Professional Documents
Culture Documents
Volume 4
Recently published in this series
Vol. 3. D.G. Toll, H. Zhu, A. Osman, W. Coombs, X. Li and M. Rouainia (Eds.), Information
Technology in Geo-Engineering – Proceedings of the 2nd International Conference
(ICITG), Durham, UK
Vol. 2. Y.-J. Cui, F. Emeriault, F. Cuira, S. Ghabezloo, J.-M. Pereira, M. Reboul, H. Ravel
and A.M. Tang (Eds.), Proceedings of the 5th International Young Geotechnical
Engineers’ Conference – 5th iYGEC 2013
Vol. 1. P. Arnold, G.A. Fenton, M.A. Hicks, T. Schweckendiek and B. Simpson (Eds.),
Modern Geotechnical Design Codes of Practice – Implementation, Application and
Development
Edited by
y
V..M. Ulitssky
hael B. Lisyuk
Mich
and
Alexeey G. Shaashkin
All rights reserved. No part of this book may be reproduced, stored in a retrieval system,
or transmitted, in any form or by any means, without prior written permission from the publisher.
Publisher
IOS Press BV
Nieuwe Hemweg 6B
1013 BG Amsterdam
Netherlands
fax: +31 20 687 0019
e-mail: order@iospress.nl
LEGAL NOTICE
The publisher is not responsible for the use which might be made of the following information.
Here I have in front of me a copy of G.P. Tschebotarioff’s inscription and signature on a book on
soil mechanics published in the USA and presented in 1973 to professor Vladimir Petrovich Sipidin at
the Department of subsoils and foundations of St. Petersburg Transport University (Figures 1 and 2).
Therein Gregory Tschebotarioff designates himself a student who entered the Transport University in
1918. The inscription states that the book is being presented to be read and subsequently donated to the
university library.
The full title of the book is “Foundations, retaining and earth structures”. I am a member of the in-
ternational community of geotechnical engineers united today by the ISSMGE, and I am really happy
about the fact that G.P. Tschebotarioff had chosen his professional career at our Transport University
and that amongst his colleagues, professors of soil mechanics, to whom he would subsequently present
his books, were representatives of our department at St. Petersburg. Moreover, he held Russian special-
ists, and particularly geotechnical engineers, in high regard. Having worked in the USA for a long time
he condemned politically engaged distortions of Russian history published by the USA media during
the McCarthy era at the time of the so-called “cold war”. It is officially known that during exchange of
scientific delegations between the USA and the USSR Professor Tschebotarioff declined certain
advances on the part of the CIA. Moreover, as a sign of protest he turned down the Professor Emeritus
rank at Princeton. Another reason for that reaction were cases of persecution of professors of Slavic
nationalities, particularly those teaching the Russian language and literature. This gives us a man of a
very broad range of attention in the areas of both science and humanities.
All this does not quite endorse the image of a leader of Whites’ opposition to the Reds on the river
Don. At the age of 19 in 1918 he opted to come to St. Petersburg and enter the University of Transport
which he could not graduate from for purely political reasons, being a member of a family close to the
Russian Emperor, who were at the time being persecuted by the Soviet law enforcement agencies.
Having spoken to my senior colleagues from the field of soil mechanics and construction I confi-
dently can bring to your attention the following moments from Gregory Tschebotarioff’s youth. After
entering the University of Transport he understood quite clearly that his social origins could lead to
quite dramatic consequences. He was aware of what was being done to his colleagues – officers from
famous St. Petersburg families, but the wish to become a professional civil engineer prevailed over the
sense of self preservation danger. According to the famous professor of St. Petersburg University of
Architecture and Construction (St. Petersburg Institute of Civil Engineering) and St. Petersburg Trans-
port University Vladimir Alexeevich Gastev and professor Victor Anatolievich Florin, Gregory Tsche-
botarioff was interested in soil testing research in the Laboratory of Soil Mechanics at the Transport
University. At that time the laboratory supervisor was the famous Russian and subsequently American
professor S.P.Timoshenko. I am convinced that their ways crossed whilst still in St. Petersburg: it was
in that oldest mechanical laboratory that N.M. Gersevanov conducted his pile tests (a graduate of the
Transport University of 1902 and son of the rector of the University who served in that capacity for
more than 25 years). Subsequently, the leading and one of the largest specialized underground con-
struction and foundation engineering institutes in Russia was named after N.M. Gersevanov (known
today as Moscow NIIOSP). Among his student contemporaries there were N.N. Maslow and
V.A. Florin who subsequently having become leading geotechnical specialists of world renown served
as translators of papers and monographs by G.P. Tschebotarioff published in the USA. One can
say quite confidently that since the very first days of his study at the University of Transport, being
in contact with a constellation of future geotechnical gurus (S.P. Timoshenko, V.A. Gastev,
N.N. Maslow, N.M. Gersevanov), he could not help getting engaged into geotechnical science which
was at the time at the breaking point in terms of its importance for construction practice, and not only
in Russia. He continued to maintain his ties with those bright minds also in his late years. The present
writer was not spared the “geotechnical bug” that tied him to the circles of people engaged in soil
mechanics after attending lectures by professor N.N. Maslov in 1957 in St. Petersburg. I was at the
time a cadet of the military faculty at the University of Architecture and Construction (former Institute
of Civil Engineers, later known as LISI).
vii
According to Prof. N.N. Maslov young G.P. Tschebotarioff arrived to Berlin Technical School
with notes on lectures by S.P. Timoshenko on theory of elasticity and books including publications by
the Transport University Press presreved in his personal library. He profoundly impressed the exami-
nation board having presented to them his Russian knowledge in the German and English languages.
According to his own testimony, Gregory Tschebotarioff had free and lengthy conversations in those
languages with the members of the board and the invited leading civil engineers from the Berlin
Technical School.
Sadly, fear for his life never left G.P. Tschebotarioff during the war years after the October Revo-
lution in 1917, even after his departure to the south of Russia. Those fears were especially reinforced
after he found out that some people known to him were imprisoned after their arrests and the military
officers arrested in those years were loaded on ferry boats and drowned in the Gulf of Finland. This
information is contained in his memoirs. I read about a lot of similar facts in Tschebotarioff’s book
entitled “Russia, My Native Land” published in New York in 1964 by McGrow-Hill Book Company.
From this book I learned that his grandfather, whose name was also Gregory Tschebotarioff, was of
Cossack stock and, a graduate of the Paris Institute of Technology, was in charge of railway construc-
tion in the South-East of Russia connecting the cities of Rostov and Voronezh in late 19th – early 20th
centuries, whereas his mother Valentina Ivanovna during the war was a nurse in the military hospital at
Tsarskoe Selo, where at this time the Empress Alexandra Fyodorovna was also engaged in a similar
capacity. Gregory Tschebotarioff doubtless was in contact with them – his mother and his god-mother,
during his sojourn while on leave in 1917 at Tsarskoe Selo, where his family resided at the time.
Initially it was with a certain degree of reluctance that I read sections on G.P. Tschebotarioff’s life
not connected to soil mechanics and foundation construction. But those five chapters read like an
adventure story resembling “The Road to Calvary” by Alexey Tolstoy. G.P. Tschebotarioff was fre-
quently arrested in the south of Russia but he was lucky “not to have been shot” as he himself put it in
the book. Once he was mistaken by the Reds’ patrol to be a “Whites’ guerrilla fighter” due to a typical
white officer’s knapsack he was wearing, but he was spared by the timely benevolent intervention of a
high-ranking official of the Reds who happened to be a former officer of the Imperial Russian Army,
and a native St. Petersburger. He quietly talked to Gregory Tschebotarioff and in spite of violent
protests from the blood-thirsty revolutionaries let him go. The second arrest was even more dangerous
but he had time to conceal himself from the arresting brigade in the huge crowd of pro-revolutionary
populace greeting the arrival of the Red Leaders to the city of Novorossiysk. In such environment of
constant threats the only solution left for him was emigration. He was evacuated to Egypt together with
the College of the Don Cadets where he worked as instructor since 1921, acting as an aide to the
Artillery Inspector of the Don Army. We will not be far from the truth supposing he instructed his
officers in the matters of construction science because construction was the only practical field where
disciplined Russian officers were in high demand, organizing and conducting building activities – there
were simply no other activities ongoing in Egypt at that time.
After graduating as a civil engineer in Berlin, Gregory Tschebotarioff worked in Egypt. Demand-
ing ground conditions of that country alerted him to the issue and importance of soil mechanics in
general, and to complicacy and responsibility of foundation construction in particular. He served as a
consultant in these areas for some time in France, Germany and the USA. As of 1937 he became a full-
time professor at Princeton, holding a tenure in the art of construction. It is interesting to point out that
directions connected to soil mechanics and foundations, including stability of retaining structures were
quite rightly regarded as construction art. Gregory Tschebotarioff was involved in projects related to
construction of bridges, high dams, tunnels and other civil and military structures, some of them being
unique.
I would like to point out the following moments from Gregory Tschebotarioff’s notes which in-
fused our work in ISSMGE ТС 207 “Soil-Structure Interaction and Retaining Walls” for eight years
(2005–2013):
viii
elephants in question are the architect, the superstructure engineer, and the geotechnical engineer. The
reports at ТС207 SSI sessions in Paris presented bold achievements implemented in design of super-
tall structures (over 1000 m) constructed in Dubai, as well as in other unique structures using state-of-
the-art software engineering solutions and contemporary construction codes (Proceedings of the
TC207 workshop on soil-structure interaction and retaining walls – September 2013, Paris –
www.paris2013-icsmge.org).
The geotechnical engineers of the new century have reached a new stage in numerical calculations
using advantages of contemporary computing, i.e. such as were simply not there before. This disadvan-
taged level of calculation methods troubled Professor Gregory Tschebotarioff and burdened him
greatly, which he openly wrote and spoke about.
In conclusion I would like to quote our newly elected ISSMGE President Professor Roger Frank:
“The perspective of contemporary codes for geotechnical design will be grounded in three words:
“SOIL-STRUCTURE-INTERACTION”!”
ORGANIZED BY
ORGANIZING COMMITTEE
V.M. Ulitsky, Chairman of TC207, Head of Petersburg Branch of RSSMGFI, St. Petersburg
M.B. Lisyuk, Co-Chairman of TC207, Deputy Director of “Georeconstruction” Institute,
St. Petersburg
A.G. Shashkin, General Director of “Georeconstruction” Institute, St. Petersburg
Ö. Bilgin (USA)
W. Bilfinger (Brazil)
J.-L. Briaud (USA)
R.E. Dashko (Russia)
Fang Liu (China)
R. Finno (USA)
C. Haberfield (Australia)
H. Hazarika (Japan)
R. Katzenbach (Germany)
M. Mets (Estonia)
I. Sokolić (Croatia)
C. Shashkin (Russia)
V. Paramonov (Russia)
W. Van Impe (Belgium)
Yasser El-Mossallamy (Egypt)
CONFERENCE SECRETARIAT
Contents
G.P. Tschebotarioff Address v
V.М. Ulitsky
Conference Organization xi
Keynote Lectures
Effect of Ground Motions and Tsunami Impact Force on the Performance of a Damaged
River Dike 132
Hemanta Hazarika, Tadashi Hara, Shunichi Kuroda, Kentaro Kuribayashi,
Tsuyoshi Nishi, Hideo Furuichi, Koichiro Takezawa and Tsuneo Ohsumi
Influence of Thermal Condition on the Extraction Characteristics of Steel Strip Reinforcing
Members 140
Hijiri Hashimoto, Takahiro Yamanashi, Hirochika Hayashi and Masahiko Yamaki
Methods of Research on Frost Heave of Soil and Foundation – Freezing Soil Bulk Interaction
in Terms of Seasonal Deep Frost Penetration Including Areas of Pressure Migration 147
O.Z. Khalimov
Oblique Pullout Capacity of a Single Drilled Pile in Sandy Grains Cemented by Methane
Hydrates: DEM Analyses 155
Fang Liu and Mingjing Jiang
Modeling of Interaction of Polymeric Sheet Reinforcement and Backfill 161
P.V.S.N. Pavan Kumar and M.R. Madhav
Soft Grouting for the Seismic Protection of Existing Buildings 169
D. Lombardi, V. Nappa and A. Flora
Soil Structure Interaction: An Analysis of the Effect of Creep and Shrinkage of Concrete 176
Luciana de Moraes Pereira da Rosa, Eliane Maria Lopes Carvalho
and Bernadete Ragoni Danziger
Specifics of Finite Element Modelling of Freezing and Thawing Deformations in Soils 184
V.N. Paramonov and A.V. Nabokov
Comparative Analysis of Piled Foundation Design of a Highrise Building in Bangkok Subsoils 189
N. Phienwej, K. Amornfa and W. Cheang
Foundation System Design for a Tall Structure on Mortar-Column-Improved
Ground 195
Salah Sadek and Samer Hasan
Interaction of Foundations and Bearing Elements of a Building with Thawing Subsoil 203
I.I. Sakharov and M.V. Paramonov
Design of Pile Group System for “Most Sava – Zeleni” Railway Bridge
in Zagreb 208
I. Sokolić and B. Vukadinović
Soil-Structure Interaction Analyses for the Study of Nuclear Power Plant
Foundation Alternatives Under Static and Seismic Loading 214
Martin Vaníček and Ivan Vaníček
Analysis of Soil-Structure Interaction by Ménard Pressuremeter Tests and Ground
Improvement Case Histories 220
S. Varaksin and B. Hamidi
Piled Embankments in Soft Estuary Clay – Experience from Design and Field Measurements
for Redevelopment of Harbour Areas in Northern Germany 227
Lars Vollmert and Juri Schlee
An Efficient Method for Estimating the Dynamic Response of Base-Isolated
Structure with SSI Effect 235
Zhuang Haiyang, Yu Xu and Zhu Chao
xv
An Investigation into the Vertical Axial Capacities and Groundwater Cut-Off Capabilities
of Secant Pile Walls 251
Abid Adekunte
3D Modelling in Deep Excavations – Case Studies 259
Önder Akçakal, Turan Durgunoğlu, Berkay Koçak and Taşkın Tarı
Pseudo-Static Analysis of Flexible Retaining Structures Including Seismic Thrust Dependency
on Wall Deformation 265
Bruno Becci and Marco Carni
Soil-Structure Interaction for Sheet Pile Walls Considering Ground Surface and Sub-Wall Soil
Conditions 273
Ö. Bilgin
Monitoring Results for Construction of a Building with a Complex Underground Space in Soft
Soils in St. Petersburg 279
S.G. Bogov, N.P. Bochkarev and V.Y. Smolak
Seismic Soil-Interaction on Earth Retaining Structures. Performance-Based Seismic Design
of a Retaining Wall 284
Carlos Jurado Cabañes
Shoring-Structure Interaction in Stabilization of Excavation Adjacent to Historic Buildings –
A Case Study 290
Hamidreza Elahi, Mohsen Sabermahani and Hesam Vahidifard
Verification of Deck on Piles Quay Wall Structures for Operational and Seismic Load
Conditions 296
H.J. Lengkeek and F. Besseling
A Bold Constructive Method for the Retaining Structures of the Excavation
of the Underground Parking Clinics, in Brazil 302
C.E.M. Maffei, H.H.S. Gonçalves and M.C. Guazzelli
Comparison of Finite Element and Limit Equilibrium Methods in Analysis of Soil Nail Walls 310
S. Melentijevic and J.L. Fernández
Understanding of Soil Responses Due to Tunnelling – 3D Numerical Analysis and Case Study
of Bendemeer Station (DTL-3), Singapore 318
C.W. Ong, Thiri Su, K.Y. Yong and Kulaindran Ariaratnam
Assessment Analysis of Lateral Movement of Gate Shaft Structure on Fractured Rock Mass,
Jatigede Dam Area, West Java, Indonesia 329
A. Putera Agung M, S. Pramusandi, A. Ardianto and B. Sunaryo
San Pasquale Station of the Linea 6 in Napoli: Dewatering Field Tests, Measurements
and Back-Analyses 337
Gianpiero Russo, Silvia Autuori and Marco Valerio Nicotera
Frost Protection for Geotextile-Reinforced Soil Walls in Service 345
Atsuko Sato, Takahiro Yamanashi, Teruyuki Suzuki, Naoki Tatta and Kouichi Yoshida
xvi
Abstract. Soil-structure interaction calculations constitute a basis for design decision making on a structure of any building even
an ordinary housing. The importance of soil-structure interaction calculations largely increases when unique structures are
designed for which construction there is no experience in geotechnical conditions of a given area. Special interest is placed on
calculations of a high-rise building on soft subsoil where there is a necessity to estimate soil strains caused by application of
unusually high loads.
