You are on page 1of 247

Selected Papers from the Proceedings of the

Fourth Geo-China International Conference

Geotechnical Special
Publication No. 259

Behavior of Geomaterials
and Foundations for Civil
Infrastructure Applications
Edited by
Behzad Fatahi, Ph.D.
Tamer Sorour, Ph.D.
Zhen Leng, Ph.D.

GEOTECHNICAL

SPECIAL

PUBLICATION

NO.

259

GEO
O-CHINA
H A 2016
BEHAVIIOR OF GEOMATTERIALSS AND FO
OUNDAT
TIONS
FOR CIVIL INFRASTR
N
TRUCTUR
RE APPL
LICATION
NS
SE
ELECTED PAPERS FROM TH
HE PROCE
EEDINGS OF THE F
FOURTH
GE
EO-CHINA
A INTERN
NATIONA
AL CONFE
ERENCE
July 2527,
2
20116
Shandong, Chinna

SPON
NSORED BY
Y

Shando
ong Univerrsity
Shando
ong Deparrtment of T
Transportattion
University of Oklahhoma
Chineese Nation
nal Sciencee Foundatioon
Geo--Institute of
o the American Socieety of Civiil Engineerrs

ED
DITED BY

Behzad
d Fatahi, Phh.D.
Tamer Sorour, Phh.D.
Zhen Leng, Ph.D
D.

Published
P
by
b the Amerrican Societyy of Civil En
ngineers

Published by American Society of Civil Engineers


1801 Alexander Bell Drive
Reston, Virginia, 20191-4382
www.asce.org/publications | ascelibrary.org
Any statements expressed in these materials are those of the individual authors and do not
necessarily represent the views of ASCE, which takes no responsibility for any statement
made herein. No reference made in this publication to any specific method, product, process,
or service constitutes or implies an endorsement, recommendation, or warranty thereof by
ASCE. The materials are for general information only and do not represent a standard of
ASCE, nor are they intended as a reference in purchase specifications, contracts, regulations,
statutes, or any other legal document. ASCE makes no representation or warranty of any
kind, whether express or implied, concerning the accuracy, completeness, suitability, or
utility of any information, apparatus, product, or process discussed in this publication, and
assumes no liability therefor. The information contained in these materials should not be used
without first securing competent advice with respect to its suitability for any general or
specific application. Anyone utilizing such information assumes all liability arising from such
use, including but not limited to infringement of any patent or patents.
ASCE and American Society of Civil EngineersRegistered in U.S. Patent and Trademark
Office.
Photocopies and permissions. Permission to photocopy or reproduce material from ASCE
publications can be requested by sending an e-mail to permissions@asce.org or by locating a
title in ASCE's Civil Engineering Database (http://cedb.asce.org) or ASCE Library
(http://ascelibrary.org) and using the Permissions link.
Errata: Errata, if any, can be found at http://dx.doi.org/10.1061/9780784480021
Copyright 2016 by the American Society of Civil Engineers.
All Rights Reserved.
ISBN 978-0-7844-8002-1 (PDF)
Manufactured in the United States of America.

Geo-China 2016 GSP 259

iii

Preface
As a result of increasing and continuous social and infrastructural growth, appropriate
ground for living and development becomes progressively scanty. Thus, engineers
have been considering constructing buildings and infrastructure at locations with less
favorable geotechnical conditions and in some cases in seismically active regions.
Therefore, proper understanding of the soil behavior has become significantly
important to optimize the design and construction of foundations. This Geotechnical
Special Publication (GSP) contains 30 papers that were accepted and presented at the
GeoChina International Conference on Sustainable Civil Infrastructures: Innovative
Technologies for Severe Weathers and Climate Changes, held in Shandong, China on
July 25-27, 2016. Major topics covered in this GSP are dynamic behavior of soils and
foundations, and physical, numerical, constitutive modeling of soil behavior. The
overall theme of the GSP is Behaviour of Geomaterials and Foundations for Civil
Infrastructure Applications, and all papers address different research findings of this
theme. It provides an effective mean of sharing recent technological advances,
engineering applications and research results among scientists, researchers and
engineering practitioners. All abstracts and full papers have been peer-reviewed prior
to their acceptance and inclusion in this publication. We are most grateful to all
authors and reviewers who have contributed to this Geotechnical Special Publication.

ASCE

Geo-China 2016 GSP 259

iv

Contents
Offshore Foundations Applied to Renewable Energy Infrastructure (RES) ........ 1
Luisa N. Equihua-Anguiano, Marcos Orozco-Caldern, Carlos Chavez-Negrete,
and Jos Roberto Prez-Cruz
Study on the Property and Analysis of DX Pile Bearing Capacity ........................ 9
Po Lin Chen, Da-Wei Jian, De-Xin He, and Dave Ta-Teh Chang
Influence of Soft Soil Shear Strength on the Seismic Response of Concrete
Buildings Considering Soil-Structure Interaction ................................................. 17
Ruoshi Xu, Behzad Fatahi, and Aslan S. Hokmabadi
Validation of the Structural Analysis by PLAXIS2D Geotechnical Software ...... 25
Mahbubeh Mortezaee and Ali Akhtarpour
Nonlinear Behavior of a 2 2 Pile Group under Rotating Machine-Induced
Vertical Vibrations ................................................................................................... 33
S. Biswas and B. Manna
Pile-Soil Interactions under Thermo-Mechanical Conditions Imposed by
Geothermal Energy Piles in Sand ........................................................................... 41
Rajni Saggu and Tanusree Chakraborty
Experimental Study on the Dynamic Response of PHC Pipe-Piles in
Liquefiable Soil ......................................................................................................... 49
Fuyun Huang, Haimin Qian, and Yizhou Zhuang
The Elastodynamic Interaction of a Rigid Circular Foundation Embedded
in a Functionally Graded Transversely Isotropic Half-Space .............................. 57
Reza Yaghmaie and Hamidreza Asgari
Axisymmetric Vibration of an Elastic Circular Plate in an
Inhomogeneous Half-Space...................................................................................... 65
Reza Yaghmaie and Hamidreza Asgari
Numerical Study on a Novel Vibration Screening Technique Using
Intermittent Geofoam ............................................................................................... 73
M. Majumder and P. Ghosh
Seismic Resistance of Batter Pile Foundations in Peaty Soft Ground ................. 81
Koichi Tomisawa and Koichi Isobe

ASCE

Geo-China 2016 GSP 259

Variation in a Shear Modulus of Enzyme-Treated Soil under Cyclic


Loading ...................................................................................................................... 88
Ansu Thomas, R. K. Tripathi, and L. K. Yadu
Numerical Modeling of a Dynamic PileSoil Interaction in Layered Soil
Media ......................................................................................................................... 96
Mohsen Mohammadizadeh and Moein Mohammadizadeh
Study on the 3D Nonlinear Artificial Boundary of Viscoelastic Media with
the Standard Linear Solid Model .......................................................................... 104
Bo Zhang, Xueying Yang, Jie Li, Wentao Dong, Ruisong Pan, Weimin Yang,
Lei Yang, and Dunfu Zhang
Shaking Table Investigation on the Seismic Failure Mode of PHC
Pipe-Piles Considering Liquefaction ..................................................................... 108
Xingwu Wen, Jiesheng Zheng, Fuyun Huang, and Haimin Qian
Influence of Shallow Foundation Characteristics on the Seismic
Response of Mid-Rise Buildings Subjected to Strong Earthquakes .................. 117
Quoc Van Nguyen, Behzad Fatahi, and Aslan S. Hokmabadi
Characterization of Freezing Fresh Concrete by Multiple Non-Destructive
Methods ................................................................................................................... 125
Yan Liu, Junliang Tao, Xinbao Yu, Zhen Liu, and Xiong (Bill) Yu
The Influence of Pile Cap to p-y Curves under Lateral Loads .......................... 136
Shen-Kun Yu, Zhi Zhang, and Honghua Zhao
Multi-Scale Analysis of Deformation Modes in Granular Material Using a
Dynamic Hybrid Polygonal Finite Element-Discrete Element Formulation .... 144
Brandon Karchewski, Peijun Guo, and Dieter Stolle
A Novel Model to Simulate the Behaviour of Cement-Treated Clay under
Compression and Shear ......................................................................................... 152
Lam Nguyen, Behzad Fatahi, and Hadi Khabbaz
Chemo-Mechanical Approach to Modelling the Expansive Behavior of
Sulfate Bearing Soils: The Role of Crystallization Pressure in
Ettringite Formation .............................................................................................. 159
Pawan Sigdel and Liang Bo Hu
Analysis of the Effect of Stiffening at the Top of a Single Granular Pile
with a Stress-Dependent Deformation Modulus .................................................. 167
K. S. Grover and J. K. Sharma

ASCE

Geo-China 2016 GSP 259

Finite Element Analysis of Laterally Loaded Piles in Clays ............................... 175


Fath Elrahman E. Nur Eldayem and Yahia E.-A. Mohamedzein
Analysis of Ring Foundation Systems Resting on an Anisotropic Elastic Soil
Medium Subjected to Working Compressive and Tensile Loads ...................... 183
Avishek Nath, V. Srinivasan, and Priyanka Ghosh
3-D Analysis of a Piled-Raft Foundation Subjected to Vertical and Lateral
Loads ........................................................................................................................ 193
Nasr E. Nasr and Tamer M. Sorour
Research on a Calculation Method and Three-Dimensional Simulation of a
High-Filled Embankment Rheological Settlement .............................................. 201
Zhi-Chao Wang, Du-Min Kuang, Tao Zhao, Ying-She Luo, and Wei-guo Wang
Dynamic Characteristics Study of Geosynthetic-Reinforced Soil under
Cyclic Loading ........................................................................................................ 210
Wei Shi, Tao Lu, Longlong Zhang, and Yue Pan
Transitional Plasticity Compression Model for Clays ........................................ 217
Gyan Vikash
Numerical Behavior of Reinforced Soil by Rigid Inclusion ................................ 225
Samia Boussetta, Mounir Bouassida, and Mondher Zouabi
Study of the Behavior of Tunis Soft Clay ............................................................. 233
Mnaouar Klai and Mounir Bouassida

ASCE

vi

Geo-China 2016 GSP 259

Offshore Foundations Applied to Renewable Energy Infrastructure (RES)


Luisa N. Equihua-Anguiano1; Marcos Orozco-Caldern2; Carlos Chavez-Negrete1;
and Jos Roberto Prez-Cruz1,3
1

Professor, Civil Engineering, UMSNH Univ., Felicitas del Ro, Morelia, Mich. Mexico. E-mail:
nicteea@yahoo.com.mx; cachavez@umich.mx
2
Specialist Engineer, Cd. del Carmen, Campeche, Mexico. E-mail: marcos.orozco.c@outlook.com
3
Conacyt Research Fellow. E-mail: jrperezcr@conacyt.mx

Abstract: New sources of energy are necessary today to totally avoid the oil
dependence. People around the world are researching on renewable energy systems
(RES). In this sense, we present an overview of the current status of the energy
production, based on RES technologies, in Mexico. In addition, a typical yield failure
envelope obtained with RS3 in 3D is included. The main objective is to present this
envelope, which has the special soil characteristics, found in sea sediments.
Additionally, we present a response spectrum, since earthquakes are events that
frequently impact the infrastructure, making the design always dependant on transient
events The time histories of acceleration, used to obtain the spectrum, correspond to
earthquakes localized in the area of the Gulf of Mexico.
INTRODUCTION
Worldwide oil dependence involves high economic and environmental costs due to
rising use of fuels and derivative and the declining global reserves. One of the main
disadvantages for oil exploration is the greater water depth, which is reflected in
difficulties to achieve special infrastructure. Renewable energies (RES) are a
challenge for economics and clean energy sources in the world. Countries like Mexico
require future developments of this kind of energy to support the demand growth. For
this, it is necessary to consider the offshore wave and wind potential besides to its
onshore potential. Developing RES require of many engineering disciplines and
Geotechnics is closely linked to the development of these. Suction caissons are an
excellent alternative for offshore wind turbines. These foundations are widely used for
oil and gas industry. There are many studies of the suction caisson used in deep
waters, nevertheless, few studies have been developed for wind turbines.
RENEWABLE ENERGY
Renewable energies are defined as those practically limitless sources respect to the
human lifetime and whose use is technically feasible. General RES classification can
be divided in two types: tidal and wind energies.

ASCE

Geo-China 2016 GSP 259

Types of wind turbines


There are two main systems of wind turbines (WT), based on axes structures:
horizontal (HAWTs) and vertical (VAWTs), being HAWTs the most frequently used.
From the point of view of Civil Engineering, the WT design needs a study of soilfoundation interaction, by using the environmental forces. Figure 1a depicts the typical
loads used in the design.
Foundations for offshore wind turbines
Offshore turbines require special foundations for stability. Worldwide, the monopile
foundations have dominated offshore wind structures (Fig 1b), followed only by the
gravity bases (Fig 1e). Monopod or tripod/tetrapod are used as shown in Figures 1c
and 1d, with piles as a foundation system. Suction caissons may be used as a
foundation for offshore wind turbines due to its simple installation procedure and
other advantages. They have been mainly used in clays for oil/gas industry, whose
efficiency in offshore areas is widely known.
Mexico Production
Power generation in Mexico is dominated by thermoelectric, using fossil fuels. Up
today, the power generation based in renewable energy covers only 2.3% of the total
electricity generation capacity (SENER, 2011) with an installed electrical capacity of
18,716 MW, while the wind potential is estimated at 71,000 MW (Economy, SD,
2012).
GEOTECHNICAL AND STRUCTURAL CONSIDERATIONS
Soil parameters
Geotechnical parameters were obtained from a shear strength profile su=1.6z that
matches a CPT profile (Fig 2a), where z is the vertical depth. To obtain effective
parameters c and s, numerical simulations were done by varying both parameters,
until the real shear strength profile was reproduced, using Plaxis 2D. Shear profile is
considered for a linear increase and the elastic modulus E is constant with depth (z).
A linear perfect plastic Mohr-Coulomb model was used by considering a plastic
behavior and the water table in the surface of the numeric model. The geotechnical
parameters are showed in Table 1.
Anchor characteristics
Anchor's geometry is a cylinder, top closed-bottom opened, with a relationship
L/d=0.5 (Fig. 2b). The walls and the top anchor were simulated using liner elements
according to RS3, by considering a high stiffness to guarantee a rigid behavior. The
shaft friction and the stiffeners, inside the anchor, were not modeled. Points A and B
were taken as the references to obtain results of stresses and displacements.

ASCE

Geo-China 2016 GSP 259

(a)

(b)

(c)

(d)

(e)

FIG. 1. (a) Typical loads and dimensions for a 3.5 MW turbine (Villalobos-Jara,
2006) and Typical foundations for wind offshore turbines (b) monopod (c)
tripod/tetrapod (d) piled jacket tower and (e) gravity base as reported by
Theengineer (2015).

(a)

(b)

FIG. 2. (a) CPT profile (Colliat, 1999) and (b) scheme of suction caisson
(M=moment load, V=vertical load, H=horizontal load).
Table 1. Geotechnical effective parameters using Mohr Coulomb criterion
c (kN/m2) E (kN/m2)
OCR
Soil
s ()

sat (kN/m3)
Clay
17
27
1.6
9049
0.35
1.0
NUMERICAL MODELING
Mesh
Figure 3a shows the mesh used in all calculations, where it is observed a more
refinement in the anchor area to ensure a good accuracy. Boundaries conditions are
zero-displacement in the vertical plane (z axis), while the base of the mesh is fixed in
three coordinates (x, y and z). The mesh has an extension of 66 m in depth and has
150m x 150m. Figure 3b shows the Major principal strain-Sigma xz effective curve
obtained at the point A (Fig 5a), using H=5.9 MN, M=0 MN and V=0 MN.
ASCE

Geo-China 2016 GSP 259

(a)

(b)

FIG. 3. (a) RS3 mesh for L/d=0.5 (10951 elements) and (b) Major principal
strain - xz curve, H=5.9 MN, M=0 MN, L/d=0.5.
Yield failure envelope
The yield points obtained of numerical simulations were plotted in a plane and fitted
using equation (1) (Villalobos-Jara, 2006).
2

H M
H M
y=
1 = 0
+
2e
hV
i 0
i 0
i 0 dmV
i 0
hV
dmV

(1)

Where hi and mi represent the intersection of each ellipse (with the axis H/Vo and
M/dVo respectively); e is the eccentricity of the ellipse; H is the horizontal load; M is
the momentum; V0 corresponds to the maximum vertical load; and d is caisson
diameter.

FIG. 4. Numerical points and analytical yield failures envelope.

ASCE

Geo-China 2016 GSP 259

The yield envelope, obtained under different horizontal (H) and moment (M) loads,
was normalized respect to the maximum vertical load (Vo) and the caisson diameter (d)
as shown in Figure 4. The filled points are the numerical results obtained for different
H-M combinations and V=0 MN. The plot of the yield surface, using Eq. (1),
corresponds to the solid line. There is a good agreement between the two results.
Figure 5b shows the failure kinematics obtained in RS3, with load H and with H-M
loads combination.

(a)

(b)

Fig. 5. (a)Failure kinematics in 2D of a horizontal (H) loaded suction caisson load


applied at A and B points and (b) total suction caisson displacement.
GROUND SHAKING RESPONSE

Seismic performance of civil structures should be reviewed carefully. Recent


research in full scale models of monopiles in granular marine soils, subjected to a
lateral load (Hokmabady, et al., 2012), show results that may be applied in the design
practice. In the same way, effects of the seismic soil-pile-structure interaction in clays,
shown by Hokmabady, et al., 2014, determine the seismic response at a pile group
foundation, tested in a shaking table of superstructures fixed on the base. To obtain a
response spectrum, we performed an analysis by using the computer program EERA
(Bardet et al., 2000).
Soil properties of site
For our purpose, we use the shear strength profile su=1.6z. The best estimated profile
of shear wave velocity (Vs) was obtained by using an empirical correlation between su
and in-situ shear modulus Gmax (Taboada et al., 2012). The lower and upper limits Vs
profiles were determined by applying a factor scale of 0.85 and 1.15 respectively. We
note that the histories of accelerations, applied to the depth (where Vs = 750 m/s),
were required to extrapolate the properties of the soil below the bottom of the survey,
considering a linear variation with depth. Modulus reduction curves of G/Gmax and
variation of damping ratio , used in this study, correspond to the proposed by Vucetic
and Dobry (1991).
Accelerograms
For this work, we used the history accelerations control movements of the horizontal
components given by the earthquake recorded at Minatitlan station in 1971/10/31 (see
Fig. 6), with epicenter coordinates lat 15.27, long 93.9, and magnitude of 4.7 on the
Richter scale (BMDSF, 1997). Scale factors were applied to the component
accelerations to obtain earthquakes motions of a return period of 100 to 500 years.
ASCE

Geo-China 2016 GSP 259

Table 2 includes the six combinations analysis, the combination number can be
increased depending of seismic sources signals.
Site response results
Throughout the depth, we calculated the results of profiles variation of normalized
stiffness modulus (G/Gmax), shear strain (), the maximum deformation of the land and
the maximum acceleration (amax) were calculated throughout the depth. The transfer
functions, the history of accelerations and response spectrum were calculated at
different depths.
Figure 7 presents accelerations results, obtained from the analysis of seismic
response of offshore site. Specifically the histories of accelerations, calculated on the
surface of the seabed and at 120 m depth, correspond to the bottom of the soil profile
on the site. Figure 7a depicts the condition of the lower bound estimated soil
properties (0.85Vs) and a maximum acceleration of 0.025g (combination 4) in the
bedrock. It can be seen that the movement of the seabed level are amplified with
respect to the calculated at 120 m depth. Figures 7b and 7c show the combinations,
considered as the central and upper limit (factor of 1.0 and 1.15 for Vs), with
maximum acceleration in the bedrock of 0.03g and 0.025g; respectively.

(a)

(b)

FIG. 6. Acceleration time histories recorded ad Minatitlan station, 1971/10/31.


Table 2. Combinations for ground surface response
Acceleration
Earthquake
Combination
maximal (g)
Direction
Vs profile
motion
No.
bedrock
0.025
LB (0.85Vs)
4
0.030
Central (Vs)
Longitudinal
8
0.025
UB (1.15Vs)
6
MINA7110311
LB (0.85Vs)
0.025
13
0.030
Central (Vs)
Transversal
17
0.025
UB (1.15Vs)
15

We compared the calculated response spectrum with the response spectrum included
at the reference NRF-003-PEMEX-2007, that considers the design spectrum for a
return period of 200 years and a damping =5% for Bay of Campeche and Northern
Region of Gulf of Mexico. Figure 8 presents that the maximum spectral accelerations
are close to 0.29 s, and in some cases the spectral accelerations exceed the plateau

ASCE

Geo-China 2016 GSP 259

spectrum of reference NRF for the Bay of Campeche, hence the importance of
calculating the design spectrum.

(a)

(b)

(c)

FIG. 7. Acceleration time histories calculated at 120 m depth and sea bed. (a)
Combination 4, (b) Combination 8 and (c) Combination 6.

Spectral acceleration, g

Respose spectra computed


1

Bay of Campeche, NRF-003-2007

0.1

0.01
0.01

0.1

10

Period, sec
FIG. 8. Response spectra computed and comparison with NRF-003.
CONCLUSIONS

In this paper, a typical yield failure envelope obtained with RS3, considering special
soil characteristics of sea sediments has been presented. The failure points obtained
using RS3 has a good agreement with the ellipse obtained from equation (1).
Adhesion factor is not possible to simulate in RS3, so that the responses shown in
this article have a higher load capacity. It is necessary obtain envelopes that take into

ASCE

Geo-China 2016 GSP 259

account adhesion factor. Laboratory tests and full scale models are a necessary to
complement the seismic response of civil structures.
A study of the seismic response of an offshore site, taking into account both the
seismic and geotechnical site was made. The program used to calculate the seismic
response takes into account dimensional equivalent linear model. Three sets of soil
properties were studied through varying the profile of Vs and using two components of
acceleration in order to calculate the response spectra at the seafloor.
Mexico has great potential to develop renewable energy systems (RES), however it
requires more research on all the related fields.
REFERENCES

Bardet, J.P., Ichii, K. and Lin, C.H. (2000). EERA A computer program for
equivalent-linear earthquake site response analyses of layered soil deposits.
University of Southern California. Department of Civil Engineering.
BMDSF (1997). Base Mexicana de Datos de Sismos Fuertes, Sociedad Mexicana de
Ingeniera Ssmica, A.C.; Instituto de Ingeniera, UNAM; CFE; Fundacin ICA;
CIRES, A.C.; CENAPRED, Volumen II, Mxico.
Colliat, J.L. (1999). Caissons succion pour lancrage de structures ptrolires en
mer profonde. Revue Francaise de Gotechnique, 88:11-19.
Economa, S.D. (2012). "Informe anual de la Secretaria de Economa." Mxico;
Secretaria de Economa, PROMXICO.
Hokmabadi, A.S., Fatahi, B. and Samali, B. (2014). Phisical Modeling of Seismic
Soil-Pile- Structure Interaction for Buildings on Soft Soils, International Journal
of Geomecanics ASCE, ISSN 1532-3641/04014046(18), Vol. 15, april.
Hokmabadi, A.S., Fakher, A. and Fatahi, B. (2012). Full scale lateral behavior of
monopoles in granular marine soils, Journal of the Marine Structures, Elsevier,
Vol. 29, december, 198-210.
NRF-003-PEMEX.2007. Diseo y evaluacin de plataformas marinas fijas en el Golfo
de Mxico. Comit de Normalizacin de Petrleos Mexicanos y Organismos
Subsidiarios, enero 2007.
Plaxis 2D Version 8, Edited by R.B.J. Brinkgreve, Delf University of Technology &
Plaxis b.v., the Netherlands.
RS3 Version 1.017 64-bit, Copyright 2013-2015 Rocscience Inc., Toronto, Ontario,
Canada.
SENER, (2012). National Energy Balance 2011. Secretary of Energy Planning and
Technological Development, Mexico.
SSN (2013). Servicio Sismolgico Nacional, catlogo de sismos. UNAM, Instituto de
Geofsica, http://www.ssn.unam.mx/.
Taboada, V.M., Gan, K.C., Cruz, D., Barrera, P., Espinosa, E., and Carrasco, D.
(2012). Field Gmax relationships for Bay of Campeche clay. Memorias XXVI
Reunin Nacional de Mecnica de Suelos e Ingeniera Geotcnica, Sociedad
Mexicana de Ingeniera Geotcnica, A.C., Cancn, Quintana Roo, Mxico.
Villalobos-Jara, F.A. (2006). Model Testing of Foundations for Offshore Wind
Turbines. Thesis for the degree of Doctor of Philosophy, Oxford University, UK.
Vucetic, M. and Dobry, R. (1991). Effect of soil plasticity on cyclic response.
Journal of the Geotechnical Eng. Div., ASCE, Vol. 111, No. 1, January, 89-107.

ASCE

Geo-China 2016 GSP 259

Study on the Property and Analysis of DX Pile Bearing Capacity


Po Lin Chen1; Da-Wei Jian2; De-Xin He3; and Dave Ta-Teh Chang4
1

Assistant Professor, Dept. of Civil Engineering, Chung Yuan Univ., 200 Chung Pei Rd., Taoyuan 32023,
Taiwan, R.O.C. E-mail: hauer45680@yahoo.com.tw
2
Ph.D. Candidate, Dept. of Civil Engineering, Chung Yuan Univ., 200 Chung Pei Rd., Taoyuan 32023,
Taiwan, R.O.C. E-mail: davidjian520@yahoo.com.tw
3
The Inventor of DX Pile Technology, Beijing ZhongKuo Foundation Technology CO., LTD. E-mail:
dx-13@sohu.com
4
Professor, Dept. of Civil Engineering, Chung Yuan Univ., 200 Chung Pei Rd., Taoyuan 32023, Taiwan,
R.O.C. E-mail: ttc514ttc@yahoo.com.tw

Abstract: In this study, three sets of vertical static loading test for the performance of
DX piles (=600 mm) are implemented in Fangcheng Port Steel Plant Construction
Project (in Guangdong Province, PRC). Test results showed that three sets of DX pile
(600 mm), that were even though not to be plugged into the Moderately weathered
rock bed, but their ultimate bearing capacity are greater than or equal to () 9,800 kN
and tested settlements ranged in 21~24.54 mm, are better than the requirements of
original design of straight hole embedded Pile (800, plug-into-rock). Through the
Stress - Strain Gauge installed along the pile and pile bottom, we found that the major
bearing capacity of DX pile came from the bearing plate at highly weathered rock layer;
while the shared bearing capacity of pile bottom is less than 25% of total capacity.
Therefore, by using the DX pile, the piles are no longer need to be plugged into deeper
rock stratum. Through the above results, we can also infer that the impact of pile bottom
sediments of DX pile is less than cast-in-place pile.
INTRODUCTION
DX pile is a new type of variable cross-section pile technology, which is a multi-nodal
rotary drilling, squeezing and expanding cast-in-situ pile invented by the Chairman Mr.
He De-Xin of Beijing Zhongkuo Foundation Technology Co., Ltd. (named after the
inventor Mr. He De-Xin, hereinafter referred to as DX pile). Based on the traditional
hole-drilling cast-in-situ pile technique, DX pile forms its bearing mechanism by using
a dedicated rotary drilling equipment to squeeze and expand pile near the bottom into a
bearing plate shape then pour concrete, so as to be jointly supported by its pile body,
bearing plate and pile toe. Since the bearing plate increases effective bearing area of pile,
meanwhile the squeezing and expanding equipment compact the surrounding soil to
dense, thus the bearing capacity of DX pile can be greatly improved.
ASCE

Geo-China 2016 GSP 259

10

In comparison with the traditional squeezing and expanding pile technologies, this
rotary drilling, squeezing and expanding device has significant advantages, such as:
providing bilateral relative displacement and Dual hydraulic cylinders to drive the three
pairs of rotary drilling squeezing and expanding arms, always in contact with the
surrounding earth, to cut, roll and squeeze soil with little disturbance and relatively less
sediment; furthermore the bearing stratum of bearing plate can be selected depending on
ground conditions, with the advantages of high flexibility, adaptability, fast and safety
construction.
What is important is that DX pile is a new type of pile jointly supported by
multi-section side resistance (skin friction) and multi-layer-end resistance (end bearing).
Currently, its stress mechanism and calculation method are still in the exploratory phase.
With an increasingly wide application, the DX pile can greatly shorten the demand pile
length, reduce the consumption of concrete and save the construction period, thus it
appears a very broad application prospects. Due to competitive advantages, the DX pile
has been used in various fields of pile foundation projects and created great effects and
economic benefits. This case study will focus on the bearing features of DX pile, with
its application in the highly and moderately weathered rock layer. Chen F. and Chen L.
(2012)Chen L., Zhang Q. and Yuan X. (2012)Wang M., He D. and Tang S.(2012)
Geological Conditions of the Test Site
The test site is situated in the sea-filling area of cofferdam and Reclaimed sand.
Initially, the stretch of beaches and shallow sea had been filled, the soil layers are
described in a descending way as follows: (as shown in Table 1):
Table 1. Soil Parameter on Site

Soil

No

Classification

Clayey fill earth

Silty fine sand

Clayey soil

Loose
Moderately
compacted
Fluid plastic
and soft plastic
Completely
weathered
Strongly
weathered
Moderately
weathered

Completely weathered
muddy sandstone
highly weathered
muddy sandstone
Moderately weathered
muddy sandstone

11
12
13

Thickn
ess (m)

2.4-2.7

Max tested
skin friction
(kPa)
59

6.3-8.1

83

0.3-2.8

100

0.5-2.5

102

6.2-9.9

114

134

Max tested
bottom
friction (kPa)

12973

(1) Clayey fill earth: Purple-brown-gray, with plant roots, slightly wet, loose state with
a thickness of 2.7m.

ASCE

Geo-China 2016 GSP 259

11

(2) Silty fine sand: Yellow gray-gray-brownish gray, mechanical reclaimed seabed
sand, partly containing round gravel, coarse sand with shells, shell fragments;
rendering saturated, loose-slightly tight state, and with a thickness of 6.3m.
(3) Reclaimed clayey soil: Gray- brownish gray, mechanical reclaimed seabed formed
mud deposits, partly containing silty fine sand with shells, shell fragments;
rendering saturated, plastic flow-soft plastic state, and with a thickness of 2.8m.
(4) Completely weathered muddy sandstone: purple-grey purple and partially light grey
and grey black; folded with silty mudstone, shale, the main mineral components
including quartz, feldspar, clay minerals and sericite etc.; Cementation of iron and
mud; rock core rendered a soil-like state, and with a thickness of 2.5m.
(5) Highly weathered muddy sandstone: Purple-purple gray-partial light gray-dark
gray; folded with silty mudstone, shale, the main mineral components including
quartz, feldspar, clay minerals and sericite etc.; Cementation of iron and mud;
sand-like and moderately thick layered structure; rock core rendered earth-like and
clastic state, dehydration easy to crack into a chunky, and with a thickness of 9.9m;
(6) Moderately weathered muddy sandstone: Purple-purple gray-partial light gray;
folded with silty mudstone, shale, the main mineral components including quartz,
feldspar, clay minerals and sericite etc.; Cementation of iron and mud; sand-like
and moderately thick layered structure; rock core rendered long, short columnar or
block state (broken-intact state); dehydration easy to crack into a chunky; not
drilled through.
Allocation of Test Piles and Matters Concerned
As a reference, the Maximum Test Load of DX pile testing was scheduled not to
exceed 10000 kN; and the bearing capacity calculation and type design of DX pile is
conducted on the basis of geological exploration data to select a highly weathered rock
layer to act as bearing stratum of bearing plate and ensure that pile bottom is embedded
in the moderately weathered rock, and area from the bottom of bearing plate to the pile
toe should be embedded in the highly weathered rock at least 2.0 m (as shown in Figure
1). There were Six DX test piles to be set up on site, numbered #9-1, #9-2, #9-3 and
#9-7, #9-8, #9-9, to carry out this ultimate bearing capacity test of axially loaded
vertical pile, its relevant dimensions are as detailed in Table 2.
Table 2. Details of DX test pile
No.
#9-1
#9-2
#9-3
#9-7
#9-8
#9-9

ASCE

Pile
length(m)
29.1
29.1
23.0
36.6
31.0
25.3

Pile
diameter(m)
0.6
0.6
0.6
0.8
0.8
0.8

Bulbs
depth(m)
19
18
19
16
16
18

Bulbs
diameter(m)
1.4
1.4
1.4
2
2
2

Maximum test
load (kN)
pile failure or not
exceed 10000kN
not exceed
10000kN

Geo-China 2016 GSP 259

12

Precautions for the layout of test pile are as follows:


Name of soil layer
(1) The horizontal interval between two neighboring piles in x
& y direction are 20m and 8m. The pile top use stirrup
(10mm) with 150mm spacing (i.e. @ 150mm) and Fragmented stone filling soil
welded with wire fabrics and filled with concrete to make
smooth flat at the top of the pile.
Bowing and filling
powdered fine sand
(2) Attention should be paid to hole cleaning. Before hole
filling cohesive
cleaning the pump must be turned off to grind to smash Blowing andsoil
mud masses for 10 minutes, and then pumped it out of the
Powdered fine sand
ground surface.
Strongly weathered
(3) When the head of rotary drilling squeezing and expanding
sandstone
machine is laid down, it shall not be allowed to apply force
Moderately
if there is any obstacle. Instead, the head should be pulled
weathered sandstone
out, and inspect the causes for a solution to avoid the
hole to collapse and being buried.
FIG. 1. Layout of DX pile
(4) After each step of rotary drilling, squeezing and
bearing plate and pile toe
expanding, slurry should be added to maintain the
water head pressure in the hole.
(5) Filling factor for measuring pouring concrete quantity of DX pile should be greater
than 1.
(6) In each test pile, rebar stress gauges were installed with layers of 2m spacing (two
sets put in a layer); then a Load Cell was also installed at bottom of the pile.
(7) Construction quality should be controlled in accordance with the procedures,
regulations and technical requirements. (Refer to Table3)
(8) The pile testing adopted heaped load to provide reaction force.
(9) The 800 test piles are loaded according to the 1,000kN grade standard; the 600
test piles loaded according to the 700kN grade standard; twice of the grade load
was used in the first-level loading.
(10) Terminate test loading in case one of the following situations occurs:
a. If there is a steep decline segment of the decidable limit bearing force on the
Q-S Curve and the amount of settlement exceeds 40mm;
b. The amount of pile settlement is over twice of the condition of previous-level
load, and it cannot keep relatively stable after 24 hours;
c. When it has not reached the ultimate load but has reached the maximum test
load, also the settlement rate of the pile top has achieved the relatively stable
criteria (convergence).
d. Estimated ultimate bearing capacity has been reached.
Table 3. Quality control standards
Item
Number
1
2
3

ASCE

Item

Quality Control Standards

Hole location deviation

d/6-d/4 mm (d referring to pile diameter)


and not over 100mm

Hole diameter
deviation
Hole depth deviation

50mm
+300mm

Geo-China 2016 GSP 259

Plate diameter
deviation
Hole deviation

Slurry proportion

Slurry viscosity

8
9

Sand content of Slurry


Settlement thickness

13

-4%
<1%
<1.25g/cm3 (Within 500mm to bottom of
hole)
Funnel viscosity 28s (Within 500mm to
bottom of hole); Marsh funnel 30-45s
8% (Within 500mm to bottom of hole)
100mm

Analysis of Pile Testing Results


As shown in FIG. 2., FIG. 3. and FIG. 4., the testing results of three trial piles
(diameter 0.6 m, numbered # 9-1, # 9-2 and # 9-3) were plotted into axial force and
depth relationship diagram. It can be observed that pile axial force appeared plunged at
the location near the bearing plate, that bearing plate share a great part of the load, while
the bottom of the pile withstand only about 25% to 38% of load which can be speculated
that a small amount of displacement in pile bottom incurred. While observed from the
average pile skin friction and depth relationship diagram (as shown in FIG. 5, FIG. 6
and FIG. 7) it was found that nearby area of bearing plate provided more than 40% of
the pile skin friction resistance, but the friction resistance of the remaining area was
developed less than 60%.

ASCE

FIG. 2. Relationship between the pile


axial force and depth of DX pile
#9-1(600)

FIG. 5. Relationship between the


average pile skin friction and depth of
DX pile #9-1(600)

FIG. 3. Relationship between the pile


axial force and depth of DX pile
#9-2(600)

FIG. 6. Relationship between the


average pile skin friction and depth of
DX Pile #9-2 (600)

Geo-China 2016 GSP 259

FIG. 4. Relationship between the Pile


Axial Force and Depth of DX Pile #9-3
(600)

14

FIG. 7. Relationship between the


Average Pile Skin Friction and Depth
of DX Pile #9-3 (600)

Since the measured settlement of all the three piles was lower than 24.5mm, far lower
than the damage limit 40mm, and reached the piles bearing of 9,800kN, extremely close
to the maximum test load (10,000kN), the pile testing was suspended. And based on the
test results, the bearing of the test pile had reached 9,800kN, thus the axial
stress-resistance bearing value of the DX pile can be set as over 4,900kN. Accordingly,
the bearing plate installed in the highly weathered rock layer can effectively enhance the
bearing of the DX piles and reduce settlement.
With the results of three test piles (diameter 0.8 m, numbered # 9-7, # 9-8, # 9-9)
plotted as axial force and depth relationship diagram, their bearing capacity envelops
(as shown in FIG. 8, FIG. 9 and FIG. 10) are consistent with 0.6 m test pile (numbered #
9-1, # 9-2, # 9-3), thus pile axial force appeared plunged at the location near the bearing
plate, bearing plate shared a great part of the load, while the bottom of the pile bearing
further reduced to 18% ~ 25% of load which would cause a small amount of
displacement in pile bottom. While observed from the average Pile Skin Friction and
Depth relationship diagram (as shown in FIG. 11, FIG. 12 and FIG. 13), it can also be
found that the nearby area of bearing plate shared a great proportion of the Pile Skin
Friction which produces a small amount of displacement.
The settlement of all the these three piles were lower than 18.31mm, far lower than the
damage limit 40mm, and the piles had reached the maximum testing load of 10,000kN,
thus the tests were suspended. Whereas, according to these test results, the limit axial
stress-resistance bearing of the test pile can be estimated more than 10,000kN, and the
practice bearing capacity of the DX pile should be set as over 10,000kN.
The analysis of the soil layers of the piles testing showed that a bearing plate installed
on the strongly weathered rock layer can effectively enhance the bearing of pile and
reduce settlement. The reason is that the strongly weathered rock and the bearing plate
provide most of the bearing, which greatly reduces the skin friction development of the
pile body and thus suppress settlement.

ASCE

Geo-China 2016 GSP 259

15

FIG. 8. Relationship between the pile


axial force and depth of DX Pile
#9-7(800)

FIG. 11. Relationship between the


average pile skin friction and depth of
the DX Pile #9-7 (800)

FIG. 9. Relationship between the pile


axial force and depth of DX pile #
9-8(800)

FIG. 12. Relationship between the


average pile skin friction and depth of
DX pile #9-8 (800)

FIG. 10. Relationship between the Pile


Axial Force and Depth of DX Pile #9-9
(800)

FIG. 13. Relationship between the


Average Pile Skin Friction and Depth
of the Pile #9-9 (800)

CONCLUSIONS
1. According to the observed settlement and the axial force development, all the six test
piles did not really reach the limit bearing, so there was still large amount potential

ASCE

Geo-China 2016 GSP 259

bearing existed in the piles. However, the piles' bearing capacity based on the
theoretical calculation were lower than that of testing results, thus it is suggested that
the limit axial stress-resistance bearing of the test pile can be consistent with the pile
body strength based on the theoretical calculation.
2. Based on the comparison analysis between the two diameters of piles under the same
bearing condition, the bearing plate of 800mm pile provided stronger skin friction
force and had a smaller range of the bearing plate impact than the bearing plate of
600mm pile. It is speculated that the ratio between the diameter of the bearing plate
and the pile is related.
3. With the same bearing, it is possible to shorten the pile length or reduce pile diameter
by adopting DX pile, such as: using 600mm DX pile to meet the requirement of
10,000kN design load and replace the 800mm straight hole pile to achieve the
benefit of high-quality and low-cost construction.
4. The bearing plate installed in the highly weathered rock layer can significantly
enhance the bearing, and its pile end embedded on the moderately weathered rock
bed can meet the requirement of bearing and suppress settlement.
5. Based on the measurement of stress gauges installed in the pile body and the pile toe
of the DX pile, it is accounted that the bearing plate situated in the highly weathered
rock layer provided over 50% bearing, and contribution of the bearing at the pile toe
was less than 25% of the total bearing. Thus, there is no need to install large part of
DX pile in the moderately weathered rock bed. Meanwhile, the settlement of the pile
toe plays only little impact on the bearing of the DX pile, it is entirely different from
that of the straight-hole pile.
Acknowledgments
Thanks to President Mr. He De-Xin of Beijing Zhongkuo Foundation Technology Co.,
Ltd. for providing precious construction experience and professional advice.
References
Chen F. and Chen L. (2012), "Discussion on design method of DX pile." Engineering
sciences, vol. 14, No1, 60-64.
Chen L., Zhang Q. and Yuan X. (2012), "Numerical analysis on uplift bearing
characteristic of cast-in-situ DX pile." Engineering sciences, vol. 14, No1, 31-36.
Wang M., He D. and Tang S.(2012) "New pile foundation technology of 21 century:
DX pile." Engineering sciences, vol. 14, No1, 4-11.

ASCE

16

Geo-China 2016 GSP 259

17

Influence of Soft Soil Shear Strength on the Seismic Response of Concrete Buildings
Considering Soil-Structure Interaction
Ruoshi Xu1; Behzad Fatahi2; and Aslan S. Hokmabadi3
1

Ph.D. Candidate, School of Civil and Environmental Engineering, Univ. of Technology Sydney (UTS),
Sydney, Australia. E-mail: Ruoshi.Xu@student.uts.edu.au
2
Senior Lecturer, Geotechnical Engineering (Ph.D., CP.Eng.), School of Civil and Environmental
Engineering, Univ. of Technology Sydney (UTS), Sydney, Australia. E-mail: Behzad.fatahi@uts.edu.au
3
Geotechnical Engineer (Ph.D., CP.Eng.), Ove Arup & Partners, Hong Kong. E-mail: aslan.shokmabadi@arup.com

Abstract: Influences of undrained shear strength on seismic response of moment


resisting concrete building considering soil-structure interaction (SSI) have been studied.
A 15-storey building model resting on class Ee soil with different values of undrained
shear strength has been simulated through FLAC3D. Fully nonlinear dynamic analysis
under four different earthquakes including two far-field and two near-field recordings
has been conducted by direct method and results in terms of base shear, maximum
lateral displacement, inter-storey drift and spectral acceleration have been compared and
discussed. Results indicate that by increasing the undrained shear strength of the subsoil,
the superstructure experiences extra base shear under earthquake excitations due to SSI.
Furthermore, the maximum lateral displacements and inter-storey drifts of the
superstructure increase by adopting higher values for the undraied shear strength of the
subsoil. It is concluded that practicing engineers should treat soil properties gained from
field or laboratory tests with extreme care when dealing with numerical based seismic
design of the soil-structure systems.
INTRODUCTION
In recent decades, several studies have shown that superstructures resting on soft soil
are vulnerable to the effect of SSI (e.g. Hokmabadi et al. 2014; Fatahi et al. 2014). Also
fully nonlinear analysis of seismic soil-structure interaction (SSI) becomes feasible due
to the development of commercial computer analysis packages. During strong
earthquake excitations, significant soil damping is induced by both soil modulus
degradation and soil plastic flow, which influences the response of the superstructure.
Undrained shear strength is one of the essential soil properties and it is commonly
required when performing nonlinear numerical analysis as it governs soil plastic
behaviour. To obtain accurate undrained shear strength, either field or laboratory tests is
required. However, Kamei and Iwasaki (1995) pointed out that undrained shear strength
gained could be impacted by the mode of testing, boundary conditions, strain rate,
confining stress level, initial consolidation, state in-situ testing device and other

ASCE

Geo-China 2016 GSP 259

variables and therefore it is expected different test types produce different test results for
soil shear strength. This paper aims to investigate the influence of soft soils undrained
shear strength on the seismic response of moment resisting structures considering soilstructure interaction.
DYNAMIC ANALYSIS OF SOIL-STRUCTURE SYSTEM
Byrne et al (2006) highlighted the advantages of the fully nonlinear analytical methods
in simulating the dynamic SSI problems. These methods are capable to precisely model
nonlinearity in dynamic analysis of soil-structure systems and can follow the prescribed
constitutive model. Structural nonlinearity can also be captured in the fully nonlinear
methods. Accordingly, fully nonlinear analytical methods have been adopted in this
study to investigate the influence of undrained shear strength on seismic response of a
moment resisting concrete building. During solution process, specified limiting plastic
moment is assigned to the structural elements to simulate the elastic-perfectly plastic
structural behaviour. In addition, the damping of the system in numerical simulation
should be able to reproduce energy losses in natural system subjected to dynamic
loading. Wegel and Walther (1935) mentioned that energy dissipation is largely
hysteretic in nature soil and rock; therefore, hysteretic damping algorithm is necessary to
enable strain-dependent modulus and damping functions to be incorporated directly into
the numerical simulation. Apart from above mentioned, damping in soil arising from
plastic flow, especially for soft soil medium subjected to large seismic loading, should
also be taken into account by employing appropriate constitutive model in numerical
simulation.
NUMERICAL MODELLING
In this study, seismic soil-structure interaction analysis is conducted adopting direct
method which evaluates the dynamic response in a single step and can capture more
realistic behaviour for the soil-structure system in comparison with the substructure
method. Furthermore, time domain analysis is utilised to consider the nonlinear dynamic
responses, as recommended by Chu (2006). A three dimensional explicit finite
difference based program FLAC3D (Fast Lagrangian Analysis of Continua) version 5.0
has been employed in this study.
Structure and Soil Medium Model
A fifteen-storey moment resisting concrete building which comprises beam and shell
structural elements has been adopted. SAP2000 V14 has been utilised for the structural
design purpose. The specified compressive strength of concrete and concrete unit weight
are assumed to be 40MPa and 23.5kN/m3, respectively. All the structural sections of the
model have been designed based on inelastic method assuming elastic-perfectly plastic
behaviour by specifying a limiting plastic moment of the structural element. According
to AS/NZS1170.1-2002 (Permanent, imposed and other actions), permanent and
imposed loads are determined and applied to the structure model. It should be noted that
referring to ACI318-08 (2008), cracked sections for the reinforced concrete sections are
taken into account through multiplying second moment of area of the uncracked sections

ASCE

18

Geo-China 2016 GSP 259

19

(Ig) by
b cracked section
s
coeffficients (0.35
5Ig for beam
ms, 0.70Ig for columns aand 0.25Ig foor
slabss). Designed
d foundation comprises solid
s
elemennts with 16 m
meters in botth length and
width
h, and 1 meeter in depth
h. Four real earthquake rrecords show
wed in FIG 3 have beenn
appliied at the baase of the fo
oundation to perform dyynamic analyysis for fixedd-base casess.
Finallly, the strructural meembers are designed iin accordannce with A
AS3600-2009
(Ausstralian Stand
dard for Con
ncrete Structtures) in a w
way that the performancee level of thhe
struccture stays in
n the life saffe zone by limiting
l
the maximum innelastic inteer-storey drifft
to 1.5%. FIG 1a shows the dimensions
d
of
o the structture adoptedd in this studdy possessingg
the fundamental
fu
natural perio
od of 1.28 seeconds.

FIG 1 (a) Fixed


d-base system; (b) Soil--structure in
nteraction ssystem
Intterface elem
ments were placed betw
ween the fo
foundation aand the soiil surface to
incorrporate the different
d
mecchanical chaaracteristics of them, andd to capture any possiblle
uplifft and slidin
ng of the foundation du
uring shakinng excitationn. The adoppted interface
elem
ments are rep
presented by
y normal (kn) and shearr (ks) springss between tw
wo planes inn
contaact, where the
t shear strrength of th
he interface elements arre defined bby the MohrrCoullomb failuree criterion. The
T relativee interface m
movement iss controlled by interfacce
Normal andd shear spring
stiffn
ness values in
i the normaal and tangen
ntial (shear) directions. N
stiffn
ness values for
f interfacee elements of the soil-strructure moddel are set too ten times oof
the equivalent
e
sttiffness of th
he neighbourring zone (E
Eq. 1) basedd on the recoommendationn
by Rayhani
R
and Naggar
N
(200
08).
k

10

(1)

wherre, K and G are the bulk


k and shear modulus
m
of the neighboouring zone, respectivelyy,
and z
min is the smallest wid
dth of an adjoining zone in normal ddirection.
Acccording to the
t classificaation of AS1
1170.4-2007 (Earthquakke actions in Australia), a
soft soil
s represen
nting subsoill class Ee haas been seleccted in this sstudy. During simulationn,
Moh
hr-Coulomb model
m
has been
b
adopted
d to simulatee plastic flow
w in the soill elements aas
used by many reesearchers fo
or SSI analy
ysis (e.g. Connniff and Kiiousis, 2007, Hokmabaddi
et al.. 2014; Fataahi et al. 201
14). FIG 1b shows the ccharacteristiccs and dimennsions of thhe
soil medium
m
ado
opted in this study. The undrained
u
shhear strengthh has been aapplied in thhe

ASCE

Geo-China 2016 GSP 259

20

way that the strength increases with the depth of soil medium at a rate of 2.3 kPa/m (e.g
50 kPa on the surface and 120 kPa at the bottom of the soil medium). Thus, the average
undrained shear strengths for the soil medium are 65, 75, 85, 95 and 105 kPa. In addition,
a benchmark case, in which only the hysteretic damping of the soil is modelled (i.e. no
plastic flow), has been considered for the comparison purposes.
Hysteretic damping has been adopted to capture the energy losses in the soil medium
during shaking excitations by reproducing the actual hysteretic behaviour of the soil
elements in the nonlinear dynamic analysis. The hysteretic damping in FLAC3D is based
on the defined secant modulus reduction curve (FIG 2) for primary loading and a
Masing rule assumption for loading/unloading (Itasca, 2014). The Mohr-Coulomb model
has a constant elastic shear modulus (Gmax) and a constant yield stress (Su). The
hysteretic damping model is used to provide energy dissipation in the elastic range of the
adopted constitutive model by reducing the shear modulus from the initial value of Gmax
and increase the damping ratio. Thus, when hysteretic damping is used with MohrCoulomb model in FLAC3D, the modulus-reduction technique is applied in elastic range,
and natural damping induced by constitutive model applies in the plastic range.
Free field boundaries have been employed at the side boundaries of the soil medium
generating one-dimensional free-field wave propagation in parallel with the main-grid
analysis (Itasca, 2014). Thus, plane waves propagating upward undergo no distortion at
the boundaries as the free-field grid supplies conditions identical to those in an infinite
model. In addition, a rigid boundary has been adopted at the bedrock level to simulate
the large dynamic impedance (e.g. low velocity sediments sitting on high velocity
bedrock).
20

G\Gmax

Damping Ratio (%)

1
0.8

(a)

0.6
0.4
0.2
0
0.001

After Vucetic and Dobry (1991)


Applied Modulus reduction curves in this study
0.01

0.1

Cyclic Shear Strain, c (%)

15

After Vucetic and Dobry (1991)


Applied damping curves in this study

(b)

10

0
0.001

0.01

0.1

Cyclic Shear Strain, c (%)

FIG 2 (a) Relations between G/Gmax versus cyclic shear strain; (b) Relations
between damping ratio versus cyclic shear strain
Earthquake Ground Motions
In order to perform a comprehensive study on the seismic response of the structural
models, two near field seismic accelerations including Northridge earthquake, 1994
(FIG 3a) and Kobe earthquake, 1995 (FIG 3b), and two far field seismic accelerations
including El-Centro earthquake, 1940 (FIG 3c), and Hachinohe earthquake, 1968 (FIG
3d), are utilised for the dynamic time-history analysis in this study,which are selected by
the International Association for Structural control and Monitoring for benchmark
seismic studies (Karamodin and Kazemi 2008). Also it is well known that near field
earthquakes contain significant forward directivity pulse.

ASCE

Geo-China 2016 GSP 259

21

Name: Northridge
PGA: 0.887 g
Mw (R): 6.7
Dominant frequency: 0.4-0.63 Hz

0.5

Acceleration (g)

Acceleration (g)

-0.5

(a)

Name: Kobe
PGA: 0.833 g
Mw (R): 6.8
Dominant frequency: 1.0-1.5 Hz

0.5

-0.5

-1

(b)
-1

10

15

20

25

30

10

20

Time (Sec)

0.2

Acceleration (g)

Acceleration (g)

0.4

Name: El Centro
PGA: 0.349 g
Mw (R): 6.9
Dominant frequency: 1.0-2 .0Hz

0.4

0
-0.2

(c)
5

10

15

40

50

20

Time (Sec)

25

Name: Hachinohe
PGA: 0.229 g
Mw (R): 7.5
Dominant frequency: 0.3-2.1 Hz

0.2

-0.2

(d)
-0.4

-0.4
0

30

Time (Sec)

30

10

20

30

40

Time (Sec)

FIG 3 (a) Northridge Earthquake 1994; (b) Kobe Earthquake 1995; (c) El Centro
Earthquake 1940; (d) Hachinohe Earthquake 1968
NUMERICAL ANALYSIS
Dynamic analysis is carried out for a 15-storey model in conjunction with class Ee soil
possessing geotechnical characteristics (Vs = 150 m/s and density = 1470 kg/m3)
presented in FIG 1 for two different systems: fixed-base system (structure without soil
medium) and SSI system (structure resting on soil medium considering the interaction
between foundation and soil). For modelling fixed-base system, the four mentioned
earthquakes are applied at the structural foundation; while for SSI system, the
earthquakes are applied at the bedrock level directly. The Australian seismic code
evaluates local site effects based on the properties of the top 30 meters of the soil profile,
due to the fact that the main part of the amplification and attenuation occurs within the
first 30 meters of the soil profile. Therefore, a 30-meter soil medium has been
considered in this study.
RESULTS AND DISCUSSION
The results of the dynamic analyses for a 15-storey model in terms of base shear,
maximum lateral displacement, inter-storey drift and spectral acceleration under the
influence of four earthquake records for fixed-base system and SSI system with different
average undrained shear strength are derived from FLAC3D history records and are
presented and discussed in this section. According to Table 1, the base shear of the
structure accounting for SSI is always less than the base shear of the fixed-base cases
under the corresponding earthquake excitation. For SSI system, it is evident that with the
increase in the average undrained shear strength the structural system experiences
excessive base shear. Take the Northridge earthquake as an example, for average
undrained shear strength of 65 and 105 kPa and the case considering hysteretic damping
only, the maximum base shear were 11200 kN, 18000 kN and 25000 kN, respectively.

ASCE

Geo-China 2016 GSP 259

22

This means that the structure sitting on the soil with undrained shear strength of 105 kPa
experiences up to 60% more base shear in comparison with structure sitting on the soil
with undrained shear strength of 65 kPa. The reason is that less plastic flow occurs
during the excitation of the cases with higher undrained shear strength resulting in less
soil damping. Consequently, more energy transfers through the subsoil to the
superstructure and eventually leaded to the increase in the base shear of the structure
resting on the soft soil medium.
Table 1 Maximum base shear of the structure for different cases
Earthquake
Northridge
Kobe
El-Centro
Hachinohe

Su=65 kPa
11200
14400
10100
9450

Su=75 kPa
13000
15800
11800
11300

Su=85 kPa
14900
17700
12700
12500

Base Shear (kN)


Su=95 kPa Su=105 kPa
16200
18000
19600
21000
13000
13100
13200
13600

Excluding Su
25000
23000
13700
13600

Fixed-base
36653
36127
14345
13727

Table 2 and Figure 4 present the maximum inter-storey drifts and lateral displacements
of the superstructure under the applied earthquakes, respectively. This data was based on
the lateral deformation of each storey when maximum defection at the top level occurs
because, as Hokmabadi et al. (2012) stated, this approach gives a more reasonable
pattern of structural deformation compared to the approach where maximum absolute
storey deformation irrespective of the time they occurred were recorded. Accordingly,
the seismic response of the SSI system amplifies in comparison with the fixed-based
structure, in terms of maximum lateral displacements and inter-storey drifts due to the
rocking in the soil-structure system. Take the Northridge earthquake as an example, for
average undrained shear strength of 65 and 105 kPa and the case considering hysteretic
damping only, the maximum inter-storey drifts were 1.43%, 2.06% and 2.84%,
respectively. This means that the structures sitting on the soil with undrained shear
strength of 105 kPa experiences almost one and half times of inter-storey drifts in
comparison with structure sitting on the same soil with less undrained shear strength.
Although the shear wave velocity (or Gmax) is constant, the performance of soil is
significantly impacted by the average undrained shear strength due to the additional
amount of natural damping triggered by the plastic flow.
Table 2 Maximum inter-storey drift reported for different cases
Earthquake
Northridge
Kobe
El-Centro
Hachinohe

Su=65 kPa
1.43
1.46
0.83
0.77

Su=75 kPa
1.67
1.57
0.84
0.81

Maximum Inter-storey Drift (%)


Su=85 kPa Su=95 kPa Su=105 kPa
1.84
1.94
2.06
1.65
1.83
1.84
0.84
0.89
0.92
0.83
0.84
0.84

Excluding Su
2.84
2.00
0.92
0.85

Fixed-base
1.10
1.42
0.48
0.36

The acceleration response spectrum of Kobe earthquake presented in FIG 5, as an


example. It should be noted that all the other three earthquakes present similar pattern.
According to FIG 5, by increasing the average undrained shear strength, the maximum
spectral acceleration has been shifted to the longer period range. Moreover, in the range
beyond period of 1 second, it is clear that the spectral accelerations for all cases

ASCE

Geo-China 2016 GSP 259

23

considering SSI have been amplified, which further indicates that structures having
relatively long natural periods are vulnerable to the effect of SSI. This can be explained
referring to the conservation of energy concept. Accordingly, in relatively strong soil
medium (i.e. high Su value), soil experiences less plastic deformation and eventually less
natural damping has been created. Therefore, more energy transfers into the foundation
of the superstructure resulting in the amplification of the developed base shear and in
turn the foundation rocking and structural lateral displacements.
15

15

(b)

10

Fixed-base
65 kPa
75 kPa
85 kPa
95 kPa
105 kPa
Excluding Su

Northridge
0
0

200

400

600

Storey Level

Storey Level

(a)

800

1000

1200

10

Kobe
0

1400

200

Maximum Lateral Displacement (mm)

400

600

800

1000

Maximum Lateral Displacement (mm)

15

15

(d)

10

Fixed-base
65 kPa
75 kPa
85 kPa
95 kPa
105 kPa
Excluding Su

El Centro
0
0

50

100

150

200

Storey Level

(c)
Storey Level

Fixed-base
65 kPa
75 kPa
85 kPa
95 kPa
105 kPa
Excluding Su

250

300

350

10

Fixed-base
65 kPa
75 kPa
85 kPa
95 kPa
105 kPa
Excluding Su

Hachinohe
0

400

Maximum Lateral Displacement (mm)

50

100

150

200

250

300

350

Maximum Lateral Displacement (mm)

Spectral Acceleration (g)

FIG 4 Maximum lateral displacement: (a) Northridge Earthquake 1994; (b) Kobe
Earthquake 1995; (c) El Centro Earthquake 1940; (d) Hachinohe Earthquake 1968
3

65 kPa
75 kpa
85 kPa

95 kPa
105 kPa
Original

0
0

0.5

1.5

2.5

Period (s)

FIG 5 Spectral accelerations for Kobe earthquake 1995 with 5% damping ratio
under the influence of soil undrained shear strength variation
CONCLUSIONS
In this study, influence of the shear strength of soil on the seismic response of a
moment resisting concrete building has been numerically investigated. A three
dimensional numerical soil-structure model has been developed adopting direct method
of analysis. In order to capture realistic soil nonlinearities for different shear strength,
hysteretic damping has been implemented to simulate backbone curve of the shear
modulus and the damping ratio versus shear strain for soil. Numerical results show that

ASCE

Geo-China 2016 GSP 259

as the shear strength of subsoil increases, while other soil input properties are constant,
the base shear, the maximum lateral displacement and the inter-storey drift of the
structure amplifies by up to 60%, 100% and 44%, respectively. It can be concluded that
soil undrained shear strength has considerable effects on the seismic response of midrise building frames resting on soft soil deposits due to soil-structure interaction, and
considering unrealistic shear strength for the soil can considerably alters the predicted
response of the structure. Thus, in order to obtain a reliable numerical prediction, the soil
characteristics (undrained shear strength and plasticity) should be treated with extreme
care while conducting soil-structure interaction analysis.
REFERENCES
ACI318-08 (2008). Building code requirements for structural concrete and commentary
American Concrete Institute.
AS/NZS1170.1 (2002). Structural design actions - part 1: permanent, imposed and
other actions Standards Australia/Standards New Zealand, Sydney, Australia.
AS1170.4 (2007). Structural design actions - part 4: earthquake actions in Australia.
Standards Australian, Sydney, Australia.
AS3600 (2009). Concrete structures Standards Australia, Sydney, Australia.
Chu, D. (2006). Three-dimentional nonlinear dynamic analysis of soil-pile-structure
interaction PhD, Washington University.
Fatahi, B. and Tabatabaiefar, S. (2014). Fully nonlinear versus equivalent linear
computation method for seismic analysis of midrise buildings on soft soils
International Journal of Geomechanics, Vol. 14 (4): 1-15.
Hokmabadi, A.S. and Fatahi, B., 2015. Influence of foundation type on seismic
performance of buildings considering soilstructure interaction International
Journal of Structural Stability and Dynamics, p.1550043.
Hokmabadi, A.S., Fatahi, B. and Samali, B. (2014). Assessment of soil-pile-structure
interaction influencing seismic response of mid-rise buildings sitting on floating pile
foundations Computers and Geotechnics, Vol. 55: 172-186.
Hokmabadi, A.S., Fatahi, B. and Samali, B., 2015. Physical Modeling of Seismic SoilPile-Structure Interaction for Buildings on Soft Soils International Journal of
Geomechanics, 15(2), doi: 10.1061/(ASCE)GM.1943-5622.0000396.
ITASCA (2014). FLAC3D version 5.01 Fast Lagrangian Analysis of Continua in three
dimentions, User's Manual Minneapolis, Minnesota, USA: Itasca Consulting
Group, Inc.
Kamei, T. and Iwasaki, K. (1995). Evaluation of undrained shear strength of cohesive
soils using a flat dilatometer Soils Found, Vol. 35 (2): 111116.
Karamodin, A.K. and Kazemi, H.H. (2008). Semi-active control of structures using
neuro-predictive algorithm for MR dampers Structural Control and Health
Monitoring, Vol. 17 (3): 237253.
Vucetic, M. and Dobry, R. (1991). Effects of soil plasticity on cyclic response.
Journal of Geotechnical Engineering, ASCE, Vol. 117 (1): 89107.
Wegel, R. L. and Walther, H. (1935). Internal dissipation in solids for small cyclic
strains Physics, Vol. 6: 141157.

