DAMPING
STIFFNESS PARAMETERS
E
Eoed
VS
VP
[kN/m2]
[kN/m2]
[m/s]
[m/s]
1.88x105
6.581x105
3615
6763
"
[kN/m3]
"
[kN/m2]
141
4.889x105
A(f)=
1
H
HD
cos 2
f + 2
f
Vs
Vs
14
Plaxis Practice
Numerical modeling
3.1 Input signal
In numerical computation, the earthquake loading was often imposed as an acceleration
time-history at the base of the model.
Here, the input signal chosen for numerical analyses is the accelerometer registration of
Tolmezzo Station (Friuli Earthquake, Italy, May 6th, 1976). The sampling frequency is 200
Hz, the duration is 36.39 s and the peak acceleration is 0.315 g.
Accelerations time-history and Fourier Spectrum of the signal are reported in Figure 2.
3.2 Finite element model
The finite element model is plotted in Figure 3. It is constituted by a rectangular domain
80 m wide and 16 m high and two additional similar lateral domains, in order to place
far enough the lateral boundaries (total width 240 m). This should help minimizing the
influence of the boundaries on the obtained results, even though no clear indications
exist in literature on this aspect. Recently, Amorosi et al. (2007) have shown a case of
site response analysis in which they have extended the width of the mesh eight times its
height, in order to obtain acceptable results.
The medium is schematized as a Linear Elastic layer that is implemented in the Plaxis
code. Its parameters are indicated in Table 1.
The initial stress generation was obtained by the k0-procedure in which the value of the
earth pressure at rest, k0 was chosen by means of the well-known formula for the elastic
medium:
80m
80m
The mesh generation in Plaxis is fully automatic and based on a robust triangulation
procedure, which results in an unstructured mesh. In the meshes used in the present
analyses, the basic type of element is the 15-node triangular element. The dimensions
of any triangle can be controlled by local element size. By subdividing the homogeneous
layer in sub-layers with a fixed thickness and by using the local element size, it is possible
to assign to the triangles a maximum size.
An average dimension that is representative for refinement degree of the mesh is the
Average Element Size (AES) that represents an average length of the side of the elements employed.
Every time a numerical analysis is performed, the mesh influence must be tested.
Kuhlmeyer & Lysmer (1973) suggested to assume a size of element not larger than /8,
where is the wavelength corresponding to the maximum frequency f of interest. In this
case /8 = VS/8 f = 1.81 m, being VS= 361.5 m/s and f = 25 Hz. In the analyses of the
present work an AES=1.58 m was used.
80m
16m
Different criteria exist to evaluate the Rayleigh coefficients (see for instance Lanzo et al.,
2004; Park & Hashash, 2004; Amorosi et al., 2007). In terms of frequency, the dynamic
response of a system is affected by the choice of these parameters to a large extent.
In the numerical implementation of dynamic problems, the formulation of the time integration constitutes an important factor for stability and accuracy of the calculation process. Explicit and implicit integration are two commonly used time integration schemes.
In Plaxis, the Newmark type implicit time integration scheme is implemented. With this
method, the displacement and the velocity at the point in time t+t are expressed respectively as:
15
Plaxis Practice
1
12
2
The coefficients N and N, which should not be confused with Rayleigh coefficients,
determine the accuracy of numerical time integration. For determining these parameters,
different suggestions are proposed, too. Typical values are (Barrios et al., 2005):
a) N =1/6 and N =1/2, which lead to a linear acceleration approximation (conditionally
stable scheme);
b) N =1/4 and N =1/2, which lead to a constant average acceleration (unconditionally
stable scheme);
c) N =1/12 and N =1/2, the Fox-Goodwin method, which is fourth order accurate (conditionally stable scheme);
In order to keep a second order accurate scheme and to introduce numerical dissipation,
a modification of the initial Newmark scheme was proposed by Hilber et al. (LUSAS, 2000),
introducing a new parameter ( in the notation of the author), which is a numerical
dissipation parameter. The original Newmark scheme becomes the -method or Newmark HHT modification. The -method leads to an unconditionally stable integration time
scheme and the new Newmark parameters are expressed as a function of the parameter
, according to:
where the value of belongs to the interval [0, 1/3]. By assuming =0 the modified
Newmark methods coincides with the original Newmark method with constant average
acceleration.