There is an example of design of a 400-m-high #
) /(st storey it decreased
skyscraper consisting of 78 above ground and 3 (5/ #
# 9(st storey – down to
! " 12.1 m.
building on more than 200-m-thick bulk of The thickness of outside wall of the core
sedimentary deposits. made of reinforced concrete (concrete class ;80)
The skyscraper represented a spiral screwing changed from 2 m in the lower part down to 0.4
"
! ! "#! $ % m in the upper part of the building. Internal
“petals” connected with a central round core. bearing reinforced concrete walls of the core of
With building height increasing square “petals” at least 300 mm thickness ensured additional
turned counter-clockwise around the building rigidity of the core structures.
axis and reduced &
" #' Due to a spiral screwing form of the building
towards the axis of the round central core of the there emerged horizontal loads applied to floor-
building. ings and the beam carcass which tended to turn
In terms of structure the building followed a floorings around the core axis. In order to take
core and frame scheme. Its general stability was these horizontal loads creation of rigid discs of
ensured by joint work of a central monolith floorings was required. It was envisaged to make
reinforced-concrete core ($ # "# floorings of the above part in office floors of
located along the perimeter and united by metal " ""
"" " ;/*
"k-
)# ""
" " (5* ##
"" "
!
!" )# "#! +$* ##
three levels of technical storeys providing signif- Exterior metal columns represented an im-
icant increase of total horizontal rigidity of the portant element of the core and frame system of
building against wind loads. Outriggers repre- the building. Steel columns were located along
sented metal console beams with one end being the building perimeter and provided resting for
fixed in the central reinforced concrete core and the system of beams of floorings in the span
the other linked with exterior metal columns. An from the central reinforced concrete beam to
)# ! $* ##-thick wall of exterior walls of the building. The columns had a
altering section with a system of rigidity ribs. box-shape cross section with exterior planes of
The core served the main element which pro- 5$*=5$* ## " # (,* (**
vided transfer of vertical and horizontal load mm. At the level of the last flooring loads from
components in the given structural scheme. The metal structures of approximately 76-m-high
central reinforced concrete core diameter re- spire were applied to metal columns of the main
duced with height. At the level of lower floors up frame.
to the level of the 61st "#! +,$
4 V.M. Ulitsky et al. / Soil-Structure Interaction Calculations of a High-Rise Building
The structural scheme of the building foresaw Upper Proterozoic Kotlin sediments (Vkt2)
considerable concentration of loads transferred to are characterized by comparative permanency of
subsoil near the central part – the core. cross section. In the given site venda bedrock
One of the important principles of construct- deposits were found at absolute benchmarks of
ing a high-rise building on soft soil is installation minus 43.80 m – # $**H # J#
of a developed underground part which has two depth o ,$* # ! $B* # !
functions: 1) load distribution from a high-rise composes of layered clays with thin seams (up to
building to a large area; 2) reduction of pressure 1 mm) of poorly dislocated sandstone. A dislo-
on soil by a weight value of soil extracted from a cated area is characterized by excessive fractur-
volume of an underground part. For an under-
)
# )&
ground part of a high-rise building to be able to layering and excessive hydration of deposits.
fulfill the function of load redistribution it should J# ! $B* ! (+9$
have the required spatial rigidity. It means that m the section is represented by layered clays
underground spaces should be crossed by radial with thin seams (up to 1 mm) of sandstone. In
rigidity diaphragms providing even transfer of ! /$*L5H* # '
forces from the core and exterior columns found a seam of dislocated clays having low
(connected with outrigger beams in technical values of #"" !! !
storeys of the above ground part) throughout a (+9$ # -clay assise of the
square of underground space of the building. upper subset of Kotlin set represented by a bulk
Certainly this underground space becomes of uneven over- " "
uncomfortable for utilization: rigidity dia- fine-
! "#
!#
@#!
" "
lites.
# ! !" "onstruction of In hydro-geological terms the territory was
skyscrapers on sedimentary soils a priori is not characterized by presence of several aquifers.
economically well-grounded investment; it is Aquifer I is non-!
% MM
MMM M
more likely to reflect ambitions of a developer. – pressure ones. Statistical level of aquifer I is
" " ""
" @ )"# *5L(5 # N
geotechnical engineers designing the considered aquifer II – )"# *$L(9 # N
high-rise building started at the stage of devel- for aquifer III – )"# *(L*, #
!# "" ""!
!## N % M
– at the benchmarks of
structural solution for the underground space *(L*/ #
providing the required spatial rigidity was find at In order to obtain accurate distribution of
design stage. It was envisaged to install 800-mm- forces in barrettes structures and account of their
thick radial traverse wall at the level of under- interaction with soil simulation of barrettes and
ground floors. At the level of underground floors the piled raft was carried out with a help of
the load from exterior metal columns of the '#" # ' @
building was carried by reinforced concrete walls of the core were also modeled as volumet-
"# (B$*=(B$* ## " " " #
"
Vertical reinforced concrete bearing structures of were simulated as shell elements (interaction
the underground part were supposed to be made between external and internal walls of the core
of concrete of ;5* "
;60 – for floorings. was modeled with a help of special contact
Installation of barrettes (of ;60 class con- elements allowing to set different shear rigidity
"G ! (*$ # # in contact zones). Floorings in the internal
surface was envisaged to transfer loads of the volume of the core were simulated as shell
high-rise building to subsoil. On the top barrettes #
#e columns and reinforced
were united by a 3.6–m-thick powerful monolith concrete columns of the underground part – as
raft made of ;60 concrete. ! #
)# "" )
According to investigation data the geotech- the rigidity core was set by spatial rod elements
nical profile of the site within the drilling depth (rigidity of reinforced concrete slabs of above
of 170.0 m consists of quaternary deposits of ground floorings beyond the core was not taken
various origin and venda bedrock. "
) #
@) G
# )#
V.M. Ulitsky et al. / Soil-Structure Interaction Calculations of a High-Rise Building 5
behavior was made by shell elements with interactio "#! (
'
equivalent rigidity. structures of the designed building is given in
Fig. 1 shows the general view of the compu- fig. 2.
tation scheme for the problem of soil-structure
The design company proposed barrettes of by designers in order to limit the expected
(*$ # ! # " ) # ) ' ) (* "# #t-
the layer of venda hard soil as a foundation for ing the value of absolute settlements the project
the high-rise building. This option was selected authors tried to restrict their difference as well.
6 V.M. Ulitsky et al. / Soil-Structure Interaction Calculations of a High-Rise Building
Figure 3. Contours of settlements (m) of the designed high-rise building taking into account non-linear behavior of the subsoil
according to visco-plastic model
Contours of vertical displacements of sepa- ($** X ) #
) $ M " loaded that is linked to special nature of soil
area near the core the values of normative load behavior and rigidity of radial traverse walls
on barrettes do not exceed 1310 tf (fig. 6). distributing pressure from the core and external
@ " appear columns throughout the area of the underground
) #
#@## ) structure. The maximum value of normative
2400 tf. The values of normative load on bar- load does not exceed 3700 tf.
rettes under radial traverse walls do not exceed
Figure 5. Contours (m) of heads of barrettes based on calculation data on non-linear subsoil according to visco-plastic model
V.M. Ulitsky et al. / Soil-Structure Interaction Calculations of a High-Rise Building 9
Figure 6. Forces (kN) in barrettes (normative) based on results of calculation of the building taking into account non-linear
behavior of the soil according to visco-plastic model
10 V.M. Ulitsky et al. / Soil-Structure Interaction Calculations of a High-Rise Building
Figure 7. The deformation scheme of the underground part of the designed high-rise building based on non-linear subsoil
according to visco-plastic model. There are given values of vertical displacements (m)
Abstract. In this paper the behavior of single piles under lateral impact is investigated by using the performed pendulum and
bogie tests. Three types of soil deposits were used in the experimental tests including hard clay, loose sand and crushed lime-
stone. The acceleration and displacement of the striking mass are measured. The results show that the soil response to impact is a
function of soil stiffness; however the inertia and damping resistance are also important contributors. In order to model the
system, a Single Degree of Freedom (SDF) model with two different basic materials is used to find the best simple model to
predict the behavior of piles under impact. Finally an advanced model with multi degree of freedom is used to analyze a single
pile under the pickup truck impact with 60 miles per hour. The results of the simple model and the experimental tests are
compared with the LS-DYNA simulations.
match the test results material model, there is a in figure 2. In addition to the properties of each
good chance that the model can be used for soil soil including the Pressuremeter Test (PMT) and
in multi degree of freedom. First a number of the Standard Penetration Test (SPT) as in situ
scaled test were done with a pendulum and a tests, and sieve analysis water content, unite
bogie to provide a calibration for the SDF model. weight and direct shear test as lab tests, series of
static horizontal load tests were done.
1. Experimental tests Table 1. Details of single pile tests
Plan
30cm
100cm
180cm
Section
2. Test results and discussion the pile is obtained. The displacement and the
acceleration for all eight impact tests are present-
The acceleration of the striking mass and the ed in Figure 3. The static lateral load test results
displacement of the pile are two major outputs of for each soil type are also shown in figure 4.
each test. It is also possible to obtain the rotation As it is seen from the acceleration output, for
of the pile head from the high speed movies. By all of tests there are three steps: 1- a peak at the
multiplying the acceleration of the pendulum or beginning, 2- a flat part in the middle, 3- a drop
the bogie by the mass, the load-time history on at the end of the impact.
20
dH 2
dH V EH1 E*
V EH E *
dt
dt H1 H 2 H
1 2
3 4
5 6
7 8
900 6. References
800
700
Barghouthi, (1984). "Pile response to seismic waves," thesis
presented Univ. of Wisconsin, at Madison, Wisconsin.
600
Displacement (mm)
Abstract: A 5 m high earth fill embankment comprising vertical side-walls constructed from precast concrete wall panels tied
together using steel straps was constructed on a 10 m thick deposit of soft clay. Due to the relatively small specified settlement
tolerances of the embankment/wall structure, ground improvement in the form of drilled displacement columns (unreinforced
concrete columns) was used to improve the engineering properties of the embankment foundation. This paper discusses the
design and construction of the embankment and foundation system and compares the predicted settlements with those measured
after construction. Discussion on the quality control measures adopted during construction is also provided.
Keywords. ground improvement, soft clay, embankment, settlements, drilled displacement columns
Options considered included a fully-piled struc- tive advantage. The authors therefore consider
ture or semi-rigid inclusions such as soil mixed the generic term “Drilled Displacement Col-
columns or drilled displacement columns umns”, “DDCs” or “geotechnical elements” to be
(DDCs). a more appropriate description when applied to
Piling and soil mixing are both well- ground improvement.
established forms of ground support and are not In saying this, it is also recognised that col-
further discussed in this paper. DDCs are not as umns are correctly designed as structural ele-
widely known but were ultimately used to ments on many projects. It is therefore essential
improve the strength and stiffness of the em- that a designer knows when it is appropriate to
bankment foundation. They are described in consider columns as ground improvement,
greater detail below. thereby designing them economically as ge-
otechnical elements, or when they should be
considered as structural elements (i.e. as piles).
Some of the clearest guidance on this matter
is provided by Wong and Muttuvel (2012). They
suggest that, for the case where the columns act
semi-rigidly by sharing load with surrounding
soil as a result of (a) column compressibility, (b)
yielding of the column toe, and (c) the use of a
load-transfer platform (they also postulate that a
load transfer platform is not always warranted
depending on embankment height and fill quali-
ty), they may be treated as geotechnical ele-
ments. This is also the authors’ experience.