ASCE

24

Geo-China 2016 GSP 259

25

2D

Validation of the Structural Analysis by PLAXIS

Geotechnical Software

Mahbubeh Mortezaee1 and Ali Akhtarpour2


1
2

M.Sc. Student, Structural Engineering, Azad Islamic Univ., Shahroud.


Assistant Professor, Dept. of Engineering, Ferdowsi Univ. of Mashhad.

Abstract: In soil-structure interaction, modeling the soil and structure set is one of the
most important issues in geotechnical and structural field. In this regard, application of
soil mechanic software such as PLAXIS is of particular importance. In this paper, the
performance of PLAXIS in soil-structure interaction is addressed. Frist, beam analysis
results with different support condition were investigated by PLAXIS and compared
with classic structural analysis methods results. Then two building frames, a simple
one story frame and a 17-floor, 5 spans frame were modeled and analyzed by PLAXIS
and ETABS. Results of PLAXIS had good agreement with the results from classic
equations of structural analysis and ETABS software, suggesting the reliability of
PLAXIS in soil-structure interaction analysis.
Keywords: Soil-structure interaction; Building frame modeling; Validation;
PLAXIS2D; ETABS.
INTRODUCTION
In conventional analysis of structure dynamic, it is usually supposed that the
foundation soil is rigid and elasticity of soil is neglected. In this case, structure
response is under the influence of its own dynamic properties free from elasticity of
soil. This is true when the building is constructed on hard rock, but in cases where the
building is on soft soil, the response is totally different. In this case the structure has
some interaction with soil resulting in some changes in basis movements; this
interaction is mutual and known as soil-structure interaction. This interaction
significantly changes the structure response. Lamb investigated elastic vibrations in
1904 which is considered as a start for soil dynamics. Based on Lamb studies,
Reissner (1936) studied circular rigid foundation vibrations on elastic soil which is
known as the beginning for soil-structure interaction. Hokmabadi et al. (2014) studied
the effects of the seismic soil-pile-structure interaction (SSPSI) on the dynamic
response of buildings. it is observed that the SSPSI amplifies the maximum lateral
deflections and in turn interstory drifts of the structures supported by end-bearing pile
foundations in comparison with the fixed-base structures. Fatahi et al. (2014)
investigated the soil-pile interactive performance under lateral loads. In case of soft
ASCE

Geo-China 2016 GSP 259

26

clay, it has been observed that lateral loads on piles can initiate the formation of a gap,
soil heave and the tension crack in the vicinity of the soil surface and the interface,
whereas in medium dense sand, a semi-elliptical depression zone can develop.
The analysis of soil-structure interaction can be classified into two main categories,
including direct and multistep methods. In the direct method, the whole structurefoundation-soil system is analyzed in a single step. An example of it is finite element
method. The main advantages of this method are the capability of considering the
nonlinear behavior of materials of soil and structures, and modeling the complex
geometries (Wolf 1985).
Investigation of soil-structure interaction is an important issue which could not be
neglected; therefore soil-structure set modeling is a crucial item in geotechnical and
structural field. In this content, soil mechanic software for numerical modeling which
has also geotechnical and structural modeling features, are in priority. In this study,
the performance of PLAXIS software is investigated. First its performance was
evaluated for modeling a simple building frame and then a 17-floor frame was
modeled.
1.

VALIDATION OF SOFTWARE IN ONE-FRAME BEAM MODELING

In order to validate the software in modeling one span beam, two problems were
considered: 1) beam with simple support condition and 2) beam with double cantilever
support condition. Two different loading including load concentrated at the mid-span
and distributed loading were also considered. For comparison of the settlements and
moment values, classical equations of structural analysis were used. The
characteristics of the beam are listed in table 1.
Table 1. Beam Properties for Modeling and Analysis
Concrete
Load
Inertial Cross
size
elastic
Length
Beam type
moment section
(cm) modulus
L (m) Concentrat Distribute
I (m4) A (m2)
E (Pa)
ed (KN) d (KN/m)
Rectangular
0.00265
0.1575
2
100
100
concrete 3545 2.51010
78125
beam
The properties of table 1 were considered for both types of beams. For modeling the
beams with different support conditions in PLAIXIS, an individual region must be
created. As the beams could not been used alone and the soil is necessary for meshing.
A coarse mesh is sufficient. As the initial condition, soil analysis was deactivated,
therefore, only the beams were activated.
1-1 Beam with simply supported condition
Tthe classical equations of structure analysis used for calculation of maximum
settlement and moment with concentrated and distributed load, according to formulas
(1) to (4).
Fl3
max =
(1) Maximum settlement with concentrated loading
48EI
ASCE

Geo-China 2016 GSP 259

27

5ql4
max =
(2) Maximum settlement with distributed loading
384EI
Fl
(3) Maximum moment with concentrated loading
M max =
4
ql2
(4) Maximum moment with distributed loading
M max =
8
In these equations, max is maximum settlement of beam, Mmax is maximum moment,
F donates concentrated load, q is distributed load, E is elastic modulus, I indicates
inertial moment and L is beam length. Comparison of results of PLAXIS with the
values obtained from classic equations for maximum settlement of the beam with
simple support condition and different loadings are shown in table 2.
Table 2. Comparing the Maximum Settlement under Simple Support Condition
from PLAXIS and Classical Equations of Structural Analysis
max (mm)
Numerical

Analytical

Settlement
difference (mm)

Concentrated

0.281

0.250

0.031

Distributed

0.344

0.313

0.031

Support

Loading state

Simple
Simple

Comparison of results of PLAXIS with the values obtained from classic equations
for maximum moment of the beam with simple support condition and different
loadings are shown in table 3.
Table 3. Comparing the Maximum Moment under Simple Support Condition
from PLAXIS and Classical Equations of Structure Analysis
Support

Loading state

Simple
Simple

Mmax (KN.m/m)

Difference (%)

Numerical

Analytical

Concentrated

50.00

50.00

0.00

Distributed

50.00

50.00

0.00

1-2 Beam with double cantilever support condition

The classical equations of structure analysis used for calculation of maximum


settlement and moment with concentrated and distributed load, according to formulas
(5) to (8) (Hick 2009).
(5) Maximum settlement with concentrated loading
(6) Maximum settlement with distributed loading
(7) Maximum moment with concentrated loading

ASCE

Fl3
192 EI
ql4
max =
384 EI
Fl
M max =
8
max =

Geo-China 2016 GSP 259

28

ql2
M max =
(8) Maximum moment with distributed loading
12
The parameters were previously introduced. Comparison of results of PLAXIS with
the values obtained from classic equations for maximum settlement of the beam with
double cantilever support condition and different loadings are shown in table 4.
Comparison of results of PLAXIS with the values obtained from classic equations for
maximum moment of the beam with double cantilever support condition and different
loadings are shown in table 5.
Table 4. Comparing the Maximum Settlement under Double cantilever Support
Condition from PLAXIS and Classical Equations of Structure Analysis
max (mm)
Numerical

Analytical

Settlement
difference (mm)

Concentrated

0.093

0.062

0.031

Distributed

0.093

0.062

0.031

Support

Loading state

Double cantilever
Double cantilever

Table 5. Comparing the Maximum Moment under Double cantilever Support


Condition from PLAXIS and Classical Equations of Structure Analysis
Support

Loading state

Double cantilever
Double cantilever

Mmax (KN.m/m)

Difference (%)

Numerical

Analytical

Concentrated

25.00

25.00

0.00

Distributed

33.33

33.33

0.00

For example, plot of bending moment and deformation of double cantilever beam
according to numerical equations are illustrated in figure 1 and 2.

FIG. 1. Double cantilever beam


moment under Distributed load
derived from PLAXIS

FIG. 2. Double cantilever beam


settlement under Distributed load
derived from PLAXIS

By comparing the results of the software with classical equations of structure, for
both types of beams and different support conditions little difference can be seen so
accuracy of the results of the software is confirmed.
.

ASCE

Geo-China 2016 GSP 259

29

2.
INVESTIGATING THE PERFORMANCE OF PLAXIS IN BUILDING
FRAME MODELING
After assurance about the software performance in one-span beam modeling, a
building frame was modeled and the results were compared with ETABS data. For
applying the effect of structure, the most accurate method is modeling in the form of
building frame, as the real results depend on the weight, geometry and stiffness of the
structure.
2-1 One-floor one span bending frame
The properties of the concrete bending frame are listed in table 6. Internal forces of
the frame components are obtained for predefined nodes.
Table 6. Frame Properties for Modeling and Analysis
Column size (cm)
5050
2.51010

Beam size (cm)


3545
Elastic modulus(Pa)

Frame geometry
Repeating distances, perpendicular to
Width (m)
Height (m)
plane (m)
8.2
8
6
Floor system: concrete slab
Load distribute
Dead load(Kg/m2)
One way
400

The important point in building frame modeling with PLAXIS is that according to
plain strain, the components stiffness must be divided on the framed distance, and the
stiffness versus width unit should be included in calculation. It is evident that the alive
and dead loads should also be entered to calculation as the weight per unit width.
In modeling the frame with PLAXIS to create geometry the space between the
components must be defined as a cluster and a material (soil) would be allocated to
this cluster. In continue and at the beginning of analysis this cluster is disabled, so it
has no effect on the result of analysis. Also the method of meshing this cluster affects
the results accuracy and proper elements should be used for initial meshing.
a)
Bending moment of frame components
Bending moments of the specified 5 nodes, obtained from two softwares are listed in
table 7. Their distributions are shown in figure 3 and 4. Good agreement of the results
indicates the ability of PLAXIS in analysis of building frames.

FIG. 4. Bending moment of


frame obtained from PLAXIS
ASCE

FIG. 3. Bending moment of


frame obtained from ETABS

Geo-China 2016 GSP 259

30

Table 7. Bending Moments (KN.m/m)


Node
PLAXIS
ETABS
Difference (%)

1
80.320
80.350
-0.03

2
163.931
163.941
-6.1E-3

3
129.973
129.959
0.01

4
163.931
163.941
-6.1E-3

5
80.320
80.350
-0.03

b)
Shearing force of frame components
Shearing forces of the specified 4 nodes obtained by two software are listed in table
8. Their distributions are shown in figure 5 and 6. Good agreement of the results
indicates the ability of PLAXIS software in analyzing the building frames.

FIG. 6. Shearing force of frame


obtained from PLAXIS

FIG. 5. Shearing force of frame


obtained from ETABS

Table 8. Shearing Forces (KN/m)


Node

PLAXIS
ETABS
Difference (%)

40.71
40.72
-0.02

2
Beam
146.952
146.950
1.36E-3

3
Column
40.71
40.72
-0.02

Beam
146.952
146.950
1.36E-3

Column
40.71
40.72
-0.02

40.71
40.72
-0.02

c)
Deformations
Maximum settlement was observed in the middle of beam equal to 9.89 mm and 10
mm for PLAXIS and ETABS, respectively. The difference is negligible (0.11 mm).
Deflection diagrams are shown in figure 7 and 8.

FIG. 8. Frame deflection


PLAXIS software

FIG. 7. Frame deflection under


dead load in ETABS software

2-2 Static modeling of 17-floored 5-span bending frame


After validation of software performance for a simple frame, a 5 span concrete
bending frame whose first 3 spans include 17 floors (68 m height) and the last 2 spans
contain 4 floor (23 m height) was created and static analysis was performed. The
properties are listed in table 9.
ASCE

Geo-China 2016 GSP 259

31

Table 9. Properties of Analyzed Building Frame


Beam size
(cm)
3545

Column
size (cm)
5050

Regular repeated distances perpendicular on the


plane (m)
8.2

Loading due to concrete slab with the thickness of 15 cm was considered 400 kg/m2.
At the presence of alive loads, they could be considered as additional weight of the
beams. Here, the applied loads are expressed as weight per width unit and the bending
stiffness of the beams and columns were divided on frames distances. In static analysis
there is no difference between loads modeling in the form of overload or beams
weight increase, but in dynamic analysis the over loads should be modeled as beams
mass increase. Also for creation of rigid floor condition, the axial stiffness of the
beams should be increased significantly. In structural analysis, the effect of (P-)
should also be considered which is formed due to perpendicular loads on the deformed
structure and result in lateral movement increase of the floors, shear and moment of
the components. In PLAXIS the option of update mesh is selected to include this
effect. Because, this option updates coordinates of the nodes in each step regarding
deformations. There was a comparison made on the results with or without
consideration of (P-) effect. For validation of PLAXIS results for modeling of
building frame a beam and two columns were selected as shown in figure 9.

FIG. 9. Building frame analyzed by PLAXIS and ETABS

Maximum bending moment, shear force and settlements of the analyzed beam with
PLAXIS and ETABS can be observed in table 10. The comparison of the results
verifies the capability of PLAXIS in analysis of building frames.

ASCE

Geo-China 2016 GSP 259

32

Table 10. Maximum Bending Moment, Shear Force and Settlement the Beam

PLAXIS
ETABS
PLAXIS with (P-)
ETABS with (P-)

Bending moment
(KN.m)
157.88
156.40
158.13
156.528

Shearing force
(KN)
134.09
134.72
134.14
134.75

Settlement
(m)
0.031
0.032
0.030
0.032

Maximum bending moment, shear force and deflections of the analyzed columns
with PLAXIS and ETABS can be observed in table 11. The comparison of the results
indicates accepted accuracy of PLAXIS in analysis of building frames.
Table 11.Maximum Bending Moment, Shear Force and Deflections for Columns
Axial force (KN)
Bending moment (KN.m)
Column 1
Column 2
Column 1
Column 2
PLAXIS
2460
4290
45.50
17.52
ETABS
2531
4196
44.32
17.78
PLAXIS with (P-)
2450
4290
45.56
17.73
ETABS with (P-)
2530
4196
44.47
18.27

Comparing the Values of static analysis in PLAXIS and ETABS software in Tables
10 and 11 indicates the sufficient accuracy of the software.
CONCLUSIONS
For application of soil-structure interaction effect, generally soil mechanic software
are used. In this study, the verification of the results obtained by PLAXIS geotechnical
software in modeling of structural frames based on theoretical equation of structural
analysis and ETABS software were investigated. By studying the results, it can be said
that PLAXIS has the capability to analyze building frames. This is highly important in
soil-structure analysis.
REFERENCES

ASCE/SEI 7-10 (2010). "Minimum Design Loads for Buildings and Other Structures."
American Society of Civil Engineers, Virginia.
Fatahi, B., Basack, S., Ryan, P., Zhou, W.H., and Khabbaz, H. (2014). "Performance
of laterally loaded piles considering soil and interface parameters." Geomechanics and
Engineering, vol. 7, no. 5, pp. 495-524.
Hick, T.G. (2009). "Civil Engineering Formulas." McGraw-Hill, USA.
Hokmabadi, A., Fatahi, B., and Samali, B. (2014). "Physical modeling of seismic soilpile-structure interaction for buildings on soft soils." International Journal of
Geomechanics, vol. 15, no. 2; 10.1061/(ASCE)GM.1943-5622.0000396).
Lamb, H. (1904). "On the Propagation of Tremors over the Surface of an Elastic
Solid." Philosophical Transactions of the Royal Society of London. Series A,
Containing Papers of aMathematical or Physical Character., Vol. 203: p.1.
Liao, H.J., Liu, J., Zhao, Y.G. and Xiao, Z.H. (2007). "Analysis of Soil-Structure
Interaction with Finite Element Method." Key Engineering Materials Vols. 340-341:
1279-1284.
Wolf, J.P. (1985). "Dynamic Soil- Structure Interaction." Prentice-Hall Inc, USA.
ASCE

Geo-China 2016 GSP 259

33

Nonlinear Behavior of a 2 2 Pile Group under Rotating Machine-Induced


Vertical Vibrations
S. Biswas1 and B. Manna, M.ASCE2
1

Research Scholar, Dept. of Civil Engineering, Indian Institute of Technology Delhi, New Delhi
110016, India. E-mail: sanjit.jal@gmail.com
2
Assistant Professor, Dept. of Civil Engineering, Indian Institute of Technology Delhi, New Delhi
110016, India. E-mail: bmanna@civil.iitd.ac.in

Abstract: Forced vertical vibration tests have been performed in the field on the 2 2
pile group (l/d = 35, s/d = 3) made of hollow steel pipes embedded in clayey silt and
nonlinear frequency-amplitude curves are obtained for four different eccentric
moments. The soil-pile separation lengths of the pile group are also measured for
different eccentric moments. The numerical analysis has been performed by
continuum approach of Novak and the analytical nonlinear frequency-amplitude
responses of the pile group are compared with the experimental results. It is observed
from the comparison curves that the nonlinear frequency-amplitude response obtained
from the analysis have a very close match with the field nonlinear response curves.
INTRODUCTION
Prediction of the nonlinear characteristics of the pile groups under machine induced
vibratory force is still a challenging task to the designers. Most of the previous designs
of the pile groups have been limited to small-amplitude vibration or linear vibration.
However, many failures of the machine foundations have raised concerns about linear
approaches used for the design of pile foundations. Therefore, there is need to include
the effect of soil nonlinearities and soil-pile gaping in the design of pile foundations
subjected to rotating machine induced vibrations.
Among different methods proposed in the literature to estimate the nonlinear
response of the pile foundations, the continuum approach is the most popular and
reliably used method. Novak et al. (1978) estimated the dynamic soil stiffness
considering plane strain case of an embedded cylindrical body under different modes
of harmonic motion. To include the nonlinear behavior of soil, Novak and Sheta
(1980) proposed a cylindrical boundary zone around the pile having lesser shear
modulus and more damping compared to the outer free field. Based on the
superposition method, Novak and Mitwally (1990) obtained an approximate but
efficient solution for determination of the complex stiffness of pile groups for different
ASCE

Geo-China 2016 GSP 259

modes of vibration. The effectiveness of the proposed analytical methods are also
verified by many researchers with the help of experimental investigations. Novak and
Grigg (1976) conducted dynamic tests on small pile foundations in the field and the
results are compared with the analytical results. Han and Novak (1988) conducted
dynamic field tests on large scale model piles under strong horizontal and vertical
excitations. Similar kind of field study was also performed by Manna and Baidya
(2010) on small pile groups to determine the nonlinear dynamic characteristics of the
pile foundations under vertical vibration. The behavior of the pile group supported
buildings under seismic loading was also studied by Hokmabadi et. al. (2014, 2015)
using the experimental results of shaking table test and numerical analysis.
From the literature review, it is observed that studies related to the estimation of soil
parameters responsible for the nonlinear response of the pile foundation are very
rarely performed. Therefore in this study, vertical vibration field tests are performed
on a 2 2 pile group (l/d = 35, s/d = 3) under a static load of 15 kN to measure the
soil-pile separation lengths for different excitation forces. Theoretical study has been
performed by continuum approach using experimental results to predict the possible
variation of nonlinear soil parameters of the 2 2 pile group.
EXPERIMENTAL INVESTIGATION
Site Characterization
The test site is located in between Block II and III at Indian Institute of Technology
Delhi, Hauz Khas, New Delhi, India. The standard penetration tests (SPT) are
performed in the field at every 1 m interval up to a depth of 6 m and simultaneously
disturbed and undisturbed soil samples are collected for laboratory testing. Various
laboratory tests such as bulk density, natural water content, particle size distribution
and Atterbergs limits tests are carried out on collected soil samples. Based on all the
soil data, the vertical soil profile has been classified into two different soil layers as
per Unified Soil Classification System. The first layer is classified as inorganic clayey
silt with low plasticity (ML-CL) up to a depth of 3 m followed by inorganic gravely
silt with low plasticity (ML) soil. Seismic crosshole tests are also conducted in the
field at every 1 m interval depth to obtain the shear wave velocity of soil. All the soil
properties obtained from the field and laboratory tests are listed in Table 1.
Construction of Pile Group
Hollow steel pipes of 6.0 m in length and outer diameter (d) of 0.166 m (thickness =
7 mm) has been used as piles for dynamic field testing. For end bearing resistance,
circular plates of 0.16 m diameter are welded to the bottom of the pipes. For installing
the piles undersize boreholes are made at 3d spacing with a helical auger of 0.15 m
diameter to ensure a good contact between the pile and soil. The modulus of elasticity
and Poissons ratio of the steel piles are 200 GPa and 0.2 respectively. To measure the
separation between the pile and soil, four pressure sensors of 4 mm diameter and 0.14
mm thickness are attached at the outer surface of a pile with adhesive tapes at an

ASCE

34

Geo-China 2016 GSP 259

35

interval of 0.1 m. After attaching the sensors, the piles are driven up to a depth of (l)
5.8 m into the boreholes with a hammer. A steel pile cap of dimension 0.8 m 0.8 m
0.03 m is connected with the pile by specially fabricated pile cap connectors.
Table 1. Properties of Soil with Depth

Water
Content
(%)
11.33

Wet
Density
(kN/m3)
17.05

13.27

3
4

LL
(%)

PL
(%)

---

---

18.03

25.1

19.0

13.39
13.70

18.85
18.78

-----

-----

14.62

19.03

24.5

21.1

14.17

19.68

---

---

Depth
(m)

Grain Size
Corrected Vs
Distribution
SPT-N (m/s)
(%)
--11
84
Sand = 30.3, Silt =
11
87
58.8, Clay = 10.9
--16
92
--29
103
Gravel = 4.5, Sand
= 35.6, Silt = 51.2,
33
109
Clay = 8.7
--35
113

Vertical Vibration Tests


In order to construct the vertical vibration test setup, total 19 steel plates (weight =
0.5 kN 19 = 9.5 kN) are placed on top of the pile cap to provide static load. Then a
mechanical oscillator assembly (weight = 4 kN) is mounted on top of the steel plates
to produce different magnitudes of sinusoidal excitation forces on the pile foundation.
The oscillator is connected with a 10 hp AC motor by means of a flexible shaft. The
speed of the AC motor can be varied by a speed control unit. A piezoelectric
accelerometer is placed at the centre of the top most plate and it is further connected to
a data logger assembly (data logger and laptop) to record the steady-state amplitude of
vibration. The pressure sensors are connected to a data collection hub which is further
connected with a laptop to measure the contact pressure between the pile and soil with
time. The field dynamic pile testing setup is shown in Fig. 1. Steady-state vertical
responses of the pile group is measured with time for different eccentric moments
(W.e = 0.735, 1.448, 2.117 and 2.721 Nm) under the static load (Ws) of 15 kN. The
oscillator is running in a controlled way within the frequency range of 5.0 Hz to 45.0
Hz to produce the steady-state vibration. The time versus acceleration responses are
recorded for different frequencies by accelerometer-data logger system. A typical time
versus acceleration response of the 2 2 pile group is shown in Fig. 2. From the timeacceleration responses the frequency versus vertical amplitude curves have been
derived for different eccentric moments.
Soil-Pile Separation
During the dynamic field testing the contact pressure between the pile and soil is
also measured with time for different eccentric moments. Due to the slippage and the
application of high magnitude of repetitive stress, the soil adjacent to pile become

ASCE

Geo-China 2016 GSP 259

loose and gradually separated from the pile during the vibration. The contact pressure
between the pile and soil become zero under dynamic loading up to the separated
length. A typical time versus contact pressure curves are presented in Fig. 3. From the
time-contact pressure data the frequency versus maximum contact pressure has been
obtained for all the sensors at steady-state to find the separation length between the
pile and soil under a particular eccentric force. A typical frequency versus maximum
contact pressure curve of the pile group is presented in Fig. 4. From the figure, it is
observed that the maximum contact pressure reading of the 0.1 m sensor is vanished at
the resonant frequency of 38.75 Hz. From this observation, it can be concluded that
the 0.1 m pressure sensor is separated from the adjacent soil under that particular
eccentric moment. Hence, the soil-pile separation length is considered 0.1 m under
that eccentric force (W.e = 2.721 Nm).

FIG. 1. Dynamic test setup of the pile group for vertical vibration.

FIG. 2. Typical time versus acceleration response for different frequencies.

ASCE

36

Geo-China 2016 GSP 259

FIG. 3. Typical time versus contact pressure response for different frequencies.

FIG. 4. Frequency versus maximum contact pressure curves for different depths.
THEORETICAL STUDY
A computer software called DYNA5 (Novak et al. 1999) is used to determine the
nonlinear response of pile foundation based on the continuum approach of Novak and
superposition method. The response of the single pile is calculated using the
continuum approach (Novak and Aboul-Ella 1978). The dynamic response of the pile
group is evaluated by superposition method of Novak and Mitwally (1990) using the
dynamic interaction factors of individual piles (Kaynia and Kausel 1982). To obtain
the nonlinear response of the soil-pile system an equivalent nonlinear soil model is
incorporated in the analysis (Novak and Sheta 1980) by considering a cylindrical
annulus boundary zone around the pile with lesser soil modulus and higher damping
than outer zone. In this soil model the linear-elastic behavior of soil is considered
although the stiffness of the soil model can be reduced by incorporating the boundary
zone around the pile. Therefore, the effect of soil nonlinearly on the dynamic pile
response can be effectively simulated using this soil model. In this study the shear

ASCE

37

Geo-China 2016 GSP 259

modulus, thickness, damping ratio of the boundary zone is considered as Gm, tm, Dm
respectively. The shear modulus of soil is represented as G and the radius of the pile is
represented as R.
Boundary Zone Parameters
The equivalent nonlinear analyses have been performed by incorporating the
parameters like shear modulus reduction factor (Gm/G), weak zone soil damping (Dm),
thickness ratio (tm/R) and most importantly soil-pile separation lengths. These
parametric values are kept constant during the analysis for a particular eccentric
moment. Based on the literatures (Elkasabgy and El Naggar 2013, Han and Novak
1988, Manna and Baidya 2010) the boundary zone parameters are adjusted for
different eccentric moments in a way that the theoretical response curves approach the
vertical vibration test results. The soil-pile separation lengths obtained from the
dynamic field tests are used in the analysis. The shear modulus reduction ratio (Gm/G)
is considered as zero up to the separation lengths from the ground level. The natural
soil damping factor is assumed as 0.1 for all the soil layers. The variations of the
boundary zone parameters with depth for different excitation levels are shown in Fig.
5. From the figure, it is found that the shear modulus reduction ratio (Gm/G) are
reduced whereas the weak zone soil damping (Dm) and the thickness ratio (tm/R) are
increased as the excitation intensity increases due to the progression of the boundary
zone with the increase of eccentric moments. It is also noticed that the shear modulus
reduction ratio (Gm/G) is increased but the weak zone soil damping (Dm) and the
thickness ratio (tm/R) are decreased with depth for all excitation level.

FIG. 5. Variations of boundary zone parameters with depth.


ASCE

38

Geo-China 2016 GSP 259

COMPARISON OF TEST RESULTS AND NUMERICAL ANALYSIS


Nonlinear analysis is performed to obtain the frequency versus amplitude response
of 2 2 pile group for different eccentric intensities. The results obtained from the
analysis have been compared with vertical vibration test results which are presented in
Fig. 6. From the comparison curves, it is observed that the analytical resonant
frequency and amplitude values are very well matched with the vertical vibration tests.
From the responses, single peaks are observed which implies that the pile foundation
behaves as a single degree of freedom system under rotating machine induced vertical
vibrations. From the analytical results, the nonlinear phenomenon is also ensured as
the resonant frequency values are decreased with the increase in eccentric moments
and the resonant amplitude values are also disproportionally increased with the
eccentric intensities. This is observed because of the reduction in soil-pile system
stiffness due to the growth of the weak boundary zone around the pile with the
increase in eccentric intensities.

FIG. 6. Comparison of frequency versus amplitude response curves


CONCLUSIONS
In this present study, the parameters responsible for the nonlinear vertical response
of the pile group are predicted by performing experimental and numerical study. The
major conclusions of this study are stated below.
1.
2.
3.

ASCE

Both experimental and analytical responses of the 2 2 pile group are found
nonlinear under rotating machine induced vertical vibration.
The pile transfers the applied load to the adjacent soil in the sinusoidal pattern as
the contact pressure between the pile and soil is observed sinusoidal.
In this study the maximum separation length of the 2 2 pile group is found 0.1
m under the maximum dynamic force.

39

Geo-China 2016 GSP 259

4.
5.

The nonlinear response of the pile group can be efficiently predicted by the
continuum approach with proper variation of the boundary zone parameters.
The complex vertical stiffness of the pile foundation is decreased with the
increase in excitation intensity due to the expansion of the weak boundary zone
with the excitation levels.

REFERENCES
Elkasabgy, M. and El Naggar, M.H. (2013). Dynamic response of vertically loaded
helical and driven steel piles. Can. Geotech. J., Vol. 50: 521-535.
Han, Y. and Novak, M. (1988). Dynamic behavior of single piles under strong
harmonic excitation. Can. Geotech. J., Vol. 25: 523-534.
Hokmabadi, A.S., Fatahi, B., and Samali, B. (2014). Assessment of soil-pile-structure
interaction influencing seismic response of mid-rise buildings sitting on floating
pile foundations. Computers and Geotechnics, Vol. 55: 172-186.
Hokmabadi, A.S., Fatahi, B., and Samali, B. (2015). Physical modeling of seismic
soil-pile-structure interaction for buildings on soft soils. Int. J. of Geomech., Vol.
15 (2): 04014046.
Kaynia, A.M. and Kausel, E. (1982). Dynamic behavior of pile groups. Procedings
of 2nd International Conference on Numerical Methods of Offshore Piling, Texas:
509-532.
Manna, B. and Baidya, D.K. (2010). Dynamic nonlinear response of pile foundations
under vertical vibration - theory versus experiment. Soil Dyn. and Earthquake
Eng., Vol. 30: 456-469.
Novak, M. and Aboul-Ella, F. (1978). Impedance functions for piles embedded in
layered medium. J. of Eng. Mech., Vol. 104(3): 643-661.
Novak, M., El Naggar, M.H., Sheta, M., El Hifnawy, L., El Marsafawi, H. and
Ramadan, O. (1999). DYNA5 - A computer program for calculation of foundation
response to dynamic loads. Geotechnical Research Centre, University of Western
Ontario, London.
Novak, M. and Grigg, R.F. (1976). Dynamic experiments with small pile
foundations. Can. Geotech. J., Vol. 13: 372-385.
Novak, M. and Mitwally, H. (1990). Random response of offshore towers with pilesoil-pile interaction. J. of Offshore Mech. and Arctic Eng., Vol. 112: 35-41.
Novak, M., Nogami, T. and Aboul-Ella, F. (1978). Dynamic soil reactions for plane
strain case. J. of Eng. Mech., Vol. 104(4): 953-959.
Novak, M. and Sheta, M. (1980). Approximate approach to contact problems of
piles. Proceedings of Dynamic Response of Pile Foundations: Analytical Aspects
M ONeill et al Eds., New York: 53-79.

ASCE

40

Geo-China 2016 GSP 259

41

Pile-Soil Interactions under Thermo-Mechanical Conditions Imposed by


Geothermal Energy Piles in Sand
Rajni Saggu1 and Tanusree Chakraborty2
1

Assistant Professor, Dept. of Civil Engineering, Amity School of Engineering and Technology, Amity
Univ., Sector-125, Noida-201313, UP. E-mail: rsaggu@amity.edu
2
Assistant Professor, Dept. of Civil Engineering, Indian Institute of Technology (IIT) Delhi, Hauz
Khas, New Delhi 110016. E-mail: tanusree@civil.iitd.ac.in

Abstract: The load transfer mechanism of the geothermal energy piles involves
thermal soil-structure interaction. In the present work, pile-soil interaction is
investigated for the energy piles using axisymmetric nonlinear finite element analysis
procedure. For the analyses carried out herein, the stress-strain response of piles is
considered to be linear-elastic. The stress-strain response of sand is simulated using
state parameter based constitutive model CASM. The CASM model is implemented
in finite element software Abaqus through user material subroutine. Floating
geothermal energy piles of different lengths and diameters in sand are analyzed under
allowable axial loading and thermal loading. Two types of sands namely Ottawa sand
and Toyoura sand are considered for analyses. The shear stress in soil at the pile-soil
interface is studied. In addition, parametric sensitivity studies are performed by
varying relative density of sand and coefficient of earth pressure of sand. It has been
observed from the results that pile length, relative density of the soil and the axial
load magnitude are the key parameters in deciding the load-displacement response of
geothermal energy piles subjected to thermo-mechanical loading. Moreover, the soil
particle shape and size also affect the magnitude of shear stress in soil at pile-soil
interface. Negative shear stress generates at the pile-soil interface when subjected to
thermal loading which causes the drag load on the pile.
INTRODUCTION
Reinforced concrete drilled shafts are often used more effectively as geothermal
energy piles due to easy installation of high-density polyethylene plastic pipes
attached to the reinforcement cage which circulates the heat carrier fluid. In current
geotechnical practice, the design of the geothermal energy piles depends profoundly
on empirical equations (Laloui et al. 2006) and uses very high values of factor of
safety which is truly uneconomical. Hence, there is need to understand and quantify
the load transfer mechanism of energy piles under thermal and mechanical loads
through rigorous numerical analysis. In literature, both centrifuge and full scale
experiments on energy piles have been performed by many researchers to study the
thermo-mechanical behavior of the energy piles. Peron et al. (2011) discussed that

ASCE

Geo-China 2016 GSP 259

interaction between pile and soil upon thermal loading of pile and stated that
generation of negative shear stress at the pile-soil interface should be incorporated in
the design of energy piles. Stewart and McCartney (2014) performed centrifuge
modeling of energy foundations and have reported higher thermal axial stress near the
toe of the foundation. Saggu and Chakraborty (2015a, 2015b) have investigated
numerically the thermo-mechanical response of geothermal energy piles and they
noted that the thermal load and soil internal friction angle play an important role in
deciding the pile behavior. However, rigorous finite element (FE) analysis based
predictions for the shear stresses at pile-soil interface for the floating energy piles in
sand under thermo-mechanical loading are currently rare in the literature.
The specific objectives of the present study are (i) to determine the shear stresses in
the soil at pile-soil interface for the piles under allowable axial load and (ii) effect of
soil type on the shear stress. In order to accomplish the objectives mentioned above,
axisymmetric nonlinear finite element (FE) analyses of geothermal energy piles have
been performed herein using the FE software Abaqus (Abaqus Manual Version 6.11).
Numerical investigations have been performed by varying (i) the relative density (DR)
of soil, (ii) the coefficient of lateral earth pressure at rest (K0), (iii) the length of pile
(LP), (iv) the diameter of pile (B) and (v) different sand types. The pile has been
assumed to behave linear elastically. The soil behavior has been simulated with a
critical state soil mechanics based constitutive model CASM (Yu 2006). The
constitutive model has been incorporated in Abaqus through a user defined material
subroutine, e.g. UMAT for defining the mechanical stress-strain behavior of soil and
a user defined material subroutine, e.g. UMATHT for defining the thermal soil
parameters. The temperature change on pile has been assumed to be uniform over the
entire length of the pile (Laloui et al. 2006).
FINITE ELEMENT MODEL AND MATERIAL PROPERTIES
In the present work, the axisymmetric finite element analyses have been performed
using the finite element software Abaqus. The pile and soil geometries are created as
separate parts. The pile is considered to be made of concrete and has been modeled as
linear elastic material with Youngs modulus (E) = 33 GPa and Poissons ratio () =
0.2. The thermal properties of concrete, e.g. thermal conductivity (kc), specific heat
capacity (Cc) and coefficient of thermal expansion (c) are considered as 2.1 W/mC,
800 Joule/C and 10-5/C, respectively. For sandy soil, two types of sands namely
Ottawa sand and Toyoura sand are considered in all the analyses. The sand
surrounding the pile is modeled with critical state soil mechanics based material
constitutive model CASM. The vertical far-field soil boundary in the model has been
placed at a distance of fifteen times the pile diameter from the pile-soil interface and
the bottom soil boundary is placed at a distance of one pile length from the bottom of
the pile. The far-field soil boundary has been restrained in horizontal direction and the
bottom soil boundary has been restrained in vertical and horizontal directions. Among
the thermal boundary conditions, heat flow has been allowed through the far-field
side and the bottom boundaries of the mesh. A constant temperature of 15C has been
assumed at the top soil boundary. Heat flux is assumed to be zero along the axis of
symmetry. The ambient ground temperature has been assumed to be 15C (Laloui et

ASCE

42

Geo-China 2016 GSP 259

43

al. 2006). The analysis domain has been discretized using four node axisymmetric
thermally coupled quadrilateral, bilinear displacement and temperature elements
(CAX4T) for the soil as well as for the pile. A refined mesh has been used near the
pile-soil interface with the minimum element size (ts) of 0.075 mm in soil. The
maximum element size has been taken as 0.15 mm at the far-field regions. The
Ottawa sand used in the present study (designated as ASTM C778) is a medium-sized
silica sand with round to subround particles, with D50 = 0.39 mm. Toyoura sand is
fine subangular to angular silica sand with D50 = 0.19 mm. The interface between the
pile and the soil surfaces has been modeled as frictional contact in the tangential
direction with a coefficient of friction (= tan) between soil and concrete of 0.55. In
the normal direction between the pile and soil, hard contact with zero penetration has
been considered. Heat conduction between pile and soil has been made possible by
defining a gap conductance value equal to the soil thermal conductivity. In the present
study, a FORTRAN code has been written for drained and undrained triaxial
compression test simulation for the determination of the CASM model parameters.
Herein, the model parameters have been determined for Ottawa sand with minimum
void ratio emin = 0.48 and maximum void ratio emax = 0.78 and for Toyoura sand emin =
0.6 and emax = 0.98. The values of the sand parameters are given in Table 1. The
thermal constitutive model is implemented in Abaqus through user defined material
subroutine UMATHT.
Table 1. Material parameters used for Ottawa sand and Toyoura sand
Mechanical parameters
Slope of critical state line in eln(p') space,
Slope of unloading-reloading line
in e-ln(p') space,
Critical state line intercept at 1 kPa
pressure,
Reference state parameter, R
Model parameter, n
Slope of critical state line in q-p'
space, Mcc
Poissons ratio, s
Shear modulus, G (MPa)
Bulk modulus, K (MPa)
Thermal Parameters
Thermal conductivity, ks (W/mC)
Specific heat, Cs (Joule/C)
Coefficient of thermal expansion,
s (/C)

ASCE

Sand type
Ottawa Toyoura
0.02
0.019

Reference
Loukidis (2006)

0.005

0.005

Yu (2006)

0.8

0.934

Model simulation

0.075
4.5
1.2

0.075
4.5
1.27

0.3
32.6
15

0.15
35
31.9

Yu (2006)
Model simulation
Yu (2006), Loukidis
(2006)
Yu (2006)
Yu (2006)
Murthy et al. (2006)

0.274
722
10-4

0.235
677
10-4

Tarnawski et al.
(2009)

Geo-China 2016 GSP 259

FINITE ELEMENT SIMULATION FOR SAND


Axisymmetric nonlinear finite element analyses of piles in Ottawa sand and
Toyoura sand have been carried out under thermo-mechanical loading using Abaqus
in two steps- (i) a static step to apply the gravity loading and bring the model in
geostatic equilibrium and (ii) a coupled temperature-displacement step to apply the
thermal and mechanical loading. The thermal load T = 21C with 12 days of heating
and 16 days of cooling is applied to the pile using a temperature time-history similar
to that applied by Laloui et al. (2006). The ambient temperature of soil is assumed as
15C. An allowable axial load, Q is applied on the pile head. In present analyses, the
allowable axial load has been decided by considering a factor of safety as 2.5 over the
ultimate load corresponding to pile base settlement as 10% of the pile diameter. The
analyses have been performed for the piles in loose sand of DR = 50%, medium dense
sand of DR = 70%, and very dense sand of DR = 85%. The piles considered herein are
floating piles in a single layer of sand. Two different pile lengths, e.g., 15 m and 20 m
and two different pile diameters, e.g., 0.5 m and 1 m have been considered for
analyses. The coefficient of lateral earth pressure at rest (K0) was assumed to be 0.5
and 1 to consider anisotropic and isotropic in situ stress conditions, respectively. For
the pile analyses, thermal expansive strain due to heating in the pile, tensile strain in
the pile, tensile axial load and stress and upward shaft resistance are considered as
positive; depth is considered positive in the downward direction from the ground
surface. Displacement is considered positive in the upward direction.
ANALYSIS RESULTS AND DISCUSSIONS
Figures 1(a) and 1(b) compare the shear stress generated in the soil at pile-soil
interface at the end of heating for floating piles of length LP = 15 m and LP = 20 m, for
diameter B = 0.5 m and figures 1(c) and 1(d) for diameter B = 1. The soil surrounding
the piles is considered as Ottawa sand. Results have been plotted at the end of heating
for DR = 50%, DR = 70% and DR = 85% for K0 = 0.5 and K0 = 1. Negative shear stress
generates at the pile-soil interface near the pile head due to the expansion of the pile
as a result of heating. Negative shear stress generates till one third of length of the
pile for the piles in soil with DR = 50% and K0 = 0.5. However, the depth at which
transition from negative to positive shear stress occurs is less for the piles in denser
soils except for LP = 15 m and B = 1 m in soil with DR = 85% and K0 = 1.0. In this
case, the uplift of the pile happens due to restriction to pile settlement by the highly
dense soil at pile-soil interface. Piles in denser soils can sustain higher magnitudes of
allowable axial load. Thus, these piles are under compression which causes the
generation of positive shear stress for almost the entire length of the pile. For LP = 20
m and B = 0.5 m in soil with DR = 85% and K0 = 0.5, negative shear stress occurs till
one third depth of the pile from pile top due to upward expansion of the pile which is
reasonable. The depth at which transition from negative to positive shear stress occurs
decrease if K0 = 1 as compared to that for K0 = 0.5. However, for LP = 20 m and B = 1
m, positive shear stress occurs for entire length of the pile except negligible amount
of negative shear stress near the pile head which happens due to restriction to upward
displacement of the pile from the axial load at the pile head. Thus, the pile section and

ASCE

44

Geo-China 2016 GSP 259

45

axiaal load magnitude contrrols the upw


ward movem
ment of the ppile due to hheating. The
upw
ward movem
ment of pile relative
r
to th
he soil causees the develoopment of neegative skin
fricction which produces
p
a drag
d
load. Th
he negative skin frictionn results in rreduction off
allo
owable axial load on th
he pile. The drag loadd can be high for longger piles as
obsserved in Fig
gures 1(b) an
nd 1(d). Figure 2 showss the drag looad carried bby the piles.
Forr all the casees, the drag load
l
is signifficant at DR = 50% and K0 = 0.5. Mooreover, the
axiaal load at thee pile head is not significantly high. At K0 = 1.0, the drag looad decrease
whiich is due to
o the denser packing of the soil. Draag load on thhe piles reduuces by 7779%
% if the piles are in soil with DR = 70%
7
and K0 = 0.5. Higheer decrease iin drag load
is observed
o
for longer piless for the sam
me diameter. H
However, w
with increase of DR from
70%
% to 85%, th
he drag load increases by
y 60 to 70%..
K0 = 0.5
K0 = 1.0

12
200
9
900

6
600
3
300
0

10

10

20
2
-50

6
600

50
0

100 150 200 25

5000

10
0

4000

2000
1000
0

15
5

B = 1.0
0m

20
0
-60

50 100 150 200


0 250
Shear strress (kPa)

K0 = 0.5
K0 = 1.0

3000

8
50% 70% 85%

LP = 15
5m
0

12
200

10

85%
50% 70%
7

Shea
ar stress (kPa
a)

K0 = 0.5
K0 = 1.0

18
800

15
-50

700

LP = 20 m
B = 0.5 m

(d) 0

24
400
Axial load (kN)

Depth (m)

(c)

1400

15

50 10
00 150 200 250
Shear strress (kPa)

K0 = 0.5
K0 = 1.0

2100

B = 0.5 m
0

2800

50% 70% 85%


8

LP = 15 m
15
-50

3500
Axial load (kN)

(b) 0

15
500

Axial load (kN)

0
Axial load (kN)

Depth (m)

(a)

50% 70%
85%
7

LP = 20 m
B = 1.0 m
0

60
0 120 180 240 300
Shea
ar stress (kPa
a)

DR

K0

DR

K0

DR

K0

50%
50%

0.5
1.0

70%
70%

0.5
1.0

85%
85%

0.5
1.0

FIG
G. 1. Shear stress
s
in soiil at pile-soill interface ffor the piless under
thermo-mechanical lo
oading at alllowable axiial load

ASCE

Geo-China 2016 GSP 259

46

Higher
H
increaase in the drrag load is observed
o
forr longer pilees. The increease in drag
load
d becomes significantly
s
y high, almosst 90-95% aat K0 = 1.0 w
which may bbe attributed
to the
t higher ax
xial expansion the longeer piles (Sagggu and Chaakraborty 20015a). Thus,
it iss concluded that the soill relative den
nsity and pille length conntrols the m
magnitude off
drag load. Thee drag load
d for the cases considdered hereinn is not of significant
mag
gnitude as the
t piles aree loaded to their
t
maxim
mum capacityy. Howeverr, it may be
noted that the piles loaded as per thee static requuirements oof the buildiing and the
beaaring capacitty of the un
nderlying soil may undeergo high m
magnitude off drag load.
Thee applied ax
xial load and
d drag load should be inn equilibrium
m with sum of positive
shaaft friction an
nd the toe reesistance of the pile. Forr design purrpose, the neegative skin
fricction which causes
c
the drrag load on the
t pile shouuld be taken into consideeration.
D
Drag
lload
d on th
the pile
il (kN)

(a)
( 60

LP = 15 m

(b) 50

B = 0.5 m

LP = 15
5m
B = 1.0
0m

20

40
90
60

20

30

0
40

60

0
40

80

Drag load on the pile (kN)

(d)

150
120

LP = 20 m
B = 0.5 m

350
280

90

210

60

140

30

70

0
40

80

Relative density (%)

Relative dens
sity (%)

(c)

60

LP = 20
2 m
B = 1.0
1 m

0
40
0

60
80
Relative den
nsity (%)

Drag load
K0 = 0.5

60
0
80
0
Relative
e density (%)

K0 = 1.0

FIG. 2. Drag
D
load on
o the piles in
i soils with
h different rrelative denssities
EFFECT OF SOIL
S
TYPE
E ON SHEA
AR STRESS
S
To
T understan
nd the effectt of particlee size and pparticle anguularity on shear stress,
anaalyses are performed fo
or the floating piles unnder thermo--mechanical loading at
ASCE

Geo-China 2016 GSP 259

47

axiaal working load. The piles


p
are con
nsidered to be installedd in Toyourra sand and
Ottawa sand with
w DR = 60
0% and DR = 85%. Figgures 3(a) annd 3(b) show
w the shear
streess generatio
on in soil at pile-soil
p
inteerface at the end of heatting for the ppiles with B
= 1 m and LP = 15 m and 20
0 m, respecttively. The aallowable axiial load for tthe Toyoura
nd with DR = 60% is slig
ghtly higher than
t
the loadd carried by Ottawa sandd, however,
san
the allowable axial
a
load is of same magnitude forr both the saands at DR = 85%. For
both the pile lengths in soill with DR = 60% and K0 = 0.5, negaative shear sttress occurs
neaar the pile top and positive shear streess occurs toowards the ppile base. Thhe transition
from
m negative to
t positive sh
hear stress happens
h
at 1//6th length off the pile if tthe pile is in
Ottawa sand an
nd 1/4th lengtth of the pilee if the pile is in Toyourra sand. Thee magnitude
of positive
p
sheear stress increases alon
ng the pile llength. The magnitude of positive
sheear stress is higher in the
t case of Toyoura saand. At appproximately same axial
mecchanical load, the higherr shear stress in case of Toyoura sannd at the endd of heating
may
y be attributted to the reaadjustment of
o subangulaar particles ddue to self-w
weight of the
soill and expanssion of pile due to heatiing. Howeveer, at DR = 885%, the sheear stress in
soill along the pile
p length is higher in case of Otttawa sand aas comparedd to that for
Toy
youra sand, which impllies that the dense Toyooura sand eexhibit dialattion caused
duee to the expaansion of the pile. Negatiive shear strress near the pile head is also higher
for the Ottawa sand. The higher
h
K0 and
d increased DR of soil reesults in higgher amount
of shear
s
stress which
w
is reassonable due to increasedd stiffness off the soil.
(a) 0

(b) 0

Depth (m)

LP = 15 m

LP = 20 m
5

10

10

15

15
-70

70
0

140

Shear strress (kPa)

Toyourra sand
DR = 85%

DR = 60%
K0 = 0.5
K0 = 1.0

K0 = 0.5
5
K0 = 1.0
0

210

20
-70

7
70

140

210

Shear stress (kPa)

Ottawa sand
s
DR = 85%

DR = 60%
K0 = 0.5
K0 = 1.0

K0 = 0.5
K0 = 1.0

FIG.
F
3. Shea
ar stress in soil
s at pile-ssoil interfacce for the piles in differrent sand
types
t
with DR = 60% aand 85%
CO
ONCLUSIONS
Axially
A
load
ded piles un
nder constan
nt mechaniccal load m
magnitude eqqual to the
allo
owable axiaal load and thermal loading-unloaading (heatinng-cooling) have been

ASCE

Geo-China 2016 GSP 259

analyzed herein using the finite element code Abaqus and a user defined material
subroutine (UMAT). CASM based on the concepts of the critical state soil mechanics
is used for modeling non-linear response of soil. The investigation of (i) shear stress
at pile-soil interface, (ii) drag load on the piles, and (iii) effect of soil type on shear
stress have been studied. The following conclusions are drawn from the studies
conducted herein:
1. The shear stress generation at pile-soil interface depends on the relative
movement of the pile and the soil. Soil relative density and magnitude of axial
load at the pile head controls the resultant displacement of the pile.
2. Negative skin friction is higher for longer piles of lesser diameter. Negative
skin friction varies with soil relative density and the coefficient of earth
pressure of the soil.
3. Drag load should be calculated and catered in the design of geothermal energy
piles.
4. The sand particle shape and size affect the magnitude of shear stress in soil at
pile-soil interface due to heating and cooling of the piles.
REFERENCES
Abaqus/Standard Users Manual, Version 6.11, 2011. Dassault Systmes Simulia
Corporation, Providence, Rhode Island, USA.
Laloui, L., Nuth, M. and Vulliet, L. (2006). Experimental and numerical
investigations of the behavior of a heat exchanger pile. International Journal for
Numerical and Analytical Methods in Geomechanics, Vol. 30: 763-781.
Loukidis, D. (2006). Advanced constitutive modeling of sands and applications to
foundation engineering. Ph.D. Dissertation, Purdue University, USA.
Murthy, T. G., Loukidis, D., Carraro, J. A. H., Prezzi, M. and Salgado R. (2006).
Undrained monotonic response of clean and silty sands. Gotechnique, Vol.
57(3): 273-288.
Peron, H., Knellwolf, C. and Laloui, L. (2011). A method for the geotechnical
design of heat exchanger piles. Geo-Frontiers, ASCE: 470-479.
Saggu, R. and Chakraborty, T. (2015a). Thermal analysis of energy piles in sand.
Geomechanics and Geoengineering: An International Journal, Vol. 10(1): 10-29.
[doi: 10.1080/17486025.2014.923586].
Saggu, R. and Chakraborty, T. (2015b). Cyclic thermo-mechanical analysis of
energy piles in sand. Geotechnical and Geological Engineering, Vol. 33: 321342. [doi: 10.1007/s 10706-014-9798-8].
Stewart, M. and McCartney, J. (2014). Centrifuge modeling of soil-structure
interaction in energy foundations. Journal of Geotechnical and Geoenvironmental
Engineering ASCE, Vol. 140(4): 04013044.
Tarnawski, V. R., Momose, T., Leong, W. H., Bovesecchi, G. and Coppa, P. (2009).
Thermal conductivity of standard sands. Part I. Dry state conditions.
International Journal of Thermophysics, Vol. 30(3): 949-968.
Yu, H. S. (2006). Plasticity and Geotechnics. Springer Publishers, New York, USA.

ASCE

48

Geo-China 2016 GSP 259

49

Experimental Study on the Dynamic Response of PHC Pipe-Piles in Liquefiable Soil


Fuyun Huang1; Haimin Qian2; and Yizhou Zhuang3
1

Associate Professor, College of Civil Engineering, Fuzhou Univ., No.2 Xueyuan Rd., Fuzhou, Fujian
Province 350108, Peoples R China. E-mail: huangfuyun@fzu.edu.cn
2
Graduate Student, College of Civil Engineering, Fuzhou Univ., No.2 Xueyuan Rd., Fuzhou, Fujian
Province 350108, Peoples R China. E-mail: 350635311@qq.com
3
Professor, College of Civil Engineering, Fuzhou Univ., No.2 Xueyuan Rd., Fuzhou, Fujian Province
350108, Peoples R China. E-mail: 478372092@qq.com

Abstract: Shaking table testing on seismic performance and interaction of PHC pipepiles in saturated soils had been conducted using a series of sine wave, typical
earthquake waves as well as artificial ground motions based on in-site. The testing
results involving time histories of ratios of water pore pressure and responding
displacement related to the soils liquefaction were presented. The testing results
indicate that the natural frequency of soil-pile system gradually reduced, which means
that the piles are damaged gradually. Moreover, the displacement responses of piles in
saturated soil and unsaturated soil have a little difference that the peak displacements
are larger than that in unsaturated soil. In addition, the prestressed degree is benefit to
the resist the damage. The researches can be used to calibrate FE models and provide a
data base on seismic design of PHC pipe piles.
INTRODUCTION
Prestressed high-strength concrete (PHC) pipe-piles are the relatively economic type
of pile foundation that has been commonly used in the coastal soft-soil area of China
due to the high axial bearing capacity, convenience of construction and good quality.
However, PHC pipe-pile belongs to a type of thin-walled concrete member with weak
flexural capacity and shear resistant capability. The damage of PHC pipe-piles during
construction and earthquake induced the limit use of PHC pipe-piles in liquable soil in
the code of design of building foundation in Fujian Province (Wen & Huang, 2016).
Nevertheless, there are still a number of PHC pipe-piles had been used in liquable soil
area, which should be paid more and more attention.
A number of researches on seismic performance of PHC pipe-piles had been
conducted in recent years (Takuya 2003; Chung 2007; Rong 2012; Gao 2012; Sun
2012; Li 2013; Mohamed 2013 and Hokmabadi 2014). However, the researches are
lack of considering the effect of liquefaction during earthquake. Wen & Huang (2016)
was tested 3 scaled PHC pipe-piles in saturated soil with a heavy counterweight under

ASCE

Geo-China 2016 GSP 259

50

shaking table according to a PHC pipe-pile foundation project of high-rise buildings in


Fuzhou City. However, it only simply presented the details of test and failure model of
PHC pipe-pile with the consideration of soil-pile interaction. In this paper, the dynamic
characteristics of PHC pipe-piles in liquefiable soil under Shaking-table are introduced.
Due to the limit of space, only a part of experimental results of pore pressure ratio and
displacement response are presented, while responses of acceleration, soil pressure as
well as strain do not present.
EXPERIMENTAL MODEL
A steel box with a dimension of 1.52.5 meters was fabricated and three scaled PHC
pipe-pile are produced. There are 2.75 meters long and 155mm pile diameter with
52.5mm wall thickness. Each model pile is made of C80 concrete, six 7-wires
11.1mm pre-stressed strand and 6mm spiral hoops. Details of design and similarity
relationship as well as the sensors deployment of model had been presented in Wen and
Huang (2016).

FIG. 1. Shaking-table test of PHC pipe-pile.


LOADING
Sine wave, two typical earthquake waves of Chi-chi and El-Centro as well as an
artificial wave were selected as the input excitations which were compressed by the
time scale factor. Artificial wave was generated by the seismic response spectrum
based on in-site and Chinese code (2006). PHC-2 and PHC-3 pile were covered with
overlying clay, while PHC-1 pile without overlying clay.
The loading system for the piles of PHC-1, PHC-2 and PHC-3 were identical to have
26 cases. In order to study the effect of liquefaction and compare the difference, each
case involved in the soil with unsaturated first (Condition 1) and then in saturated soil
(condition 2). Case 1-4 who were the sine wave with peak acceleration at 0.0414g and
Case 5-8 the Chi-chi wave with 0.0414g to 0.1656g mainly identified the basic
dynamic behavior and response. The peak acceleration of El Centro wave and artificial
wave had nine levels from 0.0414g to 0.3726g (1A-9A) with step of 0.0419g. At each
change of peak acceleration, white noise scanning was adopted to measure dynamic
characteristics of the system to identify the damage degree. The shaking table test was
conducted in Fuzhou University, in-site arrangement of the test is shown in Figure 1.
The details of loading case can refer to Wen and Huang (2016) due to the limit of space.

ASCE

Geo-China 2016 GSP 259

51

EXPERIMENTAL RESULTS
Dynamic Characteristics
Through white noise excitation, the acceleration response of PHC soil-pilesuperstructure system can be measured in each condition, and then the fundamental
frequency was obtained by FFT transform. It is found that the fundamental frequency
of PHC-1 in Condition 1 and Condition 2 were 3.14Hz and 2.28Hz respectively, PHC-2
in condition 1 and condition 2 respectively 3.27Hz and 2.65Hz, PHC-3 respectively
3.57Hz and 3.13Hz.
Compared with three piles, it can be seen that the different soil conditions (Condition
1 to Condition 2) have obvious different natural frequencies of soil-pile system that in
saturated soil is much less than in unsaturated soil (2.28<3.14, 2.65<3.27 and
3.13<3.57). The soil condition has a great influence on dynamic behavior of the PHC
pipe-pile model. Furthermore, the larger prestress of PHC pipe-piles, the larger natural
frequency of soil-pile system (3.14<3.27<3.57 and 2.28<2.65<3.13). The existence of
prestress can increase the basic frequency of soil-pile system and mitigate the damage.
3.2

2.6
2.4
2.2
2.0
1.8
1.6
0

1A

2A

3A

4A

5A

6A

7A

8A

9A

Peak Acceleration of Wave

PHC-1
PHC-2
PHC-3

11
10

Frequency (Hz)

2.8

12

damping ratio (%)

PHC-1
PHC-2
PHC-3

3.0

9
8
7
6
5

1A

2A

3A

4A

5A

6A

7A

8A

9A

Peak Acceleration of Wave

FIG. 2. Fundamental frequency and damping ratio under EL-centro wave


The natural frequency vibration and damping ratio of PHC-1, PHC-2 and PHC-3
pipe-piles with the increase of peak ground accelerations (PGA) of excitation (0.0404g
to 0.3726g or 1A to 9A) are shown in Figure 2. Due to limit of space, only the results in
saturated soil (condition 2) under El-centro earthquake are presented hear. From Fig.2,
we can see that with the increase of PGA, the fundamental frequency of model was
gradually reduced up to less than 50%, which means the piles has damaged greatly
especially for PHC-1. In addition, the damping ratio gradually increased that both of
PHC-1, PHC-2 and PHC-3 have a stable trend. The reasons are under the increase of
vibration excitation, the pore water pressure will be raised, which decrease and soften
dynamic shear strength and dynamic shear modulus of soil.
Displacement Response of Pile Head
Lateral displacement responses of pile head were measured by arranged two
displacement-meters at the top and bottom of PHC pipe-pile. Thus, the difference
between the two displacement meters can be seen as the relative displacement of pile.