Moreover, in order to obtain a stable solution, the following condition must apply in the
Plaxis code:
1 1
N
+ N
4 2
Neither the linear acceleration approximation or the Fox-Goodwin method does meet such
requirement.
If no damping, material and/or numerical, is introduced in a dynamic analysis, the model
reaches the resonant conditions at the natural frequencies of the system with a corresponding theoretically infinite amplification ratio. Figure 5 shows the response at a
control point on the free-surface obtained for an undamped analysis (N = N = R =
R = 0) in terms of the acceleration time-history and the Fourier spectrum as a result of
the input signal shown in figure 2. The numerical results are very close to the expected
theoretical values.
16
Plaxis Practice
The standard setting of Plaxis is the damped Newmark scheme with N = 0.3025 and N
= 0.6, that correspond to = 0.1.
Figure 6 explains the results of numerical analyses for three different values of . Rayleigh
coefficients were put equal to zero. When increases, the peaks amplification at the
natural frequencies of the layer decrease. However, the shape of amplification function is
not essentially modified. The numerical damping coefficients chosen by default in Plaxis
(black curve in Figure 6) conduct to an amplification ratio (A=7.97 at f=16.55 Hz) smaller
than the theoretical one (A=10.54 at f =16.95 Hz, see Figure 1) in correspondence to the
second natural frequency of the layer.
Note also that the value of second natural frequency of the stratum is underestimated
by the time domain analyses. This is due to the finite element formulation with lumped
masses instead of consistent mass matrices (Roesset, 1977). The natural frequencies
with a lumped masses formulation, which is implemented in Plaxis, are always smaller
than the true frequencies. Consistent mass matrices overestimate them. The accuracy of
the results decreases with the number of vibration modes.
Numerical damping has a great influence on the dynamic response of a geotechnical
system and this issue should be particularly considered when an earthquake signal needs
to be preliminarily processed. In fact, to reduce the calculation time, filtered signals at the
frequency of interest (i.e., accelerograms with a reduced number of registration points)
are often used for the input motion. In this case, users should be aware that the analysis
needs an adequate calibration of Newmark coefficients, in such a manner to avoid the
loss of important frequency contents of the signal. A comparison of the system response
to a complete signal and a 25 Hz filtered signal is represented in Figure 7.
Figure 8 shows the different amplification functions for three values of Rayleigh damping
coefficient R. The coefficient R is given equal to zero for avoiding excessive damping
of the motion at high frequencies. The results are referred to a numerical damping of
= 0.055. This value has been worked out to obtain a good agreement between numerical
and theoretical values of the amplification ratio that correspond to the second natural
frequency of the layer as shown in Fig. 10.
The solution with free horizontal displacements (FHD) on lateral boundaries is only reasonable for non-plastic material and when local site response is the objective of the study.
If a 2-D configuration of the problem should be examined, horizontal fixities on the left and
on the right hand of the model must to be applied. In these conditions, silent boundaries
are often used to simulate infinite media.
Different methods exist to apply a silent boundary (Ross, 2004). In Plaxis, viscous adsorbent boundaries can be introduced, which are based on the method described by Lysmer
& Kuhlmeyer (1969). By default, relaxation coefficients c1 and c2 are set to 1.0 and 0.25,
respectively.
By placing the lateral boundaries sufficiently far from the central zone, the effects due to
the reflection of waves on boundaries can be neglected.
A comparison of the results with Standard Earthquake Boundaries SEB (Fig. 3) and Free
Horizontal Displacements FHD (Fig. 4) on lateral boundaries is presented in Figure 10, by
using default values for c1 and c2. It seems to suggest that better results are obtained by
using FHD rather than SEB.
17
Recent Activities
References
- Amorosi A.,Elia G., Boldini D., Sasso M., Lollino P. (2007). Sullanalisi della risposta sismica
locale mediante codici di calcolo numerici. Proc. of IARG 2007 Salerno, Italy (in Italian).