Treatment of columns as structural elements
is more likely to be applicable where (a) column
spacing exceeds a maximum of about 2.5 m, or
more simply where a foundation support system
Figure 1. Typical tied-wall arrangement transitions from high redundancy to low redun-
dancy, i.e. there is potential for the performance
of a single column to detrimentally impact the
1.1. Background to construction and design performance of the overall embankment / struc-
using DDCs ture, or (b) bending, shear or tensile stresses are
DDCs typically comprise 450 mm diameter high and overall stability of the embankment
columns constructed from unreinforced concrete. may be compromised or lateral displacements
They are generally installed with an “Omega” cannot be controlled, resulting in the need for
auger using a full-displacement technique. They steel reinforcement. In either of these cases,
are often referred to by other names including adopting a ground improvement approach to
“Controlled Modulus Columns (CMCs)”, “Con- design rather than a structural approach should
crete Injected Columns (CICs)” and “Drilled be undertaken with due caution and appropriate
Displacement Piles (DDPs)”. However, these engineering judgment.
terms are considered to be confusing or inappro-
priate. Firstly, the term CMC is linked to a 1.2. Use of DDCs in Australia
proprietary product. Secondly, the authors The use of DDCs is becoming increasingly
consider the use of the term “pile” to be confus- popular in Australia due to the ability to install
ing in an industry where piles are traditionally columns quickly, effectively and with minimal
designed as structural elements and therefore generation of spoil material. This is particularly
need to satisfy structural design requirements. important for soils such as CIS, which have a
This typically involves factoring up loads and high acid-sulphate potential and spoil can be
factoring down design soil strengths. If a ground prohibitively expensive to dispose of off-site.
improvement column was required to satisfy Contractors are also becoming more proficient at
traditional pile design requirements, the compar- using the technique, particularly testing and
atively cheap technique would lose its competi- implementing quality control.
20 J. Gniel and C. Haberfield / Design, Construction and Performance of a Tied-Wall Embankment
1% of the bending capacity, further assessment undertaking this analysis as strips are often very
of the most appropriate parameters to use in the narrow and numerical errors may result if a
analysis were not undertaken. This being said, if suitably fine mesh is not adopted.
relying on bending stiffness in a DDC model, an Since undertaking the analysis described
appropriate level of engineering judgment is above, Wong and Muttuvel (2012) have provided
required. This is discussed in more detail below. additional design guidance in the form of a “new
Columns ranging in spacing from 2 m to design approach”. Many components of the
2.2 m, installed on a square grid, were ultimately work presented in this paper are consistent with
adopted. the approach proposed by Wong and Muttuvel
(2012). Subsequent work undertaken by the
authors of this paper suggest that the “new
design approach” appears reasonable and appro-
priate although would benefit from further
research and verification.
penetration of 2 m, proof load testing of the trial 30 mm and 50 mm, which was less than the
columns was undertaken to confirm an accepta- settlement calculated using finite element model-
ble load-settlement response at this minimum ling. The results of testing were therefore
penetration. Proof load testing comprised Pile considered to validate the design and provide
Driving Analyser (PDA) and Case Pile Wave confirmation that the rigs were capable of in-
Analysis Program (CAPWAP) analysis. This stalling columns that met the design intent.
required reinforcing the column and building up
the head so that PDA gauges could be attached
aboveground. The column then needed to cure
for at least one week to achieve the required test
strength. This process added significant cost to
testing. The end result was that fewer tests were
undertaken than for conventional piling but a
greater reliance was placed on rig instrumenta-
tion data and site supervision.
A photograph of the built-up head of a DDC
within a test frame (awaiting PDA testing) is
presented in Figure 5.
about 7D (where D is the column diameter) needed to be assessed numerically before addi-
in a 24 hour period. This distance was based tional columns could be installed.
on the results of inclinometer testing under-
taken near driven precast piles and DDCs as 4.3. Outcome of quality control
part of another component of ongoing re-
search, and observations of concrete heave In the authors’ opinion, the adopted approach of
from adjacent wet DDCs that suggested (a) PDA testing trial columns installed at loca-
squeezing of the column cross-section oc- tions with known stratigraphy, (b) reviewing rig
curred at closer spacing. instrumentation data for each and every CMC
x The drill stem was not to be lifted at any installed, and (c) regular site supervision
stage during the drilling process to avoid throughout construction, provided a robust
necking of columns. approach to quality control. A photograph of the
x The concrete pumps were not to be shut-off completed embankment footprint prior to con-
until the base of the drill stem was above base struction of the LTP is presented in Figure 7.
of enlarged head level to avoid necking or The enlarged heads of the DDC are evident.
contamination of the column shaft.
x The base of the enlarged head had to be
cleaned and then the enlarged head poured
while the shaft concrete was still wet.
In addition to this, the contractor provided the
design engineers with the following information
within 48 hours of each DDC being installed:
i. DDC name, date and grid reference or
identification
ii. Penetration rate vs. depth
iii. Torque vs. depth
iv. Crowd vs. depth
v. Pitch vs. depth
vi. Lifting speed during concreting vs.
depth
vii. Concrete volume vs. depth
viii. Total depth and total volume of con-
crete injected, including concrete dock-
ets showing volume of concrete deliv-
ered to site
This information was reviewed for each and
Figure 7. Completed embankment footprint prior to place-
every column installed. ment of LTP
An allowance was also made for geotechnical
engineers to make regular visits to observe DDC 5. Settlement monitoring
installation throughout the duration of the pro-
ject. This supervisory role was considered One end of the embankment supported on DDCs
critical to ensuring consistency during the con- transitions into an untreated foundation which
struction period. coincides with the edge of the CIS deposit. A
Where defects or discrepancies were ob- roughly 15 m wide transition zone comprising
served, such as column necking, insufficient toe DDCs with steadily decreasing length (but same
penetration or conflicting information regarding spacing) was adopted to reduce differential
the volume of concrete installed or delivered to settlement across the transition. Although not
site, additional measures were implemented. the subject of this paper, the transition zone
These ranged from Pile Integrity Testing, used to provides useful settlement data for comparison to
investigate whether a reduction in column cross- the DDC treated zone.
section had occurred, or more significant The section of untreated foundation is under-
measures such as installation of replacement lain by CIS of 3 m maximum thickness. This
columns. The latter was not a simple task as the area was not treated with stiffening elements but
impact of a localised increased in stiffness had been surcharged by a large, monitored
26 J. Gniel and C. Haberfield / Design, Construction and Performance of a Tied-Wall Embankment
stockpile for several months. Estimates of total construction settlement when used in the manner
settlement in this zone, as a result of embank- described in this paper.
ment construction, were in the order of 100 mm.
Approximately half of this was expected to occur 6. Conclusions
during construction with the remainder occurring
in the following six to twelve months. Using a case study of a tied-wall embankment
founded on a significant thickness of soft clay,
5.1. Construction Phase Settlement DDCs are shown to provide an effective and
economic option for reinforcing the foundation.
Settlement was not monitored as the embank- Importantly, the technique can be used to reduce
ment was constructed. Monitoring commenced post-construction settlements to acceptable
immediately after the last layers of fill were limits.
placed. Therefore it is not possible to confirm Design using finite element methods can be
the amount of settlement that occurred during used to reliably model DDC behavior by consid-
construction, except that it was likely to be ering columns as geotechnical elements (sur-
significantly less than 100 mm based on the level rounded by soil strips) with very low bending
of wall panels compared to the design level. stiffness. The approach requires considerable
engineering judgment and would benefit from
5.2. Post-construction settlement further research and modelling, some of which is
As noted above, post-construction monitoring currently being undertaken. Design using limit
commenced immediately after construction was equilibrium approaches should be treated with
completed. The results of this monitoring is caution.
presented in Figure 8, which shows the results of A method for controlling the quality of DDCs
seven (7) monitoring points located at various is discussed, which involves PDA testing of trial
points along the length of the tied-walls. columns installed over the location of known
With reference to Figure 8, points 100, 300 CPTs or boreholes (undertaken in advance of the
and 500 are positioned above the section of DDC trial). Comparison of the rig data, including
improved embankment footprint. Not surprising- torque and crowd, to the CPT/borehole data can
ly, post-construction settlements measured in the be used to confirm that the rig can detect changes
6 months following construction are a maximum in stratigraphy and, most importantly, can detect
of 5 mm and do not show signs of ongoing creep. the DDC founding layer.
This is consistent with design. Due to time and costs constraints associated
Points 600 and 900 are located at either end with building up DDC heads for PDA testing,
of the transition zone (point 600 being closest to fewer dynamic tests are undertaken on DDCs
the full-depth DCCs). They show maximum than might be undertaken on a conventional pile-
post-construction settlements of 18 mm and 28 supported embankment. This can be supplement-
mm respectively. The rate of settlement appears ed by reviewing the rig instrumentation data for
to be slowing but is ongoing. each and every DDC installed and by the design
Points 1100 and 1200 are located above the engineer undertaking regular site visits through-
section of untreated embankment footprint and out the construction phase to ensure consistency
show maximum settlements of 26 mm and with design.
28 mm, respectively. The rate of settlement Further research is currently being undertak-
appears to be slowing but some ongoing settle- en that investigates the role of the LTP in the
ment is expected, which is consistent with overall performance of a DDC supported em-
design. bankment. In particular, the research looks at the
On the basis of this data, it is apparent that impact of insitu fill overlying the highly com-
DDCs are very effective in reducing post- pressible soil unit.
J. Gniel and C. Haberfield / Design, Construction and Performance of a Tied-Wall Embankment 27
7. References
Abstract: The analysis of the soil-structure-interaction is a very important part during all stages of planning, design and
construction. Large urban construction projects create an impact not only on neighbouring structures but also on existing,
sensitive underground structures, like metro and street tunnels. But not only new constructions have an effect on existing
structures. Deconstruction projects may cause enormous displacements as well. For the analysis of the soil-structure-interaction
in most cases numerical methods are necessary. For verification of the numerical simulations geodetic and geotechnical meas-
urements according to the observational method have to be carried out. An outstanding large urban construction project on
existing underground structures of the metro of Frankfurt am Main, Germany, is presented in this paper.
Keywords. soil-structure interaction, Frankfurt limestone, uplift, settlements, hardening soil model.
Figure 2. Cross section of Figure 1. For the evaluation of the effect of the deconstruc-
tion of the Municipal Technical Administration
The sealing of the sublevels and of the un- and the construction of the new structures the
derground structures consists of several bitumen uplift and the settlements have to be predicted.
layers and copper sheets, located on the outside For the prediction three dimensional numerical
of the reinforced concrete structures. Figure 3 simulations were used.
shows the sealing detail of the underground For the correct prediction of the deformations
structures and the sublevels of the Municipal the complex geometries and the stress and time
Technical Administration. It has to be guaranteed dependent deformation behaviour of the Frank-
that the sealing of the structures is not damaged furt Clay (Breth & Stroh 1974, Reul & Krajew-
during the deconstruction of the Municipal ski 2010, Katzenbach et al. 2011a, Kolembach
Technical Administration and the construction of 2012, Katzenbach & Leppla 2013a) have to be
the new structures. considered in the numerical models.
The soil and groundwater conditions at the In the last decades the development of nu-
project area are as follows: merical models focused on the simulation of the
x fillings down to a depth of 8 m stress and deformation behaviour of the soil and
x quaternary sands and gravel down to a depth the structures, e.g. foundations systems, retaining
of 15 m systems or tunnels. The experiences from differ-
x tertiary Frankfurt Clay down to a depth of 30 ent projects show, that the numerical analysis
m using the Hardening-Soil-Model (HS-Model)
x Frankfurt Limestone produce true results. The HS-Model is an elasto-
plastic constitutive equation with an isotropic
The sublevels of the Municipal Technical double hardening. The elastic sphere with stress
Administration, the underground parking, the dependent, non-linear elastic behaviour is cov-
underground station and the tunnels are below ered by two yield surfaces. In the principle stress
the groundwater level. The groundwater level is sphere these two yield surfaces are a cone, closed
strongly influence by the river Main, which is by a cap (Figure 4). The sphere of the permitted
about 180 m south of the project area. stress is defined by the the limit condition of
Mohr-Coulomb. If the stress level reaches one of
30 R. Katzenbach et al. / Numerical Analysis and Verification of the Soil-Structure-Interaction
the yield surfaces elastic and plastic defor- including the main parts of the foundation and
mations occur. The yield surface is expanding underground structures. The walls and the
and causes a plastic solidification. The direction columns were modeled with special plane ele-
of the plastic deformation is defined by a flow ments. The raft and slabs were modeled with
rule. continuum elements. All construction parts
The numerical material parameters for the consisting of reinforced concrete are modeled
Frankfurt clay are shown in Table 1. Generally with elastic deformation behaviour. The expan-
these parameters lead to good results. sion gaps between the tunnel blocks were mod-
Regarding the current project it is important eled as well.
to consider the difference between the construc-
tion of new structures and the deconstruction of
existing structures. The stress path is extremely
different. The stress path of the described project
is influenced by the primary loading, the unload-
ing and the reloading.
tures and in the remaining sublevels avoid The investigations lead to the knowledge that
deflexions of the foundation rafts. However, in the material parameters in Table 1 do not lead to
the present case the three dimensional model correct results, if unloading effects dominate the
requires strong simplifications in modelling the present geotechnical problem. Moreover it can
remaining building structures. By these simplifi- be stated, that the proved material sets allow
cations the flexural stiffness of the structures satisfactory predictions for example foundation
could only be reproduced unsatisfactorily. problems or construction pits. In the case that the
effects of the deconstruction of high or heavy
buildings are to comprehend additional reflec-
tions are necessary. In advance to chapter 4 it has
to be pointed out that the uplift due to unloading
is nearly the same as the former settlement due to
loading. For this reason, the numerical analysis
of the uplift of Frankfurt Clay requires the same
approach of stiffness as for analysis of settle-
ments of new structures. The common distinction
for the Young´s Modulus for loading and unload-
ing like shown in Table 1 is not useful. This
Figure 6. Comparison of measured and calculated uplift awareness is qualitative and is based on the new
(3-D calculation, cross section of Figures 1 experiences. Further research is needed for the
and 2, material parameters of Table 1). development of a new constitutive law under
consideration of the stress path including pre-
For a better agreement between the results of loading and unloading and the over-
practice and analytical model additional two consolidation of the Frankfurt Clay.
dimensional simulations have been carried out, Additional to the analysis of the deformations
where the stiffness of the structures could be analysis for buoyancy were carried out. The
considered more realistically. A typical result of analysis showed that the safety against buoyancy
the Finite-Element-analysis is given in Figure 7. of the sublevels is not given, when the Municipal
The biggest uplift deformations amount is more Technical Administration is totally deconstruct-
than 5 cm. That is a difference of more than ed. For this case a temporary loading was in-
70 % compared to the results of the three- stalled and an emergency plan was developed.
dimensional calculations (Figure 6). The uneven
deformation curve, which is measured in reality
can be reproduced qualitatively. However the
total deformations are still too small compared to Figure 7. Calculated uplift deformations (2-D-FEM calcula-
the measurements of more than 6 cm tion, cross section of Figures 1 and 2, material parameters of
(Figure 6). Table 1).