ASCE

Geo-China 2016 GSP 259

52

2.5

PHC-1
PHC-2
PHC-3

2.0
1.5
1.0
0.5
0.0

Peak Displacement (mm)

Peak Displacement (mm)

Figure 3 is the diagrams of peak lateral displacements for the three piles under
different frequencies of sine wave (Case 1-4), and Fig.3a & b are their results in
unsaturated and saturated soil respectively. From Fig.3a, it can be seen that when the
frequencies of sine wave are approached to natural frequency of the pile-superstructure
system, the lateral displacement responses are the most obvious in unsaturated soil.
Nevertheless, the response of piles in saturated soil has a little difference. Their peak
displacements are significantly larger than that in unsaturated soil especially PHC-1
because the soil is liquefied (it can be seen from Fig 5&6 later).
12

PHC-1
PHC-2
PHC-3

10
8
6
4
2
0

Frequency (Hz)
(a) In unsaturated soil

Frequency (Hz)
(b) In saturated soil

40
35
30
25
20
15
10
5
0

Artificial Wave
El Centro Wave

0.0828 0.1656 0.2484 0.3312 0.4140

Acceleration (g)
(a)PHC-2

Peak Displacement (mm)

Peak Displacement (mm)

FIG. 3. Responding displacement of pile-head under sine wave


40
35
30
25
20
15
10
5
0

Artificial Wave
El Centro Wave

0.0828 0.1656 0.2484 0.3312

Acceleration (g)
(b)PHC-3

FIG. 4. Responding displacement to PGA in saturated soil


Figure 4 show the diagrams of peak lateral displacements in saturated soil under
seismic waves of Artificial wave and El-Centro wave (Case 5-26). Due to the limit of
space, the results in unsaturated soil do not present. From Figure 4, it can be seen that
the lateral peak displacements are basically linear to the increase of PGA of input
seismic waves when their PGAs are below 0.1656g, but after increasing the PGA, the
peak displacement is increasing quickly and begin to perform nonlinearly because of
the soil softening. Therefore, it indicates that lateral displacement of saturated sand can
be significantly affected when the soil is liquefied (pore pressure ratio is over 0.6).
Furthermore, the responding displacements under artificial wave are more obvious than
that under El Centro wave. Otherwise, by comparing the Figure 4a and 4b, it indicates
that the less of prestress ratio of PHC, the larger of the lateral displacement will be
induced.
ASCE

Geo-China 2016 GSP 259

53

Pore Pressure Ratio

Response of Pore Pressure Ratio


Figure 5 &6 are the pore pressure ratio time-history curve of PHC-1 &2 piles under
different frequencies of sine waves at different depths of soil, where K1 locates at the
depth of 0.5m, K2 at 0.9m and K3 at 1.3m. From the Figure 5, it can be seen that the
distribution of pore pressure ratio was basically similar under the sine wave of 0.0414g
with different frequencies. The pore pressure ratio increased to a maximum very
quickly as the excitation input, and then will be sustaining around the maximum for a
period of time until decreases to zero. Its duration is far larger than that of excitation.
1.0
0.8
0.6

0.8
0.6

0.4

0.4

0.2

0.2

0.0
0

1.2

1.0
1Hz
2Hz
4Hz

0.0
100 200 300 400 500
0

Time (s)
(a) K1

1Hz
2Hz
4Hz

1.0

1Hz
2Hz
4Hz

0.8
0.6
0.4
0.2

100 200 300 400 500

Time (s)
(b)K2

0.0

100 200 300 400 500

Time (s)
(c)K3

Pore Pressure Ratio

FIG. 5. Time-history curve of PHC-1 pore pressure ratio under sine wave
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0

1.4
1.2
1Hz
1.0
2Hz
0.8
4Hz
0.6
0.4
0.2
0.0
100 200 300 400
0

Time (s)
(a) K1

1.4
1.2
1Hz
1.0
2Hz
4Hz
0.8
0.6
0.4
0.2
0.0
0
100 200 300 400

Time (s)
(b)K2

1Hz
2Hz
4Hz

100 200 300 400

Time (s)
(c)K3

FIG. 6. Time-history curve of PHC-2 pore pressure ratio under sine wave
Observing the pore pressure ratio at different depths, pore pressure ratio increased
gradually from top to bottom. The mainly reason is that the pore pressure water at
upper layer is easier to discharge than deeper soil. With the increase of frequency of
sine waves, the pore pressure ratio will be decreased because the faster for the
frequency, the quicker for the dissipation. Comparing Figure 5 and Figure 6, it can be
found that the pore pressure ratio for PHC-2 is slightly larger than that of PHC-1, but
not much obvious. The reason is that for PHC-2, it is covered a clay layer to keep the
pore pressure to decreasing slowly.
Figure 7 and Figure 8 are the pore pressure ratio time-history curves of PHC-1 pile
under the seismic waves of artificial and El Centro, Figure 9 and Figure 10 are the
curves of PHC-2. From Figure 7 and Figure 8, it can be found that the loading pore
pressure of artificial wave was larger than El Centro wave under the same peak
acceleration. In addition, the larger the peak acceleration is, the greater the pore
pressure ratio performed correspondingly, and the pore pressure ratio increased
gradually from top to bottom. However, the pore pressure ratio is not too larger to
ASCE

Geo-China 2016 GSP 259

54

Pore Pressure Ratio

liquefy, this mainly because pore water pressure is more likely to discharge without
upper clay layer.
0.16

0.16
0.12

0.12

0.1656g
0.1242g
0.0828g

0.08

0.12

0.1656g
0.1242g
0.0828g

0.08

20

40

60

80

Time (s)
(a) K1

100

0.00

0.1656g
0.1242g
0.0828g

0.08

0.04

0.04
0.00

0.16

0.04
0

20

40

60

100 0.000

80

Time (s)
(b)K2

20

40

60

80

100

Time (s)
(c)K3

Pore Pressure Ratio

FIG. 7. Time-history curve of PHC-1 pore pressure ratio under artificial wave
0.12

0.12
0.1656g
0.1242g
0.0828g

0.08

0.08

0.04
0.00
0

0.12
0.1656g
0.1242g
0.0828g

0.04
20

40

60

Time (s)
(a) K1

80

100

0.00
0

0.1656g
0.1242g
0.0828g

0.08
0.04

20

40

60

Time (s)
(b)K2

80

100

0.00

20

40

60

80 100 120

Time (s)
(c)K3

Pore Pressure Ratio

FIG. 8. Time-history curve of PHC-1 pore pressure ratio under El Centro wave
0.8

0.8

0.1656g
0.1242g
0.0828g

0.6

0.8

0.1656g
0.1242g
0.0828g

0.6

0.4

0.4

0.4

0.2

0.2

0.2

0.0
0

50

100

150

Time (s)
(a) K1

200

0.0

50

100

Time (s)

150

0.1656g
0.1242g
0.0828g

0.6

200

0.0
0

50

100

Time (s)

(b)K2

150

200

(c)K3

Pore Pressure Ratio

FIG. 9. Time-history curve of PHC-2 pore pressure ratio under artificial wave
0.8

0.8
0.1656g
0.1242g
0.0828g

0.6

0.8

0.1656g
0.1242g
0.0828g

0.6

0.6

0.4

0.4

0.4

0.2

0.2

0.2

0.0
0

50

100

Time (s)
(a) K1

150

200

0.0
0

50

100

Time (s)
(b)K2

150

0.1656g
0.1242g
0.0828g

200

0.0
0

50

100

Time (s)
(c)K3

150

200

FIG. 10. Time-history curve of PHC-2 pore pressure ratio under El Centro wave

ASCE

Geo-China 2016 GSP 259

55

From Figure 9 and Figure 10, it can be found that when there was overlying a clay
layer, the pore pressure at K1 will be at a later arrival peak than K2 and K3. With the
shaking of the experiment, the lower water moves upward increasing the pore pressure
at K1. It is very obvious to see that the pore pressure at K1 reaches the peak and
dissipate more slowly than K2 and K3. Although the water pressure at shallower depth
is more likely to disappear, deeper water will add to shallow soil to slow the dissipation.
In addition, with the increasing of PGA, the ratio of water pore pressure is also
increasing that the ratio is close to 0.6 and begin to liquefy when the PGA reaches
0.1656g. Comparing with Figures 7 and 8, and Figures 9 and 10, the ratio value and
duration of water pore pressure of later is significantly larger than that of former, where
the reason is the same with a clay layer. Moreover, the ratio duration is also
dramatically larger that that of excitations of seismic wave.
CONCLUSIONS
1) With the continuous vibration, models are damaged, and the natural frequency of
soil-pile system gradually reduced, while damping ratio gradually increased. Otherwise,
the prestressed degree is benefit to the resist the damage.
2) The responses of piles in saturated soil and unsaturated soil have a little difference,
and the peak displacements are larger than that in unsaturated soil especially for PHC-1.
3) The testing results indicate that with the increasing of frequencies of sine wave, the
pore pressure ratio will gradually decreases correspondingly. The larger the peak
acceleration is, the greater the pore pressure ratio will be accordingly that the ratio is
close to 0.6 and begin to liquefy as the PGA reaches 0.1656g.
4) The pore pressure ratio will increase to a maximum very quickly as the excitation
input, and then will be sustaining around the maximum for a period of time until
decreases to zero. Its duration is far larger than that of excitations. Covered a clay layer
can help the pore pressure to decreasing slowly and increase the value of ratio to easier
liquefaction.
ACKNOWLEDGMENTS
The authors appreciate the support of
China(No:51578161, 51278126, 2010Y0101).

Natural

Science

Foundation

of

REFERENCES
Chung, SG. and Kim, SR. (2007). "Appraisal of true resistance of phc pipe-piles
driven in thick soft deposit." 2007 International Forum on Strategic Technology :
109-112.
DBJ13-07-2006 Code for Design of Building foundation (2006).
Gao A. (2012). Experimented research on seismic performance of PHC pipe pile and
analysis of bearing capacity. Tianjin University, Dissertation.
Hokmabadi, A. S. Fatahi, B. and Samali, B. (2014). "Physical modeling of seismic soilpile-structure interaction for buildings on soft soils." International Journal of
Geomechanics, Vol. 15(2).
Hokmabadi, A. S. Fatahi, B. and Samali, B. (2014). "Assessment of soilpilestructure
interaction influencing seismic response of mid-rise buildings sitting on floating pile
foundations. " Computers and Geotechnics, Vol. 55: 172-186.

ASCE

Geo-China 2016 GSP 259

Li, Y.C. Xing, K.Y. Liu, H. et al. (2013). "Experimental study of seismic performance
of PHC pipe pile considering soil-pile-structure interaction." Chinese Journal of
Rock Mechanics and Engineering, Vol. 32 (2): 401-410.
Motamed, R. Towhata, I. Honda, T. Tabata, K. and Abe, A. (2013). "Pile group
response to liquefaction-induced lateral spreading: e-defense large shake table test. "
Soil Dynamics and Earthquake Engineering, 51, 35-46.
Nagae T., Hayashi S. (2003). Earthquake-resistant property of prefabricated highstrength concrete pile. ASCE, High Performance Materials in Bridges, 2003:173182.
Rong, X. Xu, XZ. Li, YY. (2012). "Experimental research on aseismic behavior of
PHC pipe piles." Industrial Construction, Vol. 43 (7): 72-75.
Sun, Ye. and Liu, Zuhua. (2012). "Strain distributing disciplinarian of phc pipe-piles
under seismic action." Low Temperature Architecture Technology, Vol. 165: 24-27.
Wen, X.W. Huang, F.Y. Zheng, JS. et al. (2016). Experiment on seismic failure mode
of PHC pipe-piles in liquefiable soil under Shaking Table. GEO-China.

ASCE

56

Geo-China 2016 GSP 259

57

The Elastodynamic Interaction of a Rigid Circular Foundation Embedded in a


Functionally Graded Transversely Isotropic Half-Space
Reza Yaghmaie1 and Hamidreza Asgari2
1

Ph.D. Candidate, Dept. of Civil Engineering, Johns Hopkins Univ., 3400 N. Charles St., Baltimore,
MD 21218. E-mail: ryaghma2@jhu.edu
2
Ph.D., Dept. of Civil and Environmental Engineering, Florida International Univ., 10555 W. Flager
St., Miami, FL 33174.

Abstract: The responses of foundations in a layered transversely isotropic medium


under dynamic loadings have a wide application in geotechnical design of foundations
and piles. The present work proposes a rigorous semi-analytical-numerical solution to
a rigid circular disk problem representing the rigid foundation for a wide range of
mechanical loadings. In this regard, a relaxed treatment of the mixed boundary-value
problem yields to an analytically-derived system of coupled dual integral equations
where can be later reduced to a pair of decoupled Fredholm integral equations through
Fourier and Hankel transforms in polar directions. Furthermore, the transversely
isotropic half-space ground is a functionally graded soil following an exponential
trend in the elasticity constants. To ensure the accuracy of the proposed model and
numerical evaluation of the singular integrals involved, the present work has been
validated with reference solutions for different cases of zero foundation embedments,
different degree of the material anisotropy, frequency of excitation and loading
directions. Numerical results are also included for different cases of zero foundation
embedments, different degree of the material anisotropy, frequency of excitation and
loading directions. The results of this paper can be used for solving a wide range of
linear elastodynamic problems in anisotropic mediums.
INTRODUCTION
The wave propagation in soils has a broad range of applications in soil dynamics,
earthquake engineering, seismology, rock mechanics, oil extractions and
embankments. The medium is often assumed as homogenous elastic and isotropic (see
for example Apsel and Luco (1983) and Yaghmaie and Noorzad (2010)). The
nonlinearity effects in stress-strain relations and history dependent material laws for
plasticity in soil using reduced order computational models are investigated by some
researchers (see e.g. Fatahi and Tabatabaiefar (2014a) and (2014b) for details). The
general solution of the stress and displacement fields due to the induced stress wave
propagation for special homogenous elastic mediums can be obtained through the
ASCE

Geo-China 2016 GSP 259

58

potential functions in an analytic fashion. The potential functions are developed to


inherently decouple the partial differential equations in three dimensions (see for
example Eskandari-Ghadi et. al. (2010) and Yaghmaie et.al. (2013) and (2016)). In
real soil problems, the elasticity modulus of the ground alters in depth therefore one
needs to attenuate the model to account for this alternation. Commonly functionally
graded have been used to mimic the actual soil or medium properties behavior. There
are a few researches, where the Greens functions for either isotropic or anisotropic
functionally graded materials (FGMs) in the elastostatic cases are derived (see for
example Wang et. al., (2006), Kashtalyan and Rushchitsky (2009) and Eskandari and
Shodja (2010)). In this paper an exponential functionally graded soil is assumed to
describe the evolution of elasticity constants beneath the soil. The half-space axis of
Symmetry is in the depth direction (z-). The potential functions of the functionally
graded soil are obtained induced by the response to an arbitrary harmonic load at an
arbitrary distance from the top surface of the soil. The governing equations for the
potential functions are second order and forth order partially differential equations
(PDEs) with constant coefficients, whose solutions are determined by inverse Fourier
series transforms and Hankel integral transforms. The representative kernel integrals
are numerically evaluated for the general harmonically applied loads and are
graphically shown. The results of this paper are in agreement with special isotropic
cases reported in the literature. The general results describing the functionally graded
transversely isotropic half-space may be used as references for soil-structure
interaction problems analyzing and visualizing the effect of degree of change of
material properties with respect to depth on the displacement and stress fields through
the proposed semi-analytical-numerical scheme.
FORMULATION OF THE FUNCTIIONALLY GRADED TRANSVERSELY
ISOTROPIC DOMAIN
An exponentially functionally graded soil is assumed to follow an exponential form.
A cylindrical coordinate system (r,,z), whose z-axis is normal to the horizontal
surface and parallel to the material axis of symmetry is used. The constitutive law
relating stress to strain fields in the transversely isotropic elastic domain follows
rr = c11 ( z ) rr + c12 ( z ) + c13 ( z ) zz

= c12 ( z ) rr + c11 ( z ) + c13 ( z ) zz


rr = c13 ( z ) rr + c13 ( z ) + c13 ( z ) zz

(1)

r = ( c11 c12 )( z ) r , zr = 2c44 ( z ) zr , z = 2c44 ( z ) z


Where ij and ij refer to the stress and strain tensor components respectively. The
material constants cij ( z ) follow an exponential form in depth as
cij ( z ) = c0ij e z

(2)

Where c0ij represents the material property value at z = 0 and is a positive real
valued exponent defining the gradation of the soil in depth. The consistency condition
for the material parameters can be obtained as
c ( z ) > 0 for = 11,33, 23,12 and c11 ( z ) c33 ( z ) c132 ( z ) c33 ( z ) c12 ( z ) > 0 (3)

ASCE

Geo-China 2016 GSP 259

59

The half-space is consists of two regions denoted as region I (0 < z < s) and region II
(z > s). An arbitrary harmonic load applied on an arbitrary surface at z = s is
considered as the excitation induced by the propagating stress waves in the domain.
The representing equilibrium equation for the domain is obtained as
2u 1 ur ur ( c011 c012 ) 2ur c044 2ur ( 3c011 c012 ) u
c011 2r +
+
+ 2

+
r r r 2
2
r z 2

2r 2
2r 2
r

( c011 + c012 ) 2u

2u z
u u
+ c044 r + z + 0 2ur = 0
r
r z
r
2r
z
( c011 c012 ) 2u + 1 u u + c011 2u + c044 2u + ( 3c011 c012 ) ur +
2

2
r r r 2 r 2 2
r 2 z 2
2r 2

r
( c011 + c012 ) 2ur + ( c013 + c044 ) 2uz + c u + 1 uz + 2u = 0
044
0

r
r
z
2r
z r
2ur 1 ur 1 2u
2u 1 u z 1 2u z
+ 2
+
+
+
+
c044 2z +
c
c
+
( 013 044 )
r r r 2
r
r z r z r z
+ ( c013 + c044 )

(4)

2u z
u z
u u 1 u
c033 2 + c013 r + r +
+ 0 2u z = 0
+ c033
z
r r
z
r
where it is assumed that the density also follows an exponential form
as z ( z ) = 0 e az . Note that the temporal displacement field components are expanded

using the Fourier series bases and coefficients corresponding to the harmonic with
angular frequency are [ur , u , u z ] ( r , , z ) = [ur , u , u z ] ( r , , z , t ) e it .
Using the potential two sets of potential functions F and introduced by
Eskandari-Ghadi (2005) for elastic transversely isotropic homogeneous domains, one
can obtain the following representations in the Fourier displacement coefficients for
the functionally graded soil as
2 F
F 1
ur = 3
2

(5)
r z
r r
3 2 F 2 F
u =

+
(6)
r z
r r

2 0 2


u z = (1 + 1 ) r2 + 2 + +
F
(7)
z
z
c
c

(
)

011
012

2 ( c013 + c044 )
c011 + c012
2c044
, 2 =
and 3 =
. The governing
c011 c012
c011 c012
c011 c012
equations for the two scalar potential functions is obtained by implementing equations
(5-7) into (4) as
z
2 2 2 ( 3 2 ) 2
2z 2
2 2z
2
2
r + e
+ e
+
(8)
im 2 m +
e 1 F = 0

1 + 1
z 2
z 4

(9)
20 = 0
Where 1 =

ASCE

Geo-China 2016 GSP 259

60

Where 2im are given in Yaghmaie et. al. (2013) and Yaghmaie et. al. (2016),
2c033
. s and s2 are the roots of the polynomial
4 =
( c011 c012 ) 1
(10)
s 4 + ( 2 2 (1 + ) ) s 2 + (1 + ) = 0
4

and =

( c011 c012 ) ( c011 c012 )


2 0
1 4

1 + .
2
2
2c011s1
1 + 1 2
( c011 c012 ) 2c044 s2

The boundary conditions, loads and problem geometry is axisymmetric and


independent of values. Therefore reducing the circumferential dimension and
utilizing the Hankel integral transform of order -m over equations (8-9), the one
dimensional wave equations in terms of z-coordinate can be obtained as
z
2 2 2 ( 3 2 ) 2
z 2
2 2z
+ 2 e 2 2 + e 2
+
(11)
1m 2 m +
e 1 Fm = 0
1 + 1
z
z
4

2
(12)
0m m = 0
Where Fm and m are the Hankel transformed potential functions and is the
independent variable representing the Hankel transformed radial coordinate r-. The
homogeneous solution to equations (11-12) are
F1m ( , z ) = e

z
2

A1m ( ) e 1z + B1m ( ) e 2 z + C1m ( ) e1z + D1m ( ) e2 z

1m ( , z ) = e

z
2

(13)

E1m ( ) e 3 z + F1m ( ) e3 z for region I

F2 m ( , z ) = e

z
2

A2 m ( ) e 1z + B2 m ( ) e 2 z

2 m ( , z ) = E1m ( ) e

3 z
2

(14)

for region II

Where i are defined in Yaghmaie and Noorzad (2010).

FIG. 1. The schematic of the considered problem in the functionally graded soil

ASCE

Geo-China 2016 GSP 259

61

The potential function coefficients will be obtained from the traction-displacement


boundary conditions. As the material is elastic, the temporal scales in the stress fields
are decoupled from the spatial terms similar to the assumption adopted earlier over the
displacement fields. In general the boundary conditions are of the form
(15)
u ( r , , z = s ) = u ( r , , z = s + )

v ( r , , z = s ) = v ( r , , z = s + )

(16)

w ( r , , z = s ) = w ( r , , z = s + )

(17)

rz ( r , , z = s ) = rz ( r , , z = s + ) + P ( r , )

(18)

z ( r , , z = s ) = z ( r , , z = s + ) + Q ( r , )

(19)

zz ( r , , z = s ) = zz ( r , , z = s + ) + R ( r , )

(20)

zz ( r , , z = 0 ) = 0

(21)

rz ( r , , z = 0 ) = 0

(22)

(23)
z ( r , , z = 0 ) = 0
The above boundary value problem for Fourier coefficients of displacement fields
are solved for various harmonic excitations. The singular kernel integrals in terms of
the scalar potentials in the Hankel transformed domain requires calculation of residues
of the kernel function in the singular points and inverse Hankel transformation over
the nonsingular paths (for details see Yaghmaie et. al (2013) and Yaghmaie et. al
(2016)). The gauss quadrature technique is utilized for integrations of the kernel
function. The next presents the description of the results and comparisons with
existing solutions.
RESULTS AND DISCUSSIONS

For the homogeneous transversely isotropic domain the displacement and stress
P ( r , ) = P0 ( r , )
and
fields
regarding
a
vertical
excitation
(i.e.

Q ( r , ) = R ( r , ) = 0 ) are obtained by Khojasteh et. al. (2008). The comparison of the

results of this manuscript are compared with the existing solution (by
assigning = 0 in the formulations) in Fig. 2. The results of the functionally graded
half-space match the homogeneous case. Then the elastodynamic boundary value
problem in (13-23) is solved for various s and values and three transversely
isotropic materials whose properties are given in Table.1.
Table 1. Properties of the Transversely Isotropic Materials

ASCE

Material

E ( MPa )

E ' ( MPa )

G ( MPa )

G ' ( MPa )

'

I
II
III

50000
100000
150000

150000
50000
50000

20000
40000
60000

20000
20000
20000

0.25
0.25
0.25

0.25
0.25
0.25

Geo-China 2016 GSP 259

62

Validation with existing solutions

z/s
FIG. 2. Validation of the homogeneous soil for radial displacement field along
the dimensionless depth z normalized with s coordinate with existing solution

The vertical and radial displacements along z/s for the three materials given in
Table.1 are plotted in Fig. 3 (a) and (b) respectively. The dimensionless angular
frequency is chosen as 0 =

= 3 and the soil inhomogeneity parameter is = 0.5 .


s
The loading in this case is a singular Dirac function point load at the plane z=s in the
(r )
n where n is an arbitrary vector in the plane z=s. The singularity in the
form of
2 r
real part of the displacement response at z=s is due to the singular load being applied
at the depth z=s.

z/s

z/s

(a)
(b)
FIG. 3. The dimensionless (a) radial and (b) vertical displacement fields along z/s
for 0 = 3 and = 0.5

ASCE

Geo-China 2016 GSP 259

63

The real and imaginary parts of the radial dimensionless displacement component
for the same dynamic loading and inhomogeneity parameter along r/s are illustrated in
Fig. 4 (a) and (b) respectively. As expected the results are oscillatory due to the
harmonic mode being imposed however the real and imaginary parts of Fourier
coefficients decay rapidly to zero along the radial coordinate.

r/s

r/s

(a)

(b)
FIG. 4. The dimensionless (a) real and (b) imaginary parts of radial
displacement fields along r/s for 0 = 3 and = 0.5

The real and imaginary parts of the stress component zz for the same dynamic
loading and inhomogeneity parameter along z/s are illustrated in Fig. 5 (a) and (b)
respectively. The real part of the stress field is singular near z=s and oscillatory. Both
of the real and imaginary components of stress field amplify for Material I however
for the other two materials they exhibit a decaying trajectory.

z/s

z/s

(a)
(b)
FIG. 5. The dimensionless (a) real and (b) imaginary parts of stress field zz along
z/s for 0 = 3 and = 0.5
CONCLUSIONS

The potentials functions governing the response of elastodynamic problem in an


elastic functionally graded transversely isotropic half-space soil subjected to harmonic
excitations is derived. The semi-numerical-analytical methodology in calculating the
resulting singular kernel integrals are validated by comparing the results with the

ASCE

Geo-China 2016 GSP 259

homogeneous ground existing solutions. The exponential behavior of density and


modulus of elasticity is capable of predicting real soil properties behaviors in many
engineering application problems. The computational mathematical results of this
paper can be utilized in engineering design problems in predicting the soil behavior
under drilling effects over the surface or beneath the soil with a high degree of fidelity.
REFERENCES

Apsel, R.J., Luco, J.E., (1983). "On the Greens functions for a layered half space. part
II". Bull. Seism. Soc. Am. 73 (4), 931951.
Eskandari, M., Shodja, H.M., (2010). "Greens functions of an exponentially graded
transversely isotropic half-space". Int. J. Solids Struct. 47, 15371545.
Eskandari-Ghadi, M., (2005). A complete solutions of the wave equations for
transversely isotropic media. J. Elast. 81, 119.
Eskandari-Ghadi, M., Ardeshir-Behrestaghi, A., (2010). "Forced vertical vibration of
rigid circular disc buried in an arbitrary depth of a transversely isotropic halfspace". Soil Dyn. Earthquake Eng. 30 (7), 547560.
Fatahi, B., Tabatabaiefar S., (2014). "Effects of soil plasticity on seismic performance
of mid-rise building frames resting on soft soils". Adv. Struct. Eng. 17 (10), 1387
1402.
Fatahi, B., Tabatabaiefar S., (2014). "Fully Nonlinear versus Equivalent Linear
Computation Method for Seismic Analysis of Midrise Buildings on Soft Soils". Int.
J. Geomech. 14 (4), 115.
Kashtalyan, M., Rushchitsky, J.J., (2009). "Revisiting displacement functions in threedimensional elasticity of inhomogeneous media". Int. J. Solids Struct. 46, 3463
3470.
Khojasteh, A., Rahimian, M., Eskandari, M., Pak, R.Y.S., (2008). "Asymmetric wave
propagation in a transversely isotropic half-space in displacement potentials". Int. J.
Eng. Sci. 46, 690710
Wang, C.D., Pan, E., Tzeng, C.S., Han, F., Liao, J.J., (2006). "Displacements and
stresses due to a uniform vertical circular load in an inhomogeneous crossanisotropic half-space". Int. J. Geomech. 6 (1), 110.
Yaghmaie, R. and Guo, S. and Ghosh, S. (2013). "Simulating coupled transient
electromagnetic-large deformation dynamical mechanical systems using a novel
multi-time scaling method." Engineering Mechanics Institute Conference,
doi:
Northwestern
Univerity,
Evanston,
IL.
August
4-7,
2013.
10.13140/RG.2.1.4744.3041
Yaghmaie, R. and Guo, S. and Ghosh, S. (2016). "Wavelet transformation induced
multi-time scaling (WATMUS) model for coupled transient electro-magnetic and
structural dynamics finite element analysis." Comput. Methods Appl. Mech. Eng.
(in Press) doi:10.1016/j.cma.2016.01.016
Yaghmaie, R. and Noorzad, A. (2010). "3D dynamic soil-foundations interactions in a
transversely isotropic medium-impedance characteristics." Engineering Mechanics
Institute Conference, University of Southern California, Los Angeles, CA., August
8--11, 2010. doi: 10.13140/RG.2.1.1988.0803

ASCE

64

Geo-China 2016 GSP 259

65

Axisymmetric Vibration of an Elastic Circular Plate in an Inhomogeneous HalfSpace


Reza Yaghmaie1 and Hamidreza Asgari2
1

Ph.D. Candidate, Dept. of Civil Engineering, Johns Hopkins Univ., 3400 N. Charles St., Baltimore,
MD 21218. E-mail: ryaghma2@jhu.edu
2
Ph.D., Dept. of Civil and Environmental Engineering, Florida International Univ., 10555 W. Flager
St., Miami, FL 33174.

Abstract: The response of foundations in soil is a complicated problem due to the


inhomogeneity of the soil and inertial effects of the ground. A wide range of
applications exist in geotechnical design of foundations and piles. The present work
introduces a rigorous solution methodology to obtain the response of a rigid circular
foundation under the effect of a torque. Due to the action of the evolving torsional
moment, the ground beneath the foundation undergoes a complicated state of stress
and displacement. It is shown mathematically that the initial boundary value problem
can be formulated to a set of time dependent displacement-stress boundary conditions.
A novel treatment of the arising dual integrals describing the state of displacement and
stress leads to an evolving set of linear algebraic Fredholm integral equations of
second kind where the problem is solved numerically. A rigorous parametric study is
carried out to illuminate the effect of inhomogeneity and the interaction of the
elastostatic vs elastodynamic simulations. The proposed analytical and numerical
methodology serves as a unified guide for solving a wide class of foundation structure
interaction problems.
INTRODUCTION
The soil foundation interaction is important in the field of soil-dynamics and earthquake engineering. Most earth-quake engineering studies focus on the effect of the
ground motion on the foundation or the structure on top. The topic of structural
loading on the soil has received a lot of attention in the past. For details see Ai and Liu
(2014), Lin et.al. (2013) and Eskandari-Ghadi et al. (2014). Most but not all of these
studies assume that the soils are homogeneous elastic, transversely-isotropic elastic or
orthotropic and elastic. The main gap in the literature is that actual soils beneath
structures are seldom homogeneous and inertial plays an important role in the
redistribution of the stresses that might affect the neighboring structures via the
ground. In this study we fill this gap by analyzing the inhomogeneous, elastodynamic
problem. The torsional problem was first introduced in 1944 by Reissner and Sagoci.
As in the original Reissner-Sagoci problem, the analysis was restricted to elastostatic

ASCE

Geo-China 2016 GSP 259

66

analyses For details see Rahman (2000), Brzoza and Pauk (2008) and Rahimian et al.
(2006). The elastodynamic Reisnner-Sagoci problem for an inhomogeneous ground
has not been studied and forms as the primary motivation of this study. The effect of
inertia plays an important role as the wave propagation might amplify the stress when
the stress wave transmits thought the ground. The inhomogeneity of the ground is
parameterized into a generic power-law form that allows one to alter the gradation in
the modulus and density of the medium. In section formulation and solution of the
problem, we detail the derivation of the governing elastodynamic scalar wave equation
for the circumferential displacement field associated with the rotation of a rigid
circular disk about its axis of symmetry under the action of a torque T. The considered
problem is axisymmetric and its formulation in the cylindrical coordinate (r,,z) is
independent of variable . The scalar wave-equation is solved using the Hankeltransform in the radial direction and the time-dependence is reduced using the Laplace
transform. The resultant ordinary differential equation for the displacement is solved
in the depth (z-) direction as a modified-Bessel-equation. The inverse-transforms to
the radial and time-domains result in dual-integral equations in terms of
circumferential displacement and stress fields that lead to a set of Fredholms integral
equations of the second kind which are further numerically solved. In section
numerical results and discussions, we have presented the validation of our solutions
methodology with the original elastostatic and elastodynamic homogeneous ReissnerSagoci problem. Results comparing the effect of inhomogeneity and inertia on the
stress-wave propagation in the inhomogeneous ground are also investigated in this
section.
FORMULATION AND SOLUTION OF THE PROBLEM
Consider a cylindrical coordinate system (r,,z) such that the functionally graded
material (FGM) half-space occupies the region 0r<, 0z< and 02. Assume
that a circular rigid punch of radius a is bonded to the boundary plane z=0 and is
rotated by the torque T. Let (z) and (z) denote the shear modulus and density of the
FGM half-space defined as

(1)
( z ) = 0 (1 + z )

( z ) = 0 (1 + z )

(2)

where 0 and 0 are the shear modulus, density at z=0 and and are real
positive constants. Let the axis z be the symmetry axis of the punch. The scheme of
the considered problem is shown in Fig.1.
The investigated problem is axisymmetric, independent on the variable .
Let v ( r , z , t ) , z ( r , z , t ) and r ( r , z , t ) be the displacement and stresses in the FGM
half-space. The constitutive relations are expressed as
(3)
v ( r , z , t )
z ( r , z, t ) = ( z )
z
(4)
v ( r , z , t ) v ( r , z , t )
r ( r , z, t ) = ( z )

r
r

ASCE

Geo-China 2016 GSP 259

67

The dynamic governing equilibrium can then be written as


(5)
2 v 1 v v 2 v ( z ) v ( z ) 2 v
+
2+ 2+
=
2
2
( z ) z z ( z ) t
r r r
r
z
The initial and boundary conditions for the FGM half-space are assumed to be
(6)
v ( r , z , 0 )
=0
and
ICs: v ( r , z , 0 ) = 0
t
(7)
BCs: v ( r , 0, t ) = r for 0 r < a
and
z ( r , 0, t ) = 0 for r > a
where is constant and corresponds to the rigid rotation of the disk. Moreover, the
regularity conditions at infinity should be satisfied as:
(8)
lim v ( r , z , t ) = lim z ( r , z , t ) = 0
z

To solve the formulated above problem the Hankel transforms of the first order and
Laplace transform will be applied described as

(9)
v ( s, z , t ) = rv ( r , z , t ) J1 ( sr ) dr
0

( s, z , p ) = rv ( s, z, t ) e pt dt

(10)

where J1 is the Bessel function of the first kind. v ( s, z , t ) and ( s, z , p ) represent


the Hankel transformed and both Hankel and Laplace transformed displacement fields
respectively. From equation (5) the equilibrium problem at discrete time-transformed
points follows
(11)
2
p 2 2
0
+

+
s

z 2 1 + z z cs2

where cs = 0 / 0 is the constant shear-wave speed in the functionally graded


half-space. Introduce the variable = 1 + z into (11) then it follows that

FIG. 1. The schematic of the considered axisymmetric problem

ASCE

Geo-China 2016 GSP 259

68

(12)
2 p 2
s2
0
+

2 2 cs2 2
The homogenous solution to equation (12) can be obtained by assuming the form
(13)
= y
1
and y is an unknown function. Applying equation (13) to (12)
where =
2
yields the modified Bessels equation as
(14)
2 y 1 y 2
p2
s2
+

+
+
y
=
0

2 2 2 cs2 2
By applying the regularity conditions in (8) the homogeneous solution to equation (14)
can be written in the form
(15)
( s, z, p ) = s 1 A ( s ) K ( s p )

where s p =

1+

p2
and K is the modified Bessels function of second kind and
s 2 cs2

A ( s ) is an unknown function which will be determined from boundary conditions.

The inverse Hankel transform of equation (15) follows

( r , z, p ) = A ( s )(1 + z ) K ( (1 + z ) s p )J1 ( sr ) ds

(16)

where corresponds to the Laplace transformed displacement field. The unknown


coefficient A ( s ) is obtained through satisfying the boundary conditions in (7). The
resulting dual integral equations are expressed as

r
0 ( s ) (1 + h ( s ) )J1 ( sr ) ds = p for 0 r < a

(18)

s ( s )J1 ( sr ) ds = 0 for r > a


0

where ( s ) =

(17)

K ( s p )
s ( p)
s
A( s)
1 . Let function
A ( s ) K 1 ( s p ) and h ( s ) =
s ( p)
s
K 1 ( s p )

( s ) be given in the form

( s ) = g ( q ) sin ( sq )dq

(19)

where g ( q ) is an unknown function. Substituting (19) into (18) it can be observed


that this equations is completely satisfied. The substitution of (19) into (17) leads to
the following Fredholms integral equation of the second kind
a
(20)
1
4r
g ( r ) + ( G ( q r ) G ( q + r ) ) g ( q )dq =
for 0 r < a
p
0

ASCE

Geo-China 2016 GSP 259

69

where G ( x ) = h ( s ) cos ( sx )ds .


0

Note that the above integral is evolving with the time-transformed variable p and is
required to be obtained in every p increment. Similarly the right hand side vector in
equation (20) is updated at every p step. Once the solution to function g ( r ) in
equation (20) is obtained, the unknown coefficient A ( s ) can be calculated.
Surface Contact Stress Distribution

The stress distribution under the rigid punch is of great interest in fracture mechanics
where shear cracks may initiate from near the edges of the punch. The stress
distribution under the punch may be obtained using equations (3), (19), and (20):
a
(21)
0
z ( r , 0, p ) =
J
sr
g
q
sin
sq
dqds
(
)
(
)
(
)
0
0
2 r 0
Equation (21) can be written as
(22)
g (q)
a
z ( r , 0, p ) = 0 J 0 ( sr )
dq
2 r r
q2 r 2
Applying the integration by parts in equation (22) yields
a
0 g ( a )
r
a g ( q ) dq
+
z ( r , 0, p ) =

2a a 2 r 2 g ( a ) r r q q q 2 r 2
The applied torque T due to the rigid rotation of the punch then follows
a 2
a
r 0 g ( a )
r
a g ( q ) dq
T ( p ) = 2
+

dr


2a
a 2 r 2 g ( a ) r r q q q 2 r 2
0

(23)

(24)

NUMERICAL RESULTS AND DISCUSSIONS


Validation with Elastostatic Case

The elastostatic homogenous results may be obtained by replacing the variables p


and in the initial boundary value problem in equation (20) with 1 and 0
respectively. In order to avoid the singularity of s ( p ) , = 10 E 6 is utilized to
account for the homogenous case. The results of the simulation are compared with the
existing elastostatic problem given in Rahimian et al. (2006) for the circumferential
displacement field v for various spatial points in the radial and vertical coordinate
system in Fig. 2. The results are in good agreement with the reference solution. For
more details see Yaghmaie and Noorzad (2010), Yaghmaie et al. (2013) and
Yaghmaie et al. (2016).

ASCE

Geo-China 2016 GSP 259

FIG. 2. The comparison of elastostatic results obtained from the proposed model
with existing results for displacement component v with radial coordinate r.
Parametric Study for FGM Half-Space

Consider now the shear modulus and density of the non-homogeneous half-space
given in equations (1) and (2) evolve in a quadratic form, i.e. = 2 . Using equation
(13) it follows that = 0.5 .

(a)
(b)
FIG. 3. The contours of displacement field at t = 0.1s for (a) homogenous half-space
elastodynamic problem (b) FGM half-space with = 5 elastodynamic problem.

The elastodynamic problem is solved for various values. The spatial contours of
displacement response for the elastodynamic homogenous half-space ( = 10 E 6 )
problem and elastodynamic inhomogeneous case with = 5 corresponding to the time

ASCE

70

Geo-China 2016 GSP 259

71

point t = 0.1s are illustrated in Figs. 3(a) and (b), respectively. Accordingly shear
stress fields z corresponding to the elastodynamic homogenous and inhomogeneous
problem at time t = 0.1s are illustrated in Figs. 4(a) and (b), respectively. It is clear
that the FGM soil with the assumed exponent coefficients leads to a higher shear stress
and correspondingly a lower state of displacement in the half-space.

(a)

(b)
FIG. 4. Shear stress z contours at t = 0.1s for (a) homogenous half-space
( = 10 E 6 ) elastodynamic problem (b) FGM half-space with = 5 elastodynamic
problem.
CONCLUSIONS

The elastodynamic Reissner-Sagoci problem accounting for the inhomogeneity of


the ground beneath the foundation is derived. The equations are solved employing
Hankel-transforms and Laplace-transforms. The resulting dual-integral equations are
solved numerically with gauss-quadrature technique. The inhomogeneity of the soil
can easily be altered through the proposed power-law formula. The accuracy of the
proposed model is verified by existing elastostatic results in the literature. The results
of this manuscript can be used for solving engineering problems dealing with inertial
effects and inhomogeneous domains.
REFERENCES

Ai, Z.Y. and Liu, C.L. (2014). "Axisymmetric vibration of an elastic circular plate
bonded on a transversely isotropic multilayered half-space." Soil Dyn. Earthquake
Eng., Vol. 67: 257-263.
Brzoza, A. and Pauk, V. (2008). "Torsion of rough elastic half-space by rigid punch."
Arch. Appl. Mech., Vol. 78(7): 531-542.
Eskandari-Ghadi, M., Nabizedeh, S.M. and Ardeshir-Behresteghi, A. (2014). "Vertical
and horizontal vibrations of a rigid disc on a multilayered transversely isotropic
half-space." Soil Dyn. Earthquake Eng., Vol. 61-62: 135-139.
ASCE

Geo-China 2016 GSP 259

Lin, G., Han, Z and Li, J. (2013). "An efficient approach for dynamic impedance of
surface footing on layered half-space." Soil Dyn. Earthquake Eng., Vol. 49: 39-51.
Rahimian, M., Ghorbani-Tanha, A.K. and Eskandari-Ghadi, M. (2006). "The
Reissner-Sagoci problem for a transversely isotropic half-space." Int. J. Numer.
Anal. Methods Geomech., Vol. 30(11): 1063-1074.
Rahman, M. (2000). "The Reissner-Sagoci problem for a half-space under buried
torsional forces." Int. J. Solids Struc., Vol. 37(8): 1119-1132.
Reissner, E. and Sagoci, H.F. (1944). "Forced torsional oscillations of an elastic halfspace I." J. Appl. Phys., Vol. 15: 652-654.
Sagoci, H.F. (1944). "Forced torsional oscillations of an elastic half-space II." J. Appl.
Phys., Vol. 15: 655-662.
Yaghmaie, R. and Guo, S. and Ghosh, S. (2013). "Simulating coupled transient
electromagnetic-large deformation dynamical mechanical systems using a novel
multi-time scaling method." Engineering Mechanics Institute Conference,
August
4-7,
2013. doi:
Northwestern
Univerity,
Evanston,
IL.
10.13140/RG.2.1.4744.3041
Yaghmaie, R. and Guo, S. and Ghosh, S. (2016). "Wavelet transformation induced
multi-time scaling (WATMUS) model for coupled transient electro-magnetic and
structural dynamics finite element analysis." Comput. Methods Appl. Mech. Eng.
(in Press) doi:10.1016/j.cma.2016.01.016
Yaghmaie, R. and Noorzad, A. (2010). "3D dynamic soil-foundations interactions in a
transversely isotropic medium-impedance characteristics." Engineering Mechanics
Institute Conference, University of Southern California, Los Angeles, CA., August
8--11, 2010. doi: 10.13140/RG.2.1.1988.0803

ASCE

72

Geo-China 2016 GSP 259

73

Numerical Study on a Novel Vibration Screening Technique Using Intermittent


Geofoam
M. Majumder1 and P. Ghosh2
1

Research Scholar, Dept. of Civil Engineering, Indian Institute of Technology Kanpur, Kanpur 208016,
India. E-mail: mainakm@iitk.ac.in
2
Associate Professor, Dept. of Civil Engineering, Indian Institute of Technology Kanpur, Kanpur 208016,
India. E-mail: priyog@iitk.ac.in

Abstract: In the present study, vibration screening efficiency of intermittent geofoam


(IF) in-filled trench, which is practically a combination of geofoam and air pocket
placed alternately throughout the depth of the trench, has been explored in detail using
finite element analysis. The influence of different geometric parameters of the trench,
frequency of the excitation, and geofoam density on the screening efficiency of IF has
been studied thoroughly considering non-homogeneous layered soil deposit. The
screening efficiency of IF has been compared with that of continuous geofoam (CF) infilled trench.
INTRODUCTION
The awareness of environmental issues such as ground vibration due to the operation
of machines in urban areas, has led to increasing interest in developing methods for
vibration screening. Different types of vibration screening methods may be used
depending on the construction feasibility and geological situation. Woods (1968) first
conducted several full-scale tests in layered sandy deposit with different trench
geometries to explore the effectiveness of open trench as the vibration barrier. Later,
Haupt (1981) reported the study on the effectiveness of the open trench. However, due
to construction difficulty and other associated problems an open trench for vibrations
screening purpose in the urban areas is almost infeasible. Therefore, several in-filled
trenches with different construction materials have been tried to serve the purpose such
as gas membranes (Massarsch 2005), polystyrol fillings (Francois et al. 2012), sand bags
(Al-Hunaidi and Rainer 1991) etc. In recent times, several alternative barrier materials
are getting used to retard the ground vibrations; one such material is geofoam, which can
be used as an in-filled vibration barrier. Geofoam has three dimensional closed cell
structural arrangements whose 95% of the volume gets occupied by air, which makes the

ASCE

Geo-China 2016 GSP 259

74

material very lightweight (1% of the weight of traditional earth materials) with low
acoustic impedance and hence, is suitable for effective vibration barrier. It has been
reported that geofoam can play an effective role in screening the vibrations (Wang et al.
2006; Alzawi and El Naggar 2009), but not much attention is given on the optimization
of the usage of geofoam to achieve better screening efficiency. In the present study,
giving emphasis on this optimization issue a new screening technique, intermittent
geofoam (IF) in-filled trench is proposed considering a strip foundation resting on nonhomogeneous layered soil deposit under dynamic loading analyzing with twodimensional plane strain finite element analysis. The active vibration screening
efficiency of IF is estimated and compared with that of continuous geofoam (CF) infilled trench and open trench (OT).
PROBLEM DEFINITION
In the present analysis, a rigid embedded strip foundation subjected to a vertical
dynamic load intensity of p(t) = p0sin(t) is considered. The foundation is placed on
dry, linear elastic non-homogeneous layered soil deposit with an embedment factor
(Df/B) of 1.0. One side trench of depth, d and width, w is placed at a center-to-center
distance of l from the foundation as shown in Fig. 1. Considering a high-speed machine
a constant force amplitude of sinusoidal dynamic load, P0 of 1 kN with an operating
frequency, F of 10 Hz for a duration, t of 10 seconds is applied on the foundation. In
addition to that, a static load intensity of 10 kN/m2 is considered as the self-weight of the
machine and other accessories, which is found to be well below the ultimate failure load
of the foundation under static condition. The objective is to determine the active
vibration screening efficiency of IF in terms of amplitude reduction factor (Arf) by
measuring the reduction in the displacement amplitude at different pick-up points caused
by the sinusoidal dynamic excitation on the foundation. The Arf can be expressed as
(Woods 1968)

Arf =

(UV ) After
(UV ) Before

(1)

Where, (UV)After and (UV)Before are the vertical displacement at different pick-up points
after and before the installation of the trench, respectively. The Arf value is calculated
considering the peak vertical displacement amplitude of the dynamic response.
MATERIALS AND METHOD OF ANALSYS
Material properties
As per Standard specification for rigid cellular polystyrene geofoam (ASTM
D6817), the available density of expanded polystyrene (EPS) geofoam varies from 11.2
kg/m3 (EPS12) to 45.7 kg/m3 (EPS46). In the present study, EPS 12, having the lowest

ASCE

Geo-China 2016 GSP 259

available density, is considered as in-filled material. It has elastic modulus and Poissons
ration of 3.3 MPa and 0.1, respectively. The acoustic impedance of EPS 12 is
determined as ZG = 7.72 kN-s/m3. The soil deposit considered in the present study is
same as that considered by Ghosh and Kumari (2012), where the subsoil strata is
extended up to a depth of 11.7 m followed by the bedrock. The dynamic properties of
the layered deposit are obtained and listed in Table 1. The water table is assumed to be
at great depth and hence, it is assumed to have no significant impact on the dynamic
response analysis. The embedded concrete foundation has the bulk and the shear
modulus of 1.39 107 kN/m2 and 1.04 107 kN/m2, respectively.

FIG. 1. Simplified diagram of influence domain and geometric parameters of IF.

The intermittent geofoam is formed by alternating arrangement of geofoam and


emptying pockets. The height of each pocket along the depth of the trench is determined
based on the concept of unsupported vertical cut in soil, which is determined following
the recommendation of classical soil mechanics. By keeping the pocket height lesser
than the unsupported vertical depth, the intermittent geofoam is placed within the trench
in three layers with a depth of 0.2R, where R is the Rayleigh wavelength (Fig. 1). The
total volume of geofoam (0.033R) is kept constant along the depth of the trench by
varying the height of pockets without violating the concept of unsupported vertical
depth.

ASCE

75

Geo-China 2016 GSP 259

76

Table 1. Dynamic properties of soil deposit


Properties
Static elastic
modulus , Es(kN/m2)
Dynamic elastic
modulus, Ed (kN/m2)
Shear modulus, G
(kN/m2)
Shear wave velocity,
Vs (m/s)
Longitudinal wave
velocity, VP (m/s)
Rayleigh wave
velocity, VR (m/s)
Rayleigh wavelength
(R = VR/F)
corresponding to
F = 10Hz (m)
Acoustic Impedance,
Zs (kN-s/m3)

Layer-1
2.06103

Layer-2
8.02103

Layer-3
1.12104

Layer-4
4.52104

4.20104

7.21104

8.68104

1.80105

1.61104

2.77104

3.33104

6.95104

96.72

123.00

131.29

184.71

180.87

230.01

245.51

345.41

89.47

114.02

121.71

171.23

8.94

11.40

12.17

17.12

307.48

414.02

466.47

690.82

Finite element analysis procedure


The efficiency of IF in reducing the machine borne vibration is studied using
commercially available finite element code PLAXIS V8.5 (Plaxis 2002). Sensitivity
analysis is carried out to determine the optimum domain size based on the average
displacement along the vertical boundary (QR) of the influence domain (PQRS) as
shown in Fig. 1. The domain size is kept fixed as 2.5R along the horizontal direction
(based on the sensitivity analysis) and 11.7 m along the vertical direction (bedrock
level). The entire domain is discretized with six-noded triangular elements. The average
element size (0.121R) is chosen by satisfying the criteria of wave propagation as
proposed by Kramer (1996). The distance between the vibration source and the vibration
barrier (l), width (w) and depth (d) of the vibration barrier are reported as nondimensional terms by normalizing with respect to R of the soil deposit. The interface
element of virtual thickness is placed between the soil and the foundation base.
However, being a small strain problem (Fatahi et al. 2013) no interface element is
considered between the soil and the geofoam. Total fixities are applied at the base of the
model, whereas horizontal fixities are applied at the extreme vertical boundaries
restraining the motion along the horizontal direction. In addition, the absorbent boundary
condition is considered at the extreme boundaries to absorb the increment of stresses on
the boundaries caused by the dynamic loading and to avoid the reflection of waves back

ASCE

Geo-China 2016 GSP 259

to the soil body as described by Lysmer and Kuhlmeyer (1969). The vertical
displacements at different pick-up points along the ground surface are captured during
the dynamic analysis.
RESULTS AND DISCUSSION
Effect of depth of trench
The effect of depth of trench has been studied by varying d/R from 0.6 to 1.0, keeping
other parameter constant (l/R = 0.4, w/R = 0.06 and F = 10 Hz). From Fig. 2 it is clear
that the screening efficiency of IF is higher than CF, whereas the screening efficiency
closer to OT can be achieved by reducing the volume of geofoam or increasing the
height of pockets in the intermittent geofoam. The value of Arf is reported for x/R = 1.85
as because of a local maximum does exist at this pick-up point (Woods 1968).

Effect of geofoam density


The effect of geofoam density on the screening efficiency of IF is presented by
varying the density of geofoam in the trench as 28.8 kg/m3 (EPS29), 18.4 kg/m3
(EPS19), 14.4 kg/m3 (EPS15), 11.2 kg/m3 (EPS12), as per relevant ASTM code. It can
be noted from Fig. 3 that the screening efficiency becomes better as the density of the
geofoam decreases.
Effect of location of trench
To study the suitable location of the trench, l /R ratio is varied from 0.30 to 0.10. It
can be seen from Fig. 4 that the magnitude of Arf decreases with decrease in l/R, which
indicates the enhancement in the screening efficiency with decrease in the distance
between the source and the trench.
Effect of excitation of frequency
The excitation frequency of the machine generally ranges between 10 Hz to 50 Hz
depending on the type of applications. The value of Arf decreases with increase in the
frequency as shown in Fig. 5. As the frequency of the machine excitation increases, the
wavelength of the propagating waves becomes smaller and hence, there exists an
increase in the screening efficiency.
CONCLUSIONS

The present numerical study explores the effectiveness of IF as vibration barrier in


comparison to CF and OT. Keeping the economic issue in mind it can be concluded that
properly designed IF could be an effective alternative for vibration screening technique.

ASCE

77

Geo-China 2016 GSP 259

FIG. 2. Variation of Arf with depth of the trench.

FIG. 3. Variation of Arf with geofoam density.

ASCE

78

Geo-China 2016 GSP 259

FIG. 4. Variation of Arf with location of trench.

FIG. 5. Variation of Arf with excitation frequency.


REFERENCES

Al-Hunaidi, M.O. and Rainer, J.H. (1991). Remedial measures for traffic induced
vibrations at a residential site Part1: field test. Canadian Acoustics/Acoustique
Canadiennne, Vol. 19(1): 3-13.
Alzawi, A. and El Naggar, M.H. (2009). Vibration scattering using geofoam material as
vibration wave barriers. Proceedings of 62nd Canadian Geotechnical Conference,
NB, Canada, September 20-14.p. 997-1104.

ASCE

79

Geo-China 2016 GSP 259

ASTM D6817-06. (2006). Standard specification for rigid cellular polystyrene


geofoam. American society of testing and materials, West Conshohocken,
Pennsylvania, USA.
Davies, M.C. (1994). Dynamic soil-structure interaction resulting from blast loading.
Proceedings of the Int. Conf. Centrifuge Modelling (Centrifuge 94). p. 319-324.
Fatahi, B., Fatahi, B., Le, T.M. and Khabbaz, H. (2013). Small-strain properties of soft
clay treated with fibre and cement. Geosynth Int., Vol. 20 (4): 286-300.
Francois, S., Schevenels, M., Thyssen, J., Borgions, G. and Degrande, G. (2012).
Design and efficiency of a vibration isolating screen in soil. Soil Dyn Earthq Eng.,
Vol. 39: 13-127.
Ghosh, P. and Kumari, R. (2012). Seismic interference of two nearby horizontal strip
anchors in layered soil. Nat. Hazards., Vol. 63: 789-804.
Haupt, W.A. (1981). Model tests on screening of surface waves. Proceedings of 10th
Int. Conf. Soil. Mech. Found. Eng., Stockholm, Sweden. p. 215-222.
Inglis, D., Macleod, G., Naesgaard, E. and Zergoun, M. (1996). Basement wall with
seismic earth pressure and novel expanded polystyrene from buffer layer.
Proceedings of the Tenth Annual Symposium of the Vancouver Geotechnical Society,
Vancouver, British Columbia, Canada. p. 18.
Kattis, S.E., Polyzos, D. and Beskos, D.E. (1999). Modelling of pile barriers by
effective trenches and their screening effectiveness. Soil Dyn Earthq Eng., Vol. 18
(1): 1-10.
Kramer, S.L. (1996). Geotechnical earthquake engineering. Prentice-Hall international
series in Civil Engineering and Engineering Mechanics, PrenticeHall, New Jersey.
p. 281-283.
Lysmer, J. and Kuhlmeyer, R.L. (1969). Finite dynamic model for infinite media. J
Eng. Mech. Division, Proceedings of ASCE, Vol. 95(4): 859-877.
Massarsch, K.R. (2005). Vibration isolation using gas-filled cushions. Soil Dyn
Symposium to Honor Prof. Richard D. Woods (Invited Paper) Geo-Frontiers 2005
Austin, Texas, January 24-26, 2005, p. 22.
PLAXIS 2-D- Version 8.5. (2002). Geotechnical code for soil and rock analysis. Users
Manual, Brinkgreev, R.B.J. ed., A.A., Balkema Publishers, Netherlands.
Wang, Z.L., Li, Y.C. and Wang, J.G. (2006). Numerical analysis of attenuation effect
of EPS geofoam on stress waves in civil defense engineering. Geotext.
Geomembranes., Vol. 2(5): 265-273.
Woods, R.D. (1968). Screening of surface waves in soils. J. Soil Mechanics & Fdns.,
Vol. 94: 951-979.
Zarnani, S. and Bathurst, R.J. (2007). Experimental investigation of EPS geofoam
seismic buffers using shaking table tests. Geosynth Int., Vol. 3: 165-1

ASCE

80

Geo-China 2016 GSP 259

81

Seismic Resistance of Batter Pile Foundations in Peaty Soft Ground


Koichi Tomisawa1 and Koichi Isobe2
1

Senior Researcher, Civil Engineering Research Institute for Cold Region, Sapporo 0608602, Japan. Email: ko-tomsw@ceri.go.jp
2
Associate Professor, Graduate School of Engineering, Hokkaido Univ., Sapporo 0608628, Japan. Email: kisobe@eng.hokudai.ac.jp

Abstract: Foundations in soft ground which support structures hold the potential for
excessive deformation caused by a large-scale earthquake. Such foundations need to
be designed in order to meet sufficient seismic resistance and avoid excessive
deformation against a large-scale earthquake. One method to achieve this objective is
the use of batter pile foundations, which have slanted piles to improve the maximum
bearing capacity and permit the reduction of construction costs as compared to those
for vertical pile foundations. Batter pile foundations has been regarded as a rational
foundation in terms of lateral resistance and recently tended to be reapplied in Japan in
particular, because various piling methods (such as rotary penetration pile application)
have been developed. However, seismic behavior of batter pile foundations in peaty
soft ground has not yet been clarified. In this study, dynamic centrifugal model tests
were conducted for various pile specifications to elucidate the seismic behavior of
batter pile foundations in peaty soft ground.
INTRODUCTION
A number of studies have examined the mechanism of battered pile foundations in
light of their potential to provide higher horizontal resistance and maximum bearing
capacity than vertical pile foundations. In previous studies conducted by Satoh et al.
(Satoh et al. 1969a and 1969b), the mechanical behavior and usefulness of battered
pile foundations were verified through theoretical analysis. Design methods based on
these studies (e.g., Japan Road Association 2007) are currently applied. However, as
the seismic behavior of battered pile foundations in soft ground under Level 1 and
Level 2 earthquake motions has not been fully verified, centrifuge model tests were
conducted to confirm the suitability of such foundations for soft peat ground where
consolidation settlement may occur. In the tests, seismic performance in soft ground
for both Level 1 and Level 2 earthquake motions was examined.

ASCE

Geo-China 2016 GSP 259

82

OVERVIEW OF CENTRIFUGE MODEL TESTS


Centrifuge model tests were performed to clarify the static and dynamic behavior of
battered pile foundations. In the tests, dynamic excitations were applied to the vertical
pile foundation and the battered pile foundation in a 50 G centrifugal field. The model
setups and instrumentations of the pile foundations are shown in Figure 1. The model
pile was made of steel (JIS G 3101:2004. Rolled Steels for General Structure.
Grade:SS400) with an outer diameter of 10 mm, a thickness of 0.2 mm and a length of
400 mm to simulate a prototype scale steel pipe pile with a diameter of 500 mm, a
wall thickness of 10 mm and a length of 20 m. In both foundation models, the piles
were fixed at the bottom to represent supported pile boundary conditions. As the main
purpose of this study was to use a battered pile foundation as an abutment foundation
of road bridges, an arrangement involving four coupled piles (2 2) was adopted, and
a weight of 800 g which is equivalent to 980 kN in prototype scale was installed to the
pile heads to represent the mass of the superstructure. To simulate soft peat ground as
a typical soft soil, model ground was prepared by mixing peat moss and kaolin clay at
a dry weight ratio of 1:1 to represent saturated ground with an initial water content of
w = 300%. Table 1 shows the physical property of the soil.
Horizontal load

Vertical pile

A Accelerometer
P Strain gauge
d : Laser transducer

Mass 800g
d1
A15
A16
A14
A13
A12

P6
P5
P4

d2
A8
A7
A6

A11

P3

A5

A10

P2

A4

A9

P1

A3
A2

Peat ground

Base ground

A1
45

700
790

45

Instrumented pile

(a) Vertical pile foundation

(b) Battered pile foundation

FIG. 1. Model setup and instrumentation.


To confirm the effects of the pile spacing (distance between pile centre =D, D:
pile diameter) and the inclination angle () on the seismic behavior of the battered pile
foundations. The combination of three pile spacing (3D, 4.5D and 6D) and three
inclination angle (9, 12 and 15 degrees) was adopted for tests with battered pile
foundation. Together with the test for vertical pile foundation, a total number of 6 test
cases were carried out. Table 2 shows the test conditions. For each model the
following loading steps were followed: (1) static horizontal loading; (2) Level 1
dynamic excitation; (3) static horizontal loading; (4) Level 2 dynamic excitation; and
(5) static horizontal loading. In dynamic excitation tests, a sine wave with 20 cycles
and a frequency of 1.5 Hz was used as the input wave form to simulate the earthquake
motion in Type III grounds including soft peat ground, and the acceleration amplitudes

ASCE

Geo-China 2016 GSP 259

83

of 150 gal and 750 gal were set for Level 1 and Level 2 earthquake motions,
respectively (Japan Road Association 2002a, Japan Society of Civil Engineers 2002).
Table 1. Soil physical properties
Value
Water content w (%)

300
3

Soil particle density (g/cm )

2.01

Void ratio e

6.45

Coefficient of permeability (cm/s)

4.610-4

Compression index Cc

2.08
2

Coefficient of consolidation Cv (cm /d)

7300
2

Coefficient of volume compressibility mv (cm /kN)

0.0015

Table 2. Test conditions


Test No.

Pile type

Pile spacing

Pile inclining angle

Case 1

Vertical

3D

Case 2

Battered

3D

12 degrees

Case 3

Battered

4.5D

12 degrees

Case 4

Battered

6D

12 degrees

Case 5

Battered

3D

9 degrees

Case 6

Battered

3D

15 degrees

SEISMIC PERFORMANCE OF BATTERED PILE FOUNDATION


The seismic performance of the battered pile foundation in soft peat ground was
clarified through centrifuge excitation tests. Focus was placed on its seismic
deformation performance in relation to that of the vertical pile foundation, and on its
soundness under different earthquake motions.
Horizontal displacement response of pile head
The acceleration inputs for Level 1 and Level 2 earthquake motions in the
centrifuge excitation tests are shown in Figure 2. The wave form was input from the
base ground.
Figure 3 shows the horizontal displacement response of the pile head for the vertical
pile foundation and the battered pile foundation in Level 1 and Level 2 excitations
which were measured using a laser displacement transducer. Since there was almost
no difference in static horizontal resistance for battered pile foundation when the pile
spacing was changed from 3D to 6D, focus was mainly placed on the inclination angle
(Cases 2, 5 and 6). As shown in Figures 3 (a), the pile head of the battered pile
foundation showed reducing trend in horizontal displacement response in comparison
to that of the vertical pile foundation (Case 1) under Level 1 earthquake motion. The
decrease value was almost proportional to the increase in the inclination angle. The

ASCE

Geo-China 2016 GSP 259

84

maximum horizontal displacement was approximately 2 cm for vertical pile


foundation, whereas those for battered pile foundations were approximately 1.5 cm,
1.0 cm and 0.5 cm with the inclination angles of 9 degrees, 12 degrees and 15 degrees
respectively. As shown in Figures 3 (b), the maximum horizontal displacement of the
pile head under Level 2 earthquake motion was approximately 6 cm for the case with
an inclination angle of 9 degrees, which was slightly higher than that for the vertical
pile foundation (approximately 5 cm). However, the horizontal displacements were
significantly lower for the cases with inclination angles of 12 degrees and 15 degrees
(approximately 2cm and 1 cm). Since the piles are slightly eccentrically-deformed by
the large input acceleration, the amplitude of the response acceleration has become
asymmetric in Case 1 and 5. However, the deformation during the vibration is
suppressed in accordance with the oblique angle. The mechanism can be explained by
increase in resistance against horizontal force resulting from the axial rigidity of the
batter pile, horizontal displacement of the batter pile is suppressed, comparing to the
straight pile.
Based on these test results, it can be concluded that the seismic performance of
battered pile foundations improves with increased inclination angle in comparison to
that of vertical pile foundation.
Seismic resistance of battered piles before and after vibration
To examine changes in the horizontal resistance of battered pile foundations under
Level 1 and Level 2 earthquake motions, static horizontal loading tests were
performed on coupled battered piles after the dynamic excitation with each level.
Figure 4 shows the static loading curves (horizontal load H versus horizontal
displacement at loading point y) before and after the dynamic excitation for the all 6
test cases (Case 1 (Figure 4 (a) for vertical pile foundation and Case 2 - 5 (Figure 4 (b)
- (f) for battered pile foundation). From the figures, it is confirmed that there almost no
difference in the loading curves before and after the dynamic excitation for all the test
cases.
1000
800
600
400
200
0
-200
-400
-600
-800
-1000

Acceleration (gal)

Acceleration (gal)

1000
800
600
400
200
0
-200
-400
-600
-800
-1000
0

10

20
Time (sec)

(a) Level 1

30

40

10

20
Time (sec)

30

40

(b) Level 2

FIG. 2. Input earthquake motions (in prototype scale).