- Barrios D.B., Angelo E., Gonalves E., (2005). Finite Element Shot Peening Simulation.
Analysis and comparison with experimental results, MECOM 2005, VIII Congreso Argentino de Mecnica Computacional, Ed. A. Larreteguy, vol. XXIV, Buenos Aires, Argentina,
Noviembre 2005
- Bilotta E., Lanzano G., Russo G., Santucci de Magistris F., Silvestri F. (2007). Methods
for the seismic analysis of transverse section of circular tunnels in soft ground, Workshop of ERTC12 - Evaluation Committee for the Application of EC8, Special Session XIV
ECSMGE, Madrid, 2007.
Figure 10. Comparison between SEB and FHD on lateral boundaries solutions
Conclusions
The use of dynamic analyses to calculate the seismic response of a geotechnical system is
dependent on advanced site characterization and numerical knowledge.
It is necessary a good calibration of the numerical model before conducting a dynamic
analysis for any type of 2-D problem. Some parameters (equivalent stiffness, numerical
and material damping, etc.) can be chosen by comparing the dynamic response of model
under vertical shear waves propagation to the theoretical solutions. In the present article,
an example of procedure to calibrate the finite element model parameters has been presented in order to control the system damping.
Material damping is often modelled by Rayleigh formulation. Moreover, numerical damping is also needed in order to attain a stable calculation. This leads to some difficulty
to control the actual damping of the numerical model. A possible choice in order to limit
such uncertainty is to set the minimum value for Newmark which allows stability, then
fit the theoretical solution. This can be achieved by assuming Rayleigh =0 and changing
Rayleigh only, in order to model the material damping with reasonable approximation
in the desired range of frequencies. The best-fit criterion can be, for instance, reproducing the amplification of the seismic signal over the first and second natural frequency
of the system. Modelling lateral boundaries and filtering input signal need to be carefully
considered when performing such calibration.
The proposed approach was preliminarily used for the analysis of some geotechnical
earthquake problems as the seismic response of flexible earth retaining structures (Visone & Santucci de Magistris, 2007) and the transverse section of a circular tunnel in soft
ground (Bilotta et al., 2007).
Acknowledgments
This work is a part of a Research Project funded by ReLUIS (Italian University Network of
Seismic Engineering Laboratories) Consortium. The Authors wish to thank the coordinator,
prof. Stefano Aversa, for his continuous support and the fruitful discussions.
18
- Brinkgreve R.B.J. (2002) , Plaxis 2D version8. A.A. Balkema Publisher, Lisse, 2002.
Christian J.T., Roesset J.M., Desai C.S., (1977). Two- or Three-Dimensional Dynamic
Analyses, Numerical Methods in Geotechnical Engineering, Chapter 20, pp. 683-718,
Ed. Desai C.S., Christian J.T. - McGraw-Hill
- Kuhlmeyer R.L, Lysmer J. (1973). Finite Element Method Accuracy for Wave Propagation Problems, Journal of the Soil Mechanics and Foundation Division, vol.99 n.5, pp.
421-427
- Lanzo G., Pagliaroli A., DElia B. (2004). Linfluenza della modellazione di Rayleigh dello
smorzamento viscoso nelle analisi di risposta sismica locale, ANIDIS, XI Congresso Nazionale LIngegneria Sismica in Italia, Genova 25-29 Gennaio 2004 (in Italian)
LUSAS (2000). Theory Manual, FEA Ltd., United Kingdom
- Lysmer J., Kuhlmeyer R.L. (1969). Finite Dynamic Model for Infinite Media, ASCE, Journal
of Engineering and Mechanical Division, pp. 859-877
- Park D., Hashash Y.M.A. (2004). Soil Damping Formulation in Nonlinear Time Domain
Site Response Analysis, Journal of Earthquake Engineering, vol.8 n.2, pp.249-274
- Roesset, J.M. (1970). Fundamentals of Soil Amplification, in: Seismic Design for. Nuclear Power Plants (R.J. Hansen, ed.), The MIT Press, Cambridge, MA, pp. 183-244.