32 R. Katzenbach et al. / Numerical Analysis and Verification of the Soil-Structure-Interaction
Since October 2012 the loads were not bach et al 2010 and 2013b, Paul 1998,
changed. The further construction of the new Ulitsky 2003).
structures will start in 2014 when the building x For indemnification of the 4-eye-principle an
permission is finished. independent, publicly certified geotechnical
expert has to be involved in an early planning
stage (Fuchs & Haugwitz 2009, Katzenbach
et al. 2011b).
x For extreme situations an emergency plan has
to be developed considering the stability and
the serviceability.
x For the correct simulation of the time de-
pendent deformation behaviour of soils due
to reduced stress levels further research is
necessary. The verification of constitutive
laws describing the time dependent defor-
mation behaviour due to reduced stress levels
Figure 11. Comparison of measured and calculated uplift has to consider measurement data of real con-
(2-D-FEM calculation, cross section of struction projects or load tests.
Figures 1 and 2, material parameters of Table 1).
6. References
Under consideration of the stiffness of the
structures and an improved approach for the Breth, H.; Stroh, D. 1974. Das Verformungsverhalten des
deformation behaviour of the Frankfurt Clay the Frankfurter Tons beim Aushub einer tiefen Baugrube
numerical simulations show a good accordance und bei anschließender Wiederbelastung durch ein
to the measurement results. For example the Hochhaus. 13. Baugrundtagung der Deutschen Gesell-
calculated and measured deformations of the schaft für Geotechnik in Frankfurt am Main, 10.-12.
September, Germany, 51.70.
cross section of the Figures 1 and 2 are drawn in CEN European Committee of Standardisation 2008a.
Figure 11. The comparison shows that for diffi- Eurocode 7: Geotechnical design – Part 1: General
cult analysis of the soil-structure-interaction the rules.
stiffness of the structures and of the soil have to CEN European Committee of Standardisation 2008b.
be modeled close to reality. Eurocode 7: Geotechnical design – Part 2: Ground in-
vestigation and testing.
Fuchs, B.; Haugwitz, G. 2009. Duty to check and to report
5. Conclusions problems: Practical geotechnical experiences. Darm-
stadt Geotechnics No. 17, Technische Universität
The presented soil deformation of the current Darmstadt, Institute and Laboratory of Geotechnics,
large construction project show, that the uplift Germany, 165-181.
Gravert, F.W.; Buch, G. 1974. Probleme bei der Gründung
due to the deconstruction are nearly as big as the der Dom-Römerberg-Bebauung in Frankfurt am Main.
settlements occurred due to the construction. The Bautechnik 51, Heft 2, Ernst & Sohn Verlag, Berlin, 52-
uplifts cannot be avoided. 58.
To guarantee the stability and the serviceabil- Katzenbach, R.; Bachmann, G.; Leppla, S.; Ramm, H. 2010.
ity of large construction projects in difficult soil Chances and limitations of the observational method in
geotechnical monitoring. 14th Danube-European Con-
and groundwater conditions in the area of under- ference on Geotechnical Engineering, 2.-4. June, Brati-
ground structures the following aspects have to slava, Slovakia, 13 p.
be considered: Katzenbach, R.; Leppla, S.; Seip, M. 2011a. Das Verfor-
mungsverhalten des Frankfurter Tons infolge Baugrun-
x For the precise determination of the soil dentlastung. Bauingenieur, Band 86, Heft 5/2011,
Springer VDI Verlag, Düsseldorf, Germany, 233-240.
deformation comprehensive numerical analy-
Katzenbach, R.; Leppla, S.; Weidle, A.; Werner, A. 2011b.
sis covering the non-linear deformation be- Das Vier-Augen-Prinzip in der Geotechnik: Der Prüf-
haviour are necessary (Wörner & Pfeiffer sachverständige für Erd- und Grundbau. Geotechnik-
1998). Kolloquium anlässlich 60. Geburtstag von Prof. Dr.-
x Regarding the observational method a geo- Ing. Dietmar Placzek, Universität Duisburg-Essen, 26.
detic and geotechnical measurement program May, Essen, Germany, 255-267.
Katzenbach, R.; Leppla, S. 2013a. Deformation behaviour of
has to be developed and installed (Katzen- clay due to unloading and the consequences on con-
struction projects in inner cities. 18th Conference of the
34 R. Katzenbach et al. / Numerical Analysis and Verification of the Soil-Structure-Interaction
Abstract: With rapid development in metropolitan areas, deep excavation and tunnelling are often carried out close to existing
buildings or infrastructure. The soil movement due to excavation may cause damages to adjacent foundation, worse for shallow
foundations as compared to deep foundations. In this keynote paper, the latest development and understanding of soil-structure
interaction involving foundation subject to adjacent excavation are presented; with references to successfully implemented
projects or research work based on finite element modelling, centrifuge experiments and field monitoring, observations and
interpretations. The novel concept of limiting soil pressure due to excavation stress relief is also presented.
Retaining wall
Pile Laser sensors
Zinc LVDTs
Solenoid valve
chloride
470
PPTs
PPT 1 Clay
210
PPT 4 PPT 2 PPT 3
40
Sand
ground settlement continues to increase over Figure 9 shows that the maximum induced
time after the completion of excavation, as pile bending moment is located at 8.75 m below
shown in Figure 8. The long term time dependent the ground level. The induced bending moment
wall deflection and settlement troughs have been initially increases with increase in excavation
further investigated in detail by Ong et al. (2003, depth. A maximum value of 236 kNm is
2004).
recorded at an excavation depth of 1.2 m. The
0 bending moment then decreases with increase in
excavation depth. At the maximum excavation
depth of 1.8 m, the bending moment reduces to
185.8 kNm. Thereafter, the bending moment
2 profile is found to decrease further over time.
Excavation
Depth (m)
Symbol
depth (m) Bending moment (kNm)
0.6 0 50 100 150 200 250
4 1.0 0
1.2 Symbol
1.4 Excavation depth (m) 1.0 1.2 1.4 1.6 1.8 1.8 1.8
1.8
2.5 Days after completion
0 0 0 0 0 22 240
Time after of excavation
6 excavation (days)
50
Depth (m)
200
5
300
8
0 0.4 0.8 1.2 1.6
7.5
Lateral wall movement (m)
0 12.5
Excavation
Figure 9. Development of pile bending moment profile over
Symbol depth (m) time (Ong et al., 2004)
Settlement (m)
0.6
0.8
0.3 High-resolution photographs were taken
1.0
1.2
during various excavation stages of the test, as
1.4
shown in the left-hand side photographs of
1.6 Figure 10. It is evident that tension cracks have
1.8 developed when the excavation depth exceeds
0.6 Time after
excavation (days)
1.0 m.
50
200
300
0.9
0 4 8 12 16
Distance from wall (m)
Pile head
100
50
0
0 0.3 0.6 0.9 1.2 1.5 1.8
Excavation depth (m)
deduced from the corresponding bending exists. The interception of the two straight lines
moment profiles shown in Figure 9. It is evident denotes the occurrence of wall failure. As such,
that the limiting maximum soil pressure values the excavation had failed at an excavation depth
have been reached at an excavation depth of 1.2 of approximately 5.0 m.
m. Figure 14 shows the induced pile bending
Thereafter, the soil pressures do not increase moment profiles of Test WC1 The location of
further with increasing excavation depth. This maximum pile bending moment is noted to be
observation further reinforces the postulation that about 7.5 m below the ground surface.
when the soil flows past the pile. In addition,
with the presence of the tension cracks in front of 0.0
excavation depth
the pile as described earlier, the soil could not 1m
Deflection
2.5
Rotation
Wall Head Deflection (mm)
10.0
10 1.0 12.5
-80 -60 -40 -20 0 20 40 60 80 -
(a) Shear force (kN)
d
1 0.1
1 2 3 4 5 6 7 8
Excavation Depth (m)
Figure 16(a) shows the outcome of the pile case, the pile bending moment tends to be over-
responses for Case 1 based on Method 1. The predicted, but the deflection is under-predicted as
front pile is located nearer to the slope than the shown in Figure 17(c).
rear pile. In this case, the natural behaviour of This is due to the pile being assumed to be
soil deformation is negated by the presence of uncracked (much stiffer) thus attracting high
the ‘equivalent wall’ (no longer a pile) in a 2-D bending moment and low deflection. The above
environment, resulting in relatively smaller does not simulate the behaviour on site where the
magnitudes than measurements taken on site. pile has experienced lateral movement of more
Consequently, the predicted pile responses than 100mm. This highlights the importance of
(bending moment and deflection) are both very knowing the condition of the pile on site when
much under-predicted, leading to inappropriate performing back-analysis.
design of pile to resist lateral soil movement. In Case 4, if the back-analysis is carried out
In Case 2, if both Icr and py are correctly without having prior knowledge of estimating
adopted, Figure 17(b) shows that the prediction limiting soil pressure and transformed pile
of pile responses is very reasonable. This moment of inertia, I on site due to cracking, the
simulates the available and appropriate level of predicted pile bending moment will be grossly
understanding of the back-analysis carried out over-predicted as shown in Figure 17.
considering the development on site. However, in this case, the ‘reasonable’
Case 3 simulates the situation where estimation of pile deflection is merely a
knowledge on limiting soil pressure is available coincidence.
but not on the pile moment of inertia, I. In such a
Figure 16. Profiles of measured and predicted rear and front pile bending moment and rear pile deflection
(Ong et al., 2010)
C.F. Leung and D.E.L. Ong / Effects of Deep Excavations on Adjacent Foundations 43
Figure 17. Profiles of measured and predicted rear and front pile bending moment and rear pile deflection (Ong et al., 2010)
movement increased with increasing soil Figure 19 shows the computed axial force
movement until the soil movement reached about and bending moment in the pile and clearly
200 mm. It was postulated that at this stage, the reveals that the structural capacity of the pile
soil might have just flowed past the pile and thus section is far exceeded, even for the case where
the pile did not experience any further ultimate steel and concrete strengths are used.
movement. Figure 20 compares the maximum bending
moment and axial load computed from the
horizontal soil movement acting on the capped
pile of the office building with the structural
capacity of the pile section. Again, the pile
capacity was found to have been far exceeded.
detail. The centrifuge pile group test set-ups are position with its surrounding piles. It is observed
identical to the single pile set-ups as reported by that the induced bending moment for the front
Ong et al. (2006) except that structural pile caps peripheral pile is greater than that of the front
are introduced to tie the individual pile heads for centre pile at the same distance behind the wall.
cases with capped pile head conditions. Similarly, the bending moment developed at the
It is found that the induced maximum rear peripheral piles is also greater than that of
bending moment is always smaller than that of a the rear centre pile at the same distance behind
corresponding single pile (Ong et al., 2006) at an the wall. As the number of piles in a group
identical location. If the free-head piles are becomes larger, the shadowing and reinforcing
located at the same distance, the measured effects become more prominent.
bending moment is higher for the front pile as The immediate effect of pile shadowing and
opposed to the rear pile of the pile group. In a reinforcing effect is to reduce the detrimental
pile group, each individual front pile (3m behind effects of excavation-induced soil movement on
the wall) will provide shadowing and reinforcing the pile group. By capping a pile group, the
effects to the other the rear piles (5m behind the individual pile heads are forced to act in unison
wall), thus reducing the magnitudes of pile when subject to different magnitudes of soil
deflection and bending moment. This is evident movement, depending on the distance of the piles
from Figures 21 and 23. from the wall.
The induced bending moment of the front
pile, which experiences a greater soil movement,
is moderated by the rear pile through the pile
cap. The interaction between the front and rear
piles induces negative bending moment at the
pile head, but reduces the magnitude of bending
moment developed along the pile and the pile
group deflection as observed in Figure 22.
Soil arching and “separation” of soil as
shown in Figure 23, have been observed to occur
between the front piles of a pile group when the
soil moves upon excavation, in the 4-pile group.
The arch is formed between the rows of piles
when the yielded soil gets detached from its
Figure 21. Predicted and measured pile (a) bending moment
and (b) deflection profiles for free-head 4-pile group (Test
surrounding. The detached soil is then forced to
12) (Ong et al., 2009, 2011) squeeze between the row of piles but without
significantly increasing the pressure acting on the
piles.
The degree of shadowing experienced by Generally, the observed long term maximum
each individual pile depends on its relative positive bending moment would increase after
46 C.F. Leung and D.E.L. Ong / Effects of Deep Excavations on Adjacent Foundations
excavation until about 50 days later and distribution profiles along the entire pile shaft in
subsequently reduce with time. It is believed that the soft clay.
progressive wall and soil deformations are the
reasons for such observed time dependent pile
behaviour. On the contrary, the maximum
negative bending moment generally reduces
slightly over time after excavation. This
behaviour could be the result of pile-pile cap
interaction as the maximum negative pile
bending moment is located nearer to the pile cap.