Same as the results of the static horizontal loading tests, the results of the dynamic
excitation tests indicate that the battered pile foundation in soft ground has higher
horizontal resistance to both Level 1 and Level 2 earthquake motions than the vertical
pile foundation and provides sufficient seismic performance.
ASCE

Geo-China 2016 GSP 259

85

A similar trend was observed for both cases under Level 1 and Level 2 earthquake
motions. Battered pile foundation with larger pile inclination angle generates potential
restraining ability to the earthquake induced horizontal displacement.
There was no difference in the horizontal resistance of battered pile foundation
before and after dynamic excitation under Level 1 and Level 2 earthquake motions,
and there was also no damage to the piles. It can therefore be judged that battered pile
foundations provide required earthquake resistance of the pile structure as specified in
the design standard (Japan Road Association 2002b).
12
Horizontal displacement of pile
head (cm)

Horizontal displacement of pile


head (cm)

3
Case 1

2
1
0
-1
-2
-3
0

10

20
Time (sec)

30

-8
-12
0

Horizontal displacement of pile


head (cm)

Horizontal displacement of pile


head (cm)

0
-4

10

20
Time (sec)

30

40

12
Case 5

2
1
0
-1
-2
-3
0

10

20
Time (sec)

30

Case 5

8
4
0
-4
-8
-12
0

40

10

20
Time (sec)

30

40

12

Horizontal displacement of pile


head (cm)

Horizontal displacement of pile


head (cm)

40

Case 2

2
1
0
-1
-2
-3
0

10

20
Time (sec)

30

Case 2

8
4
0
-4
-8
-12
0

40

10

20
Time (sec)

30

40

12
Horizontal displacement of pile
head (cm)

Horizontal displacement of pile


head (cm)

Case 1

Case 6

2
1
0
-1
-2
-3
0

10

20
Time (sec)

30

(a) Level 1 earthquake motion

40

Case 6

8
4
0
-4
-8
-12
0

10

20
Time (sec)

30

40

(b) Level 2 earthquake motion

FIG. 3. Horizontal displacement response of pile head under Level 1 and Level 2
earthquake motions (in prototype scale).

ASCE

Geo-China 2016 GSP 259

86

500

500
Case 1

After Level 1 excitation

400

400

After Level 2 excitation

Horizontal Load H (kN)

Horizontal Load H (kN)

Case 2

Before excitation

300

200

100

300

200

100

0
0

10

15

20

25

30

Horizontal displacement at loading point y (mm)

10

(a) Case 1

25

30

500
Case 4

Case 3

400
Horizontal Load H (kN)

400
Horizontal Load H (kN)

20

(b) Case 2

500

300

200

100

300

200

100

0
0

10

15

20

25

30

10

15

20

25

30

Horizontal displacement at loading point y (mm)

Horizontal displacement at loading point y (mm)

(c) Case 3

(d) Case 4

500

500

Case 5

Case 6

400

400

Horizontal Load H (kN)

Horizontal Load H (kN)

15

Horizontal displacement at loading point y (mm)

300

200

100

300

200

100

10

15

20

25

Horizontal displacement at loading point y (mm)

(e) Case 5

30

10

15

20

25

30

Horizontal displacement at loading point y (mm)

(f) Case 6

FIG. 4. Static horizontal loading curves of pile before and after Level 1 and Level
2 dynamic excitations (in prototype scale).

ASCE

Geo-China 2016 GSP 259

CONCLUSIONS
The results of centrifuge model tests performed to clarify the mechanism of battered
pile foundations in soft ground confirmed the following findings in regard to dynamic
behavior:
1) In the centrifuge excitation tests, the battered pile foundation in soft ground showed
a trend of reducing pile head horizontal displacement response under both Level 1
and Level 2 earthquake motions in comparison to the vertical pile foundation. The
trend was similar to that confirmed for static horizontal resistance. Seismic
performance was therefore considered generally to be better in the battered pile
foundation. It was also found that a larger inclination angle of the battered pile
resulted in a greater pile displacement restraining ability.
2) Based on the results of the static horizontal loading tests on coupled piles
conducted after the dynamic excitation tests under Level 1 and Level 2 earthquake
motions, it is considered that battered pile foundation provide the required seismic
structural resistance, as the horizontal resistance of the foundation in soft ground
was generally the same as that confirmed in the horizontal loading tests performed
before dynamic excitation was applied.
REFERENCES
Japan Road Association. (2002a). "Specifications for highway bridges, part V: seismic
design": 210-228.
Japan Road Association. (2002b). "Specification for highway bridges, part IV:
substructures": 243-265.
Japan Road Association. (2007). "Pile foundation design handbook, 2006 Edition":
414-429.
Japan Society of Civil Engineers. (2002). "Standard specifications for concrete -seismic performance verification": 107-112.
Satoh A., Akai K. and Funasaki T. (1969a). "Study on calculation method for
negative skin friction and bending of batter piles -part 1: calculation method."
Report of the Japan Highway Public Corporation Laboratory: 76-82.
Satoh A., Akai K. and Funasaki T. (1969b). "Negative skin friction and bending
moment of batter piles -strain measurement of steel pipe piles in the Fukuroi
District", Proceedings of the Japan Highway Public Corporation Laboratory
Conference: 357-379.

ASCE

87

Geo-China 2016 GSP 259

88

Variation in a Shear Modulus of Enzyme-Treated Soil under Cyclic Loading


Ansu Thomas1; R. K. Tripathi2; and L. K. Yadu2
1
2

Dept. of Civil Engineering, National Institute of Technology Raipur. E-mail: athomas.ce@nitrr.ac.in


Dept. of Civil Engineering, National Institute of Technology Raipur.

Abstract: Enzymatic emulsions contain enzymes (protein molecules) that react with
soil molecules to form a cementing bond that stabilizes the soil structure and reduces
the soils affinity for water (Velasquez et al., 2005). The present study presents an
investigation to evaluate the enzyme treated soil under cyclic loading. The study
presents the variation in shear modulus on a selected dosage of enzyme with 28 days
of curing period. Strain controlled cyclic triaxial tests were performed with different
confining pressure, loading frequency and cyclic shear stain amplitude. It is found that
the shear modulus values decreases with number of cycles and increases with increase
in confining pressure. The shear modulus decreases with the increase in loading
frequency and with the increase in cyclic shear strain. Rate of degradation in shear
modulus due to increment in cyclic shear strain is found to be more with higher
confining pressure. It may be because the enzyme treated soil behaves stiffer under
=0.15% due to the lateral support. At higher cyclic shear strain lateral support may
not be able to make it stiff.
INTRODUCTION
In this study, performance of enzyme treated soil is evaluated under cyclic loading
by varying loading rate, loading frequency and confining pressure. From the literature,
the behaviour of the enzyme treated soil generally depends upon the type of the soil,
its mineralogical composition, type of enzyme, dosage of enzyme and curing period
(Tingle et al., 2007, Velasquez, 2005; Khan and Taha, 2015; Rauch et al., 2003).
Though dynamic properties of soils are required in advanced numerical analyses of
geotechnical engineering problems, determination of dynamic properties of enzyme
treated soil is found to be very limited. Skilled procedure and high costs in conducting
tests under cyclic loading can be the possible reasons.
The shear modulus degradation of cement treated clay during cyclic loading is
investigated from a series of cyclic triaxial tests. The parameters considered for the
study are cement content (2.57.5%), curing days (728), cyclic shear strain amplitude
(0.31%), number of loading cycles (1100) and loading frequency (0.10.5 Hz).
They found that the shear modulus degradation decreases with increase in the shear

ASCE

Geo-China 2016 GSP 259

strain amplitude, cement content and curing days. It is also noted that irrespective of
the mix ratio and curing conditions, the degradation decreases with increase in loading
frequency (Subramaniam and Banerjee, 2014). Nonlinear dynamic analyses under the
influence of different earthquake records are conducted using FLAC 2D software, and
the results in terms of the maximum lateral displacements and base shears for both
elastic and inelastic behaviours of the -rise moment-resisting building frames (for
fixed and flexible boundary conditions) are obtained. A comprehensive empirical
relationship is also proposed for determining the lateral displacements of the mid-rise
moment-resisting building frames under earthquake and the influence of soilstructure
interaction. (Tabatabaiefar et al., 2014)
Based on the cyclic tests done on fully saturated sands conducted in undrained
conditions and cyclic tests conducted in the constant- volume equivalent-undrained
conditions in the cyclic strain- controlled mode, with the constant cyclic shear strain
amplitude, it is concluded that the variation of the secant shear modulus and associated
stiffness index with the number of cycles N is distinctly different at different levels of
shear strain amplitude (Vucetic and Mortezaie, 2015). Resonant column tests were
performed to study influences of void ratio, overconsolidation ratio, plasticity,
confining stress, and degree of weathering. In-situ test results showed effects of
increased confining stress with depth, and OCR, mass density, and void ratio were
estimated using correlations to assess current in-situ state. In-situ measurements of
low-amplitude shear moduli from seismic piezocones, seismic flat dilatometers,
spectral analysis of surface waves (SASW), and crosshole tests were found to be in
good agreement with laboratory values using the resonant column (Schneider, 1999).
A novel and enhanced soilstructure model is analysed in the direct method using
state-of-the-art capabilities of FLAC 2D simulating the complex dynamic soil
structure interaction for three types of soil with shear wave velocity 600m/s (Ce),
320m/s (De) and 150 m/s (Ee). The effects of dynamic soilstructure interaction for
seismic design of mid-rise moment resisting building frames resting on soil class Ce is
found to be insignificant while significant effects has been observed on soil class De
and Ee. (Tabatabaiefar and Fatahi, 2014)
MATERIAL USED
Soil Preparation
The soil used in this study is from Tilda, Chhattisgarh state, India. Collected soil
sample (D0) was oven dried and crushed with wooden mallet and sieved from 75 IS
sieve. Properties of the soil are tabulated in Table 1. Terrazyme is a natural, non-toxic
liquid, formulated using vegetable extracts. It is a surfactant (an ionic surface active
agent) which changes the hydrophilic nature of clay and lime materials to
hydrophobic. Its application not only assists in the expulsion of water from soils, but
also aids the lubrication of soil particles and increases the compatibility of many soils.
The reaction of Terrazyme on soil is particularly effective because of the ion-exchange
capacity of clay minerals. TerraZyme changes the plastic characteristics of the soil due
to a reduction in its water absorbing capacity (RaviShankar et al, 2009).

ASCE

89

Geo-China 2016 GSP 259

90

Table 1. Properties oof soil


Prop
perty
Liquid Limit
L
(%)
Plastic Limit
L
(%)
Plasticity Index (%)
OMC
C (%)
3
MDD (kN/m
(
)

Valuee
42.255
18.6
23.655
14.5
13.2

To
o study the variation
v
in shear
s
modullus, optimum
m dosage of T
Terrazyme w
was selectedd
by conducting UCS tests on
o different dosages as shown in ttable 2. Terrrazyme wass
mix
xed with oveen dried soil in differentt proportionss i.e. D1 (2000ml/ 1.5m3 of soil), D22
(200
0ml/ 2m3 off soil). Uncon
nfined comp
pressive strenngth gain forr D2 samplees at 28 dayss
of curing
c
is foun
nd to be morre compared
d to D1.
EXP
PERIMENT
TAL STUD
DY
Cycclic Triaxiall Tests
A series of cyclic
c
undraiined triaxiall tests were conducted on enzyme treated soill
(D2
2) specimens, by using
g a cyclic trriaxial testinng apparatuus. The equiipment unitt
con
nsists of a triiaxial cell, a load frame,, the controll and data accquisition haardware andd
the control softw
ware as show
wn in fig.1 (a) with electtro-mechaniical actuator. Air is usedd
m
Straain- controllled cyclic teests were perrformed undder differentt
as a chamber medium.
con
nfining pressu
ure, loading frequency and
a shear strrain amplitudde. The Enzzyme treatedd
soil specimen (3
38 X 76 cm)) of D2 dosaage with 28 ddays of curinng period is subjected too
the laboratory cyclic
c
triaxiaal tests as giv
ven in Tablee 3. The typpical stress sttrain loop iss
show
wn in Fig. 1 (b)

g system, (b
b) A typical stress strain
n loop
FIG 1 (a)) Cyclic triaxial testing

ASCE

Geo-China 2016 GSP 259

91

Table 2 UCS test reesults


Dosa
age
D0
0
D1
D2
2

UCS (kP
Pa)
213
768
1073

Table 3 Details of Cyclic Triaxiaal Tests Con


nditions
Confining pressure (k
kPa) P
Loading
g frequency (Hz) -f
Cyclic
C
shear strain ampliitude (%) -
En
nzyme dosag
ge
Curring (days) d
No. of cycles (N
N)

50, 100, 150


0.5, 1, 1.5
0.155% ,0.3%
D2
28
10

c
strain approach, it is required to
t input the single-ampllitude axial sstrain in mm
m
In cyclic
whiich can be calculated
c
from
f
the cycclic shear s trains. Cycllic loading iis generallyy
expressed in teerms of cycclic shear sttrains. For a constant ccyclic deforrmation, thee
requ
uired axial strain
s
is callculated as per
p ASTM D3999/D3999M-11 andd the singlee
amp
plitude deforrmation is fo
ound out by using
u
the exp
xpression givven in equatiion 1.
LSAA = SA X LS

(1)

Wh
here: LSA = single ampliitude deform
mation, mm, SA = singlee amplitude axial strain
(dimensionle
(
ess) and Ls = length of test specimeen (mm).
Sheear Moduluss Degradatiion
Thee cyclic stresss-strain beh
haviour of treated soil (D
D2) is analyzzed to deterrmine secantt
sheaar modulus (G). The secant shear modulus
m
is eestimated foor each cyclee of loadingg
from
m the hysterresis loops ob
btained in sttrain controllled cyclic looading as shhown in Fig.
2 (Z
Zhang et. al, 2005).

FIG
G 2 Hystereesis loop forr one cycle of
o loading sh
howing G (Z
Zhang et. all, 2005)

ASCE

Geo-China 2016 GSP 259

Effe
fect of Confin
ning Pressu
ure
Thee shear mod
dulus values decreases with numbber of cyclles and increases withh
incrrease in confining pressure as show
wn in Figs.3 to 5. In this tests D2 ssample weree
takeen after 28 days curing
g and the sample was subjected to cyclic loaading underr
variious cyclic loading.
l
From
m the resultts the shear modulus off the specimeen increasess
with
h the increasse in confinin
ng pressure.

FIG
G 3 Variatio
on in Shear modulus with
w numberr of cycles (P
P= 50kPa, 1100kPa,
150
0kPa; = 0.1
15% and f= 0.5 Hz)

FIG
G 4 Variatio
on in Shear modulus with
w numberr of cycles (P
P= 50kPa, 1100kPa,
150
0kPa; = 0.1
15% and f= 1 Hz)

G 5 Variatio
on in Shear modulus with
w numberr of cycles (P
P= 50kPa, 1100kPa,
FIG
150
0kPa; = 0.1
15% and f= 1.5 Hz)
ASCE

92

Geo-China 2016 GSP 259

Effe
fect of Loadiing Frequen
ncy
Figss 6 to 8 sho
ow the sheaar modulus variations
v
w
with numberr of cycles ffor differentt
load
ding frequen
ncy. D2 speccimen with 28 days curring period w
was consideered for thiss
test and it wass subjected to cyclic shear
s
strain amplitude of 0.15% aand loadingg
quency 0.5 Hz,
H 1 Hz and
d 1.5 Hz. Fig
gs shows thaat the shear m
modulus deccreases withh
freq
the increase in loading
l
freq
quency. The rate of degrradation is foound to be less with thee
incrrease in conffining pressu
ure.

G 6 Variatio
on in Shear modulus with
w numberr of cycles (P
P= 50kPa, = 0.15%
FIG
and
d f= 0.5 Hz, 1Hz, 1.5 Hzz)

G 7 Variatio
on in Shear modulus with
w numberr of cycles (P
P= 100kPa, = 0.15%
FIG
and
d f= 0.5 Hz, 1Hz, 1.5 Hzz)

FIG
G 8 Variatio
on in Shear modulus with
w numberr of cycles (P
P= 150kPa, = 0.15%
and
d f= 0.5 Hz, 1Hz, 1.5 Hzz)

ASCE

93

Geo-China 2016 GSP 259

Effe
fect of Loadiing Rate
Fig.. 9 shows th
he shear mo
odulus variattions with nno of cycles for differennt values off
sheaar strain. In this test specimen
s
were subjecteed to cyclic shear strainn amplitudee
(0.1
15% and 0.3%
%) and load
ding frequenccy 1.0 Hz. Itt can be inferrred from thhe graph thatt
the shear mod
dulus decreaases with the
t
increasee in cyclic shear straiin. Rate off
degradation in shear modu
ulus due to increment
i
inn cyclic sheaar strain is found to bee
morre with high
her confinin
ng pressure.. It may bee because thhe enzyme treated soill
beh
haves stiffer under =0.15% due to the
t lateral suupport. At hiigher cyclic shear strainn
lateral support may
m not be able to mak
ke the soil sttiff. With thee increase inn number off
cyclles the degraadation rate decreases.

FIG
G 9 Variatio
on in Shear modulus with
w numberr of cycles ( = 0.15% aand 0.3%
and
d f= 1Hz.)
CO
ONCLUSION
NS
Bassed on the sttudy, followiing conclusiions are draw
wn:
1) The shear modulus values decreeases with nnumber of cyycles and inccreases withh
increasee in confining
g pressure.
2) The shear modulus decreases with
w the increease in loadding frequenccy. The ratee
of degraadation is fou
und to be lesss with the inncrease in coonfining pressure.
3) Rate of degradation in shear mo
odulus due too increment in cyclic shhear strain iss
found to
o be more with higherr confining pressure. IIt may be bbecause thee
enzyme treated soil behaves stiiffer under =0.15% duee to the lateeral support.
At higheer cyclic sheear strain lateeral support may not be able to makee it stiff.
ACKNOWLED
DGMENTS
S
Thee authors app
preciate the support
s
of Avijeet
A
agenccies for provviding the Teerrazyme.
REF
FERENCES
AST
TM D3999/D
D3999M-11 (2011). Sttandard Testt Methods fo
for the Deterrmination off
the
t
Modulu
us and Dam
mping Prop
perties of Soils Usingg the Cycllic Triaxiall
Apparatus.
A
Annual Book
B
of AST
TM Standarrds, ASTM
M Internatioonal, Westt
Conshohock
C
ken, PA.
Khaan, T. A. an
nd Taha, M. R.,(2015). Effect of T
Three Bioennzymes on C
Compaction,,
Limits, and
Consistency
C
d Strength Characteristiccs of a Sedim
mentary Ressidual Soil,,
Advances
A
in Material Sccience and Engineering,
E
1-9.

ASCE

94

Geo-China 2016 GSP 259

Rauch, A.F., Katz, L. E., and Liljestrand, H. M., (2003). An Analysis of the
Mechanisms and Efficacy of Three Liquid Chemical Soil Stabilizers, Research
Report 1993-1. Center for Transportation Research, University of Texas at Austin.
RaviShankar, A.U., H. K. Rai, and R. Mithanthaya, (2009). Bio-enzyme stabilized
lateritic soil as a highway material, Indian Roads Congress Journal, Vol.70 (2):
144-151.
Schneider, J.A., Hoyos, L., Jr., Mayne, P.W., Macari, E.J., and Rix, G.J. (1999).
Field and laboratory measurements of dynamic shear modulus of Piedmont
residual soils, Behavioral Characteristics of Residual Soils, GSP 92, ASCE,
Reston, VA, pp. 12-25.
Subramaniam, P. and Banerjee, S., (2014). Factors affecting shear modulus
degradation of cement treated clay, Soil Dynamics and Earthquake Engineering,
Vol. 65: 181188.
Tabatabaiefar, H. R., and Fatahi, B., (2014). Idealisation of soilstructure system to
determine inelastic seismic response of mid-rise building frames, Soil Dynamics
and Earthquake Engineering, Vol. 66: 339-351.
Tabatabaiefar, S. H. R., Fatahi, B. and Samali, B., (2014). An empirical relationship
to determine lateral seismic response of mid-rise building frames under influence of
soilstructure interaction, The structural design of tall and special buildings, Vol.
23 (7): 526-548
Tingle, J.S., Newman, J.K., Larson, S.L., Weiss, C.A., and Rushing, J.F., (2007).
Stabilization Mechanisms of Nontraditional Additives, Journal of the
Transportation Research Board, Transportation Research Record, No. 1989, Vol.
2: 5967.
Velasquez, R., Marasteanu, O.M., Hozalski, R. and Clyne, T., (2005). Preliminary
laboratory investigation of enzyme solutions as a soil stabilizer, Report No.
MN/RC 2005-25, Department of Civil Engineering, Minnesota Department of
Transportation Research, Minneapolis.
Vucetic M. and Mortezaie . A., (2015)Cyclic secant shear modulus versus pore water
pressure in sands at small cyclic strains, Soil Dynamics and Earthquake
Engineering, Vol. (70), 60-72.
Zhang, J., Andrus, R. D. and Juang, C. H., (2005). Normalized Shear Modulus and
Material Damping Ratio Relationships, Journal of Geotechnical and
Geoenvironmental Engineering, Vol. 131 (4): 453-464.

ASCE

95

Geo-China 2016 GSP 259

96

Numerical Modeling of a Dynamic PileSoil Interaction in Layered Soil Media


Mohsen Mohammadizadeh1 and Moein Mohammadizadeh2
1

Assistant Professor, Dept. of Civil Engineering, Islamic Azad Univ. Sirjan Brench, Sirjan, Iran. Email: mohammadizadeh@iausirjan.ac.ir
2
Young Researchers and Elite Club, Islamic Azad Univ. Sirjan Branch, Sirjan, Iran. E-mail:
mohammadizadeh@yahoo.com

Abstract: Due to use of pile foundations for important structures such as tall
chimneys, television towers, high rise buildings, high retaining walls, offshore
structures, safe and economic design of types of foundations under dynamic loading is
especially important. First of all, it is necessary to investigated complexity of the
issues related to the behavior of pile and pile-soil interaction evaluated the effects of
the earthquake and methods of foundations analysis. Interactions between the
components of the piled raft foundation have a significant influence on the analysis.
Therefore, the interaction effects have been fully considered. ABAQUS 6.10 Software
has been used for analysis models. The models have been calibrated against available
certified models. Researchers have rarely studied the effects of seismic stations for
piled raft foundations. To get comparable results, the accelerograms have been scaled
before performance of the piled raft foundation studied in two layers of soil media.
The effect of raft foundation thickness, length, and diameter of the piles have been
studied under the Bam earthquake accelerograms in the three of seismic stations. The
results show that maximum and differential settlement decreases and moment is
transferred to the cap pile increase when increase raft thickness.
INTRODUCTION
By increasing the population, requirements and the needs of human increase in the
construction of high-rise buildings. In many cases, engineers are required to build
these structures on the loose soil. In order to transfer the weight of the structure to the
ground using the appropriate arrangements are required. One of the oldest methods to
overcome this problem use the pile foundations for improving bearing capacity of soil
and take advantage of bearing capacity of lower layers of soil that generally has more
bearing capacity than the upper layers. Lack of sufficient knowledge in the fields of
geotechnical has been increased value of safety factor.
Seismic pile-soil interaction can be an important consideration in evaluating the
seismic response of pile, particularly in soft soils and high-rise buildings. Different
researchs had been done in this field and has made certain important achievements but
ASCE

Geo-China 2016 GSP 259

extended field studies continue to discover the methods and expanding research to
realize the actual behavior of pile-soil interaction because it is the complexity and
scope. Near-fault ground motions often contain large long-period pulses in the ground
motion (Somerville, 2002). The first seismological evidence of the near-field
phenomenon was observed in the 1952 Kern County (Colifornia) earthquake. Benioff
reported the first near-field earthquake in 1955(Benioff, 1955). Anderson et al. stated
that non-elastic response of a structure to near-fault ground motion is a function of
resistance and period of structure for the pulse period to the fundamental period of the
record and direction the rupture propagates are important parameters in The structure
of behaviour (Anderson et al., 1978). Leung et al. investigated the optimum length of
the piles in the group and the results of this study showed that proper alignment of the
optimal length of piles can dramatically increase the raft stiffness of and reduce
settlement of differences (Leung et al., 2009). Khoury et al. By using plaxis 3d
foundation software has been analyzed the tallest building in Brooklyn and after
completing structure, the results compared with measured results. The threedimensional finite element result match measured data (Khoury et al., 2011). Eslami et
al. By using ABAQUS software Analyzed piled raft foundation in two cases nonconnected and connection head piles with raft. The results showed that the use of nonconnection head piles with raft , axial stress piles strongly reduce and also focus on
piles in the central area of foundation lead to decrease maximum and differential
settlement (Eslami et al., 2011). Li et al. showed that spectra for seismic design takes
into account the near-fault effects, especially effects of large long-period pulses in
near-fault ground motions, significantly enhance the seismic response of the
continuous girder bridge (Li et al., 2012).
Numerical Investigation
Finite Element Model, Boundary Conditions and Material Properties
Soil is a continuous porous media. The behavior of soil is in general nonlinear under
a large load at pile head. The Mohr-Coulomb model (MCM), which is an elasticperfect plastic model, can be applied for all types of soil. In this paper, by the Mohr
Coulomb failure criterion was used in modelling of interfaces behavior. The shear in
piles is significantly increased due to soil-structure interaction (SSI) effect as inertia of
the considerable foundation mass contributes to this increase in shear of pile. Thus,
neglecting SSI may lead to unsafe seismic design of piles. Since the focus of this
article is on seismic performance of group pile foundation, the superstructure was
modeled by lumped masses on pile cap (the weight of superstructure is 1000 tons) that
there are no significant difference between the results of this article with completely
modeling of superstructure. This approach used widely in different codes (Such as
Eurocode 8, 2006; API, 2006; IBC, 2009). Loads are applied in three steps. Step 1,
Geo-static step is used to establish in-situ stresses in soil. Step 2, Static step is used to
apply weight of superstructure. Step 3, Dynamic step is used to input acceleration time
history at bedrock level in an explicit approach. The model includes 8-node linear
brick, reduced integration (C3D8R) elements for soil, piles and pile cap and infinite

ASCE

97

Geo-China 2016 GSP 259

98

(CIN3D8) element for edge models. To avoid reflections from the free edge of the
soil, an absorption region was used as shown in Fig. 1.

FIG. 1. Eight-node brick elements (C3D8R) and Eight-noded infinite elements


(CIN3D8)

FIG. 2. 3D model of piles group foundation and soil


Table 1. Properties of Concrete and Soil
Material

Behavioral
Model

Elastic
Modulus
(MPa)

Concrete
Piled Raft
Foundation
Soft Clay

Elastic

25000

0.2

MohrCoulomb
MohrCoulomb

5
40

Dense Sand

Poisson's Cohesion
Ratio
(KPa)

Unit
Weight
(KN/m3)

Angle of
Internal
Friction
()
-

0.3

10

16

0.3

42

17

24

The efficiency and behavior of piled raft foundation has been investigating in vertical
two-layer soil model. The mechanical properties used for the piles and the soils are
summarized in Table 1. The soils domain are discretized by finite and infinite
elements as shown in Fig. 2. A bedrock depth 30 meters was selected that soft clay is
between depths of below ground level and 10 meters and dense sand is between
depths of 10 meters and 30 meters. Assume that soils are dry.
The Bam earthquake on 26 December 2003 with magnitude Mw=6.6 destroyed most
of the city of Bam in Iran and nearby villages, and killed more than 26,000
people(Manafpour, 2008). Large-amplitude ground-motions recorded at the Bam
accelerograph station in the center of Bam city by the Building and Housing Research
ASCE

Geo-China 2016 GSP 259

99

Center (BHRC) of Iran. We used acceleration records of the Bam earthquake at three
BHRC strong motion stations (Bam, Shahdad, and Mohammad-Abad) that they are
distance of 0 km, 37.12 km and 84.48 km from the epicenter, respectively.
Acceleration/time graphs are shown in Fig. 3. Scaled to a peak ground acceleration
similar to the all accelerograms. The 3D model shown in Figure 2 is asymmetric
because the earthquake is applied along one horizontal direction(X direction) only.
Pile-soil contact and raft-soil were applied in both normal and tangential directions
and one normal direction, respectively. whereas in the normal direction hard contact
was utilized and frictional behavior was chosen for the tangential direction. The
coefficient of friction is constant value 0.2 between clay soil and pile surface, and
2
constant value = tan( ) between sandy soil and pile surface. The effect of
3
reducing the coefficient of friction is ignored contact between the pile and the
adjacent soil.
recorded in Bam station

recorded in Shahdad station

recorded in Mohammad-Abad station

FIG. 3. Input accelerogram used in site seismic response analyses.

equivalent linear approach used to consider the effects of soil nonlinearity under
applied earthquakes. In order to perform the equivalent linear analysis, a linear
analysis was carried out with assumed initial values for the equivalent linear material
parameters including damping ratio () and shear modulus ratio (G/Gmax) of the model
(Fatahi et al., 2013). A popular spectral damping scheme that is consistent with experiment
data is Rayleigh damping, giving damping matrix [C] as the combination of mass-proportional
damping and stiffnessproportional damping, as shown in Equation 1. For the multiple degrees
of freedom system, the critical damping ratio at any frequency of mode is given as Equation 2.
The coefficients R and R can be determined from specific i and j for the ith and jth modes,

ASCE

Geo-China 2016 GSP 259

100

respectively. If both modes are assumed to have the same damping ratio , the Rayleigh
coefficients can be calculated using Equation 3.The calculated Rayleigh damping

coefficients can be directly input into ABAQUS to form a damping matrix in the
dynamic equilibrium equation(Al-Qadi et al., 2008).
[C] = R [M] + R [K]

(1)

R Ri
+
2i
2

(2)

i =

R =

2i j

i + j

, R =

2
i + j

(3)

where R and R are Rayleigh coefficients, i is the critical damping ratio at the
frequency of i, i and j are vibration frequencies of the ith and jth modes.
Verification

Pile group analysis is investigated firstly by Poulos for 9 pile case (Poulos, 2001). A
number of different methods have been used for this model by different researchers.
The details of this model are shown on Fig.4. In Fig.5, the analysis results of this
model are given. It can be seen that analysis in the finite element
software Abaqus 6.10 (SIMULIA, 2010) gives relatively small average settlement
compared to other methods.

FIG. 4. Description of the hypothetical example (AfterPoulos, 2001).

FIG. 5. Comparison of pile groups analysis result with the other methods
ASCE

Geo-China 2016 GSP 259

101

Differential settelment [mm]

Results and Discussions


The Effects of Raft Thickness
The piles are modeled with a length of 15 meters and a diameter of 0.5 meters that
placed at a distance of 4 meters from each other. The influence of the raft thickness
(the thickness of raft was varied as 0.3m, 0.5m, 0.7m and 1m) are studied on
maximum settlement, differential settlement and the bending moment.
That is where t = thickness of the foundation, D = diameter of the pile, l = distance
from each other and L = length of pile.

Normalized thickness [t/l]


FIG. 6. Variation of differential settelement with increasing in raft thickness.

Differential settelment [mm]

As shown in Fig.6, The increase in raft thickness reduces the differential settlement
in the raft foundations due to the decreased flexibility of the foundation. Differential
settlement of rafts in Bam station is more than other station because the maximum
observed intensity generally occurs near the epicenter. Accelerations have been scaled
the same in all stations, so there is no significant difference between results.

Normalized thickness [t/l]

FIG. 7. Variation of maximum settlement with increasing in raft thickness.

ASCE

Geo-China 2016 GSP 259

102

Moment (KN.m)

As shown in Fig.7, The increase in raft th ickness reduces the maximum settlement
in the foundations. The maximum settlements are almost constant when The rafts of
thickness are more than 1 m, due to their very large the radius of curvature. the
maximum settlement of a flexible loaded raft was occurred in the middle that
differential settlement between center and edge decrease with increase of raft
thickness.

Section Raft (m)


FIG. 8. Variation of bending moment in section raft for different thickness.

As shown in Fig.8, The maximum positive moment in raft along longitudinal axis
increases with increase of raft thickness. In all cases, the maximum bending moment
has occurred in the middle part of the raft but decreased at the center of raft
foundation (where placed the pile). Modulus of elasticity is higher
soil in middle part than that in edge because freedom of soil at the edges can be greater
than at the center. Soil in middle part tolerates a greater stress than others area and
produces more bending moment in middle part. Different frequency contents in three
stations due to moment in raft along longitudinal axis is slightly different
, especially at the edges.
CONCLUSIONS

In this article, we studies on a piled-raft foundation resting on two layers of sand


and clay soils under dynamic loading in different stations. Effect parameters of raft
thickness, raft thickness and pile diameter were investigated for predicting the
behaviour of piled raft foundations. Results were as follows:
1- Differential settlement reduced with an increased the raft thickness under
lateral and vertical loading that it's meant to increase the bearing capacity.
Since flexibility of the foundation reduce increase of raft thickness.
2- Change in the thickness of raft hasn't a significant effect on maximum
differential settlement of raft system.
3- freedom of soil at the edges of piled raft can be greater than at the center due to
modulus of elasticity is higher in middle part. The maximum bending moment

ASCE

Geo-China 2016 GSP 259

has occurred in the middle part of the raft. The bending moment is increased
with increasing raft thickness.
4- Different frequency contents in three stations due to the results is slightly
different because accelerations have been scaled the same in all stations.
REFERENCES

ABAQUS, ABAQUS. (2010)."6.10 Analysis Users Manual, ABAQUS/Standard


Doc." .
Al-Qadi, I., Wang, H., Yoo, P., & Dessouky, S. (2008). Dynamic analysis and in situ
validation of perpetual pavement response to vehicular loading.Transportation
Research Record: Journal of the Transportation Research Board, (2087), 29-39.
Anderson, J. C., & Bertero, V. V. (1987) ."Uncertainties in establishing design
earthquakes." Journal of Structural Engineering, 113(8), 1709-1724.
Benioff H (1955) ."Mechanism and strain characteristics of the white wolf fault as
indicated by aftershocks sequence." Calif. Div. Mines Bull, 171, 199-202.
Eslami, A., & Malekshah, S. S. (2011) ."Analysis of non-connected piled raft
foundations (NCPRF) with cushion by finite element method." Computational
Methods in Civil Engineering, 2(2).
European Committee for Standardization (2006) Eurocode 8: design of structures for
earthquake resistance, Brussels.
Fatahi, B., & Tabatabaiefar, S. H. R. (2013) ."Fully nonlinear versus equivalent linear
computation method for seismic analysis of midrise buildings on soft
soils". International Journal of Geomechanics, 14(4), 04014016.
Institute A.P.(2006) Recommended Practice for Planning, Designing and Constructing
Fixed Offshore PlatformsWorking Stress Design, API RECOMMENDED
PRACTICE 2A-WSD.
International Building Code (2009) International code council, inc. First printing, IL.
Khoury, M., Alzamora, A., and Ciancia, A. (2011) ."A Piled-Raft Foundation for the
Tallest Building in Brooklyn." Geo-Frontiers 2011: pp. 3818-3827.
Leung, Y. F., Klar, A., & Soga, K. (2009) ."Theoretical study on pile length
optimization of pile groups and piled rafts." Journal of geotechnical and
geoenvironmental engineering, 136(2), 319-330.
Li, X., Jiang, H., & Dan, S. (2012) ."Study on Seismic Safety Performance for
Continuous Girder Bridge based on Near-fault Strong Ground Motions." Procedia
Engineering, 45, 916-922.
Manafpour, A. R. (2008) ."Bam earthquake, Iran: Lessons on the seismic behaviour of
building structures". In 14th World Conference on Earthquake Engineering,
Beijing, China.
Poulos,
H.
G.
(2001)
."Piled
raft
foundations:
design
and
applications."Geotechnique 51.2: 95-113.
Somerville, P. G. (2002)."Characterizing near fault ground motion for the design and
evaluation of bridges." Third National Conference and Workshop on Bridges and
Highways. Portland, Oregon.

ASCE

103

Geo-China 2016 GSP 259

104

Study on the 3D Nonlinear Artificial Boundary of Viscoelastic Media with the Standard Linear Solid
Model
1

Bo Zhang ; Xueying Yang ; Jie Li ; Wentao Dong ; Ruisong Pan ; Weimin Yang ; Lei Yang ; and
8
Dunfu Zhang

Associate Professor, School of Civil Engineering, Shandong Univ., Jinan, PR China. E-mail: zhangbo1977@sdu.edu.cn
Lecturer, Shandong Urban Construction Vocational College, Jinan, Shandong, PR China. E-mail: 274679821@qq.com
3
Lecturer, Shandong Urban Construction Vocational College, Jinan, Shandong, PR China. E-mail: 331403425@qq.com
4
Lecturer, Shandong Urban Construction Vocational College, Jinan, Shandong, PR China. E-mail: 18162270@qq.com
5
Lecturer, Shandong Urban Construction Vocational College, Jinan, Shandong, PR China. E-mail: 369950212@qq.com
6
Associate Professor, Research Center of Geotechnical and Structural Engineering, Shandong Univ., Jinan, PR China. E-mail:
weimin.yang@sdu.edu.cn
7
Lecturer, School of Civil Engineering, Shandong Univ., Jinan, PR China. E-mail: yanglei@sdu.edu.cn
8
Professor, School of Civil Engineering, Shandong Univ., Jinan, PR China. E-mail: zhangdf@sdu.edu.cn
2

Abstract: The three dimensional nonlinear viscoelastic artificial boundary is deduced based on
the principle that the stress on the artificial boundary equals to that of viscoelastic media with
the standard linear solid model. The numerical examples show that this proposed nonlinear
viscoelastic artificial boundary is of high accuracy in simulating the radiation damping of
viscoelastic media with the standard linear solid model.
Keywords: Artificial boundary; Viscoelastic media; Radiation damping; Standard linear solid model.
Introduction
The interior sources can excite body waves and surface waves to radiate to the far field. The wave
motion amplitudes attenuate geometrically, which is called radiation damping. The artificial boundary
should simulate the radiation damping of the far field cut from the computing field(Grange et al.2008;
Zohra and Loret.2004). The exterior source is energy propagating from the far field, such as seismic
waves and blasting waves.
By far many kinds of artificial boundary have been developed. Lysmer and Kulemeyer (1969)
introduced the classical viscous boundary, but the viscous boundary condition only consists of dashpots
without considering the spring recovery ability of the artificial boundary. Clayton and Engquist (1977)
developed an absorbing boundary for acoustic and elastic wave which was called Clayton-Engquist
boundary. Deeks and Randolph (1994) and Liu et al(2005) expressed the viscoelastic boundary based on
the principle that the stress on the artificial boundary equals to that of original media. The nonlinear
viscoelastic artificial boundary used by Liu et al(2005) are all based on the assumption that the far field
media are viscoelastic materials with the Kelvin model. To date the mostly used viscoelastic model in
solving viscoelastic solid problems is the standard linear solid model. So in this paper the three
dimensional nonlinear viscoelastic artificial boundary was deduced from the assumption that the far field
media are viscoelastic materials with the standard linear solid model. The related numerical examples
have demonstrated that the proposed nonlinear viscoelastic artificial boundary is of high accuracy in
dealing with problems of simulating the radiation damping of the far field.
The normal boundary equations of wave motion in viscoelastic media with the standard linear
solid model
Combining the elastic wave theory and viscoelastic theory , the normal stress at viscoelastic media
boundary with the standard linear solid model can be deduced as

ASCE

Geo-China 2016 GSP 259

1
R

= KN (a1 + k1i)

105

1 (a1 +k1i)R it

1
1
e
e CN k1 +( a1 + 2 )ie(a +k i)Reit
2
R
R
R
R
1

= K N u ( R, t ) C N u ( R, t )
Where K N = K

(1)

k12 (a1 1 / R) + a12 (a1 + 1 / R) 4 3 k12 (a1 1/ R) + a12 (a1 +1/ R)


+ Gr +

3 R
(a1 + 1 / R) 2 + k12
(a1 +1/ R) 2 + k12

4
1
1
+ Gi k11+ 2

3
R (a1 +1/ R)2 + k12

(2a)

1
1
1
1
+ Gr k1 1 2
C N = Kk1 1 2
2
2
2
2
R (a1 + 1 / R) + k1 3
R (a1 + 1 / R) + k1
4 3 k2(a 1/ R) + a12(a1 +1/ R)
+ Gi + 1 1
(2b)

3 R
(a1 +1/ R)2 + k12

In Eq. (1) it is clear that the normal stress and displacement in the original field are equivalent to those
of a spring-damper model . The spring stiffness coefficient KN and damper viscous coefficient CN are
expressed in Eqs.(2a) and (2b). This is the normal artificial boundary used in viscoelastic medium with
the standard linear solid model deduced in this paper.

The tangential boundary equations of wave motion in viscoelastic media with the standard linear
solid model
Combining the elastic wave theory and viscoelastic theory , the shear stress at viscoelastic media
boundary with the standard linear solid model can be deduced as
u (R,t) us (R,t)
(R, t) =G* (i)(R,t) =G*(i) s

R
R
2
2

(3)
= Gr (a 2 + ) k k 2 u s ( R, t ) Gi (a 2 + ) + Gk k 2 / u s ( R, t )
R
R

In Eq. (3) it is clear that the shear stress in the original field are equivalent to that of a spring-damper
model. The spring stiffness coefficient KT and damper viscous coefficient CT are expressed in Eqs. (4a)
and (4b). This is the tangential artificial boundary used in viscoelastic medium with the standard linear
solid model deduced in this paper.
2
(4a)
K T = G r (a 2 + ) G i k 2
R
2
(4b)
C T = G r k 2 / + G i (a 2 + )
R
Numerical example
A P wave motion problem is analyzed to verify the ability of the artificial boundary deduced in this
paper by simulating the radiation damping of the far field viscoelastic media with the standard linear
solid model. The computation model is shown in FIG.1. The parameters of the material are shown in
Table 1.
Table 1 Parameters of the materials in numerical example 1
Parameters
G (GPa)
v
(kg/m3)
(MPah)
Material constants
1
0.25
1000
20

ASCE

Geo-China 2016 GSP 259

Where the expression of T(t) and S(x,z) are expressed in two cases as follows.
In case 1, the load is a simple harmonic load.
1 x 1, z 2
T(t ) = sin(3.14 t ) S(x, z) =

else
0
In case 2, the load is a non-simple harmonic load.
t 0 t 1
1 x 1, z 2

T(t) = (2t) 1< t 2, S( x, z) =

0
else

0
else

FIG.1 Computation model for simulating radiation damping

FIG.2 Comparison of the response at observation point A(0,0,0) under simple harmonic
load in case 1

FIG.3 Comparison of the response at observation point A(0,0,0) under simple non-harmonic load
in case 2

ASCE

106

Geo-China 2016 GSP 259

In the FIG.1 the boundary of the computing field is used with fixed boundary, the artificial boundary
derived in this paper and the remote artificial boundary respectively. The remote artificial boundary
means that there is enough distance between the wave origin and the boundary to avoid the wave
reflection. So when it is difficult to get the analytical solution on wave motion problems, the wave
motion results with remote boundary can be taken as the exact solution[9]-[12]. But as the computing cost
with remote boundary is too high to use in solving engineering problems, so the remote boundary mostly
is used as exact solution in verifying some problems.
From the FIG.2 and FIG.3, it can be seen that the fixed boundary has bad precision in simulating the
radiation damping of the far field. The results with the artificial boundary deduced in this paper is almost
the same with the exact solution (remote artificial boundary), which shows that the artificial boundary
derived in this paper can simulate the radiation damping of the far field viscoelastic media with the
standard linear solid model precisely.
CONCLUSIONS

(1) In order to simulate the radiation damping of the far field in viscoelastic media with the standard
linear solid model, in this paper an artificial boundary is deduced combining the elastic wave theory and
the viscoelasticity theory.
(2) The related numerical examples have demonstrated that the proposed artificial boundary is of high
accuracy in dealing with problems of simulating the radiation damping of the far field in viscoelastic
media with the standard linear solid model.
(3) The artificial boundary derived in this paper is easily combined with general FEM software, and
can be used to analyze complex foundation-structure interaction problems.
Acknowledgements

This paper is funded by National Natural Science Foundation of China (NO.51379114, 51279095,
51479107, 51304125, 51509146), Natural Science Foundation of Shandong Province, China
(ZR2012EEQ007) .
REFERENCES

Grange, S., Kotronis, P., and Mazars, J.,2008. A macro-element for a circular foundation to simulate 3D
soilstructure interaction International Journal for Numerical and Analytical Methods in
Geomechanics , Vol. 32(10):12051227.
Zohra, Z. and Loret, B.,2004. A viscous boundary for transient analyses of fluid-saturated porous
media Earthquake Engineering and Structural Dynamics, Vol. 33(2): 89110.
Lysmer, J. and Kulemeyer, R. L.,1969. Finite Dynamic Model for Infinite Media Journal of
Engineering Mechanics Division (ASCE) , Vol. 95(4):859-877.
Clayton, R. and Engquist, B.,1977. Absorbing Boundary Conditions for Acoustic and Elastic Wave
Equations Journal Bulletin of the Seismological Society of America , Vol.76(2) : 1529-1540.
Deeks, A. J. and Randolph, M. F.,1994. Axisymmetric Time- Domain Transmitting Boundaries
Journal of Engineering Mechanics, Vol. 120(1):25 -42.
LIU, J., WANG, Z., DU, X., et al,2005. Three-dimensional viscoelastic artificial boundaries in time
domain for wave motion problems Engineering Mechanics , Vol. 22 (6):46-51.

ASCE

107

Geo-China 2016 GSP 259

Shaking Table Investigation on the Seismic Failure Mode of PHC Pipe-Piles


Considering Liquefaction
Xingwu Wen1; Jiesheng Zheng2; Fuyun Huang3; and Haimin Qian4
1

Professional Engineering, Institute of Building Science of Fujian Province China, Fuzhou, Fujian
Province 350025, Peoples R China. E-mail: 342110814@qq.com
2
Professional Engineering, Institute of Building Science of Fujian Province China, Fuzhou, Fujian
Province 350025, Peoples R China. E-mail: zjs0607@126.com
3
Associate Professor, College of Civil Engineering, Fuzhou Univ., No.2 Xueyuan Rd., Fuzhou, Fujian
Province 350108, Peoples R China (corresponding author). E-mail: huangfuyun@fzu.edu.cn
4
Graduate Student, College of Civil Engineering, Fuzhou Univ., No.2 Xueyuan Rd., Fuzhou, Fujian
Province 350108, Peoples R China. E-mail: 350635311@qq.com

Abstract: Soil liquefaction induced by earthquake will cause very large damage to
structures, especially to the pile foundations. Pre-stressed high-strength concrete (PHC)
pipe-piles were regarded as the relatively economic and effective piles in practical
engineering. However, the failure modes of PHC pipe-piles and the interaction of soilpile with or without liquefiable soils are still unclear. A testing on the seismic failure
mode of PHC pipe-piles in liquefiable soils had been conducted under Shaking Table.
The PHC pipe-piles were employed and tested under a series of sine and EI-Centro
earthquake waves as well as artificial ground motions based on in-site, and the
particular attention was paid to the failure mode of the piles and basic experimental
phenomena during soils liquefaction. The testing results demonstrate that the presence
of prestress can significantly improve anti-seismic shear forces and the failure modes of
PHC pipe-piles in liquefiable soil have a little different.
INTRODUCTION
Prestressed high-strength concrete (PHC) pipe-piles regarded as the relatively
economic type of pile foundation, is widely used in the coastal soft soil area of China,
especially in Fujian Province due to high axial bearing capacity, convenient
construction and favorable quality (Shi 2004). However, PHC pipe-pile, a typical thinwalled concrete member, is disadvantageous for the flexural capacity and weak shear
resistance (Lu 2006). The damage of PHC pipe-piles on construction and earthquake
attracted a lot of attention, which induced the limit use in the code of design of building
foundation of liquable soil in Fujian Province (Huang 2008). Nevertheless, there are
still a number of PHC pipe-piles that had been used in liquable soil area. The response
and earthquake fortification of PHC pipe-piles in liquable soil area should be taken
consideration, and they are necessary to carry out in these researches.

ASCE

108

Geo-China 2016 GSP 259

Some researches on seismic performance of PHC pipe-piles had been conducted in


recent years. Takuya (2003) conducted the static tests on PHC pipe-piles, and the
influence of residual stress on axial bearing capacity was studied. Chung (2007)
conducted the low cyclic reversed loading test and studied the influence of
reinforcement on energy-dissipating capacity of PHC pipe-piles. Rong (2012) carried
out the influence of pile diameters, effective prestressed and non-prestressed rebars on
seismic performance of PHC pipe-piles. Gao (2012) studied the influence of pile type,
stirrup spacing and diameter on the seismic performance and bearing capacity of PHC
pipe-pile. Sun (2012) and Li (2013) tested on some scaled PHC pipe-piles made of
organic glass under shaking table and researched the seismic response of PHC pipe-pile
considering the soil-pile interaction. Hokmabadi and Fatahi (2014) presented a
physically modeling method of seismic soil-pile-structure for buildings, and the
influence of soilpilestructure interaction on seismic response of mid-rise buildings
sitting on a floating pile foundation was also assessed. Tang and Ling (2014) carried
out a shaking table test on RC pile group considering the soil liquefaction, and can be
seen as the happening of liquefaction when the ratio of water pore pressure is over 0.6.
Mohamed (2013) conducted a large shaking table test on pile group and the response to
liquefaction of lateral spreading was observed.
From above review, it can be found that the PHC pipe-pile lacks the widely and
deeply study especially the experiment on the seismic performance in liquable soil
under shaking tables. Some low cyclic reversed loading tests of scaled piles made of
glass or any other material cannot reflect real response and damage as well as failure
modes accurately. It lacks response data of PHC pipe-piles in earthquake, thus more
shaking table tests on PHC pipe-piles need to be conducted.
According to a PHC pipe-pile foundation project of high-rise buildings in Fuzhou
City, Fujian Province, 3 scaled PHC pipe-piles with large inertial mass were tested in
unsaturated and saturated soil under shaking table in this paper. In addition, every pile
was damaged during the test and the failure modes of them were studied. Fuzhou City
is in the regions of 7 fortification intensity with the 0.1g earthquake acceleration for
design, III site classification and 0.55s characteristic periods of response spectra. The
scaled PHC pipe-piles are totally based on PHC-A type with 500mm diameter and C80
concrete. This paper only presents the experimental details and failure mode due to the
limit of page, and other testing results and analysis will be indicated in another paper.
EXPERIMENTAL MODEL
Design of similar ratio
The diameter of prototype of PHC pipe-piles in practical engineering is 500mm. Due
to the limit of bearing capacity of shaking table, the diameter of scaled model of PHC
pipe-pile is selected as 155mm. Therefore the geometry similarity ratio is 0.31
(Sl=0.31). The materials are identical to prototype, thus the similar ratios of elasticity
modulus and density are 1.00 (SE=1.00). All similar ratios are shown in Table 1.

ASCE

109

Geo-China 2016 GSP 259

110

Table 1. Similar Ratios in Shaking Table Test


Physical Quantity
Similarity Relation
Similar Ratio
0.31
Physical Dimension l
Sl
Elasticity Modulus E
1.00
SE
Density
1.00
S
Stress
1.00
S=SE
Time t
St=SlSE-1/2S1/2
0.31
Acceleration a
Sa=SESl-1S-1
3.23
1/2 -1/2 -l
Frequency
S=SE S Sl
3.23
Design of PHC model pile
All of three scaled PHC models are same, 2.75m length long, 52.5mm wall thickness
and 155mm pile diameter. Each pile is made of C80 concrete, five 7-wires 11.1mm
pre-stressed strands and 6mm spiral hoops with spacing of 60mm, just shown in Fig.
1 (change to 30mm at pile ends for strengthening). The models reinforcement drawing
is shown in Figure 1.
According to the area of pre-stressed strand, stretching control force (T) is equals to
243.331kN (T=243.331kN). In order to investigate the influence of prestress on seismic
performance of PHC pipe-pile, the model piles are divided into three groups according
to 0T, 0.25T and 0.5T. The average payload for the piles in prototype is about 750kN,
thus the model pile will be about 30kN based on similarity relationship. Therefore, a
superstructure as the inertial mass with 3 tons is designed and subjected to the pile head.
The detailed parameters of PHC pipe-piles are listed in Table 2.
D
D
7
11
711.1

6@60
6@60

330
330

2090
2040( 3m)/1040

2.5m

330
330

6@30
6@30

6@60
6@60

6@60
6@60

6@30
6@30

52.5
50 52.5
52.5
52.5 50
155
155

FIG. 1. Reinforcement Drawing of Scaled PHC pipe-piles (units:mm)


Soil box and soil
The steel soil box in the experiment is cylinder shape with the diameter of 1.5 meters
and height of 2.5 meters with the wall thickness of 10mm, and thickness of bottom
plate that fixed on the shaking table by bolts is 25mm (soil box is shown in Figure 2).
During the test, the soil box is rigid to the shaking table by high-strength bolts. In
addition, in order to reduce the wave reflection of steel box boundary, inner wall is
equipped with PE foam (5cm thickness).

ASCE

Geo-China 2016 GSP 259

NO.
PHC-1
PHC-2
PHC-3

111

Table 2. Parameters of Scaled PHC pipe-piles


Effective Inertial Mass
Stirrup
Stretching
Dimension Thickness
Prestress at Pile Head
Spacing
Force (T)
(mm)
(mm)
(MPa)
(Tons)
(mm)
0.00
0.0
3
155
52.5
60
0.25
3.6
3
155
52.5
60
0.50
7.2
3
155
52.5
60

Moisture Content
(%)
26.2

Table 3. Parameters of Model Soil


Density
Cohesion
Void Ratio
3
(g/cm )
(kPa)
1.8
0.7
53.4

internal friction
angle ()
14.2

Percentage of Particle Size (%)

To simulate the liquation of soil, first soil layer is selected as clay to get confining
pressure, and the second layer is mud and sand. Thickness of clay is 0.35m while that
of mud and sand is 2.1m. According to cumulative curve of grain composition of mud
and sand, its known that d60=0.57 mm and d10=0.18 mm (shown in Figure 3), which
means middle grain size accounting for larger proportion. In addition, nonuniform
coefficient Cu equals to 3.17 (<5), which means good uniformity of the soil, and more
details are shown in Table 3.
100
90
80
70
60
50
40
30
20
10
0

10

0.1

0.01

(mm)

FIG. 2. Soil Box

FIG. 3. Cumulative Curve of Grain Composition


Particle Size (mm)

SENSORS AND LOADING


Sensors deployment
Total strain gauges for each pile were placed with eight sections. The first layer (clay)
was placed with two sections (respectively on the top and bottom of layer, S8, S9 and
S7, S10) separated with 350mm, second layer (sand) was evenly placed with six
sections at 350mm interval (S1-S6 and S11-S16), and each section strain gauge was
symmetrically placed along the direction of length.
For the study of acceleration response characteristics of the entire superstructure-pilesoil system, 13 acceleration sensors were selected in this experiment. The shaking table
was arranged with one sensor (A1) for comparison with input excitation; the pile was
arranged with five acceleration sensors (A2-A6) along depth-direction; the

ASCE

Geo-China 2016 GSP 259

112

superstructure inertial mass was placed with two sensors (A7, A8) totally along the xdirection; other five acceleration sensors (A9-A13) are arranged along embedded
depth-direction to get acceleration of far-field soil.
This experiment will be judged by the liquefaction of sand through pore water
pressure ratio and liquefaction macroscopic phenomena. Therefore, pressure cells of
pore water and soil were placed. Six earth pressure gauges (T1-T6) were selected and
arranged along the pile. As the number of devices is limited, only three pore water
pressure gauges (K1-K3) and two displacement meters (D1, D2) were selected and
arranged vertically along the soil. The arrangement for all measuring points and sensors
is shown in Figure 4.
A8

Superstructure
Inertial mass

Displacement meter
Strain gage

D2

A7

Earth pressure meter

S5

S12

400

K1

S13
S14
S15

S1

S16

T4

A4

A11

T5
A10

A3
T6
A2
D1

100

T3 A12

A9

450

S2

A5
K2
K3

A13

T2

450

250 350 350

S3

T1

A6

450

350

S4

500

Void-pressure cell

S10
S11

400

350 350

S9

S7
S6

450

S8

450

350

Accelerometer

A1

Shaking table

Shaking table

Vibration Direction

Vibration Direction

FIG. 4. Test Points Arrangement

0.04

0.04

0.02

0.02

0.06

0.06

0
0

10

15

20

-0.02
-0.04

-0.06

25

30

0
0

10

15

-0.02
-0.04

-0.06

(a) Chi-Chi Wave

(b) EL-Centro Wave

FIG.5 Input Waves


Seismic wave selection and loading
Sine wave was used to identify the general dynamic performance. Typical earthquake
waves Chi-chi and El-Centro (shown as Fig. 5) as well as an artificial wave were
selected as the input excitations that were compressed by the time scale factor.
Artificial wave was generated by the seismic response spectrum based on in-site and

ASCE

Geo-China 2016 GSP 259

113

Chinese code (2006). PHC-2 and PHC-3 pile were covered with overlying clay, while
PHC-1 pile without overlying clay.
The loading system for the piles of PHC-1, PHC-2 and PHC-3 were identical to have
26 cases. In order to study the effect of liquefaction and compare the difference, each
case involved in the soil with unsaturated soil first (Condition 1) and then in saturated
soil (condition 2). Case 1-4 were the sine wave with peak acceleration at 0.0414g and
Case 5-8 are the Chi-chi wave with 0.0414g to 0.1656g mainly identified the elastic
behavior and response. The peak acceleration of El Centro wave and artificial wave had
nine levels of 0.0414g, 0.0828g, 0.1242g, 0.1656g, 0.2070g, 0.2484g, 0.2898g,
0.3312g, and 0.3726g from small to large, respectively. In addition, 0.0414g and
0.0828g means about the frequent earthquakes of VII and VIII intensity degree
respectively, while 0.1242g, 0.2484g, and 0.3726g means about the rare earthquakes of
VI, VII, and VIII intensity degree respectively.
At each change of peak acceleration, white noise scanning was adopted to measure
dynamic characteristics of the system and identify the damage levels. The testing cases
are shown in Table 4.
Cases
1-4
5-8
9-17
18-26

Table 4. Test Cases


Peak Acceleration (g)

Seismic Wave
Condition
Sine wave(1Hz, 2Hz,
0.0414
4Hz, 8 Hz)
Unsaturated/
Chi-chi
0.0414, 0.0828, 0.1242, 0.1656
Saturated
El-centro
0.0414, 0.0828, 0.1242, 0.1656, 0.2070,
0.2482, 0.2898, 0.3312, 0.3726
Artificial wave

BASIC EXPERIMENTAL PHENOMENA


When the soil is unsaturated, it can be observed that the responses of acceleration and
displacement at the cap of pile of PHC-1, PHC-2 and PHC-3 pile are not obvious under
case 1, and the pressure of soil are small. With the increasing of the frequencies of
input wave gradually, the response will be increased correspondingly especially at 4Hz,
which are close to their natural frequencies that presented in the following.
After the soil is under saturated, the responses of PHC pipe-piles are more obvious
than the unsaturated one as the saturation sand is softer and its shear strength is smaller
accordingly. Therefore, even with a slight excitation, the model piles will response
evidently.