- Roesset J.M., (1977). Soil Amplification of Earthquakes, Numerical Methods in Geotechnical Engineering, Chapter 19, pp. 639-682, Ed. Desai C.S., Christian J.T. - McGraw-Hill
- Ross M., (2004). Modelling Methods for Silent Boundaries in Infinite Media, ASEN
5519-006: Fluid-Structure Interaction, University of Colorado at Boulder
- Visone C., Santucci de Magistris F. (2007). Some aspects of seismic design methods for
flexible earth retaining structures, Workshop of ERTC12 - Evaluation Committee for the
Application of EC8, Special Session XIV ECSMGE, Madrid, 2007.
Recent Activities
Recent activities
Plaxis Asia
Plaxis Expands Plaxis Asia Office. Since 2006 Dr. William Cheang was already involved
in pre and after sales activities in Asian countries. Furthermore William assist our agents
upon request to promote Plaxis products and services via conferences, courses and seminars. From 2008 Plaxis bv appointed Mr. Eddy Tan to have the lead in the business development of all local sales and marketing activities for the companys Plaxis products
and services. The aim is to establish a strong local presence in order to provide optimum
support for all Plaxis users and prospects in the Asian market.
Plaxis regards the Asian market as strategically important because it exhibits a strong
and constantly growing demand in Plaxis products and service solutions.
Plaxis Asia activities in 2008 are;
To assist HQ in managing sales and technical support in Asia
- To provide a better service & support to agents in Asia
- To provide assistance to agents in sales & marketing
- To provide necessary technical support to agents
- To have direct contact with the local education institutions
- To assist local agents to organise seminars and courses
To be able to continue the increase of staff Plaxis bv moved to a new building. Detailed
contact information on Plaxis bv and Plaxis Asia can be found at our contact page of our
website.
Plaxis Events
Last year our worldwide expansion of Plaxis courses included also 2 fully booked courses
in Latin America (Brazil and Colombia). In 2008 we have an extended Course Program to
facilitate our knowledge transfer of the background and usage of Plaxis products. Please
visit our agenda on the back cover of this bulletin or on the website to get a full overview
of the upcoming Plaxis Events. We hope to meet you soon in our new offices or at one of
the above mentioned events.
With the establishment of Plaxis Asia, one of her major focus will be the china market
among other emerging countries in Asia. Although marketing effort has been started
since 5 years ago in China, we notice the application usage are mainly in the higher
educational institutions and research sectors. A recent marketing trip to China in Jan
2008 has elevated us beyond this horizon. We have met and presented our software to
big private corporations such as the Water Resource commission, provincial governmentowned Electrical Power design consultants and railway design institutes, whom they see
the potential needs on FEM application for their work. We also take the opportunities to
understand their projects and problem faced during the various stages of their work.
We are confident that Plaxis will soon be widely used by most private design firms in China
just like any other geotechnical design consultants around the world.
19
Activities 2008
March 9 12, 2008
GI-GeoCongress 2008
New Orleans, USA
July 2008
Course on Computational Geotechnics
Guadalajara, Mexico
April 2008
Seminars in Chongqin, Shanghai,
Beijing and Guangzhou
China
August 2008
Course on Computational Geotechnics
Houston, USA
August 2008
Course on Computational Geotechnics
Taiwan
April 2008
Seminars in New Dehli & Mumbai
India
September 3 5, 2008
AMGISS, 2nd International Workshop on
Geotechnics of Soft Soils
Glasgow, Scotland, United Kingdom
May 4 7, 2008
13th Australian Tunneling Conference 2008
Grand Hyatt, Melbourne, Australia
May 8 9, 2008
4th Conference on Advances and
Applications of GiD
Ibiza, Spain
September 2008
Course on Computational Geotechnics
New Dehli, India
Plaxis BV
PO Box 572
2600 AN Delft
The Netherlands
Tel: +31 (0)15 251 77 20
Fax: +31 (0)15 257 31 07
E-Mail: info@plaxis.nl
Website: www.plaxis.nl
November 2008
Course on Computational Geotechnics
Paris, France
8006758