To account for the pile group shadowing and
reinforcing effects, an empirical soil moderation
factor ks was introduced to “correct” the
measured free-field soil movements. By back-
analysis, the magnitude of soil moderation factor Figure 24: Shaft resistance fully mobilised at 11.1 MN
ks is established to be 0.8 for a 2-pile group, 0.7 (without stress relief) and 5.0 MN (with stress relief)
for a 4-pile group and 0.5 for a 6-pile group in respectively (Ng et al., 2014)
clay as compared to 0.9 for a 2-pile group, 0.8
for a 4-pile group and 0.6 for a 6-pile group in After the respective shaft resistances have
sand as reported by Leung et al. (2003). been fully mobilised, the pile toe which is
founded on dense sand starts to mobilise its toe
3.3. Case study 3: Pile behaviour subjected to resistance in the tests. It can thus be concluded
that the shaft resistance of the low friction piles
excavation stress relief
was proportional to the normal stress on the pile
shaft as validated by theoretical calculation of
Ng et al. (2014) performed centrifuge model tests shear stress, = n’ tan ’, where n’ is the
to determine the capacity and deformation of a effective normal stress and ’ is the interface
single pile, a 3 x 3 pile group and a piled raft friction angle. The reduction in shaft resistance
foundation, with and without stress relief effects. due to stress relief led to a lower capacity of the
The behaviour of piles in non-dilatant (soft pile subjected to excavation.
clays) and dilatant soils (stiff clay or dense sand) For a pile installed in stiff clay or dense sand
were assessed so as to have a better (with a dilatant pile-soil interface), the increase
understanding on the fundamental mechanisms in pile capacity was attributed to soil dilation at
controlling the pile capacity when subjected to the soil-pile interface, resulting in an increase in
stress relief. This study is analogous to a top- # Qn) acting on the pile shaft
down construction method consisting piles which and hence the pile capacity. The contribution of
are designed from pile load tests carried out at dilation was more significant for a pile subjected
the ground surface and installed before to stress relief since dilation was more
excavation. pronounced at lower stress. The DEM study
Discrete element modelling (DEM) was also showed that an unloading of normal stress from
used to explore the shearing mechanisms at the 400 to 100 kPa can lead to a 30% increase in
pile-soil interface and to estimate the normal n.
stress increment contributing to pile shaft It has also been found that due to excavation
resistance, when a pile undergoes stress relief stress relief, tensile force can occur at 70% to
due to deep excavation. 83% section of the shaft length from the pile
For piles installed in soft clay (non-dilatant head, as a result of the upward movement of soil
soil), it was observed that the reduction in shaft due to stress relief. Therefore, a pile should be
resistance is proportional to the normal stress reinforced appropriately to resist the induced
acting on the pile shaft and hence the depth of tensile force during deep excavation. As loads
excavation. As shown in Figures 24 (a) and (b), were applied at the pile head gradually, positive
the shaft resistance has been fully mobilised at shaft resistance was mobilized along the entire
11.1 MN (without stress relief) and 5.0 MN (with pile length, as well as toe resistance.
stress relief), respectively, as validated by the
constant slope of the measured axial load
C.F. Leung and D.E.L. Ong / Effects of Deep Excavations on Adjacent Foundations 47
4.2. Case study 2: Deep excavation supported by excavation stage and thus illustrates the overall
flexible, tied-back retaining wall behaviour of lateral soil movements. It was
observed that most movements occurred
Ground deformations affected by corner effects perpendicular to the sheet pile retaining wall
were observed during a 12.8 m deep excavation with rather limited out-of-plane movement.
in Chicago as reported by Finno et al. (2005).
The excavation was supported by sheet-piles and
regroutable ground anchors which were carried
out through soft to medium clays. Field
monitoring data consists of lateral, vertical
optical survey and lateral inclinometer data.
The construction stages include wall
installation, caisson construction, excavation and
installation of tied-back anchors on ground
movements. Figure 28 shows the field measured
ground settlements and lateral movements,
respectively. The maximum settlements behind
the wall were about 80% of the maximum lateral
movements when the excavation was within the
soft to medium clay.
Figure 28: Field measured ground settlements and lateral movements (Finno et al., 2005)
C.F. Leung and D.E.L. Ong / Effects of Deep Excavations on Adjacent Foundations 49
It was found that building response depends During the tunnelling works, ground and
on its bending stiffness whereas the horizontal building settlement markers and tape
displacement depends on its axial stiffness. extensometers were used to monitor the
Figure 34 shows the horizontal strain settlement response. Venkta et al (2008) reported
modification factors plotted against the relative that the building settlement for one shop house
axial stiffness of a typical building. An upper- reached 100 mm. The tape extensometer
bound limit can be developed to estimate the measurements were categorised as transverse and
horizontal strain modification factors for longitudinal horizontal strains due to the
building on continuous footings. For example, orientation of the framed buildings with respect
for a 100mm thick reinforced concrete slab that to the tunnelling drive.
is 100m long and continuous and for Es around Using finite element method, the frame
20MPa, the relative axial stiffness is in the order building was modelled where the frame columns
of 100. were unconnected at the footing level for a 20 m
Therefore, similarly, a building with high depth excavation in an undrained condition. The
bending and axial stiffness results in a lower frame was modelled as a weightless structure.
deflection ratio and horizontal strain. Hence, The outcome shows that the horizontal strains
considering the building stiffness for estimating are mostly compressive. Although cracks were
the deflection ratio and horizontal strains, a initiated due to the settlement on the brick walls
lower risk of building damage can be obtained and column brick wall interface, the structural
from the reduction of maximum tensile strains integrity of the building was not affected.
induced by the building. It is well-established that for most buildings
Although having the same elastic stiffness in with continuous footings, the horizontal strains
bending, Goh et al. (2012) reported that a shorter are significantly reduced. However, for buildings
building will have a stiffer deflection compared supported on individual footings such as this
to a taller building. Thus, from the case study, the approach assumes horizontal
understanding, by having the same elastic and strains in the Greenfield condition, which is
geometric properties, a building adjacent to a shown to be reasonable. Therefore, based on this
deeper excavation will have higher modification understanding, a new frame stiffness factor, as
factors for the deflection ratio and horizontal shown in Figure 35, was developed to
strains compared to a shallower excavation, thus characterise the horizontal strain modification
reflecting the nature of soil-interaction effect. factor derived from a rigidly-connected frame
structure on individual footings, using a simple
5.3. Case study 3: Responses of building pin-support portal-frame structural analysis.
supported on individual footings to movements
American Society of Civil Engineers (ASCE), Vol. 128, movement in clay. Proc. 12th Pan-American
pp. 660-671. Conference on Soil Mechanics and Geotechnical
Finno, R.J, and Roboski, J.F. (2005). Three-dimensional Engineering, Massachusetts Institute of Technology,
response of a tied-back excavation through clay. Boston, U.S.A., Vol. 2, pp. 2035-2040.
Journal of Geotechnical and Geoenvironmental Ong, D.E.L., Leung, C.F. and Chow, Y.K. (2004). Pile
Engineering, American Society of Civil Engineers behaviour behind a collapsed wall. Proc. International
(ASCE), Vol. 131, pp. 273-282. Conference on Structural and Foundation Failures,
Goh, K.H, and Mair, R.J. (2011). Building damage Singapore, pp. 410-421.
assessment for deep excavations in Singapore and the Ong, D.E.L., Leung, C.F. and Chow, Y.K. (2006). Pile
influence of building stiffness. Geotechnical behaviour due to excavation-induced soil movement in
Engineering Journal of the SEAGS & AGSSEA, Vol. 42, clay: I: Stable wall. Journal of Geoenvironmental and
No. 3, pp. 0046-5828. Geotechnical Engineering, American Society of Civil
Goh, K.H, and Mair, R.J. (2012). The response of buildings Engineers (ASCE), Vol. 132, No. 1, pp. 36-44.
to movements induced by deep excavations. Ong, D.E.L., Leung, C.F., and Chow, Y.K. (2007). Effect of
Geotechnical Aspects of Underground Construction in Horizontal Limiting Soil Pressures on Pile Behaviour.
Soft Ground-Viggiani(ed), Taylor & Francis Group, 16th South-East Asian Geotechnical Conference
London. (SEAGC), 8-11 May 2007, Kuala Lumpur, Malaysia.
Goh, K.H, and Mair, R.J. (2012). The horizontal response of pp. 427-437.
framed buildings on individual footings to excavation- Ong, D.E.L., Leung, C.F. and Chow, Y.K. (2009). Behaviour
induced movements. Geotechnical Aspects of of Pile Groups Subject to Excavation-Induced Soil
Underground Construction in Soft Ground- Movement in Very Soft Clay. Journal of
Viggiani(ed), Taylor & Francis Group, London. Geoenvironmental and Geotechnical Engineering,
Leung, C. F., Chow, Y. K. and Shen, R. F. (2000). Behaviour American Society of Civil Engineers (ASCE), Vol. 135,
of pile subject to excavation-induced soil movement. No. 10, pp. 1462-1474.
Journal of Geotechnical and Geoenvironmental Ong, D.E.L., Leung, C.F. and Chow, Y.K. (2010). Effect of
Engineering, Vol. 126, No. 11, pp. 947-954. limiting soil pressure on pile group adjacent to a failed
Leung, C.F., Lim, J.K., Shen, R.F., and Chow, Y.K. (2003). excavation. Proc. of International Conference on
Behavior of pile groups subject toexcavation-induced Geotechnical Challenges in Megacities, Vol. 3, pp. 785-
soil movement. Journalof Geotechnical and 792, 7-10 June 2010, Moscow, Russia.
Geoenvironmental Engineering, American Society of Ong, D.E.L., Leung, C.F. and Chow, Y.K. (2011). Behaviour
Civil Engineers (ASCE), Vol. 129, No. 1, pp. 58-65. of Pile Groups Subject to Excavation-Induced Soil
Leung, C.F., Ong, D.E.L. and Chow, Y.K. (2006). Pile Movement in Very Soft Clay. Discussion by F. Castelli
behaviour due to excavation-induced soil movement in & Closure, Journal of Geoenvironmental and
clay: II: Collapsed wall. Journal of Geoenvironmental Geotechnical Engineering, American Society of Civil
and Geotechnical Engineering, American Society of Engineers (ASCE), Vol. 137, No. 1, pp. 110-113.
Civil Engineers (ASCE), Vol. 132, No. 1, pp. 45-53. Poulos, H.G., (1997). Failure of a building supported on
Maugeri, M., Castelli, F. and Motta, E. (1994). Analysis of piles. Proc. Int. Conf. on Foundation Failures,
piles in sliding soil. Proc. 3rd Int. Conf. on Deep Singapore, pp. 53-66.
Foundation Practice Incorporating Piletalk, Singapore, Schofield, A.N. (1980). Cambridge geotechnical centrifuge
pp. 191-196. operation. Geotechnique, 30, 3, 227-268.
Ng, C.W.W and Peng, S.Y. (2014). Three-dimensional Venkta, R., Hoblyn, S., Mahatma, S. & Lim, H.C. (2008).
centrifuge and numerical investigation of the EPB tunnelling under 2-storey shophouses in mixed
performance of piled foundations subjected to deep face conditions. Proceedings of International
excavation. Advanced in Foundation Engineering. Conference on Deep Excavation, 10-12 Nov 2008,
Ong, D.E.L., Leung, C.F. and Chow, Y.K. (2003). Time- Singapore.
dependent pile behaviour due to excavation-induced soil
54 Soil-Structure Interaction, Underground Structures and Retaining Walls
V.M. Ulitsky et al. (Eds.)
IOS Press, 2014
© 2014 The authors and IOS Press. All rights reserved.
doi:10.3233/978-1-61499-464-0-54
Abstract: The lecture will cover mainly the importance of stiffness parameters of crushable type of foundation sands in the
perspective of the state of the art of soil testing and its new developments.
Any soil structure interaction model (SSI) can be deployed successfully or can fail depending on the reliability of the esti-
mations of the soil structure interaction stiffness parameters. The stiffness evaluation of the engineered materials’ structural
elements and their combined role in the structure itself, commonly can be evaluated rather easily. However, the key issue and the
much more problematic interaction parameters in any successful SSI analysis are linked to the soil stiffness, at the relevant strain
levels, of the specific interacting soil layers under the corresponding structural loading conditions.
Keywords. crushable sands, soil-structure interaction, sand stiffness, oil tanks, settlement
Fig. 1. The site and location of the 3 oil tanks at Ostend – Belgium.
Mainly
hydrau-
lically
deposited
dredged
material of
some 30
years old
Stiff tertiary OC
clay ~100m thick
Fig. 2. Characteristic CPTU result at the site of the oil tanks in Ostend.
56 W.F. Van Impe and P.O. Van Impe / Comments on Crushable Sand Stiffness Relevant to SSI Issues
Fig 3a (begin). Settlement increase pattern of the tanks nr 1 and 2, in the phases of the water testing
and after the full load oil filling.
W.F. Van Impe and P.O. Van Impe / Comments on Crushable Sand Stiffness Relevant to SSI Issues 57
Fig 3a (end). Settlement increase pattern of the tank nr 3 in the phases of the water testing
and after the full load oil filling.
pore water pressures due to the required squeez- higher (depending on the soil skeleton fabric),
ing out of the pore fluid does exist, but is of no fig. 5.
meaning in practice because of the extremely In carbonate sands however, this threshold
short time frame linked to such “consolidation” mean stress level triggering the relevant crushing
effects. already starts at far lower values and so well
In carbonate sands, due to crushing of the within the common geotechnical SSI-engineering
particles and increasing the fines%, this phenom- mean unit loading ranges (1 to 8 MPa), fig 5.
enon of consolidation will become more im- Particle breakage in crushable sands is a pro-
portant. It also means that ageing of a carbonate gressive process, ongoing for a long time, even
deposit will lead to continuously changing cv - under constant mean stress levels, because of the
values. It however will always remain negligible wide dispersion of the real inter-particle contact
as compared to any soft layer consolidation stresses. Such inter-particle contact stresses in
behavior which might come into the picture of the sand are varying as well all the time, due to
the SSI analysis on such construction sites. local crushing and subsequent particle displace-
Relevant sand particle crushing usually ment. The secondary deformation curves are
should be considered for the long term defor- therefore no longer continuous smooth curves,
mation behavior, as soon as the mean stress level but do show rather stepwise decaying shapes.
of the soil-structural interaction element does rise An increase for example of not more than
above a certain threshold value. For quartz sands 10% of fines in an embankment (either due to
it means: above a value of the order of some 15 crushing or due to varying grain size of the
MPa; minor crushing of quartz soil grains might “various sands” deposited on site), leads easily to
already appear from a mean stress level of some an increase of the volumetric strain level of
5 to 8 MPa. But generally this type of crushing is around 10% – 12% of carbonate sand layers.
not influencing the overall settlement behavior of Such secondary deformation behavior in
the deposit too strongly. principle may be described through effective
The level of 15 MPa is of course a mean ef- stress based models such as the one associated
fective stress level; the real interparticle contact with the Taylor rheological set up. This model is
stress levels do rise easily up to 4 to 10 times probably for this type of crushable material more
W.F. Van Impe and P.O. Van Impe / Comments on Crushable Sand Stiffness Relevant to SSI Issues 59
appropriate. Indeed, in the Taylor approach, the irrelevant. Similarly, in case of carbonate
secondary stress strain behavior keeps to be sands,(fig 7) parameters of the changing sand
uncoupled and is not linked to the stress level at granulometry, the relative density, particle shape,
the end of the primary consolidation period. For etc… could become irrelevant as well from a
sands this might be a better approach, (fig. 6). much more moderate mean effective stress level
For silica sands the long-term stress strain ( for example 5 to 8 MPa).
curve at common SSI stress level increments is However, the degree of saturation will con-
depending on all the above mentioned parame- tinue to play all over its predominant role when it
ters, up to mean stress levels of about 15 MPa. comes to influencing the stiffness of crushable
From that level on, in practice only the particle sands. Indeed, from many data in our own past
strength is left as the governing parameter for the years’ lab research – mainly on Middle East
deformations:. All other parameters become carbonate sands (bio-clastic origin), the presence
Fig. 5.