FIG. 6 Water Gushing

FIG. 7 Pile-soil Separation

The subsidence of soil surface will be happened under the shaking table. It is about
1.5cm averagely for the unsaturated soil (1.6cm for the PHC-1 pile, 1.4cm for PHC-2
ASCE

Geo-China 2016 GSP 259

114

and 1.5 cm for PHC-3), while for the saturated soil, the surface subsidence is over
6.0cm. It demonstrates that saturated sand is easier to get dense and the surface
subsidence is also larger than unsaturated soil. In addition, lots of water will be bubbled
(water spraying and sand emitting), just as shown in Figure 6.
When the soil is unsaturated, there is almost no separation between the pile and its
surround sand after all of 26 cases have been tested. While in the saturated soil, the
sand surface is observed to be covered by a layer of mud and the closer to the pile, the
thicker of the layer of mud is. When the clay layer is removed, it can be found that the
closer to the pile, the greater of the surface subsidence. The soil and the pile at the
interface have separated with the width of 1.5cm, as shown in Figure 7.
Cracks
500
0.50
1060
1.06

Interface between

250
280
0.28 0.25

Soil surface

Soil surface

Crack

1st and 2nd layer

Crack

PHC-1

PHC-2
PHC-1

PHC-2

FIG. 8 Crack Distribution of Model Piles (unit: mm)


FAILURE MODES
Scraping the sand away after the experiment and observing the specimen, it can be
found that many cracks appearing on the pile of PHC-1 and PHC-2 (PHC-3 is
destroyed by the crane by accident after testing), as is shown in Figure 8. From 8, it can
be seen that there is only one large crack that is over 2.0 mm width in the liquidation
area for PHC-1 pile, while several and small cracks less than 1.0mm for PHC-2.
Comparing the crack distribution of two models pile, cracks on PHC-1 is more obvious
than PHC-2 and damage is also more serious. It demonstrates that the presence of
prestress can significantly improve anti-seismic shear forces. The failure modes of PHC
pipe-piles in liquefiable soil have a little different.
The position of crack of PHC-1 is depth down about 1.0m from soil surface, which is
about 6.5 times of diameter. Once the crack is happened, the strain of piles below the
crack will decrease suddenly and plastic hinge appears in the crack area, which means
that the resistance of shear force decrease dramatically and almost cannot transfer
moment with depth. That also means only the upper length of pile and soil can work
effectively and interactively. However, as for PHC-2, the position of the first crack is
down to 0.5m (the junction of different layers), and due to the presence of prestressed
strands, it can redistribute the inertial force and moment to the surround soil along the

ASCE

Geo-China 2016 GSP 259

pile depth if one section cracks. Therefore, after increasing the peak ground
acceleration, more deep cracks (second and third) will be appeared.
Generally, the position of crack and the maximum position of bending moment adapt
accordingly, but for foundation in liquable soil, cracks will be occurred in the junction
of soil (interface between 1st and 2nd layer) and liquefaction region. The reason is
mainly due to the pile is in the sand liquefaction area, where the rigid of ground soil is
changed dramatically, which induce the shear force resistance reduce greatly and thus
crack forms.
CONCLUSIONS
1) Experiment shows that the responses of PHC pipe-piles of saturated soil are more
obvious than that of unsaturated soil especially the frequency of excitation is close to
the structure. The saturated sand is easier to subsidence than unsaturated soil, and
finally the phenomena of water spraying and sand emitting are happened.
2) Experiment also shows that the presence of prestress can significantly improve antiseismic shear forces. The failure modes of PHC pipe-piles in liquefiable soil have a
little different. For foundation in liquable soil, cracks occur mainly in the junction of
soil and liquefaction region due to the rigid of ground soil is changed.
ACKNOWLEDGMENTS
The authors appreciate the support of Natural Science Foundation of China (No:
51578161, 51278126, 2010Y0101).
REFERENCES
Chung, SG. and Kim, SR. (2007). "Appraisal of true resistance of phc pipe-piles
driven in thick soft deposit." 2007 International Forum on Strategic Technology :
109-112.
Gao A. (2012). Experimented research on seismic performance of PHC pipe pile and
analysis of bearing capacity. Tianjin University, Dissertation.
Hokmabadi, A. S. Fatahi, B. and Samali, B. (2014). "Physical modeling of seismic soilpile-structure interaction for buildings on soft soils." International Journal of
Geomechanics, Vol. 15(2).
Hokmabadi, A. S. Fatahi, B. and Samali, B. (2014). "Assessment of soilpilestructure
interaction influencing seismic response of mid-rise buildings sitting on floating pile
foundations. " Computers and Geotechnics, Vol. 55: 172-186.
Huang, Liangji. Lin, Yixi. Cai, Jian. and Zhou, Wanqing. (2008). "Dynamic and static
comparative analyses of settlements of over length PHC pipe piles." Rock and Soil
Mechanics, Vol. 29 (2): 507-511,516.
Li, Y.C. Xing, K.Y. Liu, H. et al. (2013). "Experimental study of seismic performance
of PHC pipe pile considering soil-pile-structure interaction." Chinese Journal of
Rock Mechanics and Engineering, Vol. 32 (2): 401-410.
Lu, WT. Wang YH. and Leng, WM. (2006). "Testing and numerical analysis of load
transfer mechanism of PHC pipe-pile." Rock and Soil Mechanics, Vol. 27 (3): 466470.
Mohamed, R. Towhata, I. Honda, T. Tabata, K. and Abe, A. (2013). "Pile group
response to liquefaction-induced lateral spreading: e-defense large shake table test. "
Soil Dynamics and Earthquake Engineering, 51, 35-46.
Nagae T., Hayashi S. (2003). Earthquake-resistant property of prefabricated highstrength concrete pile. ASCE, High Performance Materials in Bridges, 2003:173182.

ASCE

115

Geo-China 2016 GSP 259

Rong, X. Xu, XZ. Li, YY. (2012). "Experimental research on aseismic behavior of
PHC pipe piles." Industrial Construction, Vol. 43 (7): 72-75.
Shi, F. (2004). "Experimental research on load transfer mechanism of pretensioned
high strength spun concrete piles." Chinese Journal of Geotechnical Engineering,
Vol. 26 (1): 95-99.
Sun, Ye. and Liu, Zuhua. (2012). "Strain distributing disciplinarian of phc pipe-piles
under seismic action." Low Temperature Architecture Technology, Vol. 165: 24-27.
Tang, L. and Ling, XZ. (2014). "Response of a RC pile group in liquefiable soil: a
shake-table investigation." Soil Dynamics and Earthquake Engineering, Vol. 67:
301-315.

ASCE

116

Geo-China 2016 GSP 259

Influence of Shallow Foundation Characteristics on the Seismic Response of Mid-Rise


Buildings Subjected to Strong Earthquakes
Quoc Van Nguyen1; Behzad Fatahi2; and Aslan S. Hokmabadi3
1

Ph.D. Candidate, School of Civil and Environmental Engineering, Univ. of Technology Sydney (UTS),
Sydney, Australia. E-mail: Quoc.V.Nguyen-1@student.uts.edu.au
2
Senior Lecturer, Geotechnical Engineering (Ph.D., CP.Eng.), School of Civil and Environmental
Engineering, Univ. of Technology Sydney (UTS), Sydney, Australia. E-mail: Behzad.fatahi@uts.edu.au
3
Geotechnical Engineer (Ph.D., CP.Eng.), Ove Arup & Partners, Hong Kong. E-mail:
aslan.s-hokmabadi@arup.com

Abstract: Performance based seismic design is a modern approach to


earthquake-resistant design shifting emphasis from strength to performance. In this
study, the influence of the shallow foundation (footing) size on the seismic performance
of the buildings subjected to strong earthquakes is investigated considering
Soil-Structure Interaction (SSI). A fifteen storey moment resisting frame sitting on
shallow foundation over soft soil with different foundation size is simulated
numerically using ABAQUS software. The developed three dimensional numerical
simulation accounts for nonlinear behaviour of the soil medium by considering the
variation of soil stiffness and damping as a function of developed shear strain in the soil
elements during earthquake. Elastic-perfectly plastic model is adopted to simulate
foundations and structural elements. Four strong earthquake records, including El
Centro 1940, Hachinohe 1968, Northridge 1994, and Kobe 1995 have been taken as
input accelerations for time history analysis in time domain. Due to natural period
lengthening, there was a significant reduction in the base shears when the size of the
foundation was reduced. It can be concluded that the foundation size can influence the
dynamic characteristics and seismic response of the building due to SSI and should
therefore be given careful consideration in order to ensure a safe and cost effective
seismic design.
INTRODUCTION
The influence of the underlying soil on the seismic response of a structure can be
disregarded when the ground is stiff enough, and consequently, the structure can be
analysed considering the fixed-base conditions. However, the same structure will
behave differently when it is constructed on a soft soil deposit. Earthquake
characteristics, the travel path, the local soil properties, and the soilstructure
interaction are the factors affecting the seismic excitation experienced by structures.

ASCE

117

Geo-China 2016 GSP 259

The results of the first three factors can be summarised as free-field ground motion.
However, the foundation of a structure does not follow the deformation of the free-field
motion due to its stiffness, and the dynamic response of the structure itself induces the
deformation of the supporting soil (Kramer, 1996). Two key mechanisms are generally
involved during a seismic soil-foundation-structure interaction: kinematic interaction
and inertial interaction. Kinematic interaction occurs because stiff foundation elements
in the soil cause the foundation motion to deviate from the free field ground motion.
Kinematic interaction could also be due to ground motion incoherence, foundation
embedment effects, and wave scattering or inclination (Stewart et al., 1999). Inertial
interaction results from the inertia developed in the structure as its own vibration
produces base shear, moment, and torsional excitation. These loads cause displacements
and the foundation to rotate relative to the free field condition (Kramer and Stewart,
2004). Fundamentally, the size of a foundation can influence the kinematic and inertial
interactions mainly by altering the mass and stiffness of the soil foundation system
which in turn influences the seismic response of the superstructure.
Several researchers (e.g. Sbartai, 2015; Sameti and Ghannad, 2014; Chen, 2015;
Hokmabadi et al., 2014) studied the seismic soil-foundation-structure interaction (SFSI)
phenomena and its influence on the seismic response of buildings by adopting the
Winkler (substructure) methods and the numerical methods. Adopting advanced
numerical models has a number of advantages over the Winkler methods, especially
their ability to conduct time history analyses while considering effects such as the
nonlinear stressstrain behaviour of the soil and the superstructure, material and
radiation damping, advance boundary conditions, and interface elements. Another
advantage of using numerical methods is their ability to perform the analysis in a
fully-coupled manner without resorting to independent calculations of site or
superstructure response (Meymand, 1998). Consequently, numerical modelling
predictions can capture the different parameters involved in soil-foundation-structure
interaction (SFSI) that are closer to reality.
The aim of this study is to numerically investigate the influence of shallow foundation
size on the seismic response of a regular mid-rise moment resisting building frame
during earthquake excitations using ABAQUS software (version 6.12) as a fully
coupled nonlinear time history analysis.
NUMERICAL MODEL
Case study description
In this study, a fifteen storey concrete moment resisting building frame, 45 m high and
12 m wide with 16 columns consisting of three spans in each direction, and 15 slabs and
a foundation, is selected (FIG. 1). This building frame represents conventional mid-rise
moment resisting buildings. The structural sections were specified after conducting a
routine design procedure regulated in the relevant building codes (AS3600, 2009,
AS1170.4, 2007). SAP2000 V 14 (CSI, 2010) software was utilised for the structural
analysis and design of the cross sections of beams and columns. Then, a nonlinear
time-history dynamic analysis under the influence of the four earthquake ground
motions shown in FIG. 1 was carried out. In this dynamic analysis the geometric
nonlinearity and P-Delta effects were considered according to AS3600 (2009). The
fundamental frequency of the adopted building was 0.830 Hz and its total mass was
1683 tonnes.
ASCE

118

Geo-China 2016 GSP 259

Th
he adopted superstructurre sits on 30m deep softt soil that is categorised as Class Ee
acco
ording to thee Australian standard (AS
S1170.4, 20007). The subb-soil is a sofft clayey soill
with
h a density of
o 1470 kg/m
m3, a shear wave velocityy of 150 m/s,, and an-undrrained shearr
strength of 50 kPa.
k The prop
perties of thiis subsoil weere extractedd from actuaal in-situ andd
oratory testss (Rahvar, 2006),
2
so th
hese parameeters have m
merit over thhe assumedd
labo
paraameters whicch may not be
b completelly conforminng to reality.. It was assum
med that thee
watter table was below the leevel of the bedrock.
b
Th
he shallow square foun
ndations (foo
otings) weree designed tto support thhe structuree
agaiinst static an
nd dynamic loads to saatisfy the reqquirement foor bearing ccapacity andd
max
ximum settleement. All th
he shallow fo
oundations w
were 1 m thiick and weree made from
m
rein
nforced conccrete. The sh
hallow found
dations had ddifferent sizze widths to facilitate ann
inveestigation in
nto how shaallow found
dation sizes influence thhe seismic response off
buillding due to
o the soil-fou
undation-strructure interaaction. Thesse foundatioons had fivee
diffferent sizes, including:
i
1.1B, 1.3B, 1.5B, 1.7B annd 2.0B, wheere B is the w
width of thee
buillding (=12 m). All theese foundatio
on sizes weere acceptabble from ann engineerss
persspective and satisfied the requireements for bearing caapacity andd maximum
m
settllement, altho
ough the saffety factor off the smallerr foundationss was less thhan the largee
ones. Moreoverr, although th
he 1.7B and 2.0B foundaations are noot common inn practice, a
wid
der range of foundation sizes
s
was co
onsidered in this study too better undeerstand how
w
foun
ndation size affects the seismic resp
ponse of a bbuilding durring strong eearthquakes.
Thee seismic response of th
hese foundaation sizes aare compareed and discuussed in thee
following sectio
ons via a 3D finite elemeent numericaal simulationn.

FIG. 1. Pro
oblem defin
nition and modelling
m
eleement detaiils of the devveloped
so
oil-foundatiion-structurre system
Num
merical Modelling Deta
ails
ABAQUS
A
v 6.14 finite element anaalysis softwaare was useed in this sttudy for thee
num
merical simu
ulation of thee soil-foundation-structuure systems.. This software packagee
can simulate co
omplex prob
blems that require
r
largee computational memorries using a
o analysis. Beam
B
and sh
hell elementss were used tto simulate tthe columnss
direect method of
and
d floor slabs of the supersstructure in this
t numericcal model. T
The characterristics of thee

ASCE

119

Geo-China 2016 GSP 259

120

columns are presented in Table 1. The structural elements were modelled using an
elastic-viscoelastic constitutive model while considering the Rayleigh damping
according to Ryan and Polanco (2008) (Equation (1)).
[C] = [M] + [K]

(1)

where [C], [M], and [K] are the damping, mass, and stiffness matrices, respectively,
and are the model coefficients used to specify the model damping ratio in two modes.
By assuming the same damping ratio () for two modes with frequencies fi and fj, the
model coefficients and can be obtained from equations for Rayleigh damping in
Chopra (2007). In this study, a structural damping ratio () of 5% together with model
coefficients of = 0.3996 and = 0.0049, calculated based on the first and second mode
frequencies of the structure (see Table 2), was used to simulate structural damping in the
dynamic analysis.
Table 1. Adopted characteristics of designed reinforced concrete column sections
Section Type
Type I (Levels 1 3)
Type II (Levels 4 7)
Type III (Levels 8 11)
Type IV (Levels 12 15)

Ix (m4)
5.33E-3
3.64E-3
2.40E-3
1.50E-3

Iy (m4) Area (m2)


10.87E-3
0.302
7.45E-3
0.250
4.89E-3
0.203
3.05E-3
0.160

E (kPa)
2.86E7
2.86E7
2.86E7
2.86E7

0.2
0.2
0.2
0.2

Table 2. Natural frequencies of the adopt 15 storey fixed base structure


Motion mode Mode 1 (f1) Mode 2 (f2) Mode 3 (f3) Mode 4 (f4)
Frequency (Hz)
0.830
2.341
4.018
5.781
The nonlinearity of soil during an earthquake plays an important role in the dynamic
response of soil-structure systems. In this study, an equivalent linear method has been
adopted, as described by Seed and Idriss (1969). In this method, a try and error process
utilising soil nonlinear backbone curves to find the strain compatible values of
damping and modulus is used to capture the soil non-linearity during shaking
excitations. The adopted equivalent soil stiffness value for each earthquake record was
different depending on the maximum shear strain generated in the soil deposit, while
Rayleigh damping was adopted to capture variations of soil damping during each
earthquake. Table 3 presents the adopted soil properties. Table 3 presents the adopted
soil properties.
Table 3. Adopted soil parameters in numerical models
Soil Properties
Denote
Unit
Value
3
Mass density
1470

kg/m
Shear Wave Velocity
150
Vs
m/s
Poissons ratio
0.4

Plasticity Index
15%
PI
For the soil-foundation-structure interaction analysis in this study, surface-based
contacts were defined such that the master surface is the top surface of the soil and the
ASCE

Geo-China 2016 GSP 259

121

slav
ve surface is
i the botto
om surface of the fouundation. M
Moreover, finnite slidingg
form
mulation and the surfacce-to-surfacee discretisattion methodd were utiliised for thee
con
ntacts. The mechanical
m
prroperties of the contact surfaces deffining the tanngential andd
norm
mal behavio
our of the co
ontact surfaaces can inflluence the rresults of the numericall
mod
delling and should be chosen
c
with
h great rigorr. Normal bbehaviour addopts hard
con
ntact in a presssure-over cllosure relatio
onship. A FO
ORTRAN suubroutine, FR
RIC_COEF,,
is em
mbedded intto ABAQUS
S to simulatee the tangent
ntial behaviouur of the conntacts basedd
on Mohr-Coulo
M
omb failure model.
m
Fo
our strong eaarthquake input motions, including thhe 1994 Norrthridge, the 1995 Kobe,,
the 1940 El Ceentro, and th
he 1968 Hachinohe earthhquakes (refferring to FIIG. 2), weree
imp
posed onto the
t finite ellement num
merical modeel while connducting a ttime-historyy
anallysis. Due to
o the large sizze of the mo
odel (around 70 Giga-byttes for a singgle case), thee
fastt computatio
on facilities were used to
t conduct thhis time-hisstory analysiis, and evenn
then
n it took arou
und 50 hourss to run a sing
gle case undeer the applieed earthquake excitation.
Thee results of th
he 3D finite element
e
num
merical simullation are preesented and discussed inn
the following seection.
(a))

(b)

(c)

(d)

FIG.
F
2. Adop
pted earthquake record
ds: : (a) 19994 Northridgge; (b) 19955 Kobe;
(c) 1940 El Ceentro; and (d
d) 1968 Hacchinohe earrthquake
RESULTS AN
ND DISCUSS
SION
Th
he results of
o the 3D nu
umerical mo
odel developped for the fifteen-storrey buildingg
supp
ported by shallow
s
fou
undations off different ssizes and th
the fixed-baase buildingg
subjjected to thee 1994 Nortthridge, 1995 Kobe, 19440 El Centroo, and 19688 Hachinohee
eartthquakes aree summarised and compared in FIG
G. 3 and FIG
G.4. Referringg to FIG. 3,
SFS
SI amplified
d the maximu
um lateral deflection
d
off the supersttructure durring shakingg
exciitations. Forr instance, th
he maximum
m lateral defo
formation off the fixed-baase buildingg

ASCE

Geo-China 2016 GSP 259

122

(exccluding SFS
SI) under thee 1994 North
hridge earthqquake was 3395mm, whiile the samee
buillding experieenced a lateral deformattion of up too 590 mm (449% more) w
when it wass
supp
ported by a 1.1B shallow
w foundation
n that accounnts for SFSI.. Moreover, as a generall
tren
nd, by increaasing the sizee of the shalllow foundattion from 1.11B to 2.0B tthe structuree
experiences lesss lateral deeformation. For instancce, an increease in the size of thee
foun
ndation from
m 1.1B to 1.5
5B resulted in
n up to 25% less lateral ddeformationn under 19400
El Centro earth
hquake (FIG
G. 13a). Th
his is a connsiderable rreduction inn the laterall
defo
ormation of a structure subjected
s
to strong earthqquakes.
(a)

(c))

(b)

(d)

FIIG. 3. Maxim
mum lateral deflection of the fifteeen-storey strructure sup
pported by
sha
allow founda
ations with varies sizes under the iinfluence off: (a) 1994 N
Northridge;
(b) 1995
5 Kobe; (c) 1940 El Cen
ntro; (d) 19668 Hachinoohe earthquaakes
In
n order to in
nvestigate the influence foundation ssize on the energy absoorbed by thee
stru
ucture during
g earthquakes, the resultss of the deveeloped 3D nuumerical moodel in termss
of shear
s
forces were compaared for diffferent cases.. To determiine the maxximum shearr
forcce at each level, the sh
hear forces generated
g
inn every coluumn at thatt level weree
sum
mmed up in every
e
time in
ncrement du
uring the tim
me-history annalysis, and tthe absolutee
max
ximum sheaar force expeerienced at that
t
level duuring the eaarthquake is reported ass
pressented in FIG. 4. In gen
neral, consid
dering SFSI contributedd to the reduuction in thee
ASCE

Geo-China 2016 GSP 259

123

sheaar forces in the


t structuree, whereas laarger shallow
w foundationns attracted m
more inertiall
forcces from the earthquake excitations than
t
the smaaller sized foundations. F
For instance,
the maximum base
b
shear of the structu
ure supportedd by the 1.55B foundatioon under thee
194
40 El Centro earthquake was
w 4.1 MN
N, while the ccorrespondinng value for tthe structuree
supp
ported by 1.1B foundation was 3.6 MN
M (13% le ss energy abbsorption).
(a)

(b)

(c)

(d)

FIG. 4. Maximum sheear force diistribution oof the fifteen


n-storey strructure
sup
pported by shallow foun
ndations witth varies sizzes under th
he influence of: (a) 19944
No
orthridge; (b
b) 1995 Kob
be; (c) 1940 El Centro; (d) 1968 Haachinohe eaarthquakes
Decreasing
D
th
he size of a foundation
n caused thee spectral aacceleration to decreasee
con
nsiderably as the natural period
p
length
hened. As a rresult, such aan increase inn the naturall
periiod substantiially changeed the respon
nse spectral accelerationn (Sa). In thee case wheree
the mid-rise mo
oment resistiing building frames withh a shallow foundation rrests on softt
t natural period
p
lay in
i the long period regiion of the accelerationn
soil deposits, the
resp
ponse spectru
um curve. Due
D to the naatural period lengtheningg induced byy a reductionn
in th
he size of a foundation,
f
the
t spectral acceleration
a
n (Sa) tended to decrease,, which thenn
redu
uced the base shear of th
he structure.
CO
ONCLUSION
NS
Th
his study in
nvestigated the influencce of shalloow foundatioon size on the seismicc
resp
ponse of a regular
r
mid-rise momen
nt resisting building frrame during earthquakee
ASCE

Geo-China 2016 GSP 259

excitations. ABAQUS was used to numerically simulate the soil-foundation-structure


system by conducting a fully coupled nonlinear time history analysis.
According to the results obtained, the size of a shallow foundation can influence the
structural design of the building under seismic loads considering the seismic
soil-foundation-structure interaction. Larger shallow foundations can moderate the
amplifications of lateral deflection and in turn inter-storey drifts of the structure caused
by SFSI. This can be a cost effective alternative to control the performance level of
buildings. Moreover, changes in the size of shallow foundations resulted in absorbing
an amount of energy from the imposed earthquake that corresponded to the natural
frequency of a particular system. It was observed that buildings with larger shallow
foundations attracted more inertial forces from earthquake excitations than smaller
foundations.
REFERENCES
AS1170.4 2007. Structural design actionsEarthquake actions in Australia. NSW,
Australia: Standards Australia.
AS3600 2009. Concrete Structures. NSW, Australia: Standards Australia.
Chen, L. 2015. Dynamic Interaction Between Rigid Surface Foundations on
Multi-Layered Half Space. Int. J. of Structural Stability and Dynamics, 1550004.
CSI 2010. SAP2000 v14 Analysis Reference Manual. California: CSI (Computers and
Structures Inc.), Berkley.
Kramer, S. L. 1996. Geotechnical earthquake engineering, Prentice Hall.
Kramer, S. L. & Stewart, J. P. 2004. Geotechnical Aspects of Seismic Hazards. In:
Bozorgnia, Y. & Bertero, V. V. (eds.) Earthquake Engineering: From Engineering
Seismology to Performance-Based Engineering. LLC: CRC Press.
Hokmabadi, A., Fatahi, B., and Samali, B. (2014). "Physical Modeling of Seismic
Soil-Pile-Structure Interaction for Buildings on Soft Soils." Int. J. Geomech.,
10.1061/(ASCE)GM.1943-5622.0000396, 04014046.
Hokmabadi, A.S., Fatahi, B. & Samali, B. 2014, 'Assessment of soil-pile-structure
interaction influencing seismic response of mid-rise buildings sitting on floating pile
foundations', Computers and Geotechnics, vol. 55, pp. 172-186.
Meymand, P. J. 1998. Shaking table scale model tests of nonlinear
soil-pile-superstructure in soft clay. PhD thesis in Civil Engineering, University of
California, Berkley.
Rahvar 2006. Geotechnical investogation and foundation design report of Mahshahr
train station. Mahshar, Iran: P. O. Rahvar Pty Ltd, Iran Railway Authority.
Ryan, K. L. & Polanco, J. 2008. Problems with Rayleigh Damping in Base-Isolated
Buildings. Journal Of Structural Engineering, 134, 1780-1784.
Sameti, A. R. & Ghannad, M. A. 2014. Equivalent Linear Model for Existing
Soil-Structure Systems. Inter. Jour. of Structural Stability and Dynamics, 1450099.
Sbartai, B. 2015. Dynamic Interaction of Two Adjacent Foundations Embedded in a
Viscoelastic Soil. Inter. Journal of Structural Stability and Dynamics, 1450110.
Seed, H. B. & Idriss, I. 1969. Influence of Soil Conditions on Ground Motion during
Earthquakes. Journal of Soil Mechanics & Found., Division, ASCE, 95, 99-137.
Stewart, J., Fenves, G. & Seed, R. 1999. Seismic soil-structure interaction in buildings.
I: Analytical aspects. J. Geotech. & Geoenv. Engrg., 125, 26-37.

ASCE

124

Geo-China 2016 GSP 259

Characterization of Freezing Fresh Concrete by Multiple Non-Destructive


Methods
1

Yan Liu1; Junliang Tao2; Xinbao Yu3; Zhen Liu4; and Xiong (Bill) Yu5

Assistant Professor, Dept. of Engineering, Univ. of Mount Union,


1972 Clark Ave., EBB 123, Alliance, OH 44601 (corresponding author). E-mail:
liuyan@mountunion.edu
2
Assistant Professor, Dept. of Civil Engineering, The Univ. of Akron,
244 Sumner St., Akron, OH 44321-3905. E-mail: jtao2@uakron.edu
3
Assistant Professor, Dept. of Civil Engineering, Univ. of Texas, Arlington
Box 19308, 248C Nedderman Hall, Arlington, TX 76019. E-mail: xinbao@uta.edu
4
Assistant Professor, Dept. of Civil and Environmental Engineering, Michigan Technological Univ.
Dillman 201F, 1400 Townsend Dr., Houghton, MI 49931. E-mail: zhenl@mtu.edu
5
Professor, Dept. of Civil Engineering, Case Western Reserve Univ.,
2104 Adelbert Rd., Bingham Building-Room 206, Cleveland, OH 44106 -7201. E-mail:
xiong.yu@case.edu

Abstract: Seismic methods are useful tools to non-destructively assess the behaviors
of fresh concrete. They have also been applied to characterize the properties of
curing concrete to provide information for construction decision. This paper shows
that freezing of concrete significantly affects the engineering properties of concrete.
In the experimental program, ultrasonic tests were conducted on curing concrete
subjected to different freezing process. The results indicate while there exists linear
correlation between low strain seismic wave velocity and concrete strength under
normal curing conditions, such relationships do not hold if the concrete is subjected to
freezing process. A correction accounting for the effects of ice on the bulk strength
needs to be applied. This correction was found to have linear relationship with water
content. Procedures to correct the effects of freezing are proposed, which include the
use of Time Domain Reflectometry to measure the water content. Finally the strength
of concrete in frozen status can be estimated. This information could be incorporated
to determine the magnitude of Winter Load Increase in cold regions for government
agencies.
INTRODUCTION
The evaluation of mechanical properties of concrete by nondestructive techniques
is gaining popularity. Several techniques are currently in use, such as impact echo,
ultrasonic test, spectra analyses of surface wave. They are based on the information
contained in the propagation of ultrasonic waves. Different wave modes and
transceiving methods are explored. For example, Boutin1 and Arnaud used the speed
of longitudinal waves (L-waves, also known as compression waves) of low
frequencies from measuring the time of transition between fluid and solid state of
cellular cement paste. A new device for monitoring the hydration of cement mortar
that measures the transit time and the energy of an L-wave pulse propagating through
a mortar sample has been introduced by Reinhardt2 et al. With this device the setting
and hardening process of mortar can be evaluated. Other investigators have applied
both, longitudinal and transverse waves (T-waves, also known as shear waves) to
examine the hydration of cementitious materials. Sayers3 and Grenfell found a linear
relationship between the effective bulk and shear moduli determined by pulse
ASCE

125

Geo-China 2016 GSP 259

126

velocities. D'Angelo4 et al. detected a considerably higher sensitivity of T-waves to


the hydration process compared to L-waves. Boumiz5 et al. studied the development
of elastic modulus, shear modulus and Poisson's ratio as functions of time and degree
of hydration.
Most current studies are on concrete under normal service conditions. For cold
regions, freezing can have significant effects on the technologies based on seismic
waves, especially for concrete in the early stage of curing. Ohio Department of
Transportation, for example, specifies a 5-day thermal curing during winter
construction. The concrete can quickly become frozen upon removing the thermal
curing. This study aims to determine a NDT test framework to quantify the effects of
freezing on seismic wave propagation in curing concrete.
BACKGROUND
Ultrasonic Testing
Ultrasonic Testing uses high frequency sound wave propagation to measure material
and structural responses. This technology has been used for flaw detection,
dimensional measurements, material characterization and etc.
A typical UT inspection system (shown in Figure 1) consists of several functional
units, such as the pulser and receiver, transducer, and display devices. A pulser and
receiver is an electronic device that can produce high voltage electrical pulses. Driven
by the pulser, the transducer generates high frequency ultrasonic energy. The sound
energy is introduced via couplant to propagate through the materials in the form of
seismic waves. From the signal, information about the location, size, orientation of
structural features can be obtained.

FIG. 1. Ultrasonic testing inspection system


In recent years, techniques based on ultrasonic wave propagation are used to
assess the strength of early age concrete. Several methods that relate the ability of
cementitious materials to transmit ultrasonic waves and their compressive strength
gain can be found in the literature. Keating6 et al investigated the relationship
between ultrasonic longitudinal pulse velocity and cube strength for cement slurries
in the first 24 hours. For concrete cured at room temperature, it was noted that the
relative change in the pulse velocity in the first few hours is higher than the observed
rate of strength gain. However, a general correlation between these two parameters
could be deduced. Another study about the interdependence between the velocity of

ASCE

Geo-China 2016 GSP 259

127

L-waves and compressive strength has been presented by Pessiki7 and Carin. Within
the scope of this work concrete mixtures with different water-cement ratios and
aggregate contents cured at three different temperatures were examined. The L-wave
velocity was determined by using the impact-echo method in a time range of up to 28
days. At early ages the L-wave velocity increases at a faster rate when compared with
the compressive strength and at later ages the strength is the faster developing
quantity. L-wave velocity is found to be a sensitive indicator of the changes in the
compressive strength up to 3 days after mixing.
From the measured ultrasonic velocity, two different types of moduli, i.e,
longitudinal and shear modulus can be calculated as follows:
L = v L2
(1)

G = vT2
(2)
where L is the longitudinal modulus, G the shear modulus and vL and vT the velocity
of L- and T-waves respectively. The moduli L and G are related to the direction of
particle motion caused by L- and T-waves. The longitudinal modulus relates strain to
longitudinally applied stress. The shear modulus describes the elastic behavior of a
material subjected to shear strain.
Time Domain Reflectometry
Time Domain Reflectometry (TDR) is a guided radar technology. It utilizes the
propagation of electromagnetic wave to measure materials properties and structural
responses. The configuration of a typical TDR system is shown in Figure 2(a), which
includes a TDR device (pulse generator and sampler), a connection cable, and a
measurement probe. The measurement probe generally consists of multiple
conductors embedded in materials whose properties are to be measured. Figure 2(b)
shows a schematic plot of typical TDR signal. Information commonly obtained on
materials electrical properties include the apparent dielectric constant Ka, which are
related to the speed of electromagnetic wave in the material; and the electrical
conductivity ECb, which is related to the energy attenuation. Both quantities can be
easily obtained from a TDR signal. A brief introduction on determining these
quantities are summarized below: the travel velocity, v, of an electromagnetic wave
through media is calculated as follows:
c
v=
(3)
Ka
where c is the velocity of an electromagnetic wave in free space (2.988108m/s) and
Ka is the dielectric constant (for TDR measurement in soils, this quantity is generally
called apparent dielectric constant). The time for the electromagnetic wave traveling
down and back along a metallic waveguide of length, L, is given by:
2L
t=
(4)
v
Substituting equation (4) to (3) yields
ct
Ka =
2L

ASCE

(5)

Geo-China 2016 GSP 259

By defining

128

ct
as apparent length la, the apparent dielectric constant can be
2

calculated as
2

l
(6)
Ka = a
L
In the TDR signal, the apparent length is determined from analyzing the time
elapse between reflections. The determination of the apparent length la is illustrated in
Figure 2(b). The electrical conductivity of bulk soil sample can be simultaneously
obtained from the final signal level (Yu and Drnevich 2004).
Relationship between Dielectric Constant and Gravimetric Water Content
Water has dielectric constant of around 81, which is much larger than that of soil
solids (typically around 3 to 7) or air (around 1). A variety of relationships has been
developed between TDR measured dielectric constant and volumetric water content.
Siddiqui and Drnevich (1995) 12 developed an equation that relate the TDR measured
dielectric constant to the gravimetric water content. The equation accounts for the
effects of soil type and density by incorporating two calibration constants. This
equation is shown below.

1
w = w K a a
b d

(7)

1.25

L
Ka = a
L
p

Relative Voltage (V)

0.75

EC b =

Apparent Length, L a

1
C

Vs

1
V

0.25

-0.25

Vs/2

-0.75

-1.25
0

Scaled Distance (m)

FIG. 2. a) Schematic plot of TDR measurement system;


b) Typical TDR signal and information utilized

ASCE

Geo-China 2016 GSP 259

EXPERIMENT DESIGN
Experiments were designed to establish the effects of freezing on the strength of
concrete. Three representative types of concrete used in Ohio DOT projects were
used in the experiment. Concrete were obtained from commercial suppliers THE
COLLINWOOD-HORNING CONCRETE COMPANY (http://www.collinwoodhorning.com). The concrete mixes are designated as ODOT Class C (Ordinary
pavement concrete of 4000 psi), High Strength (8000 psi) and Self-consolidating
concrete (6000 psi). A list of concrete properties is shown in Table 1.
Table 1. Concrete Mixes Used in the Experimental Program
ODOT Class
High strength Self-consolidating
C
Designed 28-day strength
4000
8000
6000
(psi)
Slump (inch)
4
6
6
Designed Water-cement
0.53
0.26
0.31
ratio
Accrual w-c ratio
0.648
0.552
0.33
Sand (lb)
1380
1140
1438
#57 Limestone
Gravel(lb)
1480
1660
738
#8 Limestone 783
Cement(lb)
Type 1-5 650 Type 2-3 665
648
AE-260 7
Duraflux33 (SCC
Admixture(oz)
AE-260 4
3500N 32
Admixture) 10
Super 84
Water(lb)
76
102
240
Concrete specimens were prepared to determine the mechanical strength at
different curing ages. Sufficient numbers of specimens were prepared to ensure at
least three repetitive tests can be conducted for each test. All of these samples were
kept in the curing room before they were subjected to different freezing processes. In
TDR test, seven specimens were monitored in each group of experiment; each
subjected to different curing process. Among these, there were four different freezing
process produced by placing the specimen into temperature controlled freezer at 1st,
2nd, 4th, and 6th day of curing. The remaining three specimens are cured under
normal curing conditions but with different types of TDR probe design, including a
parallel strip design, a four-spike probe design and an insulated four-spike probe
design. TDR sensor monitors the electrical conductivity (ECb) and dielectric constant
(Ka) of concrete specimens subjected to different curing processes. For each of these
seven specimens, thermal couples were used to monitor the temperature process
(Figure 3b). The channel assignments are shown in Table 2.

ASCE

129

Geo-China 2016 GSP 259

130

FIG. 3. a) TDR sensor head with cable

b) TDR sensor and thermocouple

distribution
Table 2. Distribution of Monitoring Channels
TDR
Channel
No.
1
5
3
4
6
7
8

TDR sensor type

Sample

When was it
frozen

Thermal couple No.

Parallel
Parallel
Parallel
Parallel
Parallel
Uninsulated
spikes
Insulated spikes

4-inch
4-inch
4-inch
4-inch
4-inch

1st day
2nd day
w/o
4th day
6th day

5
2
3
4
6

6-inch

w/o

6-inch

w/o

8
1 (Environment
Temp)

The ultrasonic system includes a high power pulse generator, 1MHz ultrasonic
transducers, and a PC-osciloscope (Fig. 4). The high characteristic frequency
ultrasonic transducer ensures a high resolution in determining the ultrasonic wave
speed. They are coupled to the concrete specimens via water.

FIG. 4. Ultrasonic testing device and example photos of testing specimen


In conjunction with ultrasonic test and TDR monitoring, the mechanical and
physical properties of concrete subjected to different curing procedures are also
measured. These are used to determine properties of concrete such as the

ASCE

Geo-China 2016 GSP 259

compressive strength, ultrasonic wave speed and water content of each type of
concrete (destructive oven dry method). The details of experiment plan are shown in
Table 3. The mechanical and physical properties of virgin samples were tested every
day from the beginning to the 7th day and then at 28th day. For the specimens
subjected to freezing, the strength and water content at 3rd day, 7th day, and 28th day
are measured. The frozen specimens were first defrosted before being tested.
Table 3. Physical Properties Test Schedule of Different Type of Concrete
Sample
1
2
3
4
5
6
7
28
Virgin
*
*
*
*
*
*
*
*
st
1 frozen
*
*
*
2nd frozen
*
*
*
th
4 frozen
*
*
*
th
6 frozen
*
*
*
EXPERIMENT DATA ANALYSIS
Ultrasonic velocity - strength relationship for concrete with normal curing
procedures
Figure 5 plots the strength of concrete versus ultrasonic wave velocity for
concrete subjected to normal curing conditions. As seen from this plot, there are
reasonably linear relationships between both terms. This confirms the observations in
many existing literatures. Such relationship can be used to estimate the compressive
strength of concrete from ultrasonic measurements.

FIG. 5. Relationships between ultrasonic velocity and strength of concrete


FIG. 6. Relationships between ultrasonic velocity (after thaw) and strength of
frozen concrete (after thaw)
ASCE

131

Geo-China 2016 GSP 259

Ultrasonic velocity (after thaw) - strength (after-thaw) relationship for concrete


subjected to freezing during the curing process
In this experiment, concrete specimens were placed in freezer at different curing ages
(1-day, 2-day, 3-day, 4-day, 5-day, etc.). At the end of the 7th day, the specimens were
taken out of the freezing room. Ultrasonic test and compression tests were then
conducted on each specimen either in complete frozen status or in complete thaw
status. The frozen status is ensured by compressing the specimens immediately after
taken out from the freezer; the complete thaw is ensured by subject the specimens to
rapid thawing procedures. Figure 6 shows the plots of ultrasonic wave velocity and
strength relationship for concrete specimens that are completely thaw. This figure
indicates there exist reasonably linear relationship between ultrasonic velocity (after
thaw) and strength of concrete (after thaw) subjected to freezing during the curing
process.

FIG. 7. The ultrasonic wave velocity of frozen concrete versus strengths of


concrete specimens
Ultrasonic velocity (before thaw) strength (before-thaw) relationship for
concrete subjected to freezing during the curing process
Figure 7 plots the ultrasonic velocity (before thaw) versus strength of frozen concrete
specimens (either before thaw or after thaw). The large scattering in the data
indicates there are no reliable relationships between the ultrasonic velocity (before
thaw) and compressive strength for frozen concrete, whether it is frozen or complete
thaw. This implies if NDT testing is conducted on concrete while it is frozen, the
measured wave speed can not be used to reliabily predict the strength of concrete.
The differences observed in Figs. 6 and 7 can be clearly seen from the trend of
ultrasonic velocity and strength development in frozen concrete samples (Figs. 8 and
9). Fig. 8 shows the comparison of before-thaw strength and after-thaw strength. Fig.
9 shows Real strength development when the concrete is frozen. The thaw concrete
specimens show gradual increasing trend in both compressive strength and ultrasonic
velocity. While those of frozen concrete show the opposite trend. This leads to what
are observed in Figs. 6 and 7.

ASCE

132

Geo-China 2016 GSP 259

FIG. 8. Comparison of before-thaw strength and after-thaw strength


FIG. 9. Real strength development when the concrete is frozen
Estimation of Frozen Concrete Strength
Fig. 8 shows the frozen early stage concrete has higher strength than those completely
thaw. A plot of the differences in the strength of concrete in the complete frozen or
thaw conditions is plotted in Fig. 10. Also plotted on Fig. 10 is the water content and
strength increment development with time. It is interesting to notice that both curves
show similar trends. Fig. 11 directly plots the linear relationship between water
content and the strength increment. There appear to exist high linear relationship
between both quantities. The relationship between these quantities is probably due to
the fact that when the free water in the sample freezes, the strength of ice contributes
the bulk strength of frozen concrete specimens.

FIG. 10. Water content and strength increment development with time
FIG. 11. Linear relationship between water content and the strength increment
With the use of relationship such as shown in Fig. 11, an estimation of the
strength of frozen concrete from seismic test can be performed in the following
sequences:
1) Defreeze the concrete
2) Estimate the strength of thaw concrete from ultrasonic test (using relationship
shown in Fig. 6)
3) Use free water content to determine the differences in the strength of concrete
between complete frozen and complete thaw status (using relationship shown
in Fig. 11)
ASCE

133

Geo-China 2016 GSP 259

4) Estimate the strength of frozen concrete by adding strength from steps 2) and
3).
A final issue to implement such a procedure is the estimation of the free water
content in concrete. This can be achieved by use of NDT technology such as Time
Domain Reflectometry (TDR). Previous research by the author shows TDR can be
used to reliably estimate the free water content in curing concrete (15). The
combination of ultrasonic testing and TDR testing thus provides a way to estimate the
strength of frozen concrete in the field. This information can be valuable for
pavement open-road decisions in cold regions, where the Winter Load Increases is
used to improve freight efficiency. The measured strength of frozen concrete can be
incorporated to determine the magnitude of overload that can be allowed on
pavements.
CONCLUSION
The relationship between seismic wave velocity and strength is the basis for
application of many NDT technologies based on seismic wave principles. Our
experimental data shows while there exist good linear relationships between the
compressive strength and the ultrasonic velocity, such relationship does not hold
when the concrete is frozen. The strength of frozen concrete does not have direct
correlations with ultrasonic wave velocity. The strength of frozen concrete, however,
can be estimated from that of thaw concrete and a correction related to the amount of
ice (free water). Combination of ultrasonic method and method for water content
measurement (such as TDR) provides a nondestructive method to estimate the
strength of curing concrete subjected to freezing procedures. This can potentially be
used for implement Winter Load Increases in cold regions. More studies are needed
to further validate and improve this technology.
REFERENCES
Boumiz, A., Vernet, C. and Tenoudji, F.C., "Mechanical properties of cement pastes
and mortars at early ages, evolution with time and degree of hydration." Advanced
Cement Based Materials 3 (3/4) (1996) 94-106.
Boutin, C. and Amaud, L., "Mechanical characterization of heterogeneous materials
during setting." European Journal of Mechanics, A/Solids 14 (4) (1995) 633-656.
D'Angelo, R., Plona, T.J., Schwartz, L.M. and Coveney, P., "Ultrasonic
measurements on hydrating cement slurries." Advanced Cement Based Materials
2 (l) (1995) 8-14.
Drnevich, V.P., Siddiqui, S.I., Lovell, J., and Yi, Q. (2001). "Water Content And
Density Of Soil Insitu by the Purdue TDR Method." TDR 2001: Innovative
Applications of TDR Technology, Infrastructure Technology Institute,
Northwestern University, Evanston, IL, September.
Keating, J., Hannant, O.J. and Hibbert, A.P., "Correlation between cube strength,
Ultrasonic pulse velocity and volume change for oil well cement slurries."
Cement and Concrete Research 19 (5) (1989) 715-726.
Klute et al, 1986. "Method of Soil Analysis: Physical and Mineralogical Methods."
NO.9, Part 1. 2nd Edition.
ASCE

134

Geo-China 2016 GSP 259

Pessiki, S.P. and Carino, N.J., "Setting time and strength of concrete using the
impact-echo method." ACI Materials Journal85 (5) (1988) 389-399.
Minnesota Department of Transportation, 1996. Frost Resistivity Probe. User Guide
Reinhardt, H.-W., Grobe, C. and Herb, A., "Ultrasonic monitoring of setting and
hardening of cement mortar - a new device." Mater.Struet. 33 (233) (2000) 580583.
Roberson R. L. and Siekmeier, J., 2000. "Determining Frost Depth in Pavement
System Using a Mult-Segment Time Domain Reflectometry Probe." TRB Session
Sayers, C.M. and Grenfell, R.L., "Ultrasonic propagation through hydrating cements:
onset of shear wave propagation." Ultrasonics 31 (3)(1993) 147-153.
Siddiqui, S.I., and Drnevich, V.P., (1995). "A New Method of Measuring Density and
Moisture Content of Soil Using the Technique of Time Domain Reflectometry."
Report No.: FHWA/IN/JTRP-95/9, Joint Transportation Research Program,
Indiana Department of Transportation - Purdue University, February, 271 p.
Topp, G. C., Davis, J. L., and Annan, A. P. (1980), "Electromagnetic Determination
of Soil Water Content and Electrical Conductivity Measurement Using Time
Domain Reflectometry." Water Resources Research, Vol. 16, pp. 574-582.
Yu, X. and Drnevich, V.P. 2004. "Soil Water Content and Dry Density by Time
Domain Reflectometry." Journal of Geotechnical and Geoenvironmental
Engineering, ASCE, Vol. 130, No.9, September, pp 922-934.
Yu, X., Drnevich, V. P. and Olek J., 2004, "Time Domain Reflectometry for
Measuring Water-Cement-Ratio of Concrete." the Proceedings of the 2004
International Symposium: Advances in Concrete through Science and
Engineering, Evanston, Illinois, March 21-24, 2004

ASCE

135

Geo-China 2016 GSP 259

136

The Influence of Pile Cap to p-y Curves under Lateral Loads


Shen-Kun Yu1; Zhi Zhang2; and Honghua Zhao3
1

Graduate Student, State Key Laboratory of Structural Analysis for Industrial Equipment,
Dept. of Engineering Mechanics, Dalian Univ. of Technology, Dalian 116024, China. E-mail:
815633498@qq.com
2
Graduate Student, State Key Laboratory of Structural Analysis for Industrial Equipment,
Dept. of Engineering Mechanics, Dalian Univ. of Technology, Dalian 116024, China. E-mail:
296175834@qq.com
3
Associate Professor, State Key Laboratory of Structural Analysis for Industrial Equipment, Dept. of
Engineering Mechanics, Dalian Univ. of Technology, Dalian 116024, China. E-mail: zhaoh@dlut.edu.cn

Abstract: The pile-soil-cap interaction model was established based on the threedimensional fast Lagrangian analysis software Flac3D. The concrete pile under lateral
loads with different sizes of cap and different lengths are investigated through numerical
simulations. Then the curves of lateral resistance of lateral soils surrounding the pile
with lateral displacement (p-y curves) was calculated and the influences of pile caps on
p-y curves were analyzed. The results show that the cap will affect the location of the
maximum soil resistance for the single pile, and if the critical load is reached, the
maximum soil resistance will increase with the sizes of cap. The influences of the cap to
the soil are mainly within the first quarter of the pile body. This paper also discussed the
influences of cap on the soil resistance with different pile lengths. These conclusions
provide the basis for the design of the pile-soil-cap system at the condition of lateral
loads.
INTRODUCTION
Laterally-loaded piles are widely used in ocean engineering, harbor engineering,
bridge engineering and landslide preventions etc. We commonly use linear elastic
subgrade reaction method to solve the lateral displacement when the pile deflection is
small. But considering the pile with pile cap, it allows larger displacement to happen, the
elastic-plastic analysis is used generally, also known as the elastic-plastic subgrade
reaction method. The p-y curve method is the most widely used in the practice, and is

ASCE

Geo-China 2016 GSP 259

currently one of the important methods to calculate laterally-loaded piles(Wang et al.


2005). Matlock (1970) and Reese et al.(1975) first put forward p-y curve method,
according to the in-situ tests and laboratory tests and his method is adopted by the API
specification. Scholars in China also did a lot of research on p-y curve method (e.g.
Wang et al. 1991; Tian et al. 1993; Zhang 2001; Yang et al. 2002). The basic idea of p-y
method is to express the stress-strain relationship of soil in a set of curves along the pile
depth direction(Pile foundation manual writing committee 2009). Under the action of
lateral force, the relationship curve of soil resistance per unit pile length at depth z can
comprehensively reflect the pile stiffness and soil nonlinea- rity, as well as the nature of
the pile under the loads.
The best way to determine the p-y curve is through in-situ field testing, but this is very
difficult. Now most researchers mainly use the laboratory experimental method.
According to the different soil conditions and pile foundation conditions, they choose
the calculation method, and estimate the p-y curve with soil parameters obtained through
indoor triaxial tests(Yang 2004). The development of modern computer technology
provides the convenience to use numerical analysis software for estimating p-y curves.
Many studies are conducted through numerical methods. Qi et al.(2009) used threedimensional finite element method (FEM) to build the pile-soil interaction model and
analyzed the influences of pile diameter on the lateral limit resistance and limit
displacement of soil, the results showed that the soil resistance increases gradually with
the increase of pile diameter. Some scholars have discussed the differences of transverse
response whether or not considering the pile-soil interface parameters (e.g. Hong et al.
2007; Fatahi et al. 2014). Wang et al.(2011) analyzed the rigid pile-soil interaction using
three dimensional finite element analysis, and studied the soil reaction force with the
changing lateral loads. Using the numerical method, Emilios (2013) investigated the
pile group model under lateral loads, and analyzed the differences of p-y curve when
comparing with the monopiles. Also, full-scale lateral loading tests are indispensable,
Hokmabadi et al.(2012) investigated the behavior of monopiles in granular marine soils.
With the increasing of high-rise buildings, and the development of deepwater ports, it
requires the pile foundation to have a higher lateral load capacity. However, p-y curve
studies for single pile did not or rarely consider the influence of the pile caps in the past.
The pile cap is one of the important factors influencing the lateral bearing capacity of
single pile. There were some field experimental studies of pile caps effects earlier. Kim
and Brungraber begun to study the pile-soil-caps system in the 1970s (Kim et al. 1979).
Rollins et al.(2002)studied the pile caps role on the lateral bearing resistance, and Nath
et al.(2013) also did full-scale tests on the lateral resistance of the pile cap. However,
numerical simulation work is still very limited. Based on the above analysis, in this
paper, the three-dimensional fast Lagrangian analysis software Flac3D was used to study
the influence of pile cap on the p-y curve under lateral loads.

ASCE

137

Geo-China 2016 GSP 259

138

3D FINITE DIFFERENCE MODEL FOR THE PILE-SOIL-CAP SYSTEM


A 3D finite difference model is built for a pile-soil-cap system. The soil depth is 20m;
it is divided into grids per 0.5 m thick. Both width and length of the soil mass is 32m.
For the soil mass, the soil near the pile use radcylinder grid, away from the pile using
radtunnel cell division. The concrete pile body was meshed with cylinder type grid
every 0.5 m. The embedded depth for the pile cap was considered, flush on the lateral
plane and radcylinder element was used to mesh grid, and no thickness of triangular
elements was used to mesh the pile-soil contact surface. The ratio of the soil and pile
elements was 1:1:1:1.15. It is worth mentioning that foundation soil is considered as an
ideal elastic-plastic material, and the concrete pile is a linear elastic material. Material
parameters are shown in Table 1. The model of single pile with a cap was shown in
Figure 1. In order to facilitate observation, we only show half of the actual model, and
the pile was not embedded soil absolutely.
Table 1. The Material Parameters of Piles and Soil
Concrete Pile
2500
30000
0.167
-

Density(kg/m3)
Elasticity Modulus(MPa)
Poisson Ratio
Cohesion (kPa)
Internal Friction Angle

Clay
1770
45.0
0.30
16
7.7

32m
2.5m

clay
20m

interface

10m

16m

pile

1m

FIG. 1. 3D Finite Difference Model for the Pile-soil-cap System


THE INFLUENCES OF PILE CAP ON P-Y CURVE
P-y Curve with Different Pile Cap Sizes
First, the influences of pile cap sizes on p-y curves are investigated. The selected pile
length is 10m, the pile body is cylindrical, and the diameter is 1m. The caps have four
kinds of specifications, which named no cap, small cap (1.5m1.5m), medium
cap(2.5m2.5m), and large cap(4m4m). According to the regulation of Building pile

ASCE

Geo-China 2016 GSP 259

foundation technical specification of China (JGJ94-2008)( China Academy of Building


Research 2008), for the reinforced concrete pile, we can select 75% of the loads as the
single pile bearing capacity when the lateral displacement at the pile head reaches 10mm
in the static lateral loading tests.
In the numeric calculation, we can continue to apply the lateral load until the lateral
displacement of pile head reaches 10mm, and we named the load at this time critical
load. The corresponding soil resistance at 10mm can be obtained , then we drawn the p-y
curve at the depth of 0m, 2m, 4m, 6m, 8m, 10m along the pile in Figure 2.

FIG. 2. P-y curves at different depth of the 10m pile with cap sizes in
(a) no cap; (b) small cap; (c) medium cap; (d) large cap
According to the Figure 2, we can see that the value of the soil reaction force on the
pile body sharply decreases with the soil depth. We would find that every curve has an
deflection point, at which the soil lateral displacement is zero, and then below would
have small soil reaction force in the opposite direction. Another obvious change is the
force at the pile bottom which gradually decreases with the increase of cap sizes. The
existence of the pile caps can change the point of the maximum soil reaction force under
the critical load. With no cap, the maximum soil resistance is at about a quarter along the
pile length, and the existence of the pile caps will make the maximum soil resistance
acting at the pile head. After adding the pile cap, the soil reaction force produced by the
lateral loads is mainly beard by the caps, and with the increase of cap size, the pile cap
bears more percentage of the total lateral loads. The soil resistances at the pile cap for
piles with three different caps are 4.55 times, 5.32 times, 7.19 times of those at the
embedded depth of 2m, respectively.

ASCE

139

Geo-China 2016 GSP 259

140

FIG. 3. The p - y curve of pile head at critical load with different cap sizes
Figure 3 shows the lateral soil resistance of the pile with different size of caps under
the critical load. From no cap to large cap the maximum soil resistance:78.6 kPa341
kPa, 479 kPa, and 719 kPa. The cap will significantly increase the lateral loads bearing
capacity of the pile. Based on these results, the commonly used method of calculating
the p-y curve will give the displacement greater than the real displacement for those with
caps. The design will be rather conservative. The ratio of the largest soil resistance of
three types of caps is 1:1.4:2.1 and the caps sizes ratio is 1:1.7:2.7. Presumably, when
reaching the critical load, the soil resistance is directly proportional to the size of pile
caps, but the efficiency of bearing the loads would decay with the increasing the cap
sizes.

FIG. 4. Lateral soil resistance along pile length at the


critical load for piles with different caps
Figure 4 reflects the lateral soil resistance of the pile foundation at different depth at
the critical load. The effects on the soil resistance of the pile caps are mainly limited
within the first quarter of pile length. Based on these analyses, in the reinforcement of
the soils for the pile foundation, we only need to reinforce the soil along the first quarter
depth of pile to achieve the goal of improving the lateral bearing capacity of pile.

ASCE

Geo-China 2016 GSP 259

P-y Curve for Piles with Different Length


In order to further study the effects of pile caps on p-y curves for piles with different
length, three different pile length 6m, 10m, 15m were selected for the numerical
simulation analysis. The p-y curves are considered for the piles with no cap, small cap,
medium cap and large cap, respectively. The p-y curve at the pile head are plotted in
Figure 5.

FIG. 5. The pile head p-y curves of different pile length(6m, 10m, 15m):
(a) no cap; (b) small cap; (c) medium cap; (d) large cap
Based on the data from the numerical simulation, when the pile length is 6m, the soil
lateral resistance value at the pile head would 78 kPa, 310 kPa, 531 kPa, and 944kPa for
four types of pile caps. When the pile length is 10m, the resistance is 78.9 kPa, 341 kPa,
479 kPa, and 719 kPa, respectively. When the pile length is 15m, the resistance is 91.2
kPa,335 kPa, 452 kPa, and 650 kPa respectively. Through analyzing the data, with
the increase of pile length, pile head soil resistance increases for the pile has no cap; For
the pile with small cap, with the increase of pile length, pilehead soil resistance increases
first then decreases; and for the pile with medium or large cap, the pile head soil
resistance decreases with the increase of pile length. These indicate the gradual changing
influences of pile cap size on the pile head soil resistance. To sum up, under the
condition of the pile cap, the shorter the single pile, the larger the soil resistance at pile
head when reaching the critical load, and the greater the pile could bear the lateral loads.

ASCE

141

Geo-China 2016 GSP 259

CONCLUSIONS
For the concrete piles in clay, we analyzed the influences of caps on the p-y curves
under lateral loads with the numerical analysis based on Flac3D. The following
conclusions can be drawn:

For the single pile with no cap under lateral loads, the maximum soil resistance of
pile side will appear at the quarter of the pile depth. Most of the lateral soil
resistance would be borne by the pile caps. Another influence is the resistance
force at the pile bottom gradually decreases with the increase of cap sizes.

When the applied lateral load reaching the critical load, the existence of the pile
caps only has a significant influence on the soil resistance with the first quarter
length of the pile. For the other part of soil along the pile length, it has little
influence on the soil resistance.

The commonly used p-y curve calculation method will give the displacement
greater than the real displacement for that with caps. The design will be rather
conservative.

With the increases of pile length, soil resistance at the pile head will change
because of the presence of pile caps and different cap sizes. With the increase of
pile length, pile head soil resistance increases when the pile has no cap. Under the
condition of the pile cap, the shorter the single pile, the larger the soil resistance at
pile head when reaching the critical load, and the greater the pile could bear the
lateral loads.
.
ACKNOWLEDGMENTS
This research is supported by the National Natural Science Foundation of China (No.
51308091), and the Fundamental Research Funds for the Central Universities
(DUT15QY48).
REFERENCES
China Academy of Building Research (2008). Pile foundations JGJ94-2008 specification. Beijing: China Building Industry Press.
Emilios, M., Comodromos, Mello, C., and Papadopoulou (2013). Explicit extension of
the py method to pile groups in cohesive soils. Computers and Geotechnics, Vol.
47: 28-41.
Fatahi, B., Basack, S., Ryan, P., Zhou, W.H., and Khbbaz, H. (2014). Performance of
laterally loaded piles considering soil and interface parameters. Geomechanics and
Engineering, Vol. 7(5):495-524.
Hong, Y., Xie, Y., Zhang, S., and Yan, J. (2007). Horizontally loaded piles finite
element simulation with an analysis of p-y curve method. China Harbour
Engineering, Vol. 3:5-9.

ASCE

142

Geo-China 2016 GSP 259

Hokmabadi, A.S., Fakher, A., and Fatahi, B. (2012). Full scale lateral behaviour of
monopoles in granular marine soils. Marine Structures, Vol. 29(1):198-210.
Kim, J.B., Brungraber, R.J., and Singh, L.P. ( 1979). Pile cap soil interaction from fullscale lateral load tests. Journal of the Geotechnical Engineering Division, Vol.
105(5): 643-653.
Matlock, H.S. (1970). Correlation for design of laterally loaded pile in soft clay.
Proceedings of 2nd Offshore Technology Conference, Houston:[s.n.].
Nath, U.K., and Hazarika, P.J. (2013). Lateral resistance of pile capan experimental
investi-gation. International Journal of Geotechnical Engineering, Vol. 7(3):266272.
Pile foundation manual writing committee (2009). Pile and Pile Foundation Handbook. Beijing: China Building Industry Press.
Qi, C., Wang, J., and Li, S. (2009). A research about the influence of pile diameter to
the pile-soil interaction p-y curve. Journal of Water Resources and Hydropower
Technology, Vol. 40 (3):43-46.
Reese, L.C., Cox, W.R., and Koop, F.D. (1975). Field testing and analysis of laterally
loaded piles in stiff clay. Proceedings of 7th Offshore Technology Conference,
Houston: [s.n.].
Rollins, M.K., and Sparks, A. (2002). Lateral resistance of full-scale pile cap with
gravel backfill. Journal of Geotechnical and Geoenvironmental Engineering, Vol.
128(9): 711-723.
Tian, P., and Wang, C. (1993). A unified method about p-y curve of transverse periodic
loaded piles. Journal of Hehai university, Vol. 21(1):9-14.
Wang, C., Wu D.,and Tian, P. (1991). A new unified method about p-y curve of lateral
static loaded piles in clay. Journal of Hehai University, Vol. 12 (1) : 9 -17.
Wang, C., and Sun, D. (2005). Laterally loaded piles p-y curve research and
application. China Harbour Engineering, Vol. 2:135-138.
Wang, X., Feng, H., and Li, P. (2011). Soil resistance distribution research of rigid pile
with the lateral load. Highway Traffic Science and Technology, Vol. 7, 212-215.
Yang, G., and Zhang, Z. (2002). The research of p-y curve about laterally static loaded
piles on the condition of large deflection. Water Transportation Engineering, Vol.
7: 40-45.
Yang, K. (2004). Practical pile foundation engineering. Beijing: People's Traffic
Press.
Zhang, S. (2001). The research of p-y curve about horizontally static loaded piles.
NanJing: Hehai university.

ASCE

143

Geo-China 2016 GSP 259

144

Multi-Scale Analysis of Deformation Modes in Granular Material Using a


Dynamic Hybrid Polygonal Finite Element-Discrete Element Formulation
Brandon Karchewski1; Peijun Guo2; and Dieter Stolle3
1

Instructor, Dept. of Geoscience, Univ. of Calgary, 2500 University Dr. NW, Calgary, AB, T2N 1N4,
Canada. E-mail: brandon.karchewski@ucalgary.ca
2
Professor, Dept. of Civil Engineering, McMaster Univ., 1280 Main St. West, Hamilton, ON, L8S 4L7,
Canada. E-mail: guop@mcmaster.ca
3
Professor, Dept. of Civil Engineering, McMaster Univ., 1280 Main St. West, Hamilton, ON, L8S 4L7,
Canada. E-mail: stolle@mcmaster.ca

Abstract: We are interested in capturing the multi-scale behaviour of granular


materials; that is, how micro-scale particle interactions influence the macroscopic
behaviour of granular materials. We present a novel plane strain dynamic formulation
for a multi-scale hybrid finite element-discrete element analysis. The formulation
consists of two basis elements: hybrid polygonal body elements representing grains
with linear elastic behaviour and interface elements representing the nonlinear
interactions between grains. Combining the two elements provides a convenient
technique for obtaining results akin to a discrete element simulation, but within a
continuum-based finite element framework. We apply the model to simulate biaxial
compression tests with an initial consolidation phase under uniform pressure followed
by strain-controlled deviatoric loading. The model captures the stress-strain
relationship of a typical compression test on granular material, including the post-peak
softening regime. Eigen-analysis of the granular structure reveals that this modelling
approach captures the rich bifurcation space associated with the failure of granular
materials.
INTRODUCTION
To understand the behaviour of granular materials, it is important to realize that they
differ fundamentally from solids such as metals. Whereas one may consider metals to
be continuous solids even at the microscopic scale, granular materials consist of
discrete solid grains often visible to the naked eye that interact by frictional
sliding. Indeed, the manner in which granular materials behave when compared with
solids, liquids and gasses has led some to classify them as a distinct form of matter (cf.
Jaeger et al. 1996). We direct the reader to Cambou et al. (2009) and Chau and Zhao
(2015) and the references therein for a more complete review of the literature on
granular material behaviour.