Fig. 6.
60 W.F. Van Impe and P.O. Van Impe / Comments on Crushable Sand Stiffness Relevant to SSI Issues
of pore water clearly promotes unexpected fully (in small steps) draw the “e – ln p’ ” curve
additional crushing of particles! as well as the tangent constrained modulus:
Bio-clastic sands contain particles of skeletal Moed = 'V’v / 'Hv , on the vertical axis
origin, such as shells, coral reefs or similar, versus lnp’ on the horizontal axis.
together with broken organic material fragments At the “yield point” of the “e – ln p’ “graph,
(pictures, fig.7a,b ). Such particles are structural- one might expect a minimum value of the Moed
ly weak due to their shape and the weakness of to correspond with this yield point level. The “e”
the composing carbonate materials. It results for is commonly linked to the relative density Dr
calcareous sands (fig 8a,b). from conventional lab testing. The concept of a
It’s obvious (cfr Been et al 1985) that due to relative density Dr has been developed to allow
the fines content, increasing during the crushing for comparing the behavior of different materials
of particles process, the secondary sand matrix at equivalent “density”.
stiffness is decreasing. However, the parameter of relative density Dr
We therefore would recommend to very care- is not an inherent soil property. It depends on the
Fig. 8a.
W.F. Van Impe and P.O. Van Impe / Comments on Crushable Sand Stiffness Relevant to SSI Issues 61
'v = 1+e
'e ~ A – B log p’
Fig. 8b.
determination of the minimum and maximum tions mentioned, (fig 9a,b), were showing very
density of the soil, which themselves are the good reliable correspondence with the predicted
result of index tests, very sensitive to the meth- values according to the equations above.
odology applied for the input energy. In the case of crushable soils, we encounter
the additional problem of the validity and the
3. Relevance of relative density much stronger influence of the test-method to
determine maximum density. Values of maxi-
The accuracy at which Dr can be determined is mum density depend very much on the chosen
considerably low. We need to consider errors on densification method (densification energy).
the determination of the minimum, maximum High densification energy will lead to higher
and actual sample density. Any small error in maximum density, but will induce more im-
either of these values leads to disproportionate portant crushing. This means that we are actually
errors in the calculated value of Dr (e.g. a 1 % never dealing with identical soil material, and a
error in the minimum density could lead to more value of maximum density becomes irrelevant.
than 10 % error in the value of Dr). As reported by P. Main et al (2009) (fig. 10);
From our own research over the last 5 years, from related to calibration chamber testing, a
in carbonate Middle East sands, it became clear very general mean conversion factor would be
that we should try to deal more directly with around F = 2.3. This is of course a very much
“void ratio” (e) related to the elastic wave veloci- oversimplified approach. In the past three dec-
ties measurements on site and the corresponding ades indeed, a number of studies have been
effective stresses involved. developed to define a correlation factor between
In this respect, a proposal of such possible the cone resistance measured in silica (or quartz)
correlations (P.O. van Impe -2013), for the sand and later on as well calibration chamber
Middle East sands in hydraulically deposited type of centrifuge tests were performed as well
vibro-compacted sand masses, extensively tested, for crushable sands such as the calcareous sands.
is: As shown in the Table 1 below, for each car-
Vs = Cs (’v/pa)nv(’h/pa)nhe-d bonate sand tested, one assumes the lower and
Vp = Cp (’v/pa)mv(’h/pa)mhe-f upper bounds of a type of conversion factor F
with the correlation coefficients: defined as:
Cs = 248; d = 0.291;
nv = 0.164; nh = 0.142; F=
, when comparing crushable
Cp = 1661; f = 0.1753; ,
sands with silica sands
mv = 0; mh = 0.0323 ,
when comparing crushable
F=
,
sands with quatrz sands
The on-site measured data analysed (P.O.
Van Impe -2013), using the indicated correla-
62 W.F. Van Impe and P.O. Van Impe / Comments on Crushable Sand Stiffness Relevant to SSI Issues
Lower Upper
Sand Bound Bound Note Reference
F F
Quiou calcareous and Bellotti/Jamiolkowski
1.3 2.2 Depending on Dr
Ticino Silica sand (1991)
Quiou calcareous and
1.8 2.2 Depending on Dr M.Almeida et al (1992)
Ticino Silica sand
Patrn island calcareous and G.Gudehus, Cudmari
1.5 1.7 Depending on Dr
quartz sand (2004)
Dogs Bay calcareous and Yasufuku et al.
4.0 8.0 Depending on Dr
Mol quartz sand (1995)
Quiou calcareous and
2.0 4.0 Depending on Dr Bellotti (1991)
Mol quartz sand
Bioclastic calcareous
at depths < 7m under water Van Impe
(Banco Archimedes Uruguayan) 1.4 2.3
to 4 m above water (2010)
and Mol quartz sand
Abu Dhabi (S1)
Estimate Van Impe
Bioclastic calcareous sand Dr a60 %
1.8 (2012)
and Mol sand
It’s important to notice that most stiffness 4. The estimation of the sand stiffness
correlations appear to show a very weak depend-
ency on relative density Dr. There is moreover The Moed, mentioned before as the tangent
hardly any connection between the actual stress- oedometric (constrained) stiffness, or even the
strain behavior at the site and the determined so- simple young stiffness modulus Ey, can also be
called Dr parameter. Such stress-strain behavior derived from on-site-test results such as the CPT
is indeed strongly controlled by the stress-state, data; based on empirical correlations such as :
particle shape, particle mineralogy, stress history Moed (either Ey)= D. qc ; with the values of
etc., which are not (fully) reflected in the value D f(Dr, stress history) , decreasing with rising
of Dr nor in the cone resistance qc. The SCPT relative density Dr of the sand, (for example – fig
results on the test site of our research areas 11), and increasing remarkably with more
clearly do support this statement. pronounced O.C. and ageing.
Since the determination of relevant correla- From empirical qc-V’v- Dr relationships, vary-
tions to the Dr parameter remains a long and ing obviously with increasing depth, (fig 12a,b),
expensive lab testing process, while the final lots of commonly used approaches for various
result is not at all accurate enough to make any sand type are suggested in literature. From
kind of relevant prediction of the sand stiffness research on the corresponding sand type, lab test
and so of the deformation behavior of the site. based correlations (ex. Fig 13 for Quiou sands)
For this reason, it is in our opinion much of the Dr and Moed at increasing stress levels,
more efficient to measure these stiffness related consequently would allow the geotechnics expert
parameters directly on the site. The quality designer to also make use of in situ CPT meas-
control procedure would then consist of checking urements to evaluate reliable qc- Moed corre-
the measured parameters (e.g. shear wave veloci- spondences at that site, for that type of sand. The
ty Vs and derived small-strain shear modulus G0 above mentioned values “D” are actually the
from SASW combined with SCPT) to see if they outcome of such approach.
can fit the relevant design requirements (settle- The reliability of the empirical relationships
ment, bearing capacity, liquefaction). of such stiffness parameters when starting from
non-crushable sand material has proven to be
high and fully reliable for common geo-
engineering practice and modeling such as
required for the SSI approaches.
W.F. Van Impe and P.O. Van Impe / Comments on Crushable Sand Stiffness Relevant to SSI Issues 63
450
400
350
300
SARB1
250
-5%
5%
200
Vs measured (m/s)
100
Fig. 9a. ( P.O. van Impe – 2013) 100 150 200 250 300 350 400 450 500
P.Mayne 2008
Fig. 11.
following interacting soil parameter relations: iv. G – decay curve, based on the Bender
i. qc – Dr - V’v from calibration chamber element, the resonant column and the strain
tests in centrifuge K0 conditions monitored TX test results.
ii. I’ as a function of the strain levels This should allow the geo-engineering expert,
measured in careful strain controlled TX on the basis of CPT results on site, to derive the
iii Vs – V’v – V’h – e , as well as G0 – V’v – following parameters:
V’h – e correlations based on bender element i’. qc - Dr as a function of depth, on site
TX and resonant column tests, in the adapted lab ii’. G0 as a function of depth (from I’),
test stress conditions (K0) (from iii) at corresponding K0 (from ii)
W.F. Van Impe and P.O. Van Impe / Comments on Crushable Sand Stiffness Relevant to SSI Issues 65
iii’ G and so the constrained stiffness at the where the values of the correlation coefficients
various strain levels, from (iv). Ci,avg are given in the Table 2 below. These
The analysis (P.O. Van Impe 2012), for this correlations are significantly different from what
type of highly carbonatic Middle East sand (and can be expacted for silica sands. The deviation
going out from a series of K0 centrifuge calibra- from the “typical” silica sand behavior becomes
tion chamber tests) has allowed to establish the more significant for higher stress levels and
correlation between CPT cone resistance qc, higher relative densities.
relative density Dr and the vertical effective In the middle East carbonate sands under in-
stress level V’v for the specific soil materials vestigation, at the test site and based on the
used at the hydraulically deposited sand site,(cfr results of CPT tests, values of the relative density
also fig 12b). Dr were determined along a CPT profile using
the above correlations.
V’v > 50 kPa
",
Table 2.
= ., . , . ! (1)
Parameter Above GWT Below GWT
V’v d 50 kPa C0 25.410.375 12.95
.#$ ", C1 2.265 0.27
# ",
= ! . ., . , . ! (2) C2 2.76
#
qc
50 Dr 190 78 log 10
V ' vo
100
W.F. Van Impe -1986
150 Mol sand
200
250
300
350
400
Dr=30% 50% 70% 80% 90% 100%
Fig 12a (Van Impe -1986) silica Mol sand in very large Calibration chamber and CPT(M)
66 W.F. Van Impe and P.O. Van Impe / Comments on Crushable Sand Stiffness Relevant to SSI Issues
We could link these parameters to the void Where pa is the reference atmospheric pres-
ratio e and the parameters V’h and V’v , of the sure (100 kPa). The A, B, C and D parameters
sample. Following correlations were established for each equation are fitted to the data of the BE
for P-wave velocity VP, shear wave velocity VS test in the laboratory triaxial equipment. Addi-
and small strain shear stiffness G0 , with the tionally, local strain measurements during
proposal: triaxial tests have allowed to determine the
%& &
degradation of the stiffness as a function of
VS = A S !
! ' (3) strain-level (stiffness degration curve). It wzs
found that the values for the small-strain shear
%( (
Dr1
Dr3 Dr2
It moreover was derived from all on site test- density Dr (derived from the local qc value). One
ing results that the performing series of CPT as a only can conclude that the relative density Dr
stiffness controlling test in deposits of carbonate does not reflect the actual behavior of the materi-
sand masses is not a recommendable methodolo- al and is therefore irrelevant as a quality control
gy, for not really reliable at all. As we argued at parameter.
several occasions, our recommendation on the Attempts were made to predict the in situ
contrary would go out from the carbonate sand stiffness profile based on the common CPT test,
testing on site by means of SCPT(u), combined using the correlations determined during the
on site with SASW and DMT. laboratory testing campaigns. It is shown that,
For understanding the behavior of the car- although our predictions were significantly better
bonate sands, the depth versus Dr , G0 and Vs than those based on “typical” correlations for
prediction curves, derived from well elaborated silica sands, in some cases the carbonate sand,
testing programs, with lots of data of SCPT cone the soil stiffness was significantly underestimat-
values is not enough; it is also becoming una- ed. This is due to the fact that the reliability of
voidable to dispose of a reliable estimate of the such “qc-stiffness” correlations are strongly
on-site stress conditions (V’v and V’h). This has depending on the parameters (K0, OCR,…)
been briefly mentioned here above under i/ to iv/ which cannot be “measured” on site, but have to
and the equations 1/ to 5/ . be estimated, with the appropriate geo-
Once below the mechanically OC upper engineering judgment.
zones of hydraulically deposited sand embank- Certainly at sites where various densification
ments, the predictions on the basis of the earth- techniques were applied and so do lead to a very
quake K0-steered equations proposed, are very complex stress history in the soil, adequate in
well matching the stiffness data predicted for situ testing is the only way to obtain relevant
common carbonate sand deposits out of the information on the relevant stress-strain parame-
testing highlighted above. ters (like Vs), and so on the corresponding
In any case, geo-engineering quality control relevant sand stiffness characteristics. In our
of such densified carbonate sand mass proves opinion, a combined on-site use of SCPT(u)
that hydraulic clapping energy or/and vibro- testing, coupled to DMT and SASW testing
compaction ground improvement technique in profiles, would be the most appropriate way to
such crushable sand material, actually do not deal with the evaluation of the deformation
lead to a densification as such, but are actually behavior of such crushable sand fills and em-
only leading to a ground improvement, because bankments.
of the very strongly changing stress field condi- It is the greatest challenge for an advanced
tions (from K0 to very O.C. range). SSI methodology to model such stress conditions
influences in connection to the deformations to
6. Summarising Conclusions derive from the full soil-structure interaction
approach.