ASCE

Geo-China 2016 GSP 259

The present work uses a form of combined finite element-discrete element (FEMDEM) model (cf. Munjiza 2004) to capture multi-scale behaviour in granular
materials. The model uses Voronoi cell elements to represent the grains; hence, we
call the model VCFEM-DEM. We model the interactions between grains using
interface elements with a volumetric-deviatoric hardening plasticity formulation to
account for the nonlinear constitutive behaviour of grain interactions. We direct the
reader to the work of Goodman et al. (1968), Pietruszczak and Mrz (1981), Wang et
al. (2003), Selvadurai and Yu (2005) and the references therein for details on
modelling interfaces in geomaterials. In the following sections, we briefly summarize
the formulation and then demonstrate its performance by simulating strain localization
failure of granular material in a biaxial compression test.
FORMULATION
Governing Equations
We discretize the total domain V into a set of polygonal subdomains Ve representing
grains and interfaces Vi representing grain interactions. The discretized form of the
equation of motion is,
M
au + FI (t) = FE (t)
(1)

where M is the mass matrix for the discretized system, au is a vector of nodal
displacements, FI(t) is a vector of internal forces at time t, FE(t) is a vector of external
applied forces at time t and superposed dots represent derivatives with respect to time.
The internal force vector FI has a linear elastic component FI(g) from the grain
elements and a nonlinear component FI(i) from the interfaces between grains. We solve
equation (1) using an implicit Crank-Nicolson scheme. Next, we describe how we
form the element matrices and force vectors for the two types of element.
Hybrid Polygonal Element (Grain Element)
We assume the polygonal grain elements to be linear elastic and derive the stiffness
matrix from the modified complementary energy functional of Tong and Pian (1969).
We interpolate stress inside the grain elements as = P + 0 where P contains a
set of monomial basis functions and is a vector of unknown coefficients. The P
term represents the stress field component owing to grain deformation, interpolated to
satisfy equilibrium a priori and 0 represents a constant initial stress. Interpolation of
displacement u* on the element boundary Se is of the form u* = Nu*au where Nu*
contains linear shape functions interpolating between the adjacent nodal values of
displacement au. Substituting the stress and displacement interpolations into the
modified complementary energy functional gives the discretized form of the hybrid
functional,
(2)
MCE [ , au ] = 12 T H + T G uau aTu Ft0
where H = Ve (PTDe-1P)dV, Gu = Se (nTP)TNu*dS, Ft = St (Nu*Tt )dS,
F0 = Se Nu*T(nT0)dS and Ft 0 = Ft F0. In these expressions, De-1 represents the elastic
compliance tensor, n represents the outward pointing unit normal and t is an applied
traction on St. Stationarity of equation (2) at equilibrium requires that the variations
with respect to and au vanish, and noting that the coefficients need not be
continuous between elements leads to = H-1Guau where Tong and Pian (1969)
ASCE

145

Geo-China 2016 GSP 259

146

present the requirements for rank sufficiency and invertibility of H. This substitution
leads to the element stiffness matrix for the polygonal grain elements as
Kuu(g) = GuTH-1Gu and the component of internal force FI(t) owing to body element
deformation FI(g)(t) = Kuu(g)au(t). Pian and Wu (2006) provide a more detailed
literature review on the origin, formulation and application of hybrid finite elements
and Ghosh (2011) discusses further details regarding polygonal elements.
We compute the consistent mass matrix M for the grain elements using the wellknown relation M = Ve (NuTNu)dV where Nu is a matrix containing shape functions
that interpolate between nodal values of displacement inside the polygonal element
domain Ve (cf. Sukumar and Tabarraei 2004) and is the material density. We
perform mass lumping for computational efficiency, with the understanding that this
may introduce a small amount of numerical damping into the solution.
Interface Element (Grain Interaction Element)
We define the constitutive behaviour of the interface element using a combined
volumetric-deviatoric hardening elastoplastic formulation (cf. Pietruszczak 2010). We
base the interface element formulation on an incremental strain energy relation,
U i [ , ] = ( T ) dV + 12 ( T ) dV
Vi

Vi

(3)

where = { , }T is the local strain increment, = { , }T is the previous


effective stress state, is a finite effective stress increment and { , } are local
coordinates tangential and normal to the interface, respectively. An equivalent
expression in terms of the discretized finite element nodal displacements is,
Ui [ au ] = aTu ( BT ) dV + 12 ( BT DepB) dV au = aTu FI (i) + 12 aTu Kuu(i ) au
Vi
Vi

(4)

where au is a vector of nodal displacement increments, B is a kinematic matrix


relating global displacements to local strains, Dep is the tangent elastoplastic interface
constitutive matrix, FI(i) is the internal force vector for an interface element and Kuu(i)
is the tangent stiffness matrix for the interface.
We derive the local strain increments from the relative displacement increments
of the interface in the local coordinate system as = Liu where Liu is a
difference operator that approximates a linear differential operator assuming linear
interpolation of displacements in the direction and a representative interface
thickness e. To maintain stress continuity over the interfaces, we assume that normal
strains in the direction are negligible (Pietruszczak and Mrz 1981). Transforming
the relative displacements from local to global coordinates as = Tu and linearly
interpolating the relative displacement increments as u = uau leads to the
definition of the kinematic matrix as B = LiuTu.
We define the incremental stress-strain relation for an interface element as
= Dee = Ge + Me where De is the elastic constitutive matrix and e is the
elastic strain increment for the interface with G and M being the elastic shear and
constrained moduli. We assume additivity of elastic and plastic strain increments,
e = p and that the stress state during active loading lies on an expanding
loading surface f(, ) = 0, which we define to be elliptical, following the modified

ASCE

Geo-China 2016 GSP 259

147

critical state approach (Muir Wood 1990), where is the hardening parameter. Active
loading satisfies the consistency condition and at failure, the interface stresses satisfy
the Mohr-Coulomb yield criterion. We do not present the details of the plasticity
formulation here, but this formulation leads to the definition of the internal force
vector FI(i) = Ve (BT)dV and tangent stiffness matrix Kuu(i) = Ve (BTDepB)dV for an
interface element where Dep depends on the hardening coefficient H = He + Hp where
He = (f/)TDe(f/) and Hp = -(f/){ (f/) (1 + e0) (f/) }. Note that e0
is the void ratio of the interface and is a parameter specifying the component of the
increment to the hardening parameter owing to plastic shear strain increments, which
depends on the friction angles at failure and critical state f and c, respectively,
tan f tan c

= (1 + e0 )
tan c tan f

1
2

(5)

This presentation of the formulation was necessarily brief. We refer to Karchewski


(2015) for the complete development of the formulation.
BIAXIAL COMPRESSION SIMULATIONS
Simulation Setup and Parameters
To examine the capability of this framework, we ran a series of simulations of
biaxial compression tests. This type of simulation is a common test of analysis
frameworks for modelling strain localization in granular materials in the literature (e.g.
Borja et al. 2013, Gu et al. 2014). Our aim here is to present representative results
demonstrating its ability to capture multi-scale phenomena in granular materials.
FIG. 1 shows a sample mesh for a domain consisting of a 0.05 m 0.15 m region of
coarse sand with feldspar sand grains (E = 40.0 GPa, = 0.33 and = 2.65 g/cm3) of a
uniform 3.5 mm nominal particle diameter. The granular part of the domain has 640
grain elements and 1825 interface elements. FIG. 1 also shows the distribution of
contact normal directions for the sample mesh. We selected interface properties such
that the elastic shear modulus is the same as that of the grains, but the normal
(constrained) modulus is much higher for the range of confining pressures applied in
the simulations with G = 15.0 GPa, = 1.5010-5, = 5.0010-7, f = 30, c = 25, e0
= 0.9 and e = 2.810-4 m where and are the compression and recompression
indices, respectively, in the volumetric-deviatoric strain hardening plasticity model.
There is also a 0.05 m 0.025 m aluminum loading cap, modelled explicitly to reduce
the influence of boundary conditions on the results. Both the grain elements and
interface elements in the loading cap were treated as linear elastic with E = 70.0 GPa,
= 0.33 and = 2.70 g/cm3.
We model the biaxial compression simulation in two phases. The first phase is a
consolidation phase during which we first apply an initial seating pressure of 2 kPa.
After stabilization at the seating load, we increase the confining pressure linearly to a
maximum value of 200 kPa and hold this constant until the specimen stabilizes. The
second phase is the strain controlled shearing phase during which we hold the normal
tractions on the vertical sides of the specimen constant at a confining pressure of
200 kPa and apply a vertical strain rate of 1.0 mm/s.

ASCE

Geo-China 2016 GSP 259

FIG. 1. Sample VCFEM-DEM mesh for biaxial compression simulations (left),


distribution of contact normals (top right) and magnified view of grain and
interface elements (bottom right)
Results and Discussion
FIG. 2 shows a plot of the mobilized shear in the specimen for three different grain
elements during the strain-controlled portion of the simulation for the mesh shown in
FIG. 1. The stress-strain curves are characteristic of behaviour observed in laboratory
compression tests on granular material. During the pre-peak regime, the curve is
nonlinear owing to the strain hardening plasticity model used for the interface
elements. The peak mobilized friction angle f for this simulation was 47.5, which is
substantially higher than the interface friction angle, but is consistent with the fact that
this simulation represents a very dense granular material with a porosity of 0.25.
Following the peak, there is softening until the stress level reaches critical state at a
mobilized friction angle of approximately 33.0. Again, this is higher than the critical
state friction angle of the interface elements, but we attribute this to the density of the
specimen. We note that the stress-strain curves for grain elements in different parts of
the domain coincide in the pre-peak regime, but diverge in the post-peak regime
owing to the rich bifurcation space for the post-peak behaviour.
Post-peak softening cannot be captured at the constitutive modelling level since it
develops through macro scale inhomogeneities in stress and strain (cf. Pietruszczak
and Mrz 1981, Samaniego and Belytschko 2000, Wang et al. 2003). FIG. 3
demonstrates how this occurs in the VCFEM-DEM model by plotting the number of
interface elements in an active loading, shear failure and extension failure state.
Comparing FIG. 2 and FIG. 3, we observe that there is a linear increase in interfaces at
shear failure up to the peak, but few in extension failure. Failure occurs when enough
failed interfaces coalesce to form a macroscopic shear band mechanism. Following
this, there is a rapid increase in interfaces failed in extension corresponding to local
dilations in the shear band. This observation is consistent with the evidence of Hadda et al. (2015)

ASCE

148

Geo-China 2016 GSP 259

149

Mobilized friction angle, [deg]

60
phi_001
phi_002
phi_003

50
40
30
20
10
0
0.0E+00

1.0E-05

2.0E-05

3.0E-05

4.0E-05

Average axial strain, yy [-]

FIG. 2. Representative stress-strain curves during strain controlled shearing.


Curve numbers correspond to evaluations in different grain elements.

# of plastic, failed and tensile


interfaces

2000
1500
1000

nplas
nperplas
ntens

500
0
0.0E+00

1.0E-05

2.0E-05

3.0E-05

4.0E-05

Average axial strain, a [-]

FIG. 3. Number of plastic and failed interface elements during strain controlled
shearing. nplas = # of plastic elements (active loading, but not failed), nperplas =
# of elements failed in shear, ntens = # of elements failed in extension.
for aggregation of so-called c- contacts in the formation of shear bands.
Having briefly demonstrated that the VCFEM-DEM framework captures the
fundamental phenomena defining shear failure of a granular material, we close by
examining deformation modes obtained from tangent eigen-analysis of the granular
structure at the end of the consolidation phase. FIG. 4 demonstrates two main types of
deformation modes: global modes resulting from the granular structure deforming in
unison and local modes represent deformations owing to relative grain movement and
rotation. The regions of high grain rotation divide the specimen into sub-regions,
whose size represents a characteristic length scale for the material test, as shown in the
two failure modes in FIG. 4(b). For the specimen shown, the characteristic length is
approximately one third to half of the specimen height. Note that the mode shapes and
their corresponding eigenvalues change throughout the test, owing to the stress
dependency of the interface element stiffness. However, this set of deformation modes
already shows patterns similar to the ultimate shear band failure mode. This is
consistent with observations in recent studies that there exists an entire bifurcation domain

ASCE

Geo-China 2016 GSP 259

150

~6.0 cm
~7.5 cm

~19
~47

(a) Global modes


(b) Local modes
FIG. 4. Tangent eigenmodes in biaxial compression specimen at c = 200 kPa
for the possible diffuse failure modes in a granular material (Daouadji et al. 2011, Wan
et al. 2011). The mode that governs the failure depends on how closely the boundary
conditions correspond with the deformation pattern of each mode.
CONCLUSIONS
We formulated and implemented a novel model for granular materials that uses
hybrid polygonal elements with linear elastic behaviour to represent the grains and
interface elements with volumetric-deviatoric strain hardening plasticity to represent
the interactions between grains. The results demonstrate its ability to capture
behaviour characteristic of a strain-controlled compression test on granular material
including the post-peak softening associated with a macroscopic shear band
mechanism. Through eigen-analysis of the granular structure, we observed local
deformation patterns similar to the ultimate failure mode already present at the end of
the consolidation phase. This indicates the degree to which the granular structure may
control the failure, since the fingerprint of the shear failure is present even under
isotropic loading conditions. Further investigation of how the deformation modes
change throughout the simulation is necessary to understand how this influences the
failure of the granular specimen.
ACKNOWLEDGMENTS
This work was partially funded by the Natural Sciences and Engineering Research
Council of Canada (NSERC). The authors gratefully acknowledge this support.
REFERENCES
Borja, R.I., Song, X., Rechenmacher, A.L., Abedi, S., and Wu, W. (2013). Shear
band in sand with spatially varying density. J. Mech. Phys. Sol., 61, 219-234,
DOI: 10.1016/j.jmps.2012.07.008.
Cambou, B., Jean, M., and Radja, F. (Eds.). (2009). Micromechanics of Granular
Materials. Wiley: Hoboken.

ASCE

Geo-China 2016 GSP 259

Chau, K.-T., and Zhao, J. (Eds.). (2015). Bifurcation and Degradation of


Geomaterials in the New Millenium: Proceedings of the 10th International
Workshop on Bifurcation and Degradation in Geomaterials. Springer: New York.
Daouadji, A., Darve, F., Al Gali, H., Hicher, P.-Y., Laouafa, F., Lignon, S., Nicot, F.,
Nova, R., Pinheiro, M., Prunier, F., Sibille, L., and Wan, R. (2011). Diffuse failure
in geomaterials: Experiments, theory and modelling. Int. J. Num. Anal. Meth.
Geomech., 35, 1731-1773, DOI: 10.1002/nag.975.
Ghosh, S. (2011). Micromechanical Analysis and Multi-Scale Modeling using the
Voronoi Cell Finite Element Method. CRC Press: New York.
Goodman, R.E., Taylor, R., and Brekke, T. (1968). A model for the mechanics of
jointed rock. J. Soil Mech. Found. Div., ASCE, 94, 637-659.
Gu, X., Huang, M., and Qian, J. (2014). Discrete element modeling of shear band in
granular materials. Theor. Appl. Frac. Mech., 72, 37-49, DOI:
10.1016/j.tafmec.2014.06.008.
Hadda, N., Nicot, F., Wan, R., and Darve, F. (2015). Microstructural selforganization in granular materials during failure. Compt. Rend. Mec., 343, 143154, DOI: 10.1016/j.crme.2014.09.009.
Jaeger, H.M., Nagel, S.R., and Behringer, R.P. (1996). Granular solids, liquids, and
gases. Rev. Mod. Phys., 68, 1259-1273, DOI: 10.1103/RevModPhys.68.1259.
Karchewski, B. (2015). Multi-scale modelling of geomechanical behaviour using the
Voronoi cell finite element method (VCFEM) and finite-discrete element method
(VCFEM-DEM). Ph.D. Thesis, Department of Civil Engineering, McMaster
University.
Muir Wood D. (1990). Soil Behaviour and Critical State Soil Mechanics. Cambridge
University Press: Cambridge.
Munjiza, A. (2004). The Combined Finite-Discrete Element Method. Wiley: London.
Pian, T.H.H., and Wu, W.-C. (2006). Hybrid and Incompatible Finite Element
Methods. Chapman & Hall/CRC: New York.
Pietruszczak, S. (2010). Fundamentals of Plasticity in Geomechanics. CRC Press:
New York.
Pietruszczak, S., and Mrz, Z. (1981). Finite element analysis of deformation of
strain-softening materials. Int. J. Num. Meth. Eng., 17, 327-334, DOI:
10.1002/nme.1620170303.
Selvadurai, A.P.S., and Yu, Q. (2005). Mechanics of a discontinuity in a
geomaterial. Comp. Geotech., 32, 92-106, DOI: 10.1016/j.compgeo.2004.11.007.
Sukumar, N., and Tabarraei, A. (2004). Conforming polygonal finite elements. Int.
J. Num. Meth. Eng., 61, 20452066, DOI: 10.1002/nme.1141.
Tong, P., and Pian, T.H.H. (1969). A variational principle and the convergence of a
finite element method based on assumed stress distribution. Int. J. Solids Struct.,
5, 463472, DOI: 10.1016/0020-7683(69)90036-5.
Wan, R.G., Pinheiro, M., Guo, P.J. (2011). Elastoplastic modelling of diffuse
instability response of geomaterials. Int. J. Num. Anal. Meth. Geomech., 35, 140160, DOI: 10.1002/nag.921.
Wang, X., Chan, D., and Morgenstern, N. (2003). Kinematic modelling of shear band
localization using discrete finite elements. Int. J. Num. Anal. Meth. Geomech.,
27:289-324, DOI: 10.1002/nag.274.

ASCE

151

Geo-China 2016 GSP 259

152

A Novel Model to Simulate the Behaviour of Cement-Treated Clay under Compression


and Shear
Lam Nguyen1; Behzad Fatahi2; and Hadi Khabbaz3
1

Ph.D. Candidate (M.Eng., B.Eng.), School of Civil and Environmental Engineering, Univ. of Technology
Sydney (UTS), Sydney, Australia. E-mail: Lam.D.Nguyen@student.uts.edu.au
2
Senior Lecturer, Geotechnical Engineering (Ph.D., CP.Eng.), School of Civil and Environmental
Engineering, Univ. of Technology Sydney (UTS), Sydney, Australia. E-mail: Behzad.Fatahi@uts.edu.au
3
Associate Professor, Geotechnical Engineering (Ph.D.), School of Civil and Environmental Engineering,
Univ. of Technology Sydney (UTS), Sydney, Australia. E-mail: Hadi.Khabbaz@uts.edu.au

Abstract: Soft clay treated with cement shows an improvement on strength due to the
chemical interaction between cement and clay particles. Laboratory results showed that
the peak strength of the cement treated clay reduces as the mean effective stress
increases due to the effect of cementation degradation. Therefore, in this paper, a
constitutive model was developed to simulate the behaviour of cement treated clay.
Based on the Critical State Soil Mechanics, the model proposed the non-linear failure
envelope for the cement treated clay to merge with the Critical State Line (CSL) of the
un-reinforced clay when the reinforced samples reach a sufficiently high stress levels.
Moreover, a modified mean effective stress was proposed to include the contribution of
cementation and its cementation degradation. Furthermore, the proposed model was
evaluated by comparing the proposed model predictions against the Singapore clay
treated with 10% cement available from the literature. The validation suggested that the
proposed model can be used to predict the behaviour of cement treated clay very well.

INTRODUCTION
Stabilising soft soil with cement has become an effective ground improvement
technique to improve the properties of the soft soil, such as improvements in strength
and compressibility, while the ductility of the cement treated clay can be improved by
the inclusion of other additives, e.g. fibre (Fatahi et al., 2013a; Fatahi et al., 2013b).
Researchers such as Chew et al. (2004) and Lorenzo et al. (2006) have conducted
numerous studies to investigate the effect of cementation on the behaviour of cement
treated clay. Particularly, the shear strength is improved significantly by mixing cement
with soft clay. However, as the mean effective yield stress increases beyond the initial
yield stress, the breakage of cementation bonds occur which reduces the effectiveness of
cementation, consequently affecting the shear strength of the cement treated clay

ASCE

Geo-China 2016 GSP 259

153

(Kamruzzaman et al., 2009). Several constitutive models were developed to simulate the
behaviour of cement treated clay, e.g. the Structured Cam Clay (SCC) model developed
by Horpibulsuk et al. (2010). However, the existing constitutive models assumed linear
failure envelope in plane while the peak shear strength of the cement treated clay
reduces due to the effect of cementation degradation, particularly at high effective
confining pressures (Panda et al., 1998; and Uddin et al., 1997). Hence in this paper,
based on the framework of Critical State Soil Mechanics, a constitutive model was
proposed, considering the effect of cementation degradation under various loading
conditions. Moreover, the mean effective stress was modified to consider the beneficial
effect of cementation and its degradation. The validation of the proposed model includes
comparing the model predictions with the results on Singapore marine clay treated with
10% cement available from the literature.
DEVELOPMENT OF THE PROPOSED MODEL
The process of cementation includes hydration process, ion exchange and pozzolanic
reactions (Porbaha et al., 2000). The primary chemical interaction (hydration and ion
exchange process) creates cementation bonding between clay particles and cement
which forms large clay clusters within the soil-cement matrix. Further bonding results
from a secondary pozzolanic reactions which increases the cohesion in the mixture,
consequently leading to a strongly bonded structure (Lorenzo et al., 2004). Since the
cementation process continues over a prolonged period, time dependent behaviour of
soil should be considered as suggested by Le et al. (2015). However, for simplicity, the
main focus of this study is to examine the behaviour of cement treated clay in undrained
condition which occurs relatively over a short time period. Laboratory results showed
that the initial yield stress of the cement treated clay increases significantly due to the
effect of cementation (Horpibulsuk et al., 2010). However, as the mean effective yield
stress increases beyond the initial yield stress during isotropic consolidation, breaking of
cementation bonds occurs causing plastic deformation. At high mean effective yield
stress, significant reduction in void ratio was observed as large clay clusters collapse and
the effect of cementation gradually diminishes (Kamruzzaman et al., 2009). Hence, the
effect of cementation and its degradation are mainly related to the mean effective stress
( ). Thus, in this proposed model, inspired by Nguyen et al. (2014), the mean effective
stress was modified to consider the beneficial contribution of cementation and the effect
of cementation degradation at sufficiently high mean effective stress as follows:

1+

(1)

where is the modified mean effective stress,


describes the effect of cementation.
is the contribution of cementation to the shear strength, is the slope of the Critical
State Line (CSL) of the soil-cement mixture, is a model fitting parameter affecting the
degradation rate. Increasing the beneficial contribution of cementation ( ) results in the
and the shear strength ( ) when = 0. Moreover,
is
increase of the parameter

ASCE

Geo-China 2016 GSP 259

154

the mean effective yield stress and , is the initial mean effective yield stress. It should
be noted that, was formulated in a way that when
, , the effect of
0 , hence .
cementation significantly reduces and gradually vanishes
Furthermore, when the cement treated clay is stressed up to the initial mean effective
stress
, , it is assumed that there is no effect of cementation degradation, so the
contribution of cementation is the most effective in this range, thus = + .
In this study, the proposed yield function follows the Modified Cam Clay (MCC)
model, which assumed that the cement treated clay is an isotropic material. So, when the
cement treated clay is within the yield surface, only elastic deformation occurs. Hence,
adopting Hookes Law and the modified mean effective stress (Eq.1) the elastic
deformation of the cement treated clay can be described by the following equations:

(2)

,
,

(
(

)
)(

(3)

,
,

and
are the elastic volumetric strain increment and elastic deviatoric strain
where
increment, respectively. Moreover, is the slope of the swelling line, is the void ratio
of the cement treated clay and is the Poissons ratio.
Furthermore, to follow the basis of MCC model, the yield surface in
be formulated in the following equation:
,

=
where
plane.

1+

plane can

=0

= +
, describes the size of the yield surface when
is the effective confining pressure prior to the shearing stage..

(4)
= 0 in

When the cement treated clay is on the yield surface, plastic deformation occurs.
Hence, it is important to describe the plastic flow rule for the appropriate direction of the
plastic strain vectors. In this study, the non-associated plastic flow rule was proposed to
consider the effect of cementation as follows:
=

ASCE

(5)

Geo-China 2016 GSP 259

155

where
and
are the plastic volumetric strain increment and the plastic deviatoric
strain increment, respectively. The parameter is used for balancing the deviatoric and
volumetric strain increments. Moreover, the term is derived from taking the derivative
of with respect to as follows:
,
,

=1

(6)

Moreover, the non-associated plastic potential function can be derived using the plastic
flow rule (Eq.5) in the following form:
=

(1 + 2 ) +

=0

(7)

Fig.1 displays the effect of increasing cementation ( ) on the proposed plastic


, the size of
potential surface in plane. At a same mean effective yield stress
the plastic potential surface expands with increase in . It should be noted that, when
= 0, the proposed plastic potential function is identical to the proposed yield surface
function, hence the proposed model converts to an associated model. Moreover, when
= 0 and the effect of cementation is absence, the proposed plastic potential function
returns to the yield surface of the MCC model, as shown in Fig. 1.
General Stress-Strain Relationship of the Cement Treated Clay
The following general stress-strain relationship equations are derived to describe the
total volumetric strain increment ( ) and the total deviatoric strain increment ( ),
respectively, as follows:
=

ASCE

+( )

+
(

(
(

)
)(

(8)

(9)

Geo-China 2016 GSP 259

156

C = 120 kPa
C = 80 kPa

200

C = 40 kPa

180

C = 0 kPa

160
Critical State Line
( = )
M = 0.9
= 55 kPa
= 80 kPa
= 200 kPa
= -0.6

Deviatoric stress (q) (kPa)

140
120
100
80
60
40

When C = 0 and = 0 (Plastic potential function


coincides with MCC yield surface)

20
0

-80

-30

20

70

120

170

220

Mean effective stress (p') (kPa)

FIG. 1. Proposed plastic potential surface with variations of cementation (C ).

VALIDATION OF THE PROPOSED MODEL


In this section, the validation of the proposed model includes comparing its
predictions against the reported triaxial test results on Singapore clay from
Kamruzzaman et al. (2009). Based on Nguyen et al. (2014), the model parameters were
estimated and the model values are shown in Table 1.
Table 1. Model parameters used in the validation of the proposed model

Singapore clay

0.73

0.067

2.85

0.1

(kPa)

0.9

158

(kPa)
2500

Fig. 2 shows the triaxial test results and the model predictions on 10% cement treated
Singapore clay at 300, 500 and 1000 kPa of effective confining pressure. Since all the
triaxial tests adopted the effective confining pressure was higher than the initial mean
effective yield stress , of Singapore clay ( , = 300 kPa), so the stress state of all
samples were on the current yield surface. When shearing stage commenced, the stress
paths rose upward and bent to the left in
plane as excess pore water pressure
developed within the sample, as shown in Fig. 2(a). Moreover, all samples reached the
peak shear strength above the CSL of the reconstituted soil-cement mixture, showing the

ASCE

Geo-China 2016 GSP 259

157

contribution of cementation to the shear strength. The predictions from the proposed
model were in good agreement with the experimental results in
plane adopting
the general stress-strain relationship proposed in Eqs. (8) and (9). Furthermore, the
as shown in Fig. 2(b) for the cement
strength increased with deviatoric strain
treated Singapore clay. The peak shear strength was observed to occur at
= 2.2% for
the treated Singapore clay. There are some disparities in predicting the stress-strain
relationship for the cement treated Singapore clay at small strain, particularly for sample
tested at effective confining pressure of 300 kPa. However, when the deviatoric strain
exceeds 0.5%, the proposed model simulated the strength development with increasing
strain very well. Overall, the behaviour of the cement treated clay was captured by
considering the effect of cementation and its degradation in the proposed model.

900

900

Proposed Model Prediction


Critical State Line
(CSL)
M = 0.9

700
600
500
400
300
200
100

Deviatoric stress (q) (kPa)

Deviatoric stress (q) (kPa)

800

0
0

200

= 1000 kPa

700
600
500
400

= 500 kPa

300
200

= 300 kPa

100

= 1000 kPa
= 300 kPa

Proposed Model Prediction

800

= 500 kPa

400

600

800

1000

Mean effective stress (p) (kPa)


(a)

1200

0
0

0.5

1.5

Deviatoric strain
(b)

(%)

FIG. 2. Behaviour of Singapore clay treated with 10% cement in: (a) stress path
( plane) and (b) stress strain relationship ( plane). The triaxial results
were obtained from Kamruzzaman et al. (2009).
CONCLUSIONS
In this paper, the proposed model has captured the important features of the behaviour
of cement treated clay, particularly the effect of cementation and its degradation to the
shear strength. Special characteristics of the proposed model include the modified mean
effective stress and non-associated plastic potential function, together with the general
stress-strain relationship. Moreover, the performance of the proposed model was
evaluated by predicting the reported results on treated Singapore clay very well. Overall,
this model can be adopted in geotechnical design for treating soft clay with cement.

ASCE

Geo-China 2016 GSP 259

REFERENCES
Chew, S. H., Kamruzzaman, A. H. and Lee, F. H. (2004). "Physicochemical and
engineering behavior of cement treated clays" Journal of geotechnical and
geoenvironmental engineering, Vol. 130 (7): 696-706.
Fatahi, B., Fatahi, B., Le, T. M. and Khabbaz, H. (2013a). "Small-strain properties of
soft clay treated with fibre and cement" Geosynthetics International, Vol. 20 (4): 286300.
Fatahi, B., Le, T., Fatahi, B. and Khabbaz, H. (2013b). "Shrinkage Properties of Soft
Clay Treated with Cement and Geofibers" Geotechnical and Geological Engineering:
an international journal, Vol. 31 (5): 1421-1435.
Horpibulsuk, S., Liu, M. D., Liyanapathirana, D. S. and Suebsuk, J. (2010). "Behaviour
of cemented clay simulated via the theoretical framework of the structured cam clay
model" Computers and Geotechnics, Vol. 37 (1): 1-9.
Kamruzzaman, A. H., Chew, S. H. and Lee, F. H. (2009). "Structuration and
Destructuration Behavior of Cement-Treated Singapore Marine Clay" Journal of
geotechnical and geoenvironmental engineering, Vol. 135 (4): 573-589.
Le, T. M., Fatahi, B. and Khabbaz, H. (2015). "Numerical optimisation to obtain elastic
viscoplastic model parameters for soft clay" International Journal of Plasticity, Vol.
65: 1-21.
Lorenzo, G. A. and Bergado, D. T. (2004). "Fundamental parameters of cement-admixed
clay-new approach" Journal of geotechnical and geoenvironmental engineering, Vol.
130 (10): 1042-1050.
Lorenzo, G. A. and Bergado, D. T. (2006). "Fundamental characteristics of cementadmixed clay in deep mixing" Journal of materials in civil engineering, Vol. 18 (2):
161-174.
Nguyen, L. D., Fatahi, B. and Khabbaz, H. (2014). "A constitutive model for cemented
clays capturing cementation degradation" International Journal of Plasticity, Vol. 56:
1-18.
Panda, A. P. and Rao, S. N. (1998). "Undrained strength characteristics of an artificially
cemented marine clay" Marine georesources & geotechnology, Vol. 16 (4): 335-353.
Porbaha, A., Shibuya, S. and Kishida, T. (2000). "State of the art in deep mixing
technology. Part III: Geomaterial characterization" Proceedings of the ICE-Ground
Improvement, Vol. 4 (3): 91-110.
Uddin, K., Balasubramaniam, A. S. and Bergado, D. T. (1997). "Engineering behavior
of cement-treated Bangkok soft clay" Geotechnical Engineering, Vol. 28 (1): 89-119.

ASCE

158

Geo-China 2016 GSP 259

159

Chemo-Mechanical Approach to Modelling the Expansive Behavior of Sulfate Bearing


Soils: The Role of Crystallization Pressure in Ettringite Formation
Pawan Sigdel1 and Liang Bo Hu2
1

Graduate Research Assistant, Dept. of Civil Engineering, Univ. of Toledo, Toledo, OH


43606. E-mail: Pawan.Sigdel@rockets.utoledo.edu
2
Assistant Professor, Dept. of Civil Engineering, Univ. of Toledo, Toledo, OH 43606.
E-mail: Liangbo.Hu@utoledo.edu
Abstract: Significant amount of heaving in cement stabilized sulfate bearing soils had
been observed in recent construction projects in Northeast Ohio, USA. It is well known
that the reaction between sulfate and alumina in soils and calcium of a lime or cement
stabilizer can cause the formation of ettringite, an expansive sulfate mineral which may
contribute to heaving. The crystallization pressure during the ettringite formation plays a
significant role in the soil expansion. This paper presents a numerical approach to
modelling the potential heaving in stabilized sulfate-bearing soils with a specific focus on
the role of the crystallization pressure in ettringite formation. Evolution of kinetic rates of
involved chemical reactions is simulated for the time-dependent process of crystal growth
and crystallization pressure development addressed by Correns equation. A
micro-mechanical formulation is used to simulate the induced expansive strain. The
presented simulations are able to reveal the characteristics of the evolution of sulfate
concentration, ettringite crystal growth, crystallization pressure as well as expansive
strain. A parametric study of the effect of temperature is also presented.
INTRODUCTION
Use of lime or cement stabilizer in subgrade is widespread in transportation and
construction engineering and provides a potential benefit of the reduction of swelling
potential by substituting clay cations through the formation of calcium silicate and
aluminate hydrates (Abdi and Wild 1993). However, in natural sulfate-bearing expansive
soils, lime or cement treatment may lead to a new swelling problem as sulfates may react
with lime or cement, leading to the formation of ettringite, an expansive sulfate mineral
that contributes to heaving by hydration and/or by continuous growth of itself (Mitchell
1986; Hunter 1988; Puppala et al. 2005).

ASCE

Geo-China 2016 GSP 259

A number of construction projects in the state of Ohio, USA had experienced significant
sulfate induced heaving in cement stabilized soils (Narsavage 2011; Cutright and Wigdahl
2013). Approximately 0.5~4 inches of heaving was observed at the longitudinal joint
between the center median barrier wall and shoulder pavement at numerous locations. The
pavement consisted of 11.5 inches of asphalt concrete overlaid on 6 inches of aggregate
base on top of 12 inches of cement-stabilized subgrade. An experimental study of soils
from these construction projects in Ohio revealed the presence of gypsum and other
sulfate bearing minerals but did not show adequate correlation between sulfate content
and swelling (Weir et al. 2014).
Attributing soil heave entirely to ettringite formation is likely a simplistic
misconception of a very complex process that may involve multiple mechanisms. Early
efforts by Abdi and Wild (1993) and Wild et al. (1993) suggested that the primary
mechanism in sulfate-containing lime stabilized clays might be imbibition of water by
osmosis, rather than direct formation of solid reaction products. Nair and Little (2011)
showed that swelling was most likely caused by a synergy of multiple mechanisms,
including formation of ettringite, water absorption of formed ettringite and osmosis
induced swelling of clay minerals. Numerical investigations of the interplay of multiple
mechanisms were previously examined in a chemo-mechanical constitutive framework by
Sigdel and Hu (2015) that shows that ettringite formation and osmosis swelling may play
comparable roles in the heaving process. On the other hand, understanding of the inner
mechanisms of the ettringite formation and the induced volume expansion remains much
needed in order to better understand the precise path of the swelling strain development
and identify specific dominant mechanism under particular environmental conditions.
In the present paper the crystallization pressure in ettringite formation and its role in the
soil swelling are investigated. Key concepts and mechanisms involved are discussed with
an aim to establish mathematical formulations for numerical modelling in a
chemo-mechanical framework to address the characteristics and role of crystallization
pressure. The kinetics of the involved chemical reactions are examined and subsequently
simulated for the time-dependent process of crystal growth and crystallization pressure
development. Numerical simulations of swelling strain are based on a previously
established micro-mechanical formulation (Bary 2008).
PHONEMONOLOGICAL BACKGROUND AND KEY CONCEPTS
Although the formation of ettringite is generally believed to contribute to the expansion of
sulfate bearing soils, neither the mechanism of the volume expansion during its formation
process nor its role in soil expansion is entirely clear.
One particular intriguing implication is related to the volume change in the possible
chemical reaction leading to the formation of ettringite, which is a hydrous calcium
aluminum sulfate mineral (Ca6[Al(OH)6]2(SO4)326 H2O). Comparing the original
volume of the involved reactants with that of the product (eq. (1)), Little et al. (2010)
showed that, if the volume of each of the reactants is considered, including the water
supplied from within the matrix, the volume change is actually negative, i.e.,
transformation of all reactants into ettringite causes approximately 5% shrinkage.

ASCE

160

Geo-China 2016 GSP 259

161

3CaOAl2O3+ 3[CaSO4 2H2O] + 26 H2O Ca6[Al(OH)6]2 (SO4)3 26 H2O

(1)

However, when water consumed in ettringite formation is considered to be from outside


the matrix, the molar volume calculations indicate a volume increase of 137% expansion.
This leads to an open versus closed system argument regarding the possible scenario of
ettringite formation; the source, timing and availability of the water consumed for
ettringite formation has a profound effect in assessing the contribution of ettringite
formation to the soil heave (Little et al. 2010). Experimental studies indeed have shown
ettringite formation may be only one of the possible mechanisms responsible for
expansion in sulfate-bearing soils, in addition to other mechanisms which include osmosis
induced water absorption in particular (e.g., Abdi and Wild 1993; Wild et al. 1993; Nair
and Little 2011). Numerical investigations of the interplay of multiple mechanisms are
examined in a chemo-mechanical constitutive framework (Sigdel and Hu 2015) as well as
from a partially saturated soil modelling perspective (Hu and Sigdel 2015), these works
have shown that ettringite formation and osmosis swelling can play comparable roles in
the heaving process.
On the other hand, the intricacies of the volume expansion mechanisms during ettringite
formation deserve much attention. While the investigation on the volume change based on
involved reactants and products is useful, the formation of crystals likely generate the
pressure or stress that may alter the inner structure and interact with surrounding particles,
hence the resulting expansion or volume change may play a significant role. The
crystallization pressure from the ettringite growth is commonly believed as the major
cause of macroscopic swelling and its role in swelling sulfate bearing soils is the primary
focus of the presented numerical study. The crystallization pressure can be derived from
thermodynamic considerations and generally expressed by Correns equation (e.g.,
Correns 1949; Xie and Beaudoin 1992; Scherer 2004),
=

ln

Q reac
reac

(2)

The activity product reac is defined by reac =


, where and are the activity
and the stoichiometric coefficient of species i in solution. reac is the equilibrium
is the
constant, R and T are the gas constant and the temperature, respectively, and
molar volume of the crystal. It should be noted that the underlying mechanisms of
crystallization pressure are most likely of a multi-scale nature. The mechanical pressure
generated by the crystallization process is related to the response of the confining
surrounding media (pore wall) as well as the surface phenomena at the pore scale (e.g.,
Scherer 1999, 2004). In addition, the consequence of the crystallization pressure exerted
on the macroscopic scale also must be evaluated across the scales. This is best illustrated
in Bary (2008) that employed a micro-mechanical approach to derive for the relevant
coefficient(s) in a macroscopic constitutive relation,
=

(3)

and are the stress and strain tensor. is the stiffness tensor, is the second-order
identity tensor. is a macroscopic scalar interaction coefficient. Bary (2008) used the

ASCE

Geo-China 2016 GSP 259

162

well-known Mori-Tanaka (MT) scheme to obtain this coefficient, expressed as a function


of volume fraction and elastic constants of different material inclusions.
The process of ettringite growth and implications of the accompanying crystallization
pressure are very important for understanding the expansive behavior of the
sulfate-bearing soils. In the following the key concepts discussed above are used for a
numerical investigation.
NUMERICAL SIMULATIONS
Mathematical Formulations
We consider a typical sulfate-bearing soils that contain a significant amount of gypsum or
anhydrite which provides the needed sulfate. In the following only the most salient
components of the mathematical formulations are presented.
One of keys to effectively assessing the growth of swelling strain is an understanding of
the kinetic processes. However, these reaction processes are usually very complicated and
extremely difficult to formulate and calibrate properly. We largely adopt the formulation
used in Oldecop and Alonso (2012), which describes a first-order kinetic relationship for
the gypsum dissolution,
=

for

<1

(4)

(mol/s) is the rate of the growing gypsum crystals, negative indicating dissolution.
(mol/m2s) is the dissolution rate of gypsum.
(m2) is the surface area gypsum
crystals exposed to water action. The dissolution takes place if the concentration of
sulfate, (i.e., [SO ]), is lower than its saturated concentration
. For simplicity, in the
present study the surface area
is assumed as a constant. The kinetic rate for ettringite
precipitation is adopted also in a first-order kinetic law used in Bary (2008) but expressed
in a different equation similar to eq. (4),
=

for

>1

(5)

The parameters including the rate constant, surface area and saturated concentration, are
defined in the same manner as those for gypsum dissolution. Obviously, precipitation of
gypsum and dissolution of ettringite can be also formulated similarly with their respective
rate constants. Simple treatment of prescribing a zero rate constant for these two reactions
is applied, for the sake of simplicity, in the following discussion on which a simple
scenario of gypsum dissolution and ettringite formation (precipitation) is focused, but it is
clear that inclusion of the other two reactions can be readily achieved by prescribing
non-zero rate constants.
The crystallization pressure based on eq. (2) is applied to the chemical reaction of
ettringite formation and we adopted the formulated suggested by Bary (2008),
=

ln

[Ca] [SO ]
[Ca] [SO ]

(6)

where [Ca] and [SO ] are the calcium and sulfate concentrations at chemical
equilibrium. A simplification is made by considering only the evolution of sulfate

ASCE

Geo-China 2016 GSP 259

conccentration, = [SO ] . As such num


merical simuulations of tthe sulfate eevolution and
the crystallizatio
on pressuree are done based on the presennted kineticcs and masss
conservations.
Th
he macroscop
pic swelling
g strain can be
b simulatedd using eq. ((3) which iss applied to a
stress-free swelliing case ( = ). Utilizzing the micrro-mechaniccally based dderivation byy
Bary
y (2008), wh
hen the vollume fractio
on of ettringgite is smaall, we can simplify thhe
form
mulation for the
t interactio
on coefficien
nt and replaace with thhe volume frraction of thhe
ettrin
ngite crystalss
.
Simu
ulation Resu
ults
In th
his section we present some of th
he simulatioons results regarding tthe ettringitte
form
mation and crrystallization
n pressure in the context of a free sw
well case. First, the detaills
of the results are shown in Fig
gs.1 and 2 att the temperaature of 20C
C. The simulation is madde
with adopting th
he followin
ng primary parameters:
p
the saturatted concentrration
=
3
3
2
15.5 mol/m ,
= 1.56 mol/m , th
he rate connstant
= 9.48 10
0 mol/m ss,
= 9.48 10
1
mol/m
m2s, molar volume
v
of eettringite
= 7.25 1
10 m3/moll,
and the initial volume
v
fracction of gyp
psum
= 0.1 . A finite differencce scheme iis
implemented in Matlab
M
for simulations.

ation resultts at T=20C, (a) volu


ume fractioon of formeed ettringitte
FIG.. 1: Simula
crysttals; (b) evo
olution of th
he sulfate co
oncentration
n.
Fig
g. 1a shows the
t evolution
n of ettringitte formationn and Fig. 1bb shows the ssulfate solutte
conccentration in
n compariso
on with thee prescribedd gypsum ssolubility annd ettringitte
solub
bility. It is ev
vident that after
a
approxiimately 40 ddays the conccentration beecomes morre
or less stabilized
d, it indicatess that the new
wly dissolveed sulfate is completely cconsumed byy
ettrin
ngite formation, the oveerall sulfate concentratioon reaches a certain threeshold whichh
actuaally can be evaluated bassed on the kiinetic rate law
ws. Consequuently ettringgite developps
almo
ost linearly after
a
40 days, because thee kinetic ratees now becoome constantts.
Th
he above obsservations also explain why the sw
welling strainn shown in Fig. 2b alsoo
evolv
ves almost at
a a constantt rate after 40
4 days. As shown in Fiig. 2a whichh presents thhe
evolu
ution of crystallization pressure,
p
because the suulfate concentration beccomes almosst

ASCE

163

Geo-China 2016 GSP 259

constant, afterwards the cry


ystallization pressure allso reaches almost a thhreshold, thhe
chan
nge in swell is
i due to the growth of th
he volume frraction of etttringite whicch progressees
lineaarly with tim
me as a conseequence of constant
c
kineetic rates. Thhe negative values in thhe
initiaal stage of crrystallization
n pressure ev
volution are ddue to the faact that, the cconcentrationn
of su
ulfate solute is initially set up very lo
ow in the sim
mulation andd it takes a dday or two foor
moree gypsum disssolution to assist in ettrringite growtth (see Fig. 11b).

FIG.. 2: Simulattion results at


a T=20C, (a) evolutioon of crystalllization preessure over a
perio
od of 100 da
ays. (b) swellling arising
g from crysttallization p
pressure.
It is
i worth notiing that the state when the
t sulfate ssolute concenntration reacches constannt
does not coincid
de with the so
o-called equ
uilibrium of dissolution aand precipitation of eachh
speciies, in which
h dissolution
n and precipiitation of thee same speciies occur at tthe same ratte
simu
ultaneously. In
I the simulaated scenario
o the dissoluution of gypssum and the precipitationn
of etttringite prog
gress simultaaneously at the same raate after a ceertain periodd of time. A
As
alreaady mention
ned in the preceding section, gyypsum preccipitation annd ettringitte
disso
olution can also
a be invesstigated conv
veniently byy assigning aappropriate rrate constantts
(e.g.,, Bary 2008;; Oldecop an
nd Alonso 20
012).

FIG.. 3: (a) Ev
volution off expansion
n as affecteed by temp
perature; ((b) effect oof
temp
perature in the short-teerm and lon
ng-term devvelopment off expansive strain.

ASCE

164

Geo-China 2016 GSP 259

A parametric study of the effect of temperature is presented in Fig. 3. Figs 3a and 3b


suggest that the variations as development of expansive strain at different temperatures
become more appreciable at later stage. It should be noted that the temperature may also
affect the kinetic rates of the chemical reactions; the present simulation only considers its
effect on the crystallization pressure. A more comprehensive parametric study would
likely reveal many more intricate observations and findings that elucidate the interplay
between different mechanisms. That should include the considerations of
afore-mentioned chemical reactions of gypsum precipitation and ettringite dissolution, as
well as the inter-dependence of various parameters such as non-constant solubility
variables, specific surface area that evolve dependent on crystal growth and strain
accumulation as examined in Sigdel and Hu 2015, among many others. The most
intriguing challenge in developing further understanding of the role of ettringite formation
in sulfate-bearing soils remains the quest for multi-scale modeling of crystallization
pressure within complex pore structures of clay clusters to identify the critical,
inter-related mechanisms across the pore-, meso- and eventually macro- scales.
CONCLUSIONS
The presented study focuses on the crystallization pressure of ettringite formation and its
role in the swelling of sulfate bearing stabilized soils, an important phenomenon as part of
potentially a combination of very complex processes involved. Mathematical
formulations are established following a discussion of the key physical, chemical and
mechanical theories and concepts. The present study is aimed to study the evolution of a
closed system, i.e., a representative elementary volume (REV), considering the relevant
kinetics formulated. A more precise simulation of field problems can be only attempted in
terms of carefully formulated boundary value problems (BVPs) by considering the mass
transport and linear momentum balance equations.
The modelling framework can be extended to accommodate the considerations of a
number of variable processes including more comprehensive series of chemical reactions
and chemo-mechanical couplings. Some of them can be readily implemented as
discussed. The chemo-mechanical approach adopted in the present study is restricted to
address primarily the role of ettringite formation but can also be extended to study other
potentially swelling-inducing mechanisms (e.g., Sigdel and Hu 2015; Hu and Sigdel
2015). The underlying mechanisms of crystallization pressure and its consequences
exhibit complex multi-scale characteristics. The presented numerical approach can
benefit from properly planned experimental studies that are capable of distinguishing the
effects of diverse mechanisms. Future experimental and numerical investigations must
strive to understand the critical mechanisms across the scales and formulate them
accordingly in order to develop appropriate modelling tools.

ASCE

165

Geo-China 2016 GSP 259

REFERENCES
Abdi, M. R. and Wild, S. (1993), Sulphate expansion of lime-stabilized kaolinite: I.
Physical characteristics. Clay Minerals, 28(4): 569-584.
Bary, D. (2008). Simplified coupled chemo-mechanical modeling of cement pastes
behavior subjected to combined leaching and external sulfate attack. Int. J. Num.
Anal. Meth. Geomech., 32(14): 1791-1816.
Correns, C.W. (1949). Growth and dissolution of crystals under linear pressure, Discuss.
Faraday Soc., 5: 267- 271.
Cutright, T. and Wigdahl, J. (2013), Assessment of sulfate bearing soils in Ohio. 2013
Ohio Transportation Engineering Conference, Columbus, OH.
Hu, L. B. and Sigdel, P. (2015). Exploring a chemo-mechanical Approach to understand
expansive behavior of sulfate bearing soils. ASCE Construction Institute Special
Publication on Innovative Materials and Design for Sustainable Transportation
Infrastructure, 189-199.
Hunter, D. (1988), Lime-induced heave in sulfate-bearing clay soils. J. Geotech. Eng.
114(2): 150-167.
Little, D. N., Nair, S. and Herbert, B. (2010), Addressing sulfate-induced heave in lime
treated soils. J. Geotech. Geoenviron. Eng., 136(1): 110-118.
Mitchell, J. K. (1986), Practical problems from surprising soil behavior. J. Geotech.
Eng., 112(3): 255289.
Nair, S. and Little, D. (2011), Mechanisms of Distress Associated with Sulfate-Induced
Heaving in Lime-Treated Soils. Transport. Res. Rec.: J. Transport. Res.
Board, 2212(1): 82-90.
Narsavage, P. A. (2011). Sulfate heaving of cement stabilized soil in Ohio. 2011 Ohio
Transportation Engineering Conference, Columbus, OH.
Oldecop, L. and Alonso, E. (2012), Modelling the degradation and swelling of clayey
rocks bearing calcium-sulphate. Int. J. Rock Mech. Min. Sci., 54: 90-102.
Puppala, A. J., Intharasombat, N. and Vempati, R. K. (2005), Experimental studies on
ettringite-induced heaving in soils. J. Geotech. Geoenviron. Eng., 131(3): 325-337.
Scherer, G. W. (1999). Crystallization in pores. Cem. Concr. Res., 29(8): 1347-1358.
Scherer, G. W. (2004). Stress from crystallization of salt. Cem. Concr. Res., 34(9):
1613-1624.
Sigdel, P. and Hu, L. B. (2015). Modelling ettringite induced heaving in cement
stabilized soils within a chemo-mechanical constitutive framework. ASCE
Geotechnical Special Publication 256: IFCEE 2015, 125-132.
Weir, M., Mandell, A. and Farver, J. (2014). Role of Sulfates on Highway Heave in Lake
County, Ohio. Report # FHWA/OH-2014/4, Ohio Department of Transportation.
Wild, S., Abdi, M. R. and Leng-Ward, G. (1993), Sulphate expansion of lime-stabilized
kaolinite: II. Reaction products and expansion, Clay Minerals, 28(4): 569-584.
Xie, P. and Beaudoin, J. J. (1992.) Mechanism of sulphate expansion. I. Thermodynamic
principle of crystallization pressure. Cem. Concr. Res., 22(4): 631-640.

ASCE

166

Geo-China 2016 GSP 259

167

Analysis of the Effect of Stiffening at the Top of a Single Granular Pile with a StressDependent Deformation Modulus
K. S. Grover1 and J. K. Sharma2
1

Professor, Civil Engineering, Rajasthan Technical Univ., Kota 324010, India. E-mail:
kdsgrover@gmail.com
2
Professor, Civil Engineering, Rajasthan Technical Univ., Kota 324010, India. E-mail:
jksrtu@gmail.com

Abstract: Stone columns/granular piles (GP) are composed of compacted gravel,


sands or mixture of both. The deformation modulus of granular material is determined
by the level of stress acting on it. The depth under loading and average principle stress
at each level will affect the deformation modulus of granular pile. This paper deals
with analysis of single granular pile with stiffened top portion with stress dependent
deformation modulus. Results of GP are presented in terms of normalized load
displacement curves and variations of normalized shear stresses with depth to study
effect of stress dependent deformation modulus.
INTRODUCTION
The response of a single floating granular pile with considerations of compatibility
of radial and vertical displacements along its interface was obtained by Sharma and
Madhav in 1999. The radial and vertical displacements of granular pile when its top is
stiffened is evaluated based on Mindlins equations for horizontal and vertical
displacements due to both horizontal and vertical point forces within the elastic
continuum.
According to (Balaam & Brooker 1981 and 1985) and (Ranjan & Rao 1983)
response of a GP based on elastic analysis was determined using stiffness ratio,
defined as M = Ec/Es, where Ec (or Egp) = deformation modulus of the GP and Es =
deformation modulus of the soil. Stiffness ratio is defined as, the ratio of axial stress
experienced by the column to the vertical stress component experienced by the soil for
the same settlement (Poorooshasb & Meyerhof 1997). It was suggested that the value
of deformation modulus Ec or Egp for a granular pile varies with stress level. The
deformation modulus of granular material depends on the average principal stress at
that particular level. The deformation modulus of granular material of GP varies along
its length due to the variations of axial and lateral stresses with depth.
In this paper, the effect of stress-dependent deformation modulus on the response of
single granular pile stiffened at top is studied. The upper stiffened part has higher
stiffness Egp as compared to lower part Egp. The length of upper stiffened portion of
GP is represented by non dimensional parameter . Hence, length of stiffened portion
is L/d.

ASCE

Geo-China 2016 GSP 259

PROBLEM DEFINITION AND ANALYSIS


Fig. 1 shows a granular pile of diameter, d and length, L. Initial in situ stresses
acting on GP are shown in Fig. 2 Effective vertical and horizontal in situ stresses can
be written (Refer Fig. 3) on element i as,
= s H i
ivs
(1)
= K ivs

irs

ASCE

168

Geo-China 2016 GSP 259

169

Assume the ground is in submerged condition. Let s' = submerged unit weight of
GP and Hi = depth of the center of element i, from the ground level. K = coefficient of
lateral stress (It is between 1 and 2 as per Poorooshasb & Meyerhoff 1997), in case of
GP its value depends on the installation effects of GP and on the condition of
surrounding soft soil. In the present work, K is taken as 1.
Egps is the stress-independent deformation modulus of GP. Egps, is defined as
n
1 + 2K

E
= A '
,
(2)

gps
ivs 3
Where n= 0.4 to 1.0 and A is constant of proportionality.
In this work average value of n is taken as 0.5 (Lambe & Whitman 1969). Fig. 4
shows the granular pile with load P on its top. Due to the application of load, the
horizontal and axial stresses on the elements increase. The current vertical, ivt and
horizontal stresses, irt on element i, are (Fig. 5)
= ivs
+ iv
ivt
(3)
+ ir
irt = irs
where,
iv = axial stress increment on element i due to load P
ir = radial stress increment on element i due to load P
As the deformation modulus of GP at any depth depends on its axial and horizontal
stresses at that level, for an element i, the deformation modulus, Eigp, is expressed as
0.5
ivt + 2 irt

E
= A
(4)
igp

From Eqs 2 & 4 and rewriting the following may be obtained,


0.5
+ 2
K
igp ivt
irt
=
(5)

'
K
gps ivs(1+ 2K)
Where

ASCE

Geo-China 2016 GSP 259

170

Kigp = Eigp/Es =relative GP-soil stiffness of element i, with load P


Kgps = Egps/Es =relative GP-soil stiffness of element i, without load P
Further defining the following non-dimensional parameters
Normalized Load on GP, P * =

, Normalized settlement of GP, =

S
d

Es d 2
To calculate the relative GP-soil stiffness of all the elements consistent with their
axial and horizontal stresses, the iterative analysis is carried out until the required
convergence is obtained between the new and the old values of element stiffness, Kigp.
Results of GP are presented in terms of normalized loaddisplacement curves and
variations of normalized shear stresses with depth to study effect of stress dependent
deformation modulus.
RESULTS AND DISCUSSION
Following are the ranges of parameters considered: Kgp = 10 - 500, L/d = 10 and 20,
s = 0.5 and gp = 0.3, *s =0.1, = 0.00001, coefficient of lateral stress K=1, is used
to define Ls i.e. stiffened length of GP where, Ls = x L/d. has been varied from 0.1
to 0.4. Top portion of length Ls of GP is made more stiffened by assuming different
Kgp in the top portion of GP. Kgp of the stiffened portion is Kgp and by taking =1 to
10 Kgp is varied up to 10 times of the lower portion.
An attempt has been made to obtain the variation of normalized load with depth for
different Kgp, and , to determine the variation of pile stiffness with depths by
changing Kgp, and , and to find out the distribution of the shear stresses with depth
along GP-soil interface with effect of Kgp, and , Grover and Sharma (2015).
The results of the present analysis are verified with the available results of the
previous researchers, Sharma and Madhav (1999). The values obtained are given in
the Table 1 for comparison. The results presented in Table 1 indicate that values are in
good agreement.

Table 1 Comparison of results of present work with previous

ASCE

Results of Sharma
and Madhav (1999)

Present
Analysis

Variable considered

L/d Kgp

Granular Pile Stiffness at 0.05


depth

10

50

320

320.006

Normalized Load for =0.25


normalized displacement

10

50

1.5

1.4987

Total Normalized Radial Stress


at depth 0.95

20

50

0.08

0.081

Normalized Shear Stress at


depth 0.025

20

100

1.6

1.6045

Normalized Shear Stress at


depth 0.05

10

50

1.44

1.445

Normalized Radial Stress at


depth 0.025

20

50

0.390

0.3896

Geo-China 2016 GSP 259

171

It may be observed that there is change in the values of normalized load with
normalized displacement in GP with stiffened top as shown in Fig. 6. For a constant ,
the values increase with displacement linearly. For a constant normalized
displacement of 0.01, the values of normalized load are 0.0485, 0.0578 and 0.0620 for
=2, =5 and =10 respectively. Thus, increase in values of normalized load with
increase in the value of is observed.
Fig.7 shows that there is a small change in the values of normalized load with
normalized displacement in GP with stiffened top for different value of Kgp. For a
constant Kgp, the values of normalized load increase with displacement. At a constant
normalized displacement of 0.01, the values of normalized load obtained are 0.0328,
0.0399, 0.0509, 0.0581, 0.0633 and 0.0671 at Kgp values of 10, 20, 50, 100, 200 and
500 respectively i.e. at higher value of stiffness of GP, higher normalized load is
obtained.
Fig.8 shows the results for load displacement relation for Kgp=20, L/d=20 and =2
and different values when stress dependent modulus is considered in GP. With
increase in values of , the load values are increasing. The normalized load values for
=0.4 are the maximum at a given normalized displacement.
Fig.9 shows the normalized shear stress with normalized depth for L/d=20 and
Kgp=50 with 20% stiffened length. Comparing the values at a constant depth indicate
that the values of normalized shear stress are higher for un-stiffened GP at very less
normalized depth of 0.175. After that normalized shear stress values are more for stiffened
GP and there is increase in normalized shear stress values with increase in values of .

Fig.6 Values of normalized load-displacement

ASCE

Geo-China 2016 GSP 259

172

Fig.7 Values of normalized load-displacement

Fig.8 Values of normalized load-displacement

Fig.9 Values of normalized shear stress with depth

ASCE

Geo-China 2016 GSP 259

173

Fig. 10 Values of normalized shear stress-depth

Fig.11 Values of normalized shear stress-depth


Fig.10 shows the values of normalized shear stress with normalized depth for
L/d=20, =2 and =0.4. It may be observed that as the value of Kgp is more, there is
lesser normalized shear stress at constant depth up to 0.225. The value of normalized
shear stress for Kgp=20 supersedes to Kgp=10 from normalized depth of 0.275, Kgp=50
supersedes to Kgp=20 from normalized depth of 0.425 and Kgp=500 supersedes to
Kgp=200 from normalized depth of 0.575. Thus, with increase in Kgp, the values of
shear stress are less.
Fig.11 shows the values of normalized shear stress with normalized depth for
Kgp=100, L/d=20 and =2 for different values of . It may be observed that initially
there is lesser normalized shear stress in stiffened GP as compared to un-stiffened GP.
The value of normalized shear stress is higher for =0.1 as compared to =0 from
normalized depth of 0.075, for =0.2 from normalized depth of 0.175 as compared to
=0.1 and for =0.4 from normalized depth of 0.375 as compared to =0.3.

ASCE

Geo-China 2016 GSP 259

CONCLUSIONS
From the above results it may be concluded that with the increase in values of and
normalized load has more values at the same normalized displacement in a stiffened
GP. With increase in values of Kgp, normalized load values increase at the same
displacement in a stiffened GP with stress dependent modulus. The increase in the
normalized load at constant normalized displacement may be attributed to increase in
stiffness of GP.
When GP with stiffened top is provided and stress dependent deformation modulus
is considered, then either with stiffer GP at top ( more) or with more length of stiff
GP at top ( more) or with the higher Kgp of GP, the shear stress reduces at top. Due to
increase in stiffness of GP, the settlement of GP decreases and hence, to satisfy the
compatibility condition, the shear stress in soil reduces near the top of GP. The
increase in the shear stress at greater depths is due to rearrangement of the shear stress
and values shift from the top to bottom side to maintain compatibility at GP-soil
interface.
REFERENCES
Balaam, N. P. and Booker, J. R. 1981. Analysis of rigid raft supported by granular
piles. Int. Jl. for Num. and Anal. Methods in Geomech., Vol. 5, pp. 379-403.
Balaam, N. P. and Booker, J. R. 1985. Effect of stone column yeilding on settlement
of rigid foundation in stabilized clay. Int. Jl. for Num. and Anal. Methods in
Geomech., Vol. 9, pp. 331-351.
Grover K.S, 2015, Analysis of Single Granular Pile with Stiffened Top, A thesis
submitted for the degree of Doctor of Philosophy to the Department of Civil
Engineering, Rajasthan Technical University, Kota, India.
Lambe, T. W. and Whitman, R. V. 1969. Soil mechanis, 1st Edition, John Wiley &
Sons, p. 553.
Poorooshasb, H. B. and Meyerhof, G. G. 1997. " Analysis of behaviour of stone
columns and lime columns." Computers and Geotechnics, Vol. 20, No. 1, pp. 47-70
Poulos, H.G. 1968. The influence of a rigid pile cap on the settlement behavior of an
axially loaded pile. C.E. Trans. Inst. of Engrs. Australia, CE10, No. 2, pp. 206-208.
Poulos, H.G. and Mattes, N. S. 1969. The behaviour of axially-loaded end-bearing
piles. Geotechnique, Vol. 19, pp. 285-300.
Poulos, H. G. and Mattes, N. S. 1971. Settlement and load distribution analysis of pile
groups. Aust. Geomech. J., Vol. G2, No. 1: pp. 11-20.
Ranjan, G., and Rao, B. G. 1983. Skirted granular piles for ground improvement. Proc.
VIII European Conf. on SMFE, Helsinki.
Sharma J.K., Madhav, M.R., Chandra S, 1997,Analysis of Load Test on Stone
Columns, IGC- 97, Vadodara, Theme V, pp. 5-22.1 to 5-22.4
Sharma J.K., Madhav, M.R., Chandra S.,and Miura, N., 1998, Prediction of
Settlement of Stone Column Reinforced Ground from Results of In Situ Load
Test, 13th SEAGC, Nov. 1998, Taipei, Taiwan, ROC, pp. 421-426.
Sharma J.K., Madhav, M.R., Chandra S., 1999, Consideration of Non-Homogeneity
of a Granular Pile on Settlement, Int. Conf. on Offshore Engg. (GEOSHORE), in
Dec. 99, ONGC Bombay
Sharma J.K., 1999, Analysis and settlement of granular pile (s)- single, in group and
with raft, A thesis submitted for the degree of Doctor of Philosophy to the
Department of Civil Engineering, Indian Institute of Technology, Kanpur, India.

ASCE

174

Geo-China 2016 GSP 259

175

Finite Element Analysis of Laterally Loaded Piles in Clays


Fath Elrahman E. Nur Eldayem1 and Yahia E.-A. Mohamedzein, M.ASCE2
1

Assistant Professor, Dean of Engineering, Al-Azhari Univ., Khartoum, Sudan. E-mail:


fathelrahman_2013@hotmail.com
2
Associate Professor, Dept. of Civil and Arch Engineering, Sultan Qaboos Univ., P. O. Box 33, Alkhod
123, Oman. E-mail: yahiaz@squ.edu.om

Abstract: This study presents a finite element model for analysis of laterally loaded
piles in medium stiff clay using ABAQUS program. Three dimensional (3-D)
continuum elements are used for both pile and soil elements. The interaction between
the pile and the soil is modelled by using contact surfaces. Frictional behavior between
the pile and the soil is considered in order to model the gap formation and sliding
behavior between the pile and the soil. The response of the pile is presented in terms
of the head deflection of the pile, bending moments, shear forces, deflections along the
length of the pile, the gap formation between the pile and the soil, the p-y curves and
the stress and displacement distribution within the pilesoil system. The study showed
that the ultimate lateral pile capacity increases with the increase in soil shear strength,
the L/d ratio and the pile diameter. For the cases considered, the maximum bending
moment occurs at a depth of 0.12L to 0.18L.
INTRODUCTION
The problem of pile-soil interaction involves the interaction of two different
materials under complex loading conditions. Theoretical methods and experiments are
used to study the behavior of the laterally-loaded piles. The theoretical methods
include the limit state equilibrium methods (Broms 1964), subgrade reaction methods
(Hsiung 2003), the p-y curves methods (Reese and Welch 1975; Wu et al. 1998), the
elasticity methods (Poulos and Davis 1980) and the numerical methods (Fan and Long
2005; Ahmadi and Ahmari 2009).
The p-y method is more popular in geotechnical practice because it is a semianalytical method and it is midway between the simplistic limit state methods and the
rigorous numerical methods. Different researchers proposed different p-y curves for
different clays. For example: Matlock (1970) for soft clays, Reese and Welch (1975)
for stiff clays above water table, Dunnavart and O'Neill (1989) for stiff clays below

ASCE

Geo-China 2016 GSP 259

water table. Some authors proposed p-y curves for all clays (e.g. O'Neill and Gazioglu,
1984; Wu et al., 1998). In most of these methods the p-y curve usually consists of two
parts: an initial nonlinear steep part and a flat linear part after reaching a certain level
of pressure, p. This pressure, p, is assumed to be a fraction of or equal to the ultimate
failure pressure, pu.
The analysis of pile-soil system by finite elements techniques was conducted by
many investigators (Fan and Long 2005; Ahmadi and Ahmari 2009). The finite
element method is a good method in representing the physical nature of the pile-soil
problem. In the finite element models reported in the literature, the soil was modeled
with 2-dimensional, 3-dimensional or axi-symmetric elements with different stressstrain relationships varying from simple linear elastic, elasto-plastic or perfectly
plastic material (e.g. Mohr-Coulomb plasticity, Von Mises, extended Druger-Prager
and Cap plasticity). The pile in most cases was modeled as a linear elastic beam
element.
The objective of this study is to investigate the behavior of laterally-loaded piles
installed in a plastic clay. The effects of different factors on the pile performance such
as pile dimensions and soil shear strength were studied. This study differs from the
studies reported in the literature in the following aspects: (1) The clay material is
different and it is a medium to highly plastic clay, and (2) The study covers a wide
range of L/d ratios ranging from piers to slender piles.
THE FINITE ELEMENT MODEL
The ABAQUS software was used for the finite element analysis in this study. The
finite element mesh used for the soil was very fine close to the pile and in the upper
region of the soil close to the soil surface (Figure 1). Fixed boundary conditions at the
bottom were generated and the soil sides were fixed horizontally and left free in the
vertical directions. The lateral load was applied to the pile at a point close to the soil
surface.
The finite element type used to model the pile elements was a 3- dimensional solid
element, denoted by C3D8. It is an 8-node linear brick element. The pile behavior
was modeled as linearly elastic defined by the material elastic modulus (E) and the
Poisson's ratio () equal to 205MN/m2 and 0.25, respectively.
In modelling the soil also an 8-node linear brick element (C3D8) was adopted. In this
study the Modified Drucker-Prager model was used to model the soil behavior.
The interface between the pile and the soil was based on surface-to-surface contact as
adopted in ABAQUS program. In this analysis the Coulomb friction model was used
to model the friction behavior between the soil and the pile as the two different bodies
are in contact at their extreme surfaces.
COMPARISON WITH LABORATORY MODEL TESTS
The authors performed laboratory model tests on steel and concrete model piles for
different L/d ratios (Mohamedzein et al. 2013). The results from the laboratory model
test can be used for comparison with the finite element model. The comparison was
made for a steel pile with the following properties: d=58 mm, thickness=1.5 mm,

ASCE

176

Geo-China 2016 GSP 259

embedded length (L)=580 mm and L/d ratio = 10. Figure 1 presents the finite element
mesh for the assembly of the pile and the soil (the mesh dimensions are those of the
tank used in the laboratory model test, e.g. 1250*1000*1400 mm). Figure 2 presents
the load-deflection curves for the laboratory test result and the FE model results. Both
curves give similar trends with some deviation at smaller and larger loads. The finite
element analysis gives weaker response at small loads and stiffer response at larger
loads.
RESULTS FROM THE FINITE ELEMENT ANALYSIS
The behavior of laterally loaded piles installed in clay was presented in terms of
load-deflection curves at the pile head, distribution of the bending moment and the p-y
curves. The effects of different factors such as pile diameter, pile length and soil
strength were discussed in this section. The finite element analysis was performed on
typical piles that are used in practice. Steel piles of diameters 500 and 250 mm and
embedded lengths (L) of 5m and 8.75m (L/d ratios 10 and 35) respectively, embedded
in medium stiff clay were considered. The soil depth was 10.0m and its properties are
shown in Table 1. The pile was considered as linearly elastic of modulus of
elasticity=205MN/m2 and Poissons ratio=0.25.
Load-Deflection Curves
The response of the pile can be presented in terms of the applied load and the
deflection at the pile head as shown in Figures 3 and 4. Figure 3 shows the loaddeflection curves for steel piles with diameter of 0.5 m and installed in clays of
different shear strengths. Figure 4 shows the load-deflection for steel piles with
length/diameter (L/d) ratios of 10 and 35. The load-deflection curves are generally
nonlinear. The figures show a ductile behavior with no well-defined peak. Thus, for
these piles the ultimate load can be taken as the load at a deflection of 10% times the
diameter (i.e. 0.1d) as suggested by different investigators (e.g. Briaud, 1997).
The increase in the soil elastic modulus and the shear strength increase the ultimate
lateral pile resistance of the pile (Figure 3). Also, For a given pile diameter, the
ultimate lateral resistance of the pile increases with the increase in L/d ratio (Figure 4).
Distribution of the Bending Moment
Figure 5 shows the distribution of the bending moment with depth for different
applied loads. The maximum bending moment occurs at a depth varying from 0.12L to
0.18L (L is the embedded length of the pile). The depth of the point of the maximum

ASCE

177

Geo-China 2016 GSP 259

178

1400 mm

mom
ment increasses with thee increase in
n the appliedd lateral loaad. Also, thee maximum
m
mom
ment increasses with the increase
i
in L/d
L ratio.

1250 mm

1000 mm

FIG. 1. The finite element meesh for laborratory model test for a pile
mbedded leng
gth (L) =580
0 mm and d
diameter 588mm (L/d=110)
of em

f
FIG
G. 2. Compa
arison betw
ween the reesults of fin
nite elemen
nt
and labo
oratory moodels

ASCE

Geo-China 2016 GSP 259

179

Table 1. Soil parameters used in the present FE model analysis


Clay (cap plasticity model)
Elastic modulus (kPa)
Poisson's ratio
Cohesion (kpa)

100

Angle of friction (degrees)

21

Cap eccentricity

0.33

Initial yield surface

Transition surface radius

0.02

Flow stress ratio

300

E=10000,Cu
=150

d=500 mm
L/d=10

250

Lateral Lo ad (kN)

5000-10000
0.3

E=5000,
Cu=100

200

150

100

50

0
0

10

20

30

40

50

60

Deflection (mm)

FIG. 3. Load-deflections curves for piles in clay with different


soil strengths (E and Cu in kPa)

ASCE

Geo-China 2016 GSP 259

180

250

L/d=35
d=250mm

Lateral load (kN)

200
L/d=10
d=500mm

150
100
50
0
0

10

20

30

40

50

Deflection (mm)
FIG. 4. Load-deflection curves for piles of different L/d ratios

Moment (kN.m)
-10

10

20

30

40

50

0
1
2

Depth (m)

3
4
17.5

35
52.5

70
7
8
9

d=250mm
L/d=35
Cu=100
kP

87.5
105
122.5
140

FIG. 5. The distribution of the bending moment with depth

ASCE

Geo-China 2016 GSP 259

181

Lateral Load (kN)

d=500m

d=2250m

Deflecction (mm)
FIG. 6. The p-yy curves
Thee p-y Curves
Figure
F
6 show
ws the p-y cu
urves for steeel pile of di fferent diam
meters. The ffigure showss
thatt p-y curves are nonlineaar and follow
w similar trennd to those rreported in the literaturee
(i.e.. the p-y currve usually consists of two
t
parts: aan initial nonnlinear steepp part and a
flat linear part after
a
reachin
ng the ultimaate failure p ressure, pu).. The figuree shows thatt
the soil resistan
nt significanttly increasess with the inccrease in pille diameter ((or decreasee
L ratio for a given leng
gth).
in L/d
ONCLUSION
NS
CO
Finite
F
element analyses were perfo
ormed to sttudy the perrformance oof laterally-load
ded pile in a plastic clay. The pile
p
behavioor includes ultimate lateral load,,
max
ximum bend
ding momentt and the p-yy curves. Thhe effect of ddifferent facctors such ass
L/d ratio, pile diameter and
a
soil sheear strengthh were conssidered. Thee followingg
con
nclusions werre drawn fro
om the modeel test resultss:
1. The lo
oad-deflectio
on curves arre generallyy nonlinear specially foor large L/dd
ratios and
a large pille diameters.

ASCE

Geo-China 2016 GSP 259

2. The ultimate lateral pile capacity increases with the increase in L/d ratio and
the increase in the pile diameter. Also the maximum moment increases with
the increase in L/d ratio and the increase in the pile diameter.
3. The maximum bending moment occurs at a depth varying from 0.12L to
0.18L.
4. The p-y curves are generally nonlinear and the soil resistance increases with
the pile diameter.
REFERENCES
ABAQUS (2008). ABAQUS Standards User's Manual, Version 6.1, Hibbitt,
Karlsson and Sorensen, Inc.
Ahmadi, M. M., and Ahmari, S. (2009). Finite-element modeling of laterally loaded
piles in clay. Proceedings of the Institution of Civil Engineering, Geotechnical
Engineering, 162(GE3), 151-163.
Briaud, J-L. (1997). Simple approach for lateral loads on piles. Journal of
Geotechnical and Geoenvironmental Engineering, ASCE, 123(10), 958-964.
Broms, B.B.(1964). Lateral resistance of piles in cohesive soils. Journal of the Soil
Mechanics and Foundation Division, ASCE, 90(SM2), 27-63.
Dunnavant,T. W., and O'Neill, M. W. (1989). Experimental p-y model for submerged,
stiff clay. Journal of Geotechnical Engineering, ASCE, 115(1), 95-114.
Fan, C-C., and Long, J. H. (2005). Assessment of existing methods for predicting soil
response of laterally loaded piles in sand. Journal of Computers and Geotechnics,
Elsevier, 32, 274-289.
Fatahi, B., Basack, S., Ryan, P., Wan-Huan Zhou, W-H., and Khabbaz H. (2014).
'Performance of laterally loaded piles considering soil and interface parameters'.
Geomechanics and Engineering, vol. 7, no. 5, pp. 495- 524.
Hsiung, Y-M. (2003). Theoretical elastic-plastic solution for laterally loaded piles.
Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 129(6), 475480.
Matlock, H. (1970). Correlations for design of laterally loaded piles in soft clay.
Proceedings, 2nd Annual Offshore Technology Conference, Houston, Texas, USA,
577-594.
Mohamedzein, Yahia E-A., Nour Eldaim, Fath Alrahman E. and Abdelwahab, Abu
Bakr (2013). Laboratory model tests on laterally-loaded piles in plastic clay.
International Journal of Geotechnical Engineering, Vol. 7, N0. 3, pp. 241-250..
ONeill, M. W., and Gazioglu, S. M. (1984). An evaluation of p-y relationships in
clays. American Petroleum Institute Report PRAC 82-41-2, 193 p.
Poulos, H. G., and Davis, E. H. (1980). Pile Foundation Analysis and Design. John
Wiley and Sons, Inc.
Reese, L. C., and R. C. Welch. (1975). Lateral loading of deep foundations in stiff
clay. Journal of Geotechnical Engineering, ASCE, 101(GT7), 633649.
Wu, D., Broms, B. B., and Choa, V. (1998). Design of laterally loaded piles in
cohesive soils using p-y curves. Journal of Soils and Foundations, Japanese
Geotechnical Society, 38(2), 17-26.
ASCE

182

Geo-China 2016 GSP 259

183

Analysis of Ring Foundation Systems Resting on an Anisotropic Elastic Soil Medium


Subjected to Working Compressive and Tensile Loads
Avishek Nath1; V Srinivasan2; and Priyanka Ghosh3
1

Sophomore Undergraduate, Dept. of Civil Engineering, IIT Kanpur, Kanpur, India.


E-mail: avishekn@iitk.ac.in
2
Research Scholar, Dept. of Civil Engineering, IIT Kanpur, Kanpur, India. E-mail: vsrini@iitk.ac.in
3
Associate Professor, Dept. of Civil Engineering, IIT Kanpur, Kanpur, India. E-mail: priyog@iitk.ac.in

Abstract: In this paper the settlement behavior and load bearing capacity of shallow
ring and circular foundations resting on anisotropic soil medium are studied. The case
of an axisymmetric body subjected to symmetric tensile and compressive forces is
analyzed here. In order to avoid rigorous calculation and make the problem lucid, the
soil has been assumed to be linearly elastic and only the half domain has been
considered for the analysis. The soil domain has been discretized with small grids for
adopting finite difference technique to calculate the relevant physical quantities in
each grid. Various plots have been included for better understanding of the behavior of
ring foundations and to illustrate the effect of anisotropy.
INTRODUCTION
In the last few decades, urbanization has affected the earth and the soil beneath it, in
lashing ways. With rampant urbanization, more number of buildings gets constructed
to accommodate the incoming flux of people rushing to the cities. The marvelous
structures built nowadays do reflect the tremendous achievement of civil engineering.
A very important part in any civil engineering structure is the design of its foundation.
Two important factors need to be ensured during foundation design i.e., the bearing
capacity and the settlement behavior. Erroneous estimation of bearing capacity may
lead to catastrophic events whereas in the case of settlement, it must be within the
permissible range. Thus both these factors go hand in hand while designing a
foundation. Among different types of foundation, ring foundation is one of the crucial
sub-structures, which is widely used in silos, communication towers, chimneys, water
tanks etc. Ring foundations were first studied by Fisher (1957). Later on, other
researchers have also performed elaborate studies on this topic [Egorov (1965),
Madhav (1980), Ohri et al. (1997), Hataf and Razavi (2003), Boushehrian and Hataf
(2003), Kumar and Ghosh (2005), Ahmad et al. (2014)]. This special type of
foundation may be subjected to huge compressive loads due to its self-weight along

ASCE

Geo-China 2016 GSP 259

with
h different live
l
loads su
uch as weigh
ht of water stored in thhe tanks, winnd load etc.
Rin
ng foundation
n systems peerform betteer due to the confinemennt effect of ssoil medium
m
encllosed within
n the annularr portion of the footing enabling thee soil to beaar additionall
load
ds [Rathour (2009)]. In
n the presen
nt work, ringg foundation system em
mbedded inn
anissotropic soil medium subjected to working
w
com
mpressive andd tensile loaads has beenn
anallyzed and th
he results haave been rep
ported by varrying imporrtant parameeters such ass
ri/ro (where ri an
nd ro are the inner and outer radius oof the annulaar footing reespectively),,
anissotropy parameter (), an
nd embedmeent depth of foundation ((De).
In this study,
s
the so
oil medium is assumed to be lineaarly elastic. M
Most of thee
prev
vious studies have assum
med the sim
mple model oof isotropic soil medium
m. However,,
the soil owing to
t its formattion and variious other cooupled phennomena usuaally behavess
as an
a anisotropiic medium (Graham
(
and
d Houlsby, 11943). Moreeover, in abssence of anyy
horiizontal exterrnal force, the
t soil parrticles do haave a tendenncy to movve verticallyy
dow
wnwards ow
wing to grav
vity. Therefo
fore, it is p lausible to assume its anisotropicc
beh
havior. In thee analysis, an
a effort has been made to bridge thhe gap of knnowledge inn
this particular to
opic of foundation behav
vior in anisootropic soil.
OBLEM DE
EFINITION
N
PRO

FIG
G.1 Represeentation off soil doma
ain (surfacee footing on left and embedded
d
footting on righ
ht)
Thee current pro
oblem domaain befits to be represennted throughh cylindricall coordinatee
systtem (r--z), however
h
duee to the sym
mmetrical loaading of the domain it reesults in twoo
dim
mensional po
olar coordinaate scheme where
w
tennds to infinitty. The soill domain onn
whiich the foun
ndation is reesting is illu
ustrated in F
Fig. 1. For ssimplicity, hhalf domainn
anallysis is beeen considereed here. KL
LMN repressents the axxi-symmetric and two-dim
mensional dom
main under consideratio
on. The surfaace footing iis resting onn the groundd
leveel to carry th
he compresssive load, whereas
w
to taake care of tthe uplift thhe footing iss
emb
bedded at a depth
d
De as shown in Fig. 1. The surrface footingg is compresssed under a
statiic working load intensity q, whereas the embeedded footinng is subjectted to samee
mag
gnitude (q) of
o pullout loaad intensity. The load inntensity is appplied betweeen the innerr
(ri) and outer (r
( o) radius in
n case of riing foundatiion, while ri tends to zzero for thee
circcular foundation. The ho
orizontal disstance (W) ffrom the edgge of the footing to thee

ASCE

184

Geo-China 2016 GSP 259

185

extreme side boundary (LM) and the domain depth (H) are taken as 10D and 14D
respectively, where D is the outer diameter of the footing. It is worth mentioning here
that the domain size has been fixed based on the sensitivity analysis performed by
Ahmad et al. (2014). The soil has been idealized as linearly elastic, homogenous and
anisotropic. For finite difference formulation the domain has been discretized into
small grids and displacement of each grid point is obtained by running an iterative
program in MATLAB. The width (dr) and depth (dz) of the grids are kept as same as
0.05D as reported by Ghosh and Sharma (2010).
MATHEMATICAL FORMULATION
The theory of elasticity has been employed similar to the methodology considered
by many researchers [Ghosh and Sharma (2010), Nainegali and Basudhar (2011),
Ahmad et al. (2014)]. The stress-strain and strain-displacement relations for an
anisotropic medium in polar coordinate system could be employed in the equilibrium
equations to formulate the governing partial differential equations for determining the
horizontal and the vertical displacements in the soil domain. However, the partial
differential equations thus obtained are quite difficult to solve analytically. In similar
cases, several researchers have resorted to the finite difference technique for obtaining
a solution [Maheshwari and Viladkar (2007), Ghosh and Sharma (2010)] and hence,
the same finite difference scheme is adopted here to solve the governing differential
equations for the displacements.
Equilibrium Conditions
The equilibrium equations in presence of symmetric loading conditions which
results in symmetry along axis can be expressed as,
r rz r
rz
z rz
+
+
=0;
(1a and b)
+
+
= 0
r
r
z
r
r
z
where r, , and z are the normal stresses along r, and z directions respectively
and rz is the shear stress along r direction on the z plane.
Generalized Hookes law for anisotropic soil medium
The stress-strain relation in an anisotropic medium can be obtained using Hookes
law and the resulting equations after expressing the strains in terms of displacements
in the polar coordinate system can be expressed as,
r =
=
z =

2
* u
2 *u
* w
(1 ) r + r + z

(1 + )(1 2 )
*

(2a)

2 * u
2
* u
* w
r + (1 ) r + z

(1 + )(1 2 )
*

(2b)

* u
u w
*u
* w
r + r + (1 ) z ; rz = G z + r (2c and 2d)

(1 + )(1 2 )
*

where E*=Ev ; Eh= 2 Ev ; *=hh ; vh= * ; Gvh=G*=

E *
2(1 + * )

. is an anisotropic

parameter. Ev and Eh are the vertical and horizontal elastic modulus of soil,
respectively. The modified Poissons ratio is denoted by *. Gvh, hh and vh are the
ASCE

Geo-China 2016 GSP 259

186

shear modulus and Poissons ratio of the transversely isotropic (i.e., anisotropic) soil
medium, respectively. The displacement components along r and z directions are
denoted as u and w, respectively.
Governing Partial Differential Equations
By substituting the expressions for the normal and shear stress from Eq. 2 in Eq. 1,
the partial difference equations can be obtained to determine the horizontal and
vertical displacements anywhere in the soil domain and are expressed as
2u 1 u 1 2 u u
G * u u 1 w
G*
+
+
+
+ +
= 0;
2 r r r 2 z r 2 1 2 * r r r z

*
2 w 1 w 1 2 w
u u 1 w
+ G
G*
+
+
+ +
= 0;
2 r r r 2 z 1 2 * z r r z

(3a)
(3b)

Finite Difference Formulation


To obtain the horizontal (Uj,i) and vertical (Wj,i) displacements, the finite difference
formulation of Eqn.3a and 3b can be derived as
1
1 2 *
(1 * )
Uj,i=
(Uj,i-1 + Uj,i+1) +
(Uj-1,i + Uj+1,i) +
(Wj+1,i+1 Wj-1,i+1 Wj+1,i-1
c
2c
8c
+ Wj-1,i-1) +
Wj,i=

(1 * ) dr
2c

(Uj,i-1 Uj,i+1)

(4a)

1 *
(1 2 * )
(Wj-1,i + Wj+1,i) +
(Wj,i-1 + Wj,i+1) +
k
2k

+ Uj-1,i-1) +

dr
4k r

(Uj-1,i - Uj+1,i) +

1 2 * dr
4k

8k

(Uj+1,i+1 Uj-1,i+1 Uj+1,i-1

(Wj,i-1 - Wj,i+1)

(4b)

)(

Where, c and k are parameters given by c = 1 2 * + 1 * 2 + dr ;

k = 2 1 * + 1 2 *
Here j and i represent the row and column in the finite difference grid, respectively.

) (

Boundary Conditions
The boundary conditions in accordance with stress-strain compatibility along the soil
domain can be listed as follows: Along the plane of symmetry (PoS) KN and the right
extreme boundary (LM), the horizontal displacement (Uj,i) is kept restricted. In case of
surface footing other than the loading zone along the ground surface, the normal and
shear stress are zero (z=0; rz=0) due to the absence of any external force. Along the
base of the soil domain, it is plausible to assume zero horizontal (Uj,i) and vertical
(Wj,i) displacements. Along the footing-soil interface, the horizontal displacement is
made zero to simulate perfectly rough condition. Loading zone beneath surface and
embedded footing, along EF: z=q; and za- zb=-q respectively, where q is the
magnitude of load intensity acting on the footing, superscripts a and b denote the
stresses above and below the embedded footing level respectively. The loading
condition for the embedded footing is similar to that reported by Kumar and Naskar
(2012). These constraints yield the following equations for surface footing,
Wj,i=

ASCE

)(

q 1 + * 1 2 * dr

(1 )
*

+
*

21*

(Uj,i+1 Uj,i-1)

dr

(1 ) r U

j,i

+Wj+1,i

(5a)

Geo-China 2016 GSP 259

187

and for embedded footing,


w j 2 ,i + w j + 2 ,i
*
+
2
2 1*

W j,i =

*
+
1*

) (U

j + 1, i + 1

U j 1, i + 1 U j + 1, i 1 + U j 1, i 1

U j 1 , i U j + 1 , i dr + 2 1 qdr
*
G*
2 1

(5b)

The equations for the load carrying (bearing or uplift) capacity q for a given limiting
displacement (), on the foundation are given by the following equations where wj,i =
i) For surface footing:
q=

E*
*

(1 + )(1 2 )dz

)[(1 * )(W j ,i W j +1,i )

(U j ,i +1 U j ,i 1 ) + *

dr
U j ,i ]
r

(6a)

ii) For embedded footing:


q=

E * (1 * )
(1 + * )(1 2 * )dz

* dr
1* r

)[

W j 2,i + W j + 2,i
2

(U j 1,i U j +1,i ) W j ,i ]

*
2(1 * )

(U j +1,i +1 U j +1,i 1 U j 1,i +1 + U j 1,i 1 ) +

(6b)

RESULTS
Circular surface footings and embedded footings
The anisotropy parameter is varied for all the cases as = 0.5, 1, 2 with = 1
implying isotropic case, whereas = 0.5 means the soil is stiffer vertically than
horizontally and = 2 entails the reverse. The value of * used in this analysis is 0.3.
Fig.2 and Fig.3 represent the normalized load-settlement graphs for circular surface
and embedded (De/D = 1) footing, respectively. In both the cases, the settlement at a
constant applied loading intensity (q/E* = 0.005 in case of surface footing and q/E* =
0.01 in case of embedded footing) is determined. It can be observed that the
settlement/displacement of the footing increases with decrease in . Fig. 4 lays out the
variation of normalized stress along the base of the domain in radial direction for
circular surface footing experiencing limiting settlement (/D=0.01). The normalized
stress is found to be maximum at the plane of symmetry (PoS) having the same
magnitude for all values of and then gradually decreases with increase in the radial
distance and eventually becomes zero towards the end of the domain. Further, it can
be also noted that at a particular radial distance from PoS, the normalized stress is
found to increase as increases. Fig. 5 shows the variation of the settlement at the
ground level along the radial direction, when the surface footing is subjected to a
working load intensity of q/E*=0.005. The settlement is seen to be maximum near the
loading zone following a decreasing trend towards the farthermost area of the domain.
As explained in Fig.2, the settlement is found to decrease proportionally beneath the
surface footing with increase in . The point of inflexion i.e., the point at which the
settlement trend with respect to gets reversed is observed at r/D=1. The variation of
normal stress developed at level of the circular embedded footing subjected to
specified upward limiting displacement (/D=0.02) is presented in Fig. 6. It can be
noted that the effect of is more pronounced as the embedment depths increases. Also
it can be deduced that the influence zone for embedded footing subjected to pullout
force at De/D = 3 is greater than that at De/D = 1. Similar to the normalized stress
response of the surface footing, the anisotropic factor has a similar effect on the

ASCE

Geo-China 2016 GSP 259

188

normalized stress along the embedment depth and the normalized stresses are found to
be deviating maximum at around r/D=2 and 4 for De/D=1 and 3 respectively.

q/E*

q/E*

/D

/D

0.000
0.005
0.010
0.0000 0.0025 0.0050 0.0075
-0.08
0.000
=0.5
= 0.5
-0.07
0.005
=1.0
= 1.0
-0.06
=2.0
0.010
=2.0
-0.05
0.015
-0.04
0.020
-0.03
0.025
-0.02
0.030
-0.01
0.035
0.00
0.040
FIG. 2. Normalized load-settlement
FIG. 3. Normalized load- displacement
plot for surface footing
plot for embedded footing
5

0.0010

10

0.0006

0.0002

0.02
=0.5
=1.0
=2.0

q/E*=0.005

0.05

FIG.5. Normalized settlementradial distance from PoS at ground


level in surface footing

FIG.4. Normalized stress-radial


distance from PoS at base level in
surface footing

r/D

10

-0.08
-0.06

z/E*

/D=-0.02

De/D=3
De/D=1
=0.5
=1.0
=2.0

FIG. 6. Normalized stress-radial


distance from PoS at foundation
level in embedded footing

r/D

10
De/D=1
De/D=3
=0.5
=1.0
=2.0

/D

10

0.01

0.04

/D=0.01

0.0000

ASCE

r/D

0.03

0.0004

0.0000
0.0002
0.0004
0.0006
0.0008
0.0010
0.0012
0.0014

0.00

=0.5
=1.0
=2.0

0.0008

z/E*

/D

r/D

-0.04
-0.02
0.00

q/E*=0.01

FIG. 7. Normalized displacementradial distance from PoS at


foundation level in embedded footing

Geo-China 2016 GSP 259

189

Fig. 7 shows the variation of displacement along the embedded footing layer under a
static working load intensity (q/E*=0.01). The effect of anisotropy factor is same as
that for circular footing. In reference to Fig.3, the anisotropic factor has an effect on
vertical displacement of the foundation subjected to uplift capacity. The effects are
more pronounced for higher embedment depths. The point of inflexion occurs between
r/D=1.2 and 1.6 for both the embedment depths.
Annular surface footings and embedded footings
Fig.8 and Fig.9 illustrate the normalized load-settlement plot for annular footing and
embedded footing respectively for induced limiting settlement/displacement
[/D=0.01 in case of footing and /D= -0.02 in case of embedded footing (De/D=3),
the ve sign implying upward displacement]. For both the cases, ri/ro has been varied
as 0.4, 0.6 and 0.8. It can be noted that for a particular ri/ro, the bearing capacity
increases as increases. This is in line with the findings of Srinivasan and Ghosh
(2014) for a plane strain angular foundations. But this trend occurs only in annular
footing and not in case of annular embedded footing. Also, for all the values of , the
maximum bearing/uplift capacity occurs at maximum value of ri/ro and vice versa. To
understand the efficacy of ring foundations over circular foundations, two efficiency
factors have been defined as,
= average settlement of annular surface or embedded footing carrying the working load
average settlement of circular surface or embedded footing at the same working load

Q = Total load resisted by ring surface or embedded footing experiencing the limiting settlement
Total load resisted by circular surface or embedded footing at the limiting settlement

The values of these two factors have been determined by applying the same amount of
2
working load (Q) or limiting displacement () over the annular footing (Q/E* ro
2

=0.005 or /D = 0.01) or embedded footing (Q/E* ro =0.01 or /D = -0.02) as that


of the circular one. Fig.10 and Fig.11 shows the efficiency factor vs ri/ro plots for
annular surface and embedded footing (at De/D=1) respectively. The effect of is
nearly negligible at smaller values of ri/ro in case of footing. In footing for all values of
, has been found to decrease with the increase in ri/ro till the latter reaches a value
0.6, after which the values remain constant whereas Q increases with increase in ri/ro
until 0.6 beyond which the factor digress for different values of anisotropic factor. For
embedded footing, the effect of becomes more prominent in with increasing ri/ro
but in Q the effect is nearly same at smaller values of ri/ro. In case of embedded
footing at both embedment depths and for all values of , decreases till ri/ro=0.6 and
then increases slightly whereas Q increase till ri/ro=0.6 and then decreases marginally.
This critical value of ri/ro=0.6 have also been observed by some researchers [Ahmad et
al. (2014)]. Fig.12 and 13 shows the variation in upward and downward displacement
resulting from the applied load in the annular embedded footing and footing level
respectively. Here r again denotes the radial distance from the plane of symmetry.
Plots for three different ri/ro (0.4, 0.6, and 0.8) and in each of them at three different
values of have been furnished here. As anticipated the maximum
settlement/displacement occurs below/above the foundation, decreases as we move
farther and diminishes very far which marks the end of influence zone for the applied
load. Also it can be noted that as ri/ro increase, the settlement values tend to converge
ASCE

Geo-China 2016 GSP 259

190

for all values of . For a particular ri/ro, the effect of on settlement/displacement


across the radial direction beyond the foundation, first increases reaches a maxima
somewhere in the mid domain and then decreases towards the end.
0.020
0.025
ri/ro=0.4

q/E*

0.015
0.010

0.005

0.000
0.005

-0.02

0.01

/D

FIG. 8. Normalized load-settlement


for annular surface footing
1.2
1.0
0.6
0.4
0.2
0.0
0

0.2

0.4

ri/ro

0.6

0.8

FIG. 10. Efficiency factors plot for


annular surface footing
5 r/D

10

0.004

-0.020

/D

Q/E*
=0.005

FIG. 12. Normalized settlement-radial


distance from PoS at foundation level
in annular surface footing

0.6

r/D
5

De/D=3

/D

-0.025

0.010

0.4

ri/ro

0.002

ri/ro=0.6
ri/ro=0.4
ri/ro=0.8
=0.5
=1.0
=2.0

0.2

0.8

FIG. 11. Efficiency factors plot for


annular embedded footing (De/D=1)
-0.030

0.008

3.0
2.5
2.0
1.5
1.0
0.5
0.0

=0.5
=1.0
=2.0

0.0

0.000

0.006

ASCE

1.2
1.0
0.8
0.6
0.4
0.2
0.0

0.8

/D

FIG. 9. Normalized load-displacement


for annular embedded footing
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0

=0.5
=1.0
=2.0

0.000
0.00

-0.01

0.014

0.015
0.010

0.005

0.012

0.020

q/E*

ri/ro=0.4
ri/ro=0.6
ri/ro=0.8
=0.5
=2.0

ri/ro=0.6
ri/ro=0.8
=0.5
=1.0
=2.0

-0.015

10
ri/ro=0.6
ri/ro=0.4
ri/ro=0.8
=0.5
=1.0
=2.0

-0.010
-0.005
0.000

Q/E*

=0.01

FIG. 13. Normalized displacementradial distance from PoS at foundation


level in annular embedded footing

Geo-China 2016 GSP 259

191

z/E*

0.0008
0.0007
0.0006
0.0005
0.0004
0.0003
0.0002
0.0001
0.0000

r/D

10

/D=0.01

r/D

10

0.0000
-0.0002

/D=-0.02

-0.0004

z/E*

-0.0006

ri/ro=0.4
ri/ro=0.6
ri/ro=0.8
=0.5
=1.0
=2.0

-0.0008
-0.0010
-0.0012

De/D=3

ri/ro=0.4
ri/ro=0.6
ri/ro=0.8
=0.5
=1.0
=2.0

FIG. 15. Normalized stress-radial


FIG. 14. Normalized stress-radial
distance from PoS at foundation level
distance from PoS at base level in
in annular embedded footing
annular surface footing
The trend continues to be same for all values of ri/ro. Lastly, Fig. 14 and 15 illustrate
the trend between normalized stress at embedded footing level or base level in footing
across the radial distance from the plane of symmetry when the foundation is
experiencing a limiting settlement/displacement. For a particular ri/ro, the influence of
on normalized stress has similar effect as that of circular foundation where the
maximum deviation in the stress occurs around the middle of the domain and concurs
at the PoS and far away boundary for both embedded footing and footing. At a
particular value of , the values of normalized stresses deviate maximum at the PoS
for different ri/ro for both embedded footing and footing, and the normalized stress is
found to increase with decrease in ri/ro.
CONCLUSIONS
In this study the bearing capacity and settlement behavior of annular footings under
static compressive and tensile working loads has been analyzed. The soil medium was
assumed to be linearly elastic, as the working range of the soil under structural load
during its lifetime is generally within small stresses and displacements compared to
the ultimate bearing capacity and final settlement. At the same time owing to its
formation and various other coupled phenomena the soil may show anisotropic
properties reasonably, this is why the soil medium has been assumed to be anisotropic
in nature. From all the plots and data produced in the previous section, the following
significant conclusions can be drawn:
i)
The effect of follows the same trend in circular footing and embedded
footing. It can be observed that as decrease, the settlement/displacement
increases in both the cases. Similarly, the bearing/uplift capacity decreases
with decrease in anisotropic factor when subjected to limiting displacement.
ii)
In annular footings, the effect of on bearing capacity (q/E*) increase as the
ring becomes thinner (ri/ro) but there is not much effect of in annular
embedded footing.

ASCE

Geo-China 2016 GSP 259

iii)
iv)
v)

From the efficiency factors, it can be ascertained that the efficacy of ring
foundations is more than circular ones both in terms of bearing capacity and
uplift capacity.
The critical value of ri/ro is found to be 0.6 which is close to the value reported
in previous research works. [Ohri et al. (1997), Boushehrian and Hataf (2003),
Ahmad S K (2014)].
The normalized stress distribution and vertical deformation across the radial
direction are presented for both circular and annular footings/embedded
footings.

REFERENCES
Ahmad, S. K., Srinivasan, V. and Ghosh, P. (2014). Analysis of annular footings and
anchors lying on elastic soil medium using finite difference technique.
5th International Congress on Computational Mechanics and Simulation, December
10-13, 2014, Chennai, India.
Boushehrian, H.J. and Hataf, N. (2003). Experimental and numerical investigation of
the bearing capacity of model circular and ring footings on reinforced sand.
Geotextiles and Geomembranes 21, 241-256.
Egorov, K.E. (1965). Calculation of bed for foundation with ring footing.
Proceedings of the Sixth International Conference on Soil Mechanical Foundation
of Engineers, 2, Montreal, 4145.
Fisher, K., (1957). Zur Berechnung der setzung Von Fundamenten in der form einer
Kreisformigen Ringflache. Der Bauingenieur, Berlin, Germany, Vol. 32 (5), 172
174 (in German).
Ghosh, P and Sharma, A (2010). Interference effect of two nearly strip footings on
layered soil: theory of elasticity approach. Acta Geotechnica, 5, 189-198.
Graham, J. and Houlsby, G.T. (1983). Anisotropic elasticity of a natural clay.
Geotechnique, 33 (2), 165-180.
Hataf, N., and Razavi, M.R., (2003). Behavior of ring footing on sand. Iranian
Journal of Science and Technology, Transaction B, Vol. 27, 4756.
Kumar, J., and Ghosh, P., (2005). Bearing capacity factor N for ring footings using
the method of characteristics. Canadian Geotechnical Journal, 42(5), 1474-1484.
Kumar, J. and Naskar, T. (2012). Vertical uplift capacity of a group of two coaxial
anchors in a general c- soil. Canadian Geotechnical Journal, 49(3), 367-373.
Madhav, M.R. (1980). Settlement and allowable pressures for ring footings. Indian
Geotechnical Journal, 10 (3), 267-271.
Maheshwari, P and Viladkar, M.N (2007). Strip footings on a three layer soil system:
theory of elasticity approach. International Journal of Geotechnical Engineering,
1, 47-59.
Nainegali, L.S and Basudhar, P.K (2011). Interference of two closely spaced
footings: A finite element modeling. Geo Frontiers, ASCE, 3726-3735.
Ohri, M.L., Purhit, D.G.M. and Dubey, M.L. (1997). Behavior of ring footings on
dune sand overlaying dense sand. Proceedings of International Conference of
Civil Engineers, Tehran, Iran.
Rathour, D.S. (2009). Settlement for ring footings through strain influence diagrams
including stiffness non linearity. M Tech Thesis, IIT Kanpur.

ASCE

192

Geo-China 2016 GSP 259

193

3-D Analysis of a Piled-Raft Foundation Subjected to Vertical and Lateral Loads


Nasr E. Nasr1 and Tamer M. Sorour2
1

Assistant Professor, Structural Engineering Dept., Ain-Shams Univ., Abbasseya, Cairo, Egypt. E-mail:
nasr.eid@eng.asu.edu.eg
2
Assistant Professor, Structural Engineering Dept., Ain-Shams Univ., Abbasseya, Cairo, Egypt. E-mail:
tamer.sorour@eng.asu.edu.eg

Abstract: The objective of this research work is to study the behavior of Piled-Raft
foundations subjected to vertical and lateral loads. Study for most of the previous
research in this field shows that most of these studies analyzed Piled-Raft foundations
in 2-D plane presenting the pile as a spring which may not give very accurate results. In
this research, 3-D finite element method is suggested to represent Piled-Raft
interaction. This leads to better representation for the elements which gives more
reliable results such as (axial displacement, rotational displacement, vertical forces and
moments acting on the pile and the raft). The results from this analysis will be
compared with the results of a case study on a bridge in Basra, Iraq called (Fly Over
Bridge) in order to verify the model. An extensive parametric study will be performed
to assess the effect of various factors on the internal forces generated in the raft and
piles. From analysis, there are important recommendations concerning the maximum
inclination angle of the pile, and the diameter of the piles on the internal forces
generated in the foundation and the piles and their equilibrium.
INTRODUCTION
Batter piles have been used for a long time to resist large lateral loads from winds,
water waves, soil pressures, and impacts. Their distinct advantage over vertical piles is
that they transmit the applied lateral loads partly in axial compression, rather than only
through shear and bending. Thus, batter piles offer larger stiffness and bearing capacity
than same diameter and depth vertical piles a superiority of particular importance when
the near surface soils are soft and/or the lateral load is large. Despite these advantages,
batter piles have low resistance against seismic resistance. Following the poor
performance of batter piles in a series of earthquakes, the seismic behavior of inclined
piles has been considered detrimental, and many codes require that such piles be
avoided. For instance, the French Seismic Code (AFPS1990) states flatly that Inclined
piles should not be used to resist seismic loads. The seismic Euro code EC8/ Part5,
dealing with geotechnical and foundations, is a little less restrictive, stating: It is
recommended that no inclined piles be used for transmitting lateral loads to the soil. If,
ASCE

Geo-China 2016 GSP 259

194

in any case, such piles are used, they must be designed to carry safely axial as well as
bending loading.
FINITE ELEMENT MODEL
Finite element model using (MIDAS GTS) program is suggested to represent piledraft foundation interaction. This program features from other software that solves the
problem in 3-D and represents the shaft pile properties. This leads to better
representation for the element which gives good results such as (Axial displacement,
rotational displacement, vertical forces and moments acting on the pile and foundation).
MOHR-COULOMB MODEL
The Mohr-Coulomb model is used to simulate most terrain and it displays sufficiently
reliable results for general nonlinear analysis of the ground.
Eqn. 1
MIDAS GTS can simulate changes in the modulus of elasticity and cohesion with
height for a Mohr- Coulomb model using equation (1). If the amount of cohesion
change with height is 0 (zero), a constant value is used. If the amount of change is not
0 (zero), the cohesion can be calculated with respect to a reference height using
equation.
Eqn. 2
cref
cinc
yref

: Input cohesion value


: Cohesion increment with respect to depth
: Depth at which cref is measured

The y in equation (2) represents the integral position of the element. If the integral
position is located higher than y ref, the cohesion can be smaller than '0'. To prevent this,
the cohesion value is not decreased any further and the c ref value is used.
Failure function of Mohr-Coulomb model According to Mohr (1900), failure can be
expressed using the following equation:
Eqn. 3
c

: Cohesion
: Internal friction angle

Here, the limit shear stress


n of the same plane.

ASCE

of an arbitrary plane is only related to the normal stress

Geo-China 2016 GSP 259

195

Equation (3) shows that material failure occurs at the stress state where the largest
Mohr circle comes across the Coulomb friction failure envelope. It also shows that the
does not have an effect on the failure
intermediate principal stress
condition. Hence, the failure function of the Mohr-Coulomb failure plane is as follows:
Eqn. 4
The failure criterion of equation (4) is called the Mohr-Coulomb criterion and it is the
most widely used method for ground materials due to its simplicity and accuracy.
Expressing the Mohr-Coulomb criterion using principal stress terms
equation can be rearranged into the following equation:

Eqn. 5

FIG. 1 Mohr-Coulomb failure surface shape in principal stress space.


Mohr-Coulomb failure criterion does not change with the confining pressure (or
hydrostatic pressure). Hence, the criterion is accurate when the confining stress is
within a limited range, but it does not agree with actual physical phenomena when the
confining stress is large enough to cause compressive failure. However, this criterion
gives highly accurate results within the confining stress ranges of the field and it is easy
to use. Hence, it is the most widely used failure model.

ASCE

Geo-China 2016 GSP 259

196

VERIFICATION OF FINITE ELEMENT MODEL


The results from the FEM are compared with those resulted from a case study
obtained from the design of the Fly Over Bridge in Basra, Iraq. This project was
designed by an international consultant office. Vertical and battered piles where used to
support the foundations of the pre-mentioned bridge. Figures (2-3) show the
discretization of the FEM and the axes of loading, respectively.

FIG. 2. Finite element model discretization.

FIG. 3. Axes of loading.


Material Properties
The piles are modeled as linear elastic frame element with the following properties:
Table 1. Pile Data
Section Shape/spacing
Length
Angle of inclination
Material Type
Modulus of Elasticity
Poissons Ratio
Unit Weight
Thermal Coefficient

Round of diameter 600.0 mm/ 1.80 m


22.0 m
22.60o
Concrete
40000000 kN/m2
0.2
24.0 KN/m3
0.00001

The soil elements are modeled using Mohr-Coulomb formulations.

ASCE

Geo-China 2016 GSP 259

197

Table 2. Soil Profile and Properties


Layers
Depth (m)
Unit Weight (kN/m2)
Saturated/Submerged Unit Weight
(kN/m2)
Internal Friction Angle (o)
Cohesion (kN/m2)
Unit Ultimate Skin Friction (kN/m2)
Unit Ultimate Bearing Capacity
(kN/m2)
Strain at 50% Stress

Loose
Sand

Soft Clay

Sand &
Gravel

Stiff
Clay

5.0
14.7

12.5
17.2

17.5
20.6

25.0
18.6

19.1/9.29

9.8

11.3

11.8

32.0
45.0

25.0
30.0

40.0
95.0

75.0
100.0

1.9

10.0

4.8

200.0

150.0

562.5

The ground water level was considered on a depth of 2.50 m from the ground surface.
Foundation Data
The foundation was modeled using linear elastic concrete elements with total
dimensions of:
Width = 8.40 m, Length = 19.2 m, Thickness = 1.0 m.
Cases of Loading
Several models were established in order to simulate the effect of different loading
conditions on the behavior of the pile group in conjunction with the footing and soil
condition. A summary of the results obtained from the FEM are presented in Figs (4-7).
Table 3. Applied Loads

ASCE

Loading

P (kN)

M (kN/m2)

First

-13881

7810

Second

-12881

6810

Third

-11881

5810

Forth

-10881

4810

Fifth

-9881

3810

Geo-China 2016 GSP 259

FIG. 4. (a) Variation of max. Vertical displacement (Dz) with loading on piles; (b)
Variation of max. Shearing force with loading on piles; (c) Variation of max.
Vertical force with loading on piles; (d) Variation of max. Moment with loading on
piles
By comparing the results of the FEM with the actual results obtained from the case
study as illustrated in Figs (4-7), we can conclude that there is good agreement between
the results of both cases.
PARAMETRIC STUDY
An extended parametric study was carried out to examine different factors affecting
the distribution of loads on a group of piles. The principle parameters, considered in
this study, are: pile diameter and the angle of inclination of piles, with different types
and condition of soils. A summary of the analyses are presented in Figs (8-14)

ASCE

198

Geo-China 2016 GSP 259

The effect of pile Diameter

FIG. 5. (a) Variation of pile diameter with vertical displacement (Dz) of piles; (b)
Variation of pile diameter with max. Rotational displacement (Rz) of footing; (c)
Variation of pile diameter with max. Rotational displacement (Rx) of footing; (d)
Variation of pile diameter with max. Rotational displacement (Ry) of footing.
Effect of angle of inclination of piles

FIG. 6. (a) Variation of angle of inclination of piles with max. Vertical displacement of
piles; (b) Variation of angle of inclination of piles with max. Vertical force on piles;
Variation of angle of inclination of piles with max. Moment on piles

ASCE

199

Geo-China 2016 GSP 259

CONCLUSIONS
1- The results from finite element model using (MIDAS GTS) program are compared
with the results of a case study on a bridge in Basra, Iraq called (Fly Over Bridge)
and the results were in good agreement.
2- It could be concluded that; increasing the pile diameter results in decreasing the
vertical displacement of piles in the examined soil types.
3- By increasing the pile diameter; the maximum rotational displacement of footing in
the examined soil types is decreased.
4- It could be concluded that; by increasing the inclination angle of the pile from 0o to
25o; this leads to decrease the vertical displacement of piles in the examined soil
types.
5- Increasing the inclination angle of the pile from 0o to 25o results in decreasing
conveniently the vertical force on piles in dry sandy and clayey soils.
6- It could be concluded that; increasing the inclination angle of the pile from 0o to 10o
has small effect on moment. While, the moment on the piles is decreased by
increasing the angle from 10o to 25o.
REFERENCES
Association Franaise de Gnie Parasismique (AFPS). (1990).
Aschenbrenner, Rudolf, "Three-Dimensional Analysis of Pile Foundations," Journal of
the Structural Division, American Society of Civil Engineers, Vol. 93, February
1967, p 201.
Banerjee, P.K., and Davies, T.G. (1978). The behavior of axially and laterally loaded
piles embedded in non-homogeneous soils. Geotechnique, 28(3), 309-326.
Banerjee, P.K., and Davies, T.G. (1980). Analysis of some reported case histories of
laterally loaded pile groups. Numerical methods in off shore piling, Institution of
Civil Engineers (ICE), London, 101108.
Deng, N., Kulesza R., and Ostadan, F.2007. Seismic soil-pile group interaction
analysis of a battered pile group. Proc.,4th Int. Conf. on Earthquake Geotechnical
Engineering, Aristotle Univ. of Thessaloniki, Laboratory of Soil Mechanics,
Foundation and Geo-technical Earthquake Engineering, Greece.
Gerolymos, N., and Gazetas, G. (2006). Winkler model for lateral response of rigid
caisson foundations in linear soil. Soil Dyn. Earth-quake Eng., 26(5), 347361.
Giannakou, A. (2007). Seismic behavior of inclined piles. Ph.D. thesis, National
Technical Univ. of Athens, Athens, Greece.
Hrennikoff, A., "Analysis of Pile Foundations with Batter Piles," Transactions,
American Society of Civil Engineers, Vol. 115, Paper No. 2401, 1950, p 351.
Kavazanjian, E. (2006). A driven-pile advantage: Batter piles. Pile driver, Q4, 2125.
Padron, L. A., Aznarez, J. J., Maeso, O., and Santana, A. (2009). Dynamic stiffness of
deep foundations with inclined piles. Earthquake Eng. Struct. Dyn. 38, 10531070.
Sadek, M., and Shahrour, I. (2004). Three-dimensional finite element analysis of the
seismic behavior of inclined micro piles. Soil Dyn. Earthquake Eng., 24,473485.

ASCE

200

Geo-China 2016 GSP 259

Research on a Calculation Method and Three-Dimensional Simulation of


a High-Filled Embankment Rheological Settlement
Zhi-Chao Wang1,2; Du-Min Kuang1; Tao Zhao3; Ying-She Luo2; and Wei-guo Wang1
1

Hunan Key Laboratory of Geomechanics and Engineering Safety, Xiangtan Univ., Xiangtan 411105,
China.
2
Hunan Province Key Laboratory of Engineering Rheology, Central South Univ. of Forestry and Technology, Changsha, Hunan 410004, China.
3
State Key Laboratory of Hydraulics and Mountain River Engineering, College of Water Resource and
Hydropower, Sichuan Univ., 610065, China.

Abstract: To describe the over-consolidation, dilatancy, strain softening, and


time-dependent behavior of subgrade compacted clay inside a high-filled embankment
during layered filling process, an elastic-viscoplastic constitutive model based on
sub-loading yield surface was adopted to analyze the construction and post-construction
settlements of high-filled embankment. Illustrated by the case of the dam-like embankment of Lanhai highway between the pile numbers from K50+650 to K50+860, the
rheological settlement performance of the high-filled embankment was investigated
emphatically on the condition of complex three-dimension boundary by using the proposed model. These factors such as backfilling method, backfilling speed and the
over-consolidation stress history of backfilling clay were simulated to consider the influence on the rheological settlement of high-filled embankment. The settlement prediction result of the proposed model is in good agreement with data of onsite settlement
observation. This shows that the proposed model is suitable for analyzing the long-term
settlement of high-filled embankment.
INTRODUCTION
Along with the stretching of the express-way gradually towards the hilly and mountainous areas in the Midwestern China, many high-filled embankments have been
formed with the filling height of >20 m. Compared with general low embankment,
high-filled embankment has the features of great filling height, large post-construction
settlement, and long settlement time. Hence, how to efficiently calculate and predict the
post-construction settlement of the high-filled embankment so as to eliminate the harm
brought by the embankment settlement has become an important topic in the field of
highway construction research. As the rheological constitutive model can reflect the
time-dependent deformation of soil, it is thought as an effective method to calculate the
post-construction settlement of high-filled embankment. At present, the application of
rheological constitutive model is still in its infancy in the embankment settlement calculation. Relatively simple rheological element models are selected in most cases (e.g.
Zhao Min etc. 1999, Lu Qing etc. 2005, Duan Zhu-geng etc. 2007). However, subgrade

ASCE

201

Geo-China 2016 GSP 259

202

soil inside the embankment accumulates a complex stress history during the layered filling process, resulting in a complex settlement deformation caused by
over-consolidation, dilatancy and strain softening at the state of low external confining
pressure and high internal vertical stress. Therefore, an advanced rheological constitutive model should be proposed to describe the over-consolidation and time-dependent
deformation characteristics of compacted clay of high-filled embankment.
Scholars from all over the world have conducted many successful studies on the deformation characteristics and constitutive model of over-consolidated clay. Dafalias provided a boundary surface model. Hashiguchi suggested the concept of subloading yield
surface. Asaoka introduced the concept of superloading and subloading yield surface for
over-consolidated soil and structural soil. Nakai applied the concept of subloading yield
surface to the tij model. Yao developed an over-consolidation constitutive model and a
modified model based on the Hvorslev surface. But most of these proposed constitutive
models are elastoplastic constitutive models, which cannot describe the time-dependent
deformation characteristic of over-consolidated clay. To overcome this difficult problem, Wang proposed an advanced elastic-viscoplastic constitutive model to predict
time-dependent behavior of the over-consolidated clay based on the subloading surface
model and relative overstress relation, and its stress integration algorithms is also presented, which makes the new model convenient to be implemented into the finite element software such as ABAQUS. As an illustration of the rheological settlement performance of the high-filled embankment, the case of the dam-like embankment of Lanhai highway between the pile numbers from K50+650 to K50+860, has been investigated using the proposed new model.
BASIC EQUATIONS OF NEW ELASTIC-VISCOPLASTIC CONSTITUTIVE
MODEL
The viscoplastic strain rate component ijvp is defined as
ijvp = ( F )

f d
ij

(1)

where f d is a dynamic yield function, ij is stress component, and ( F ) is defined as


follows
0
m
c0 ( kd ku )

(F ) =

F 0
F >0

(2)

where c0 and m' are parameters related to the time-dependent properties of clay; kd is the
parameter representing both effect of strain-hardening and strain-rate; ku is a volumetric
strain-hardening parameter for overconsolidated clay. The dynamic yield function f d is
defined as
f d = q M m + ln m = kd
(3)
where M is the stress ratio at critical state; q is the deviator stress; m is the mean effective stress. According to the subloading yield surface concept proposed by Hashiguchi &
Ueno (1977), a new reference yield function fu for overconsolidated clay is adopted as
shown in Figure 1, and it is given as

ASCE

Geo-China 2016 GSP 259

203

ku =

1 + e0 vp

)
v
+ ln ( myi
1 + e0

(4)

is the initial consolidation stress; e0 is the initial void ratio; vvp is the viswhere myi
coplastic volumetric strain; is the over consolidation ration at initial stress state;
and is the compression index and swelling index, respectively; is a void ratio
difference between point A at the over-consolidated state and point B at normally consolidated state as shown in Fig.1b, and it is given by

= ( ) ln

pN1e
= ( ) ln OCR

pN1

(5)

pN1
is the over consolidation ration at initial stress state. For convenwhere OCR= pN1e

are the
and pN1
ience, OCR is substituted by in the following statement. pN1e
intercepts of normally consolidated yield locus and overconsolidated subloading yield
locus at the mean effective stress p axis, respectively.
q
e0

1
X ( p , q )

f s = ks

f d = kd
f u = ku

p0 = 1kPa

vvp pN1

pN1e

ln p

(a) Dynamic yield surface

p0 = 1kPa

pN1

pN1e

ln p

(b) One-dimensional compression curve

FIG.1. Elasto-viscoplastic constitutive model of overconsolidated clay based on a


relative overstress relation
The following equation for has been used after Hashiguchi (1989)

d
= ad dp
(1 + e0 ) 2

(6)

where a is a material parameter to control the change rate of overconsolidation ratio. The
peak strength of overconsolidated clay decreases with increasing a . When a = 0 , the
overconsolidated clay will become a normally consolidated clay. A dynamic equilibrium
equation g ( ij , vvp , dvp ) for overconsolidated clay is given as
m

c

1+ e0 vp vp
v d = 0

g(ij , , ) = 0 m exp m +


Mm myi
M
vp
v

vp
d

(7)

THREE-DIMENSIONAL NUMERICAL SIMULATION OF HIGH-FILLED


EMBANKMENT RHEOLOGICAL SETTLEMENT
To analyze the rheological settlement performance of high-filled embankment, a
dam-like embankment of Lanhai highway between the pile numbers from K50+650 to
K50+860 was investigated on the condition of complex three-dimension boundary by
using the proposed elastic-viscoplastic constitutive model.

ASCE

Geo-China 2016 GSP 259

204

Establishment of the three-dimensional finite element model


The three-dimensional geometry and finite element model of the dam-like embankment is shown in Figure 2. The embankment filling is divided into three parts as 90#
zone, 93# zone, and 96# zone according to the compacting degree of backfill clay. Figure 2a and Figure 2b denotes the vertical section profile and cross section profile, respectively. Furthermore, boundary conditions were applied to all faces of the geometric
model to limit displacement in the direction perpendicular to the faces, but the bottom
face was limited to no displacement in all directions.
230

65

69
20 10

15.3

29

45

25
1:2.

2
20
1:2

327.5
17.5 12.2517.5 2
12.25
20
45

1:2
.25

Baserock

327.5

(a)Vertical Section Profile

65

1:2
.25
Medium strength silt

.25
1:2

6.3

6.3

1:1
.75
1:2

75
1:1.

5
7.4

35.6

75.3
58.6

U-shaped slope

20 10 10
75.3

50

29

23

6.3

43.4

10

52.5

5.7

20.2

14.9
19.3

75.3

10.4

6.7

(b) Cross section Profile


U-shaped slope
96# Zone
Medium
strength silt

93# Zone
Base rock

(c) Finite element model

90# Zone

FIG.2. Three-dimensional finite element model of dam-like embankment of Lanhai


highway between the pile numbers from K50+650 to K50+860 (Unit: m)
Filling construction loading process
Layered filling is performed on the high-filled embankment according to the construction schedule as shown in Figure 3a and Table 1. The embankment model is divided
into 12 element sets during the numerical simulation. Combining with death and birth of
the elements, let them Rebirth gradually. Specific order is as shown in Figure 3b.
Filling height /m

70

Set-1

56

Set-2

42

Set-3
Set-4
Set-5

28
Set-7

14
0

Set-10

50

100

150

200

250

T im e /d

300

(a) Construction process

350

400

Set-8
Set-11
Set-12

Set-6
Set-9

(b) Loading sequence

FIG.3. Filling progress chart of Lanhai highway between the pile numbers from
K50+650 to K50+860

ASCE

Geo-China 2016 GSP 259

205

Table 1. Filling progress information of high-filled embankment of Lanhai highway


EmbankLayered
Loading Loading Time step
ment height Compacting
filling
height Monitoring
degree
point
sequence time /d
/d
/m
/m
5
5
Bottom layer
set-12
20
20
9
4
set-11
40
20
90
11
2
set-10
140
100
13
2
Second layer
set-9
160
20
18
5
set-8
200
40
20
2
set-7
220
20
20
0
-270
50
93
22
2
set-6
280
10
29
7
set-5
300
20
35
6
set-4
310
10
49
14
Fourth layer
set-3
330
20
59
10
set-2
360
30
96
69
10
set-1
390
30
Post-construction
3630
3240

Time step Total time


/s
/s
1.7280E6
1.7280E6
8.6400E6
1.7280E6
3.4560E6
1.7280E6
4.3200E6
8.6400E5
1.7280E6
8.6400E5
1.7280E6
2.5920E6
2.5920E6
2.7994E8

1.7280E6
3.4560E6
1.2096E7
1.3824E7
1.7280E7
1.9008E7
2.3328E7
2.4192E7
2.5920E7
2.6784E7
2.8512E7
3.1104E7
3.3700E7
3.1363E8

Material parameters of the finite element model


To investigate the mechanical characteristics of clay in the high-filled embankment
during successive compaction process, indoor compaction, one dimensional consolidation and triaxial shear tests are performed with the roadbed backfill clay. According to
the one dimensional consolidation tests with compacting degree of 90%, 93%, and 96%
as well as water content of 21.7% as shown in Figure 4, the pre-consolidation pressure pc , initial void ratio e0 , compression index and swelling index of compacted clay
with different compacting degree can be determined.