In carbonate sand masses, in situ quality control
testing should at least start from a series of
(seismic) SCPTu (measuring the standard CPT
cone resistance, friction angle and pore water
pressure, combined with the shear wave velocity
Vs) and zone load tests.
Based on the SCPT result and the qc – Dr cor-
relations from laboratory tests, the relative
density profile can be determined for each SCPT
profile. Additionally, one can try to correlate the
measured shear wave velocity Vs and the corre-
sponding cone penetration resistance qc. From
this, it became clear that, in carbonate sands,
there is absolutely no correlation between these
two parameters. The same conclusion can be
reached when comparing the measured shear
wave velocity Vs and the corresponding relative
68 Soil-Structure Interaction, Underground Structures and Retaining Walls
V.M. Ulitsky et al. (Eds.)
IOS Press, 2014
© 2014 The authors and IOS Press. All rights reserved.
doi:10.3233/978-1-61499-464-0-68
Abstract: Piles are powerful geotechnical foundation elements which are suitable for most subsoil conditions where structure
loads should be transmitted to deeper layers either to fulfill the required bearing capacity or to control the deformation to
acceptable values regarding the structural serviceability requirements. In most cases, the pile foundations consist of group of
piles. Realistic considerations of the pile group action regarding both the ultimate bearing capacity as well as the deformation
behavior of the pile group are necessary to achieve a reliable, efficient and economic design. This paper deals with performance
of pile groups under vertical compression loads. Simple analytical methods to estimate the settlement of pile groups will be
presented and compared with enhanced three dimensional numerical analyses. The field monitoring of a well documented pile
group under vertical compression loads and the feed-back of this information into the analyses procedure of the same pile group
applying different analytical models will be demonstrated and discussed.
Keywords. pile group effects, field tests, settlement, bearing capacity of piles, failure mechanism
The load-settlement behavior of the piles in a There are two types of pile group action which
pile group, in many cases, is completely different affect the pile group behavior. The first one
from the behavior of the corresponding single considering the ultimate bearing capacity of the
pile. The group action presents the behavior of pile group in comparison with that of the corre-
the pile group in comparison with that of the sponding single pile (pile group efficiency). The
single pile. Pile group action plays an important group efficiency is defined as the ratio of the
role for the behavior of piled foundation either ultimate load capacity of a pile group to the sum
under vertical tension and compression loads or of the individual ultimate pile load capacities.
under horizontal loads. This paper will deal only The second type of pile group action considers
with the behavior of pile groups under vertical the settlement of the pile group in comparison
compression loads. The pile group action is with that of the corresponding single pile under
considered either by adapting simple correlations the same average load. Figure 1 illustrates
or by relating the pile group to simplified foun- schematically these two types of the pile group
dation forms or by applying advanced numerical actions.
analyses. Comparison between different methods
will be presented and discussed in the presented 2.1. Pile group efficiency
paper.
As it is almost impossible to take into ac- There are two failure mechanisms that can be
count, in an adequate manner, all details of the applied to estimate the ultimate load of a pile
soil disturbance caused due to pile installation by group (Kishida and Meyerhof, 1965). Failure of
theoretical means, pile load tests on single piles the individual piles in the pile group or a block
are frequently carried out to determine the load- failure of the whole pile group can occur depend-
settlement behavior of a single pile. On the other ing on the soil conditions and on the pile spacing.
hand, it is expensive and may be not possible to In case of failure of individual piles, the ultimate
carry load tests on pile groups. Therefore, the load of a pile group is the sum of the ultimate
utilized analytical tools and soil data in this load of the piles of a pile group and the pile
research are first validated based on the results of group efficiency is equal to unity. In case of
pile load test before extrapolating the analysis to driven piles in sand, the pile efficiency may be
pile groups. larger than unity (Kézdi 1957, Fig. 2a). This
increase of the bearing capacity is due to the
densification of the sand by successive driving of
the individual piles. Liu et al. (1985) conducted
an extensive field test series on bored pile groups
Y. El-Mossallamy / Pile Group Action Under Vertical Compression Loads 69
in sandy soil with and without the cap being in group performance, failure of the individual piles
contact with the soil. The results of these field govern the pile group capacity. A pile group
tests showed that there is no evidence of a block efficiency of unity can be considered in such
failure. It can generally be concluded that in cases as a lower bound that lies in the safe side.
cases where effective stresses control the pile
a- Pile group efficiency in sand (Kézdi 1957) b- Pile group efficiency in clay (Sowers et. Al 1961)
Fig. 2: Pile group efficiency
The second failure mechanism is the block Qultimate, group = Ultimate resistance due block
failure of the pile group as a whole. In this case failure
the ultimate load of a pile group can be deter- Nc: Bearing capacity factor ranging between 5
mined by the following expression for pile and 9 (according to Skempton, 1953)
groups in cohesive soil (Sowers et. al 1961): cub = Undrained shear strength of clay at and
Qultimate, group = (cub Nc B W) + (2 (B+W) Lp cus) below pile tips
cus = Average undrained shear strength of clay
(1)
along pile shaft
where: B = Breadth of group
W = Width of group
Lp = Length of piles
70 Y. El-Mossallamy / Pile Group Action Under Vertical Compression Loads
The calculated pile group efficiency apply- and El-Mossallamy 2002) are used to study the
ing equation 1, for pile groups in soft to firm performance of pile groups under vertical
clay with pile spacing larger than three times the compression loads.
pile diameter, is almost equal to unity. There- The performance of pile groups of 2×2, 3×3
fore, the failure of the individual piles in a pile and 4×4 with different pile length to pile diame-
group governs the group behavior for these ter ratio and different pile spacing to pile diame-
conditions. The measurements of the small scale ter ratio were investigated applying a real three
load tests in soft clay (Sowers et al. 1961, Fig. dimensional analyses. The depth of compressi-
2b) agree well with these results. The calculated ble soil is taken in all these analyses correspond-
efficiency of the pile groups in stiff to hard clay ing to twice the pile length. The effect of the
is always greater than unity for all pile spacing. depth of compressible layer was studied in
Therefore, for most practical applications, the separate analyses. In the first analyses stage,
bearing capacity of the individual piles in a pile stress strain behavior of the soil as well as the
group governs the whole pile group capacity and pile-soil interaction is considered as linear
the pile group efficiency can be considered elastic. The effect of soil and pile-soil nonlinear
equal to unity. The block failure can only be a behavior at pile shaft and pile base is considered
decisive design criterion in the case of pile in a separate stage of analysis.
groups in soft clay with smaller pile spacing Figure 3 shows the results of the first stage
(smaller than three times the pile diameter) or in of analysis. The pile group action is defined as
cases where there is a relatively weak layer just the ratio between the settlement of the pile
beneath the pile tips causing punching of the group connected with a rigid pile cap to the
whole pile group as a block. settlement of the corresponding single pile
acting on the average load of the pile group.
Increasing the pile spacing decreases the pile
2.2. Pile group action group action. The pile group action controls the
group performance of floating piles in homoge-
neous soil even in case of large pile spacing.
In many cases, the absolute settlement as well as The number of piles in the pile group has a
the differential settlement is the governing recognizable effect on the pile group action. The
aspects of the foundation design. The load- pile group action increases with increasing the
settlement behavior of the pile group is totally number of piles in the pile group. This effect is
different from the behavior of the corresponding more pronounced by small pile spacing than by
single pile depending on the geometry of the large spacing.
foundation, the number of the piles within the Figure 4 shows the relation between the
pile group, the pile-pile spacing, the depth of group action and the pile length to pile diameter.
compressible layer and the applied load level. The pile group action increases with pile length
to pile diameter ratio. The effect of pile length to
pile diameter on the pile group action depends
2.2.1 Factors affect pile group action on the pile spacing and on the number of piles in
the pile group. This effect is more pronounced
Numerical analyses can be utilized to study the with increasing the number of piles in the pile
pile group behavior and to evaluate factors group.
affecting it. Three dimensional analyses apply-
ing the boundary element method (El-
Mossallamy/Franke 1997, Franke et al. 2000
Y. El-Mossallamy / Pile Group Action Under Vertical Compression Loads 71
Fig. 3 Pile group action regarding pile spacing and number of piles
small strain deformation modulus. In the later is estimated as a ratio either of the pile length or
technique the deformation modulus of the soil of the pile group breadth. It was found that the
depends on the strain level. At very deep layer effective depth of the compressible layer could
below pile tip, the strain due to pile loading is be taken in practical application in the range of
too small and hence corresponding deformation 1.5 times the pile length or at least the smallest
modulus is too high. Applying such models can group width below the pile tip. It is apparent
help to determine the effective depth in a more that this topic needs more research.
rational way. Alternatively, the effective depth
Fig. 6 Load settlement performance of single pile and the corresponding average behavior of the pile group
in different soil types.
Y. El-Mossallamy / Pile Group Action Under Vertical Compression Loads 73
The effect of nonlinear response on the the failure condition. Such a decrease affects the
pile/pile interaction was studied in the case of group settlement and leads to a more uniform
4×4 pile group with pile spacing to pile diame- distribution of load within the pile group when
ter of 3.0. The piles were connected with a failure is approached. This example shows the
relatively rigid pile cap and were assumed importance of considering the nonlinear per-
floating in homogeneous soil. Two extreme formance of the pile group in cases where the
types of soil were studied. Pure cohesive stiff effective shear parameters govern the pile group
clay with undrained shear strength of 100 kPa performance or in cases where higher load
as well as a pure cohesionless soil with angle of levels are allowed to be mobilized within the
internal friction of 35° was considered in these pile group such as piled rafts.
analyses. The effective depth of the compressi-
ble soil was taken twice the pile length. The 2.2.2 Simplified methods to determine the pile
load-settlement behavior of the pile group and group settlement
the corresponding single pile were calculated
applying a nonlinear boundary element method. Simplified methods to predict the pile group
Figure 6 shows the calculated load-settlement action under working loads are very useful at
behavior of the single pile compared with the least in the first design stage to give a quick
average one of the pile group. estimation of the foundation dimensions (e.g.
Figure 7 shows the effect of the pile group number of piles, pile lengths and diameters) that
interaction on the load-settlement behavior of can fulfill the design serviceability require-
the 4 x 4 pile group depending on the load level ments. Available simplified methods to predict
(applied load related to the ultimate one). The the settlement of pile groups may be classified
ultimate load is defined as the load correspond- as follows:
ing to settlement of 10% of the pile diameter as
detected for single pile performance. Empirical correlations:
Correlations were developed based on small
scale laboratory and in-situ tests (e.g. Skempton
1953, Meyerhof 1959 and Vesi´c 1969, 1977).
All these correlations should be used with
caution and only in cases where the conditions
are similar to those of the tests where these
correlations are developed.
Equivalent raft
An equivalent raft is commonly used to trans-
form the pile group to a simpler raft form,
which can be solved by the currently available
methods. The equivalent raft is still used to
Fig. 7 Pile group action regarding the applied load level
predict the settlement of a pile group due to its
simplicity. Terzaghi (1943) has suggested
The relation between pile group action and imposing the foundation load at a level corre-
load level depends mainly on the soil type. In sponding to the lower third point of the length
case of cohesive soil, the effect of nonlinear pile of the piles (in the case of mainly friction piles)
response is negligible at relatively low load or at the pile base in case of end bearing piles.
levels (the elastic range) till complete slip takes The additional stresses in the soil are then
place along the pile shaft. At higher load levels, determined by the theory of elasticity (e.g. using
the nonlinear pile response affects the pile/pile the equation of Steinbrenner 1934) or by the 2:1
interaction considerably. Therefore, the pile method for an equivalent raft at this level. This
group action decreases as the load approaches approach is used in many standards (e.g. DIN
failure. On the other hand, in case of cohesion- 1054) with some differences in the suggested
less soils, the pile group action depends on the procedure. However, the equivalent raft method
stress level and decreases gradually with in- does not take into consideration the number of
creasing the load level up to full mobilization of piles which may have a significant effect.
the skin friction. After that, the group action
decreases considerable as the load approaches
74 Y. El-Mossallamy / Pile Group Action Under Vertical Compression Loads
Superposition technique based on theory of calculate the vertical deformation of a rigid pile
elasticity group using the superposition technique. In a
Randolph and Wroth (1978, 1979) and Ran- more convenient way, Fleming et al. 1985
dolph (1983) suggested an approximate analyti- suggested to relate the settlement of a group of
cal model based on theory of elasticity to piles to the settlement of the corresponding
predict the vertical displacement of a single pile single pile by pile group action defined as a
and a pile group. The vertical displacement of a settlement ratio Rs.as follows:
rigid pile under axial load is given by:
1
Rs Np Z (3)
P ª 4 2 S Lº
s « + » in which Np is the number of piles in the group
G ro ¬ K (1 - Q) ] ro ¼ (2) and ¡ is an exponent depending on pile slen-
derness ratio, pile/soil relative stiffness, group
and
] lnr r m o dimensions and soil properties. ¡ ranges for
most practical conditions between 0.4 and 0.6.
where:
P: The applied load, 3. Case history: the pile foundation of a
r0: Pile radius, highway bridge (Allzey, Germany)
L: Pile length
rm: Influence radius at which the shear In the last two decades, numerical analyses have
stresses become negligible. Randolph has been remarkably developed in both theory and
suggested rm = 2.5 L (1–v) based on a paramet- applications. The more enhanced numerical
ric study using an axisymmetric finite element analyses give the designer the possibility to
analyses, determine the pile group action more accurately
G, v : Soil shear modulus and Poisson’s ratio than the above mentioned simplified methods.