Void ratio

0.7
0.6
0.5
0.4
10

=21.7% K=90%
=21.7% K=93%
=21.7% K=96%

100

1000

Consolidation pressure/kPa

FIG.4. One-dimensional compression curves of subgrade compacted clay


The consolidated drained triaxial tests of the unsaturated compacted clay with compacting degrees of 90%, 93%, 96%, and cell pressures of 50kPa, 100kPa, 200kPa under
loading rates of 0.0125%/min and 0.125%/min were carried out to study the
time-dependent deformation characteristics of subgrade compacted clay of high-filled
embankment. These tests can be used to determine the cohesive force c and
time-dependent parameters, such as c0 and m' of the compacted clay, as shown in Figure
5. The parameters for compacted clay in the proposed model and other model parameters were given in Table 2 and Table 3, respectively. Consolidated drained triaxial test of
the unsaturated compacted clay under loading rate of 0.125%/min has been simulated to
verify the validity of the model. The comparison of predicted results and test data of
compacted clay in the laboratory test was shown in Figure 6, and it shows that the pre-

ASCE

Geo-China 2016 GSP 259

206

dicted results are in accordance with the test data. So it can be used to describe the rheological settlement of embankment. As the pre-consolidation pressure of compacted clay
sample prepared by the indoor standard compaction test is always lower than that produced by the onsite road roller, the onsite model parameters pc and c were also given in
Table 2 by inverse analysis according to the observing settlement of the dam-like embankment of Lanhai highway.
Table 2. Parameters for compacted clay
K/%

e0

c0 / s 1

96
93
90

0.646
0.703
0.767

0.3
0.3
0.3

2.10
2.05
1.90

0.10057
0.11347
0.12023

0.01487
0.01683
0.02024

1.96E-13
1.96E-13
1.96E-13

22.5
22.5
22.5

490
500
510

K=96%

K=90%

10

1 /%

(a) 3 =50kPa

15

20

v/%

v/%
10

10

K=93%

K=90%

K=96%

1 = 0.125%/min
5

10

1 /%

15

(b) 3 =100kPa

20

v/%

K=93%

1 = 0.125%/min
0

1
K=90%

c/kPa
Indoor Onsite
155.0
264.6
107.8
200.6
68.3
136.6

pc /kPa
Indoor Onsite
510
1012
390
986
310
857

10

K=93%

K=96%

1 = 0.125%/min

10

15

20

1 /%

(c) 3 =200kPa

FIG.5. Stress-strain curves and stress ratio-volume change-strain curves of compacted clay

FIG.6. Comparison between predicted result and test data of compacted clay

ASCE

Geo-China 2016 GSP 259

207

Table 3. Materials mechanical properties of high-filled embankment


Density Water Content Young Modulus
Poisson Ratio
Material Types
(g/cm3)
(%)
(MPa)
Base Rock
2.2
/
30000
0.25
Elastic
U-shaped Slope
1.41
4.6
25
0.30
Elastic
Medium strength silt
1.97
/
23.5
0.27
Elastic
Embankment
1.85
12
/
0.30
Elastic-viscoplastic
Note: Medium strength silt had been thrown stone and squeezed before embankment filling, and the thickness of gravel cushion is 2 meters.
Name

NUMERICAL RESULTS
Two sets of model parameters given in Table 2 were used to calculate the settlement,
one of them was determined by indoor test, and another one was obtained by inverse
analysis with the factual settlement observation data of Lanhai highway in field as
shown in Figure 7. It shows that the greater settlement value than observing settlement is
obtained in simulation with the model parameters directly according to indoor laboratory
data without regard to the actual pre-consolidation pressure. The pre-consolidation
pressure pc and cohesive force c obtained by inverse analysis is nearly twice higher than
that according to the indoor test. When the pre-consolidation pressure is relatively high,
the amount of settlement would be small. Furthermore, the numerical results are in
agreement with the measured results, so the rheological settlement can be simulated
reasonably using the elastic-viscoplastic constitutive model based on sub-loading yield
surface.

Settlement /cm

80
60
40
20
0

Filling height /m
Construction progress /d
100

15
30
45
60

Settlement /cm

30
60
90
120

1000

Observing settlement of second layer


Observing settlement of fourth layer
Calculated settlement of bottom layer
Calculated settlement of second layer
Calculated settlement of fourth layer

(a)

80
60
40
20
0

Post-construction settlement

Construction settlement

Construction settlement

Predicted settlement

Onsite inverse analysis

Post-construction settlement

Filling height /m
Construction progress /d
100

Observing settlement of second layer


Observing settlement of fourth layer
Calculated settlement of bottom layer
Calculated settlement of second layer
Calculated settlement of fourth layer

1000

Predicted settlement

b) According to indoor test

FIG.7.Time history curve of settlement of observation and finite element calculation of dam-like embankment of Lanhai highway

ASCE

Geo-China 2016 GSP 259

208

If the embankment simulated by inverse analysis parameters (as shown in Figure 7a)
was slitting along transverse section and vertical section, displacement maps of several
layers were obtained as shown in Figure 8. It shows that the maximal observation displacement occurs in the middle set-4 layer of the embankment, reaching 64cm. The displacement in the bottom and the top layer is smaller than that occurring in the fourth
layer, which is only 20~30cm. It means that the displacement occurs at the top of or the
bottom cannot always be the maximum value, namely, itself settlement deformation of
the high-filled embankment cannot be ignored.
0

50

100

150

200

250

300

40
60
80

100

50

100

150

200

20

Settlement /cm

Settlement /cm

20

Maximal construction settlement:


Top layer
Second layer
Fourth layer
Bottom layer
Maximal settlement layer
Post-construction settlement:
Top layer
Second layer
Fourth layer
Bottom layer
Maximal settlement layer

50

100

150

200

Relative distance /m

(a) Traverse section

250

300

40
60
80

100

Maximal construction settlement:


Top layer
Second layer
Fourth layer
Bottom layer
Maximal settlement layer
Post-construction settlement:
Top layer
Second layer
Fourth layer
Bottom layer
Maximal settlement layer

50

100

150

Relative distance /m

200

(b) Longitudinal section

FIG.8. Calculating displacement of high-filled embankment of Lanhai highway


along different filling layers
CONCLUSIONS
The results show that: 1) the elastic-viscoplastic constitutive model based on the
sub-loading yield surface can be easily employed to calculate the rheological settlement
of high-filled embankment; 2) compared with the conventional elastic-plastic finite element method, time in the elastic-viscoplastic finite element method will be directly involved in the calculation but not used as a counting function. So it can present the law of
construction and post-construction settlement of high-filled embankment changing with
the time, and its physical meaning is clear; 3) the maximum mean effective stress history
has a signification impact on settlement, and the amount of settlement would be small
when the pre-consolidation pressure is relatively high.
ACKNOWLEDGEMENTS
The research reported in this paper was funded by the National Natural Science
Foundation of China (Grant No.51308485), Hunan Provincial Natural Science Foundation of China (Grant No.12JJ4006), and Open Project of Hunan Province Key Laboratory of Engineering Rheology (Grant No.14HNKLER01).

ASCE

Geo-China 2016 GSP 259

REFERENCES
Asaoka A, Nakano M, Noda T.(1997). "Soil-water Coupled Behavior of Heavily Overconsolidated Clay Near/at Critical State." Soils and Foundations, Vol.37(1): 13-28.
Asaoka A, Nakano M, Noda T. (2000). "Elastoplastic Behavior of Structured Overconsolidated Soils." Journal of Application Mechanics, (3): 335-342.
Dafalias Y F.(1986). "Bounding Surface Plasticity I: Mathematical Formulation and
Hypoplasticity." Journal of Engineering Mechanics, Vol.112(9): 966-985.
Duan Zhugeng, CHEN Xiaobin.(2007). "Analysis on stone- earth fillings embankment's
rehgological settlement surveying and calculation." Journal of Railway Science and
Engineering, Vol.4(5): 57-62.
Hashiguchi K.(1989). "Subloading Surface Model in Unconventional Plasticity." International Journal of Solids and Structures, Vol.25(8): 917-945.
Lu Qing, Shang Yue-quan, Chen Yun-fa,et al.(2005). "Back-analysis of visco-elastic
parameters of filling materials and settlement prediction for high-filled embankment."
Chinese Journal of Rock Mechanics and Engineering, Vol.24(7): 1231-1235.
Nakai T, Hinokio M.(2004). "A Simple Elastoplastic Model for Normally and Over
Consolidated Soils with Unified Material Parameters." Soils and Foundations,
Vol.44(2): 53-70.
Wang Zhi-chao, Jiang Ming-jing, et al.(2014). "A new method for establishing elasticviscoplastic constitutive model of clay." Geomechanics from Micro to Macro (Proceedings of International Symposium on Geomechanics from Micro to Macro, Cambridge, UK), 699-704.
Yao Yang-ping, Hou Wei, Zhou An-nan.(2007). "A Constitutive Model for Overconsolidated Clays Based on the Hvorslev Envelope." Science in China (Ser. E),
Vol.37(11): 1417-1429.
Zhao Min, Han Shi-lian, Chen Rong-sheng.(1999). "Settlement analysis for airport high
fill concerning material visco-elasticity." China Journal of Highway and Transport,
Vol.12(S1): 25-30.

ASCE

209

Geo-China 2016 GSP 259

210

Dynamic Characteristics Study of Geosynthetic-Reinforced Soil under Cyclic Loading


Wei Shi1; Tao Lu2; Longlong Zhang3; and Yue Pan4
1

School of Civil Engineering, Qingdao Technological Univ., Qingdao 266033, China. E-mail:
susan.sw@163.com
2
School of Civil Engineering, Qingdao Technological Univ., Qingdao 266033, China. E-mail:
313892957@qq.com
3
School of Civil Engineering, Qingdao Technological Univ., Qingdao 266033, China. E-mail:
897334337@qq.com
4
School of Civil Engineering, Qingdao Technological Univ., Qingdao 266033, China. E-mail:
852469349@qq.com
Abstract: The technology of geosynthetic reinforced soil has been widely used in the structural
strengthening of highway subgrade. It can improve the bearing and deformation resistance
capacity of the subgrade effectively, prolong the service period of the road to ensure the safety
and comfort of the vehicle. So far, the quasi-static method to design the geosynthetic reinforced
subgrade under the action of traffic load is not perfect enough. The study on dynamic property,
design theory and parameters lags far behind the application and improvement of the
geosynthetic reinforced soil technology. This paper used the dynamic triaxial test in laboratory
by taking different kinds of soil (clay and silt), different confining pressure (50kPa, 100kPa, and
150kPa) and different reinforcement layer (0 layer, 1 layer, and 2 layers) to study the change rule
and influence factors of dynamic modulus of elasticity. This study would provide the gist for the
improvement of dynamic design theory and parameter choice of geosynthetic reinforced
subgrade with the effect of traffic load.
Keywords: Dynamic triaxial test; Geosynthetics; Dynamic elasticity modulus.
INTRODUCTION
Traffic load is a kind of special cyclic load with randomness. At present, the main design
method of reinforced soil subgrade engineering under traffic load is pseudo static method, which
is not perfect. This paper adopted the dynamic triaxial test, used the window screening as
geosynthetics to study the dynamic characteristics of reinforced soil under cyclic load. Through
this test to analyze the deformation behavior of reinforced soil subgrade, which would provide
the basis for the design of the reinforced soil under the cyclic load.
Experiment scheme
In this experiment, in order to make the soil sample generate stress and shear stress of cyclic
periodic change, the dynamic stress of periodic change was applied to the soil sample, while its
horizontal axial stress was in a static state.
ASCE

Geo-China 2016 GSP 259

211

Experimental material
Because the mesh of geogrid is largeif the geogrid was used as the reinforced material, the
accuracy of the test could not be guaranteed. In the dynamic characteristics study of the
reinforced soil, the scholars of our country have already used the window screening to replace
the geogrid and achieved the desired results. In addition, the physical parameters of the window
screening and geogrid are similar, so this experiment used the window screening to replace the
geogrid. Table 1 shows the physical parameters of the window screening.
Table 1. Physical parameters of the window screening
Tensile strengthkN/m
Modulus of
Thickness(mm)

elasticity(MPa)
0.755
0.4273
0.22
This experiment used the clay and silt of Qingdao area as the soil samples. Table 2 shows the
physical properties of the two soil samples.
Table 2. Physical properties of the clay and silt
Density
kg/m

Cohesive
strength
kPa

Internal
friction
angle

Poisson
ratio

Modulus of
elasticity
MPa

Plastic
limit

Plasticity
index

clay

1950

45

32

0.35

10

18

17

silt

1800

24

22

0.30

15

26.3

7.6

1.80

1.75

1.75

1.70

Dry Density(g/cm 3 )

Dry Density(g/cm 3 )

The test sample was a kind of remodeling soil sample, its optimum water content and
maximum dry density were determined by the compaction test. Test results are shown in Figure 1
and Table 3.

1.70

1.65

1.60

1.65

1.60

1.55

1.55

1.50

10

12

14

16

18

20

22

Water Content(%)

24

26

28

30

18

20

22

24

26

28

30

32

34

Water Content(%)

a) Clay
b) Silt
Figure 1. Relationship curves of dry density and water content of two kinds of soil samples

ASCE

Geo-China 2016 GSP 259

212

Table 3. The optimum moisture content and maximum dry density of soil samples
Soil sample
Optimal water content (%)
Maximum dry density(g/cm3)

Clay
17
1.80

Silt
25
1.73

Sample preparation
Prepared the clay and silt of corresponding water content according to the test procedure and
the optimum water content measured by compaction test. The material of the reinforced material
was round of 36mm in diameter, a slightly smaller diameter than the specimen, so as not to
puncture the rubber membrane. For the 1 layer reinforced specimen, the reinforced material was
placed in the middle. 2 layers reinforced specimen, the reinforced materials were placed at the
same distance. The soil surface of the reinforced material was treated by the plane to avoid the
separation between the soil layer, it can also increase the friction between soil and reinforced
material.
Dynamic triaxial test parameters
DDS-70 type microcomputer controlled electromagnetic vibration three axis instrument was
chosen for this test. In the experiment, the dynamic load was simulated by the sine load, and the
vibration frequency was 1.0Hz. According to the "Foundation dynamic characteristics test code"
(GB/ T50269-97), the sum of elastic deformation and plastic deformation is 5%, which is the
failure criterion of the specimen. The test operationed under the condition of consolidation and
undrainage, consolidation stress ratio Kc=1.0. Table 4 shows the test parameters.
Table 4. Dynamic triaxial test parameters
ConsoliType of soil
dation
ratio
Clay
50
100 150
0
1
2
1.0
1.0
Silt
50
100 150
0
1
2
1.0
1.0
As the default load unit of the dynamic triaxial test system is N, the unit need to be convert
firstly. Due to the isotropic consolidation, under three kinds of confining pressure, the maximum
loading it can bear as follows:
Confining
pressurekPa

Reinforcement
layer

a) When the confining pressure was 50kPa :


F = 0.8 50kPa

d2
4

= 0.8 0.05

3.14 39.12
N = 48N
4

b) When the confining pressure was 100kPpa :

ASCE

Frequenc
y(Hz)

Geo-China 2016 GSP 259

213

F = 0.8 100kPa

d2
4

= 0.8 0.1

3.14 39.12
N = 96N
4

c) When the confining pressure was100kPa :


F = 0.8 200kPa

d2
4

= 0.8 0.15

3.14 39.12
N = 144 N
4

According to the calculation results, Table 5 shows the maximum and minimum values of the
test load.
Table 5. The maximum and minimum of the dynamic load rating
Confining pressure
50
100
150
(kPa)
40
90
140
Maximum (N)
4
9
14
Minimum(N)
Installed the sample according to the test requirements after the sample preparation was
finished. Carried out the vibration test after the test system was zeroed and the test soil samples
were drained and consolidated. The vibration load of 1Hz was applied to the sample until it was
damaged.
Test result analysis
Effect of different soil on dynamic elastic modulus
160

50kPa Clay
50kPa Silt
100kPa Clay
100kPa Silt
150kPa Clay
150kPa Silt

140
120

80
60

120
100
80
60

40

40

20

20

0.00001

0.0001

0.001
d

0.01

aPlain soil

ASCE

140

E d(Mpa)

Ed (Mpa)

100

0.1

50kPa Clay
50kPa Silt
100kPa Clay
100kPa Silt
150kPa Clay
150kPa Silt

160

0.00001

0.0001

0.001
d

d1 layer

0.01

0.1

Geo-China 2016 GSP 259

214

180

50kPa Clay
50kPa Silt
100kPa Clay
100kPa Silt
150kPa Clay
150kPa Silt

160
140
120

E d(Mpa)

100
80
60
40
20
0

0.0001

0.00001

0.001

0.01

0.1

c2 layers
Figure 2. Relationship curve of dynamic elastic modulus and dynamic strain under
different soil conditions
From Figure 2, it can be seen that under the same conditions, the influence factors of the
reinforcement layer and the confining pressure were the same, the dynamic elastic modulus and
dynamic strain of silt and clay were decreased with the increase of the dynamic strain. At the
same level of dynamic strain, the dynamic elastic modulus of silty soil was higher than that of
the clay in the small strain period. The Ed~d curve of two soil samples gradually tended to be
consistent when the dynamic strain was greater than 0.01.
Effect of the number of layers on the dynamic elastic modulus
Plain soil 50kPa
One layer 50kPa
Two layers 50kPa
Plain soil 100kPa
One layer 100kPa
Two layers 100kPa
Plain soil 150kPa
One layer 150kPa
Two layers 150kPa

120

E d(Mpa)

100
80
60

180

Plain soil 50kPa


One layer 50kPa
Two layers 50kPa
Plain soil 100kPa
One layer 100kPa
Two layers 100kPa
Plain soil 150kPa
One layer 150kPa
Two layers 150kPa

160
140
120
100

E d(Mpa)

140

80
60

40

40
20

20

0.00001

0.0001

0.001
d

0.01

0.1

0.00001

0.0001

0.001
d

0.01

0.1

aClay
bSilt
Figure 3. Relationship curves of dynamic elastic modulus and dynamic strain under
different reinforcement conditions
From Figure 3, it can be seen that the dynamic elastic modulus of the reinforced soil and the
plain soil all reduced with the increase of the dynamic strain. The reinforcement had little effect
on the relationship between the dynamic elastic modulus and the dynamic strain. The effect of
improving the dynamic elastic modulus of embankment fill was not obvious.

ASCE

Geo-China 2016 GSP 259

215

Effect of confining pressure on dynamic elastic modulus

120

E d(Mpa)

100
80
60

180

Plain soil 50kPa


One layer 50kPa
Two layers 50kPa
Plain soil 100kPa
One layer 100kPa
Two layers 100kPa
Plain soil 150kPa
One layer 150kPa
Two layers 150kPa

160
140
120
100

Ed (Mpa)

Plain soil 50kPa


One layer 50kPa
Two layers 50kPa
Plain soil 100kPa
One layer 100kPa
Two layers 100kPa
Plain soil 150kPa
One layer 150kPa
Two layers 150kPa

140

80
60

40

40
20

20

0.00001

0.0001

0.001

aClay

0.01

0.1

0.00001

0.0001

0.001
d

0.01

0.1

bSilt

Figure 4. Relationship curves between dynamic elastic modulus and dynamic strain
under different confining pressures
From Figure 4, it can be seen that the relationship curves of dynamic elastic modulus and
confining pressure of different kinds of soil samples were similar: The dynamic elastic modulus
of soil samples increased with the increase of confining pressure under the certain condition of
dynamic strain amplitude. In small strain range, within d 10-3, the effect of confining pressure
on soil dynamic elastic modulus was significant. After the dynamic strain of d >10-3, the effect
of confining pressure on soil dynamic elastic modulus decreased obviously.
CONCLUSIONS
(1) The dynamic elastic modulus of plain soil and reinforced soil decreased with the increase of
the dynamic strain while increased with the increase of confining pressure. The effect of
reinforcement on the dynamic elastic modulus of embankment was small.
(2) Under certain conditions, such as the reinforcement and the confining pressure, the
relationship curves of dynamic elastic modulus and dynamic strain of silt and clay were the same.
Dynamic elastic modulus of silty soil was higher than that of clay in the case of small strain, the
Ed~d relationship curves of the two kinds of soil samples gradually tended to be consistent
when the dynamic strain was greater than 0.01.
REFERENCES
Sun Jin. and Bai Xiao-hua. and Zeng Guo-hong. (2006). "Dynamic triaxial testing study on
dynamic Elasticity Modulus of reinforced soil". Chinese Hydraulic Engineering. The first
session of China's water resources and hydropower in geotechnical mechanics and engineering
academic conference on (part ii).Chinese Hydraulic Engineering.Vol.2006 (3).905-907.
Yang Yan. and Bo Shu. and Zhang Jun.(2009). "Dynamic triaxial testing study on reinforced
soil." The subgrade engineering, Vol. 2009 (5):178-179.

ASCE

Geo-China 2016 GSP 259

Xu Zhen-hua. (2014)"The dynamic strength characteristics test and numerical simulation


research of Geosynthetic materials reinforced soil". Qingdao. Dissertation for the Master
Degree of Qingdao Technological University. 2014.45-51.
Chongqing highway science research institute of the ministry of transport. (2012).Technical
Specifications for Application of Geosynthetics in HighwayJTG/T D32-2012.Beijing. China
Communications Press.
Bao Cheng-gang. "Study on interface behavior of geosynthetics and soil" Chinese Journal of
Rock Mechanics and Engineering. 2006. 25(9).
Highway science research institute of the ministry of transport. (2007). Soil testing procedures in
HighwayJTG E40-2007.Beijing. China Communications Press.
Pan Yue. (2013). "The dynamic characteristics research of Geosynthetic materials reinforced soil
under cyclic loading". Qingdao. Dissertation for the Master Degree of Qingdao Technological
University. 2013.19-20.

ASCE

216

Geo-China 2016 GSP 259

217

Transitional Plasticity Compression Model for Clays


Gyan Vikash
Assistant Professor, Dept. of Civil Engineering, Shiv Nadar Univ., Chithera Dadri, Gautam Buddha
Nagar, Uttar Pradesh, India. E-mail: gyan.vikash@snu.edu.in

Abstract: Compression response of fine grained soils is generally modeled by bilinear


relationship between void ratio (e) and logarithmic of effective stress (). In
general, compression curve of overconsolidated soil is nonlinear in nature as it
varies smoothly from linear portion of initial part of the curve to the linear
portion of the later part of the curve. The bilinear model is unable to represent
the smooth transition of the compression curve. As a result, it leads to an
overestimation of the settlement and introduces a kink in the predicted stressstrain relationship of overconsolidated soil from the elasto-plasticity model
wherein it is used as a component defining volumetric hardening. The present
paper proposes a simple three parameters transitional plasticity model to predict
the compression response of clayey soil more accurately. The proposed model
considers that the irrecoverable deformation due to relative sliding between the
adjacent particle and particle groups has possibility from beginning of the
loading in over-consolidated region as well.
INTRODUCTION
The bilinear relationship between e and ln() is widely used to describe the
compression behavior of clayey soils. It has been also incorporated in the volumetric
hardening law of most of the elastoplastic constitutive theories for clays (e.g., CamClay family of models, such as Roscoe and Burland, 1968). Butterfield (1979) and
Wesley (1988) pointed out few conceptual shortcomings of this model which cause
gap with the physical phenomenon. To overcome the shortcomings of the classical
bilinear model, Butterfield (1979) proposed a bilinear relationship in ln(v) - ln()
space, where, v is the specific volume which defines as, v = 1+e. This relationship was
further confirmed by Haan (1992), Hashiguchi (1995) and Xu & Xia (2006). Though,
ln(v) - ln() relationship has been found more suitable than e - ln() relationship to
represent the compression behavior of clayey soils, this relationship is also unable to
capture the smooth transition of the compression curve from over-consolidated region
to normally-consolidated region.

ASCE

Geo-China 2016 GSP 259

218

Both the bilinear models consider that the soil experiences purely elastic
deformation during recompression and elastoplastic deformation when the stress state
moves on to the normal compression curve. Consideration of purely elastic
deformation during recompression leads to an assumption that the deformation in a
soil mass occurs only because of elastic deformation of individual clay particles or
particle groups during this stage. As a result, possibility of the rearrangement of
particles or particle groups due to relative sliding at their contact surfaces during
recompression is completely neglected. However, occurrence of hysteresis during
unloading-reloading stage, which is observed even at low stress level (Houlsby and
Wroth, 1991; Jardin et al., 1984) indicates the possibility of a small amount of plastic
deformation during unloading as well as reloading. Consequently, irrecoverable
deformation due to relative sliding between adjacent particles or particle groups,
though to an extent only, has possibility from beginning of recompression as well. In
such a phenomenon, total strain in a soil mass will be developed partly due to elastic
deformation of individual particles or particle groups and rest of it due to
rearrangement of particles or particle groups. Considering this mechanism, a new
constitutive law for compression response of clayey soil has been proposed in the
following section.
PROPOSED COMPRESSION MODEL
Proposed compression model considers that the clayey soil experiences both elastic
as well as plastic deformation from beginning of the loading irrespective of whether
the stress state is in recompression zone or normally consolidation zone. The total
incremental strain (d) has been considered as summation of the elastic component of
strain (de) and the plastic component of strain (dp).
=

(1)

Elastic Strain
Assuming the elastic behavior of the soil to be linearly pressure dependent
corresponding to a linear relationship in ln(v) - ln() space, the elastic component of
strain, de,, can be expressed in incremental form as,

(2)

where, e is the elastic compressibility. In case of laterally confined compression, the


ratio of axial stress to axial strain is defined as the constrained modulus, D. Hence,
using Eq. 2, D can be expressed as,

ASCE

(3)

Geo-China 2016 GSP 259

219

Considering soil to be isotropic and the Poissons ratio () constant, following


relationship between the constrained modulus, D, and Youngs modulus, E, can be
established.

(
(

)(

(4)

Hence, from Eqs. (3) and (4), E can be further expressed as,
=

( )( )

(5a)

where,

( ) =

)(
(

)
)

(5b)

This relationship can be regarded as a form of Hooks law of which the modulus of
elasticity, E, is stress state dependent.
Evolution of Plastic Strain during Virgin Loading
Figure 1 presents schematic representation of compression response of the clayey
soil. In this figure, line ab represents compression response of the soil during
monotonic loading from its initial structure formation, i.e., the soil has been never
unloaded-reloaded in the past. This line can be defined as virgin compression line
(VCL). Curve cdef is the compression curve of the same soil which was unloaded
earlier from point g located on VCL to point c. The stress state at point d is defined as
preconsolidation pressure; hence cd can be called as recompression curve. The curve
def represents compression response of soil beyond preconsolidation pressure, which
can be defined as normal compression curve out of which a linear portion at high
stress level (ef) is often defined as normal compression line (NCL) in practice. During
unloading-reloading, the soil experiences certain amount of plastic deformation (gd),
although the stress state may still remain below preconsolidation pressure. This plastic
deformation may result into different spatial arrangement and enlargement of contact
areas of particles or particle groups compared to the virgin loaded soil at the same
stress level. It will carry its impression even if the soil is in normal-consolidation zone,
which leads to reduced amount of the plastic deformation along NCL (gg) than along
VCL (dd) for the same amount of stress increment. The significance of that
impression, however, may reduce with the increase of stress level. As a result, the
loading curve becomes asymptotic to VCL but never actually reaches the same. In the
proposed model, VCL is assumed to be linear in ln(v) - ln() space, which can be
expressed as
ln = ln

ln

(6)

where, n is the slope of VCL, and vn is the value of v on VCL at = 1 kPa. As along
VCL, the soil experiences utmost evolution of incremental plastic strain for a given

ASCE

Geo-China 2016 GSP 259

220

stress incremen
nt, thereforee VCL can be treated as the currve definingg maximum
m
possible evoluttion of plasttic strain an
nd the bounddary of all possible staates for 1-D
D
com
mpression.

FIG. 1. Schemattic curve forr compressiion responsee of fine graained soil


Thee plastic straiin along VC
CL, dpmax, att stress levell for a givven stress inccrement d
can be expresseed in incremeental form ass
=(

(7))

Evo
olution of Pllastic Strain
n during Reecompressioon
The
T plastic strain comp
ponent of the soil whiich underweent through unloading-relo
oading stages in the pastt can be described as a function of stress ratio i = / n,
thatt maps the current
c
stresss state to th
he corresponnding state located on VCL at thee
sam
me specific volume.
v
A material
m
parameter has been introdduced to conntrol the ratee
of elastoplastic
e
c transition of compresssion responnse of the soil. The pllastic strainn
com
mponent in in
ncremental form
f
is defin
ned as,
= (

ASCE

) ( , ))

(8))

Geo-China 2016 GSP 259

221

Considering the mechanism of evolution of the plastic strain explained earlier,


function f should be chosen such that it should lies between 0 and 1. The function f
will tend to 0, at low stress level where the contribution of plastic strain is much
smaller, and it will tend to 1 as the stress state approaches towards VCL. In the
proposed model, function f is defined as
( , ) = ( ) (9)

n will be required to determine the value of i at a given stress level, which can be
determined with Eq. (6) as

(10)

Further, using Eqs. (7), (8) and (9), the plastic strain increment dp as a function of
dpmax can be expressed as
= (

)( )

= ( )

(11)

Eq. (11) yields: for < 1, dp > dpmax, for = 1, dp = dpmax and for > 1, dp < dpmax.
As the plastic strain increment dp for the soil which underwent through unloadingreloading stages in the past, should always be less than dpmax, therefore should
always be greater than unity. Figure 2 shows the variation of eqp with i and , where
eqp = (n-e)(i)1. When i tends to zero, eqp will also tend to zero and hence the
contribution of plastic strain component will be negligible. On the other side, when i
tends to one, eqp will also tend to (n-e) and dp will be close to dpmax. Figure 2 also
depicts that the rate of elastoplastic transition is gradual for smaller value of and the
transition gets sharper as the value of increases.
MODEL EVALUATION
Determination of Input Parameter
The proposed model has been evaluated for kaolin clay specimen prepared by
consolidating the slurry of kaolin clay at 200 kPa. The Kaolin clay used in this study
has liquid limit of 65%, plastic limit of 30%, plasticity index of 35%, and specific
gravity of 2.60. The proposed model requires three input parameters (e, , and n) to
describe 1-D compression response of the soil. All the defined parameters in the
proposed model can be determined from a single 1D consolidation test.
As e represents elastic compressibility of the soil, hence the test which is used to
determine the soil parameter at small strain level will be required to determine e.
However, this is not always possible in usual practice. As the contribution of the
plastic strain component in the total strain at the beginning of the loading is very
small, hence for most of the engineering application, it is reasonable to consider the
initial linear portion of recompression curve to be elastic. Therefore, e can be
determined as the slope of the initial linear portion of the recompression curve. The
value of e for the present test specimen is obtained to be 0.0030. The parameter n,

ASCE

Geo-China 2016 GSP 259

222

slop
pe of VCL, has been deetermined as the slope oof the asympptote drawn on the laterr
partt of the com
mpression cu
urve, which is
i obtained to be 0.081 for the giveen soil. Thee
elasstoplastic paarameter has
h been obttained from best fit currve of the eexperimentall
dataa and for thee given soil itt is obtained
d to be 2.

FIG.
F
2. Varriation of eeqp with i an
nd

FIG. 3. Determina
ation of n aand vn for th
he given soill

ASCE

Geo-China 2016 GSP 259

Model Predictiion

nse of the ggiven soil through thee


FIG. 4. Prediction of compresssion respon
d model
proposed
Fiigure 4 show
ws the prediccted compresssion responnse through tthe proposedd model andd
the experimentaally observed response of
o the givenn soil. Predicction has beeen done forr
the model paraameters deteermined abo
ove. This fi
figure illustrrates that thhe proposedd
del accuratelly predicts th
he experimeentally obserrved responsse. The preddiction usingg
mod
the proposed model
m
indicaates that thee proposed model is ccapable to ddescribe thee
smo
ooth transitio
on of the com
mpression cu
urve from ovver consoliddated region to normallyy
con
nsolidated reg
gion.

ASCE

223

Geo-China 2016 GSP 259

CONCLUSIONS
In the present paper, a new compression model was proposed to capture the
compression response of clayey soils. The proposed model considers that the
elastoplastic deformation occurs from the beginning of the loading even if the current
stress is much lower than the past stress. The evolution of the plastic strain in the soil
that underwent through unloading-reloading stages in the past was described by a
simple mapping rule. A material parameter was introduced to control the rate of the
elastoplastic transition. All the parameters of the proposed model were readily
estimated by a single 1-D consolidation test. The predicted compression response of
the given soil through the proposed model showed its capability of representing the
smooth transition of the compression curve from linear portion of initial part of the
curve to linear portion of later part of the curve.
REFERENCES
Burland, J. (1989). Small is beautiful - the stiffness of soils at small strains.
Canadian Geotechnical Journal., 26, 499-516.
Burland, J. (1990). On the compressibility and shear strength of natural clays.
Geotechnique., 40(3), 329-378.
Butterfield, R. (1979). A natural compression law for soils (an advance on e-log p).
Geotechnique., 29(4), 466-480.
Haan, E. J. D. (1992). The formulation of virgin compression of soils.
Geotechnique., 42(3), 465-483.
Hashiguchi, K. (1995). On the linear relations of V - ln p and ln v - ln p for isotropic
consolidation of soils. International Journal For Numerical And Analytical
Methods in Geomechanics., 19, 367-376.
Hardin, B.O. (1989). 1-D strain in normally consolidated cohesive soils.
Geotechnique., 115(5), 689-710.
Houlsby, G.T., and Wroth, C.P. (1991). The variation of shear modulus of a clay with
pressure and overconsolidation ratio. Soils And Foundations, 31(3), 138-143.
Jardine, R.J., Symes, M.J., and Burland, J.B. (1984). The measurement of soil
stiffness in the triaxial apparatus. Geotechnique., 34(3), 323-340.
Oikawa, H. (1987). Compression curve of soft soils. Soils And Foundations, 27, 99104.
Roscoe, K. H. and Burland, J. B. (1968). On the generalized stress strain behavior of
wet clay. Engineering plasticity (eds J. Heyman and F. A. Leckie), pp. 535- 609.
Cambridge: Cambridge University Press.
Wesley, L.D. (1988). Compression Index: Misleading parameter? Journal of
Geotechnical Engineering, ASCE., 114(6), 718-723.
Xu, Y., and Xia, X. (2006). Fractal model for virgin compression of pure clays.
Mechanics research communication, Elsevier. 33, 206-216.
.

ASCE

224

Geo-China 2016 GSP 259

225

Numerical Behavior of Reinforced Soil by Rigid Inclusion


Samia Boussetta1; Mounir Bouassida2; and Mondher Zouabi3
1

Universit de Tunis El Manar, Ecole Nationale dIngnieurs de Tunis, BP 37 Le Belvdre, 1002


Tunis, LR14ES03. E-mail: samia.boussetta@gmail.com
2
Universit de Tunis El Manar, Ecole Nationale dIngnieurs de Tunis, BP 37 Le Belvdre, 1002
Tunis, LR14ES03. E-mail: mounir.bouassida@fulbrightmail.org
3
Universit de Tunis El Manar, Ecole Nationale dIngnieurs de Tunis, BP 37 Le Belvdre, 1002
Tunis, LR14ES03. E-mail: mondher.zouabi@mines-douai.fr

Abstract: This paper aims to assess the numerical predictions of settlement reduction
of compressible soil reinforced by rigid inclusion with experimental results obtained
from experiments conducted in calibration chamber. Loading tests carried out in
calibration chamber are simulated in axisymmetric condition. Numerical computation
focused on settlement variation both for the unreinforced soil and reinforced soil. It
was found the main component of settlement of reinforced soil is quickly stabilized
after loading compared to the settlement of unreinforced soil. Further, the role of
efficiency has been highlighted in showing the benefit of rigid inclusion technique.
INTRODUCTION
The principle of rigid inclusion technique consists of reinforcement of compressible
soils by end-bearing vertical inclusions. Each rigid inclusion is head-covered by mini
slab enabling the concentration of induced vertical stress on inclusions. Rigid
inclusion technique comprises three main components: mattress layer, analogic soil
and the rigid inclusion (Brianon 2002). Numerical modeling of scaled test model is
considered in regard to the behavior of mattress layer, the analogical compressible soil
and the rigid inclusion. The numerical study aims to assess the settlement and
efficiency predictions compared to measurements recorded during applied uniform
stress and imposed displacement loadings on the scaled test model. On the basis of
such a validation the behavior of foundations on reinforced soil by rigid inclusions can
be predicted.
In the framework of ASIRI project (2013), two experimental investigations in
laboratory were conducted at CERMES (ENPC, France) by Dinh (2009) and
Boussetta (2013). Those contributions were devoted to quantify the reduction of
settlement due the reinforcement of compressible soil by rigid inclusions after loading
tests performed in calibration chamber. Experimental investigations also served for
parametric studies to highlight the effect of thickness of mattress layer and the cover

ASCE

Geo-China 2016 GSP 259

226

ratio with respect to the prediction of efficiency of rigid inclusions technique in


settlement reduction. Further, focus is made on the comparison between two loading
conditions: applied uniform stress (embankment loading) and imposed displacement
(rigid foundation). The rigid foundation loading revealed better than embankment
loading.
This paper focuses on the prediction of numerical behavior of tests conducted in a
calibration chamber that served for the quantification of settlement reduction and gain
in efficiency due to the reinforcement of compressible soil by rigid inclusion.
Numerical results are presented and, then, interpreted to explain the expected benefits
from the rigid inclusion reinforcement.
INVESTIGATED NUMERICAL MODELS
Numerical simulation has been performed by Plaxis software V9.2D that became of
extensive use for the modeling of various geotechnical applications.
The numerical simulation of the composite cell made up of a compressible soil, rigid
inclusion and mattress layer is conducted in axisymmetric condition.
The justification of constitutive law and inherent geotechnical parameters for the
simulation of constituents of the physical model, i.e. load transfer mattress, analogic
oil and rigid inclusion, is made on the basis of experimental results proposed by Dinh
(2009).
Modeling of the mattress layer for load transfer is described by the Hardening Soil
Model (HSM). Model parameters of HSM were determined from triaxial tests results
as suggested by Brinkgreve and Vermeer (1998). Table 1 presents the geotechnical
parameters of the constitutive material of the mattress layer M1.
Table 1. Geotechnical parameters adopted for model the material M1 (HSM)
c (kPa)

()

()

E (MPa)

(kN/m3)

36

11.4

16.2

C = cohesion; = friction angle; = angle of dilatancy; E = Young modulus; = unit


weight.
The behavior of compressible soil is described by the Soft Soil Model (SSM). Table 2
shows the oedometer parameters adopted for the compressible soil.
Table 2. Oedometer characteristics of compressible soil (SSM)
e0
2.44

Cc
0.35

Cs
0.08

p (kPa)
15

e0 = initial void ratio; Cc = compression index; Cs = swelling index; p = preconsolidation pressure.


Numerical investigation addresses two models: the unreinforced soil as the case of
reference and of reinforced soil to show the benefits of rigid inclusions technique. The
unreinforced soil was simulated by the material M1 for the mattress layer of thickness
hm = 10 cm and the compressible soil SP30 of thickness 10.5 cm. Figure 1 illustrates

ASCE

Geo-China 2016 GSP 259

227

the axisymmetric model of calibration chamber to simulate the behavior in case of


reinforced soil by rigid inclusion.
This model of reinforced soil comprises the mattress layer made up of soil type
material M1, of thickness hm = 10 cm, the compressible soil (SP3) of thickness 10,5
cm and. The recovery rate, defined by the ratio of rigid inclusion area to that mattress
layer, is equal to 2.22 %.

Axis of
symmetry

Cell border

Mattress for load


transfer
Compressible
soil

Rigid Inclusion

Interfaces

FIG.1. Axisymmetric model of calibration chamber with reinforced soil


For the sake of simplicity, the numerical model assumes equal diameter for the rigid
inclusion and the head of inclusion. Remaining parameters in regard to the transfer
mattress, compressible soil and cell border are identical to those of unreinforced soil
case. The characteristics adopted for rigid inclusion, modelled as Mohr-Coulomb
material, are: C = 0.1 kPa; =5; E = 100 GPa; = 80 kN/m3; = 0.2 .
The cell border is also simulated to analyze its influence using interface element.
The metallic cell border is modeled as for the rigid inclusion described above.
Interfaces are characterized by parameter Rint in the range from 0 to 1 with case of
reference that is the unreinforced soil. As the parameter Rint increases the interface is
more rigid. Those interfaces should be introduced as rigid element when Rint = 1.0.
Hence, interface properties including the dilatancy angle, i, excepting Poissons ratio
i; are identical to those of compressible soil. Shear strength of interface is defined as:
Cint er = Rint er .cmat

int er = Rint er .mat

(1a)
(1b)

Cmat and mat denote the cohesion and friction angle of mattress layer respectively.
Real interactions between soil and structural elements, for soil-steel interface it is
recommended to take Rint 0.5 because of the weak stiffness of soil compared to that
of rigid inclusion.

ASCE

Geo-China 2016 GSP 259

The top of mattress layer is loaded by increased uniform vertical stress from 0 to 100
kPa. Deformed mesh that resulted from uniform applied load of 100 kPa is shown in
Figure 2.

FIG. 2. Deformed mesh of unit cell model subjected to 100 kPa uniform pressure
NUMERICAL RESULTS
Figure 3 shows the variation of settlement predicted at the unit cell axis and
soil/mattress interface. Significant increase in rate of settlement is observed up to 60
kPa load then the variation of settlement is stabilized from 80 kPa.

FIG. 3. Settlement variation at soil-mattress interface vs applied uniform stress


Figure 4 shows the variation of settlement at soil/mattress interface as a function of the
unit cells radius. Settlement is constant over the unit cell radius, then, it abruptly
decreases at the cell border that is assumed as rigid material.
ASCE

228

Geo-China 2016 GSP 259

FIG. 4. Settlement variation at soil-mattress interface (material M1)


Figure 5 shows the settlement distribution at soil/mattress interface of reinforced soil
along the radius of composite cell composite. Zero settlement is noted on the rigid
inclusion, the settlement increases up to 6 mm and, then, attains 7.8 mm. When
approaching the cell border that behaves similarly to rigid inclusion the settlement
decreases up to 5 mm (see Figure 3).

FIG. 5. Distribution of settlement at soil/mattress interface of reinforced soil

ASCE

229

Geo-China 2016 GSP 259

230

Figure 6 compares between the recorded variations of settlement in function of


applied load for the unreinforced soil and reinforced soil models. First, very
significant decrease of settlement (from 14.7 to 7.8 mm) is noticed due to the presence
of rigid inclusion and mattress layer. Second, the settlement of reinforced soil by rigid
inclusion is not marked by the threshold observed in case of unreinforced soil when
the applied load reaches 100 kPa.

FIG. 6. Variation of settlement versus applied load for unreinforced and


reinforced soil test model.
PREDICTION OF EFFICIENCY
The applied force at the head of inclusion, Finc, is calculated from Eq (2):
Finc = Fi = i Ai =

(r

2
i 1

yyi + yyi 1

(2)

yyi : is the computed vertical stress at inclusion rigid at measurement point N i;


Ai: denotes the area attributed to measurement point N i, this crown area equals:
Ai =

(r

ri 21

(3)

ri: is the radius of composite cell at measurement point N i.


The efficiency is, then, calculated from the resultant force over the head of inclusion
from Eq (4):
F
Finc
Eeff ( % ) = inc =
.100 ( % )
(4)
Ftot ( p + hm ) Atot

Atot denotes the area of cross section of calibration chamber subjected at the top of
mattress layer to uniform pressure p.
This resultant force is balanced by the distributions of vertical stress at the head of
rigid inclusion and the shear stress exerted on cylindrical border of rigid inclusion

ASCE

Geo-China 2016 GSP 259

231

(Figure 7). Those stress components are taken along section A-A* sketched in figure 8
at -2 cm depth with respect to upper level of rigid inclusion.
The variation of efficiency, as calculated from Eq (4), versus applied load is shown
in Figure 9. It is noticed beyond 80 kPa the efficiency becomes constant. The variation
of efficiency, as calculated from Eq (4), versus applied load is shown in Figure 9 (hs
denotes the thickness of compressible soil). It is noticed beyond 80 kPa the efficiency
becomes constant.
Uniform applied stress
2 cm

Shear stress

Rigid inclusion
FIG. 7. Scheme of stresses distribution around the head of rigid inclusion

FIG. 8. Distribution of vertical stress over the head of rigid inclusion

FIG. 9. Variation of efficiency versus applied load

ASCE

Geo-China 2016 GSP 259

This predicted efficiency from Figure 9 leads to conclude that almost of the load
transfer has occurred on the rigid inclusion. Hence the settlement of reinforced soil is
significantly decreased. As such the expected role of reinforcement using the rigid
inclusion technique is beneficial.
CONCLUSIONS

The numerical modeling of tests performed on soil reinforced by rigid inclusion in


calibration chamber has been investigated. Numerical study has been conducted in
axisymmetric condition by using Plaxis V9.2D.
It has been found the numerical predictions are in good agreement with recorded
measurements in terms of efficiency and settlement reduction. Therefore adopted
geometrical and physical parameters of the numerical modeling are validated.
ACKNOWLEDGMENTS

The authors appreciate the financial support provided by the French-Tunisian


Committee during the research project Studying the reinforcement of soft soils by
inclusions N CMCU n 07G 1001.
REFERENCES

Boussetta S. (2013) Etude des mcanismes de transfert de charge dans les sols
renforcs par inclusions rigides laide dun modle physique. Thse de doctorat
en gnie civil. Ecole Nationale dIngnieurs de Tunis. Tunisia.
Boussetta S., Bouassida M., Dinh A.Q., Canou J., Dupla J.C., (2012). Physical
modeling of load transfer in reinforced soil by rigid inclusions. Int. J. of Geotech.
Eng. (6), 331-341.
Boussetta S., Dinh A.Q., Canou J., Dupla J.-C. et Bouassida M. (2010). Etude
exprimentale sur modle physique dun sol compressible renforc par une
inclusion rigide. 2nd Int. Conf. on Geotech. Eng. October 25-27, 2010, Tunisia:
255-264.
Brianon, L. (2002). Renforcement des sols par inclusions rigides Etat de
lart en France et ltranger. IREX, Paris, 185 p.
Brinkgreve R.B.T., and Vermeer, P.A., (1998) Plaxis- Finite Element Code for Soil
and Rocks Analysis. Version 8, AA. Balkema, Rotterdam Brookfield.
Dinh A.Q. (2009). Etude sur modle physique des mcanismes de transfert de charge
dans les sols renforcs par inclusions rigides. Application au dimensionnement.
Thse de doctorat. Ecole des ponts Paris-Tech. France.

ASCE

232

Geo-China 2016 GSP 259

233

Study of the Behavior of Tunis Soft Clay


Mnaouar Klai1 and Mounir Bouassida1
1

Universit de Tunis El Manar Ecole Nationale dIngnieurs de Tunis. LR14ES03-Ingnierie


Gotechnique, BP 37 Le Belvdre, 1002, Tunis, Tunisia.

Abstract: The paper briefly reviews research investigations conducted on Tunis soft
clay that is classified as problematic soil. Results obtained from an experimental study
carried out on undisturbed Tunis soft clay specimens are presented and interpreted. On
the basis of experimental results the paper discusses which constitutive law can
describe at best the observed behaviour of Tunis soft clay. The elastoplastic behaviour
modelled by the hardening soil model (HSM) is then justified upon the validation of
numerical results of oedometer and triaxial tests carried out on undisturbed soft clay
specimens.
Keywords: Behavior; Characterization; Hardening; Numerical; Simulation; Soft clay.
INTRODUCTION
The soil profile of Tunis City mainly consists of a layer located between 3 and 20m
depth constituted by saturated greyish sandy clay, which is at the origin of the
contamination observed on several constructions built on this weak soil. This soil
commonly called the Tunis soft clay (TSC) is very problematic because of the
difficulty to extract undisturbed specimens for performing laboratory tests. Besides,
performing in situ tests sometimes leads to unrealistic data due to its very low stiffness
compared to that of expanded membrane to measure the limit pressure during
pressuremeter tests.
Bouassida (2006) reported the difficulty in predicting the undrained cohesion of TSC
from in situ vane shear tests due to unreasonable interpretation of these results. In
parallel, the use of reconstituted TSC to avoid disturbance of specimens does not
reflect the actual behaviour of in situ soil (Klai & Bouassida, 2009). An overview on
geotechnical parameters of TSC and related correlations were suggested by
(Bouassida and Klai, 2012). In this paper, a comparison was made between the
characteristics of reconstituted and undisturbed TSC.
Relevant contribution on numerical modeling of TSC was proposed by Tounekti et al
(2008). Those authors assessed the usefulness of soft soil model (SSM) as suitable
constitutive law for the remolded Tunis soft clay after comparisons between numerical
results (simulation of oedometer and triaxial tests) and measurements from performed
tests in laboratory. The behavior of two geotechnical infrastructures has been
ASCE

Geo-China 2016 GSP 259

discussed on the basis of numerical simulation where the SSM was adopted for TSC.
The present paper focuses on the study of behavior of TSC as observed from
experimental investigation conducted in laboratory. A set of identification tests,
oedometer and triaxial tests has been performed on samples extracted during
geotechnical investigation in Tunis City. From experimental data the soil parameters
of hardening soil model (HSM) and modified Cam Clay (CCM) constitutive law are
determined and, then, used as input data to simulate oedometer and triaxial tests. The
validation of those constitutive models is discussed based on the comparison made
between experimental and numerical results.
GEOTECHNICAL INVESTIGATIONS: SAMPLING AND LABORATORY
TESTS
In the urban area of Tunis City two bore holes namely BH1 and BH2 spaced of 10 m
were executed at the Avenue de la Rpublique. Cored specimens namely CS1 and
CS2 have been extracted respectively at 7.5 m and 9.5 m depths by a double rotary
driller of external diameter 101 mm.
- BH1 soil profile shows an upper fill layer of 7 m thickness overlaying the Tunis soft
clay layer of about 18 m thickness. Three undisturbed cored specimens (specimen 1,
specimen 2, and specimen 3) have been extracted at depths of 7.55 m, 9.85 m and
18.35 m respectively.
- BH2 soil profile shows a similar formation as that observed in BH1. Thickness of
the upper fill layer is 2.5 m. Two cored specimens (specimen 4 and specimen 5) have
been extracted at depths of 3.75 m and 7.75 m respectively.
Undisturbed samples are cored in PVC tubes of 101 mm external diameter, logged in
the rotary driller gently penetrated within soft clay layer at displacement rate of about
10 mm/minute. Extracted PVC tubes are then placed in wood boxes and transported
from the site to laboratory so that shocks are prevented.
In laboratory, undisturbed soil specimens are extracted by penetrated thin cutting shoe
in the direction of in-situ extraction. Therefore, from extracted cutting shoe, soft soil
specimens are ready for laboratory tests. Laboratory tests have been carried out at the
soil mechanics laboratory of the Higher Institute of Technological Studies of Rades
(Tunis). The soil identification tests included: grain size distribution (sieve and
hydrometer), total unit weight, specific gravity, Atterberg limits and content of organic
wastes (OM). The second group of tests included oedometer tests (compressibility and
consolidation), consolidated undrained (CU) triaxial tests and consolidated drained
(CD) triaxial tests.
Experimental investigation: tests and results
Classification tests
As part of soil identification wet sieve and hydrometer analyses were performed on
five undisturbed soft clay specimens. Grain size distributions show the average fines
content (grain size < 0.08 mm) is about 87% (Klai, 2014). Table 1 summarizes the
identification parameters of the five undisturbed soft clay specimens. The
classification of saturated Tunis soft clay is highly plastic silt with very low
consistency. For undisturbed soft clay specimens, which contain wastes of shell,

ASCE

234

Geo-China 2016 GSP 259

235

Atterbergs limits values are lower than those obtained for the reconstituted Tunis soft
clay (Bouassida, 2006).
Table 1. Identification parameters of undisturbed Tunis soft clay
Specimen N
1
2
3
4
5

Specific gravity
2.62
2.50
2.53
2.32
2.39

[kN/m3]
17.4
16.1
18
17.6
16.9

WL
46
50
51
65
79

Ic
0.31
0.50
0.72
0.50
0.50

Ip
19
5
9.5
15
29

WL = liquid limit; Ip = plasticity index; Ic = consistency index; = Unit weight


Some properties can also help for a better identification of soft clays like chemical
tests for the determination of content of organic wastes and the calcium carbonate
provides useful information about the compressibility and strength (Das, 2006). The
recorded percentage of organic content of reconstituted Tunis soft clay was about 3.12
%. Undisturbed soft clay has a higher organic content than the reconstituted soft clay
which confirms its low compressibility of about 10%.
Oedometer tests
Referring to Table 2 undisturbed soft clay specimens n1, n2, n3, n4 and n5
extracted at average depth of 8.5 m are classified as under consolidated. The preconsolidation stress of tested specimens is lower than the effective vertical stress at
extraction depth that varied from 52 kPa to 180 kPa. Compression and swelling
indices indicate that undisturbed Tunis soft clay has lower compressibility and
swelling than those of reconstituted soft clay (Klai and Bouassida, 2009). Meanwhile
recorded values of compression index are in accordance with those reported by Touiti
et al, (2009) from other geotechnical investigations data conducted on Tunis soft clay
undisturbed specimens, i.e. 0.4 Cc 0.6.
Table 2. Oedometer characteristics of undisturbed Tunis soft clay
Specimen N
1
2
3
4
5

Cc
0.43
0.485
0.35
0.385
0.384

Cs
0.057
0.056
0.057
0.057
0.057

p (kPa)
12
25
17
14
14

Cc = compression index; Cs = swelling index; p = pre-consolidation pressure


CU & CD Triaxial tests
The drained friction angle of tested specimens is found in the range = 19.2 - 23.7.
The drained cohesion is not very significant, it does not exceed C= 5 kPa (Table 3).

ASCE

Geo-China 2016 GSP 259

236

The inherent over-consolidation of tested specimens is more likely attributed to the


applied consolidation stress in triaxial test (up to 300 kPa) that exceeds the in situ
effective vertical stress of extracted specimens (depth is less than 20 m).
Table 3. Shear strength parameters of undisturbed Tunis soft clay
Specimen N
1
2
3
4

Ccu (kPa)
7.53
8.49
8.67
7.79

C(kPa)
5.0
4.0
5.1
3.6

()

22.7
23.7
20.8
19.2

Justification of the Hardening Soil Model (HSM) for Tunis Soft Clay
The HSM is selected to simulate the behaviour of Tunis soft clay since it is capable to
account for the increase in stiffness due to consolidation stress. Such parameter is
essential for the modelling of foundation that extends to relatively deep soil layers for
example underneath an embankment. Zimmerman et al (2010) recommended the
adoption of the standard HSM for normally consolidated soft clays. Relationships
between the parameters of the HSM are:
ref
ref
E ref
= 100 kPa , K 0nc = 1 sin R f = 0.9 t = 0 m = 1 ; = 0.
ur = 3E50 ; ur = 0.35 ; P
;
;
;

From recorded experimental data the input parameters of HSM adopted for Tunis soft
clay layer are given in Table 4.
Table 4. Hardening soil model parameters of Tunis Soft Clay
Specimen N
(extraction depth)
1 (7.2 to 7.9 m)
2 (9.5 to 10.2 m)
3 (18 to 18.7 m)
4 (3.3 to 4.0 m)

ref
E oed

(kPa)

1337
1186
1643
1494

ref (kPa)
E 50

1672
1482
2054
1867

C(kPa)

()

5.0
4.0
5.1
3.6

22.7
23.7
20.8
19.2

Numerical investigation is performed to simulate the oedometer and triaxial tests


carried out on TSC specimens. Aside from the HSM, the Modified Cam-Clay (CCM)
model is also considered to characterize the TSC for the purpose of numerical
predictions. The geotechnical parameters of the Modified Cam Clay model considered
for undisturbed TSC specimens extracted at the Avenue Mohamed V at depths from 3
to 20 m are: the slope of compression line: = 0.21; the slope of unloading-reloading
line: = 0.024; initial void ratio: e0 = 1.55; isotropic hydraulic conductivity: k =
1.7310-6m/day; Poissons ratio: = 0.35; drained cohesion: C = 13 kPa, drained
friction angle: = 20.8; saturated unit weight: sat = 17.4 kN/m3.

ASCE

Geo-China 2016 GSP 259

Simulation of observed behaviour of TSC


The simulation of oedometer and triaxial tests is conducted by using the software
Plaxis V9.2D in axisymmetric condition due to the cylindrical geometry of tested
specimens and applied loading. Quarter of the specimen (radius equals 17 mm; height
equals 35 mm) is considered for numerical simulation.
Oedometer tests
Numerical computations are run by Plaxis software with the assumed HSM and the
CCM input parameters. Figures 1 and 2 compare experimental data with numerical
simulation results obtained by the HSM and CCM for TSC specimens n 1 and n2.
Figures 3 and 4 show experimental and numerical results predicted by the HSM.

FIG. 1. Predicted behaviour of TSC modeled by the HSM and CCM and
experimental measurements from oedometer test (specimen 1)

FIG. 2. Predicted behaviour of TSC modeled by the HSM and CCM and
experimental measurements from oedometer test (specimen 2)

ASCE

237

Geo-China 2016 GSP 259

238

Interpretation of results
From Figures 1, 2, 3, and 4 it is noted that the numerical prediction by the HSM
during the primary consolidation phase is overall in accordance with the observed
behaviour on tested specimens. In turn, significant difference is noticed between
experimental data and the numerical predictions obtained by the Modified Cam Clay
model (CCM) that overestimates the predicted decrease in void ratio. The difference in
drained cohesion between the HSM (C = 3 to 5 kPa) and the CCM (C = 13 kPa) can
also explain the lower compression observed in oedometer test due to the higher overconsolidation considered by the CMM compared to that of HSM.

FIG. 3. Predictions by the HSM of TSC behaviour compared with data from
oedometer test (specimen 3)

BH2 Specimen4 (3.3-4 m)


BH2 Specimen5 (7.3-4 m)
HSM
CCM

1.4
1.2

Void ratio

1.0
0.8
0.6
0.4
0.2
0.0

11

10

100

Effective stress (kPa)


FIG. 4. Predicted behaviour of TSC modelled by the HSM and CCM and
experimental measurements from oedometer tests (specimens 4 & 5)

ASCE

Geo-China 2016 GSP 259

During the unloadingreloading phase in Figures 1 to 4 the numerical prediction by


the CCM slightly underestimates the swelling of specimens, whilst the HSM shows a
good agreement with experimental measurements. The overestimated consolidation by
the CCM is essentially owed to the parameters and which represent respectively
the slopes of the oedometer curve both in consolidation and during unloading reloading phases of the specimens of Tunis soft clay.
Triaxial tests
Figure 5 shows the numerical predictions of deviatoric stress versus axial strain, as
predicted by the HSM, for various isotropic consolidation stresses as well as
experimental measurements during the shear phase of CU triaxial test performed on
specimen 4.

FIG. 5. Experimental and numerical results during shear loading of CU triaxial


test (specimen 4)
The observed behaviour during shear loading is overall in fair agreement with
numerical results predicted by the HSM. This leads to the conclusion that the adopted
failure parameters (C' and ) are quite representative of the observed behaviour of
undisturbed TSC specimens. Using Plaxis software (version 9.2) the simulation of
observed behaviour of those specimens subjected to oedometer and triaxial tests
showed that the HSM predictions are in good agreement with measured data rather
than predicted results as obtained by the Modified Cam-Clay model (Klai, 2014). For
this reason the HSM can be considered to model the TSC for the prediction of
behaviour foundations built on Tunis soft clay and subjected to vertical loading.
CONCLUSIONS
This paper discussed the behaviour of Tunis soft clay as observed during an
experimental investigation carried out on undisturbed specimens. Then, the simulation
of performed oedometer and CU triaxial tests, using Plaxis 2D software, has been
considered. Two constitutive behaviour laws were tested to model the Tunis soft clay:

ASCE

239

Geo-China 2016 GSP 259

the hardening soil and modified Cam Clay models (HSM and CCM). Comparisons
between numerical results from the simulation of oedometer and triaxial tests favoured
the adoption of the HSM for TSC in order to predict the behaviour of structures
founded on typical soil profile comprising the soft clay layer when subjected to
vertical loading. Recently, Klai et al (2015) investigated on the numerical predictions
of behaviour using the hardening soil and Mohr-Coulomb models of tank resting on
Tunis soft clay reinforced by sand columns to highlight the significant reduction in
settlement.
REFERENCES
Bouassida M. and Klai M. (2012). Challenges and Improvement Solutions of Tunis
Soft Clay. Int. Journal of Geomate, Sept. 3 (1), 296-305.
Bouassida M. (2006). Modeling the behaviour of soft clays and new contributions for
soil improvement solutions. Keynote Lecture. Proc. 2nd Int. Conf. On
Problematic Soils. December 3-5th 2006. Ptain Jaya, Salengro, Malaysia.
Editors Bojan, Pinto & Jefferson, 1-12.
Klai, M., Bouassida, M. and Tabchouche S. (2015). Numerical modelling of Tunis soft
clay. Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 46 No. 4, 8795.

Klai, M. 2014, On the behaviour of Tunis soft clay Application to the study of
foundations stability (in French). PhDs Nat. Eng. School of Tunis. Tunisia.
Klai M. and Bouassida M. (2009) Comparison between behaviour of undisturbed and
reconstituted Tunis soft clay 2nd International Conference on New
Developments in Soil Mechanics and Geotechnical Engineering, 28-30 May,
Near East University, Nicosia, North Cyprus.
Touati L., Bouassida M. and Van Impe W. (2009). Discussion on the Tunis soft clay
sensitivity. Geotechnical Geological Engineering Journal. Vol 27, 631-643.
Tounekti, F., Klai, M., Bouassida, M. and Marzougi, I. (2008). Etude exprimentale
en vue dun modle de comportement pour la vase de Tunis. Revue Franaise
de Gotechnique. N114, 25-36.
Zimmermann, T., Truty E. and Podles K. (2010). Numerics in geotechnics and
structures. Elmepress international.

ASCE

240