K: represents the interaction of the upper The finite element, the finite difference as well
layer of soil along the pile shaft with the lower as the boundary element methods are very
layer beneath the pile base. Randolph and powerful tools considering more complex
Wroth (1978) have suggested values of K constitutive laws, loads and boundary condi-
ranging from 0.85 to 1.0. tions. A case history will be back-calculated
The above mentioned equation may be mod- applying different methods to examine their
ified to consider approximately the pile stiffness validity determining the pile group action and
and the soil non-homogeneity (Randolph and the behavior of individual piles of the pile
Wroth 1978). Randolph and Wroth (1979) group.
suggested an approximate analytical solution to
Fig. 8: Layout of the pile load test and the measured points (Sommer and Hambach, 1974)
Y. El-Mossallamy / Pile Group Action Under Vertical Compression Loads 75
Figure 9 shows a comparison between ob- methods at low load level. In case of piles in
servations, the results of the boundary element clay the Randolph’s method gives reasonable
method (BEM) and the results the of finite results under working conditions due to the
element method (FEM). The good agreement justified bilinear behavior of the pile. However,
between measurements and analysis can be Randolph’s method should be applied with
recognized. The pile base load share of the total caution in cases where the pile-load settlement
load ranged between 10% and 15% till the skin behavior shows high nonlinear performance
friction is fully mobilized and then it under working conditions.
reached35% at the end of the pile load test. The ultimate skin friction was determined
Therefore, the pile under working loads can be from the results of the pile load test to be about
considered mainly as a friction pile. The calcu- 60 kPa. An allowable skin friction of 30 kPa is
lated settlement according to Randolph 1983 chosen for the foundation design. The measured
(equation 2) is added for comparison. It can be settlement of the tested pile under load corres-
recognized that this simple equation gives good ponding to 30 kPa skin friction was detected to
agreement with other sophisticated be in the order of 3 mm. The bridge piers
76 Y. El-Mossallamy / Pile Group Action Under Vertical Compression Loads
consisted of 2 pillars each founded on a separate average measured settlement of the pier was
pile group. The foundation piles had a diameter about 9.0 mm. The difference between the
of 1.5 m and a length of 24.5 m with 6 piles expected settlement and the measured value
under each pillar. The pile arrangement is demonstrates the importance of considering the
shown in Figure 10a. Based on the pile load test pile group action to predict a reliable settlement
results, the settlement of the whole foundation of the whole foundation. Some of the above
without considering the pile group action was mentioned methods were applied to investigate
expected to be about 3 mm. The load-settlement their reliability in determining the pile group
behavior of the whole foundation was monito- action. The results of the boundary element
red during and after the construction to obtain method as an enhanced three dimensional
information on the group action. The load- analyses were compared with the results of
settlement relationship with time of one of the simple methods (the equivalent raft and the
monitored pillars is shown in Figure 10b. The Randolph’s method) as discussed below.
Fig. 9: Comparison between observed and calculated pile load settlement behavior
Y. El-Mossallamy / Pile Group Action Under Vertical Compression Loads 77
Fig. 10: Foundation layout and load settlement behavior (Sommer and Hambach, 1974)
The load-settlement behavior of a single reasonably the pile group action under working
foundation pile (pile length 24.5 m and pile loads corresponding to factor of safety of 2.0.
diameter 1.5 m) was detected using both the The illustrated comparison shows the good
BEM and Randolph’s method (equation 2) agreement between the different methods to
applying the same soil parameters as described detect the settlement of a pile group considering
and validated by the analysis of the pile load the pile group action. In cases of pile groups
test. The load-settlement relationship up to a with limited number of piles and pile spacing in
working load of about 3 MN is mainly linear the order of two to three pile diameters, the pile
(Fig. 11). The good agreement between the group action can simply be calculated applying
results of the enhanced analysis using the BEM either the equivalent raft or the Randolph’s
and the results of the simple method according method. There is still need for further research
to Randolph can be recognized. Figure 11a to check the ability of these simplified methods
compares the behavior of the single pile with to predict the group behavior of pile groups
the average behavior of the pile group under the with a large number of piles especially in
same average load. The calculated pile group cohesionless soil.
action using the boundary element method can Figure 12 demonstrates the distribution of
be determined to be in the order of about 3.0. pile load and pile head settlement depending on
This value agrees well with the results of the pile position in the pile group based on the
conducted measurements. results of the three dimensional BEM analysis.
The pile group action was also determined The edge piles carry a higher load than the
applying both Randolph’s method (equation 3) middle piles. The settlements of the inner piles
with ¡ value of 0.6. The ¡ value is calculated are larger than that the ones of the outer piles
using the charts of Randolph (in Fleming et al. due to the mutual interaction between the two
1985). The equivalent raft method according to pile groups. Therefore, an enhanced three
the German Standard DIN 1054 was also dimensional analyses may be necessary in the
applied to determine the pile group action as final design stage of sensitive structures to
shown in figure 11. These results demonstrate determine the stiffness of the individual piles
the ability of the simplified methods to predict within the pile group.
78 Y. El-Mossallamy / Pile Group Action Under Vertical Compression Loads
Fig. 11: Load settlement behavior of the pile group and the corresponding single pile (pile group action)
necessary to determine the pile group behavior. of the 5th International Conference on Soil Mechanics
The load-settlement behavior of the piles of a and Foundation Engineering, Vol. 2, pp. 155-159.
Steinbrenner,W. (1934) "Tafeln zur Setzungsberechnung."
pile group is totally different from the behavior Die Straße, Nr. 1, pp. 121-124.
of the corresponding single pile depending on Vesi´c, A.S. (1969) "Experiments with instrumented pile
the geometry of the foundation, the position of groups in sand." Proc. Conf. on Performance of Deep
the pile in the group and the applied load level. Foundations, ASTM, Spec. Tech. Publ. No. 444, pp.
The application of the results of a pile load 177-222.
Vesi´c, A.S. (1977) "Design of pile foundations." Publica-
test to the design of the actual foundation can tion, National Research Council, Washington, D.C.,
only be done by using a suitable and calibrated No. 42.
analytical model. The presented boundary Terzaghi, K. (1943) "Theoretical soil mechanics", 510p.
element as well as finite element methods (Wiley, New York).
proves their validity for such purpose.
5. References
DIN 1054-November 1976 "Zulaessige Belastung des
baugrunds."
El-Mossallamy, Y. und Franke, E. (1997) "Pfahl-Platten-
Gründungen: Theorie und Anwendung." Bautechnik
74, Heft 11
El-Mossallamy, Y (1999) "Load-settlement behavior of
large diameter bored piles in over-consolidated clay".
Proceeding of the 7th. International Symposium on
Numerical Models in Geotechnical Engineering, Graz,
Austria, September 1999
El-Mossallamy, Y. (2002) "Innovative application of piled
raft foundation in stiff and soft subsoil" ASCE Con-
ference on Deep Foundations in February 2002
Fleming, W.G., Welman, A.J. Randolph, M.F. and Elson,
W.K. (1985) "Piling Engineering." Surrey University
Press.
Franke, E., El-Mossallamy, Y. and Wittmann, P. (2000)
"Calculation methods for raft foundation in Germany."
Design applications of raft foundation, edited by
Hemsley, Thomas Telford, 283-322
Kézdi, A. (1957) "Bearing capacity of piles and pile
groups." Proc. 4th. ICSMFE, London, Vol. 2, pp. 46-
51.
Kishida, H. and Meyerhof, G.G. (1965) "Bearing capacity
of pile groups under eccentric loads in sand." Proc.
6th. ICSMFE, Toronto, Vol. 2, pp. 270-274.
Liu, J.L., Yuan, Z.L. and Zhang, K.P. (1985) "Cap-pile-soil
interaction of bored pile group." Proc. 11th. ICSMFE.,
San Francisco, Vol. 3, pp. 1433-1436.
Meyerhof, G.G. (1959) "Compaction of sands and bearing
capacity of piles." ASCE, Vol. 85, No. SM6, pp. 1-29.
Randolph, M.F. and Wroth, C.P. (1978) "Analysis of
deformation of vertically loaded piles." ASCE, Vol.
104, No. GT12, pp. 1485-1488.
Randolph, M.F., and Wroth, C.P. (1979) "An analysis of the
vertical deformation of pile groups." Géotechnique,
Vol. 29, No. 4, 423-439.
Randolph, M.F. (1983) "Design of piled raft foundations."
CUED/D-Soils TR. 143.
Skempton, A.W. (1953) "Discussion contribution: Piles and
pile foundations, settlement of pile foundations." Proc.
3rd. ICSMFE, Zurich, Vol. 3, pp. 172.
Sommer, H. und Hambach, P. (1974) "Großpfahlversuche
im Ton für die Gründung der Talbrücke Alzey." Der
Bauingenieur, Vol. 49, pp. 310-317.
Sowers, G.F., Martin, C.B., Wilson, L.L. and Fausold, M.,
1961. The Bearing Capacity of Friction Pile Groups in
Homogeneous Clay from Model Studies. Proceedings
80 Soil-Structure Interaction, Underground Structures and Retaining Walls
V.M. Ulitsky et al. (Eds.)
IOS Press, 2014
© 2014 The authors and IOS Press. All rights reserved.
doi:10.3233/978-1-61499-464-0-80
Abstract. The paper contains results of long-term settlement monitoring of historic buildings in St. Petersburg. Based on results
of repeated surveying, statistical processing of settlement rates of more than 2300 historic buildings was performed. Basic con-
clusions are drawn as to distribution of maximum settlement values throughout the city area. Also, the paper contains settlement
monitoring results for 30 buildings constructed on soft soils in St. Petersburg (monitoring period extending from 23 to 77 years).
Basic conclusions as to the character of settlement development are offered alongside primary comparisons of monitoring results
with engineering methods of settlement calculations. The analysis thus performed is necessary for development and verification
of engineering calculation methods, as well as of numerical rheological models of subgrade behaviour.
Soft clay deposits are rather widespread in sub- efficiency of numerical solutions which use vari-
soils of buildings and structures, both within the ous models of soil mechanics.
confines of the central part of the city, and in the Big practical interest for development of soil
areas located further away from the centre. Soft media models lies in settlement evaluation of
clay strata in subsoils of buildings and structures both new buildings in the process of construction
extend from several meters up to several tens of and the existing historic buildings, time of reach-
meters (in areas of paleocanals). It is well-known ing a final settlement, degree of settlement pro-
that settlements of buildings constructed on soft gress, as well as settlement development rates
soils develop over many decades and always and distributions of their maximum values
have non-uniform character that leads to defor- throughout the city.
mations of structural elements and development To understand how advanced through its cy-
of tilts. cle a settlement of a historic building is long-
It is necessary to point out that accuracy of term geodetic monitoring results are necessary.
calculations of buildings settlements recently has However, not a single one of historic buildings
acquired an even greater importance in connec- has monitoring results since the beginning of its
tion with a growing demand to more properly construction, and some of those structures are
assess distribution of loads in superstructures by either contemporary with or older than the Kron-
means of soil-structure interaction calculations stadt Gauge. It is obvious that the majority of
because buildings become ever so taller and subsoils of historic buildings should have devel-
more complex. Development of city territories oped an equilibrium (known formally as "cen-
requires construction in separate phases with tenary") creep. To evaluate long-term building
subsequent assessment of their interference. In settlements it is necessary to re-survey geodetic
this connection evaluation of both "final" settle- levels and wall-mounted reference monitoring
ments and their time distributions are necessary. points of the local geodetic network. It is a rela-
A very important characteristic likewise is know- tively widespread method of researching long-
ing how much settlement will have occurred over term building settlements that has been used re-
the period of construction. peatedly to evaluate deformation rates and stabil-
Thus, an even greater importance is given to ity of geodetic stations [4-6].
the issue of evaluating accuracy of both engi- In view of the fact that an analysis requires
neering methods of settlement calculation and as many variously timed geodetic measurements
as possible, plotting graphs of building settle-
V.A. Vasenin / Analysis of Long-Term Settlement of Buildings in St. Petersburg Based on Monitoring Results 81
ments it is necessary to study the benchmark sys- geodetic stations. For plotting settlement graphs
tem of the city in its entirety, and also to evaluate it was necessary to bring all data to a uniform
and designate the most important stages of its level using a special procedure. Having settle-
development. ment graphs for all points it was possible to con-
To accomplish this study the author collect- duct all verification procedures, because conclu-
ed all the catalogues the main city setting-out sions drawn for one line of the matrix, should
campaigns over 220 years, starting from 1779 [3]. hold true also for another (within the limits of the
typical considered historic buildings – houses of
The main surveying works in the city 2-7 storeys constructed on soft soils).
for the period 1779-2003 Based on technical reports on the main sur-
veying projects a network of tie-ins was plotted,
1. Surveying by von Baur – 1779 (719 stations its work on the entire absolute level matrix of
in the catalogue) – preparing a city plan, en- geodetic stations being verified. With the help of
gineering preparation of city areas for the that network numerical transformation of abso-
purposes of flood protection. lute levels of monitoring points according to the
2. Surveying by Captain Savitsky – 1872-1874 main setting-out projects was carried out.
(more than 2000 monitoring points) – pre- The general analysis of long-term building
paring a topographical plan for the city sew- settlements of historic buildings was carried out
er project. in the chronological order of the setting-out
3. Surveying by Heinz – 1892 (57 points) – works (i.e. along the lines of the absolute levels
comparison of zero monitoring points on the matrix of geodetic monitoring points as con-
Neva gauges. tained in the setting-out network). The first set-
4. Surveying by prof. Witram – 1892 – mount- ting-out project, whose points are present in the
ing monitoring points on major railway sta- setting-out network, is the project of captain Sa-
tions, a tie-in with the zero mark of setting- vitsky conducted in 1872-1874. Based on that
out by Savitsky (2 above). verification settlement plots on 66 monitoring
5. Surveying by Griboedov, the Ivanins – 1900, points were drawn.
1910 - drawing up topographical plans of Figures 1–11 contain combined graphs of
Krestovsky, Yelagin and Stone Islands. monitoring points according to captain Savit-
6. Surveying by prof. Sergievsky (nearly 1000 sky’s setting-out project (1872-1874) and moni-
monitoring points) – 1911 – drawing up a toring points of the setting-out conducted in
topographical plan of the city for a sewer 1927-1932, mounted on the same buildings. The
construction project. figures present a rather good match of settlement
7. Surveying by Gureev (145 monitoring points) development trends for two types of points
– 1927 – developing a sewer construction mounted at different times. Such procedure reli-
project for Vasilievsky Island. ably restores settlement development graphs of
8. Surveying by SOOGGU – 1927-1932 (more historic buildings over the last 130-132 years of
than 2500 monitoring points) – a construc- their existence.
tion related setting-out project. Figures 12–14 contain Settlement develop-
9. Surveying by GRII 1953-1956 (4800 moni- ment curves for the remaining buildings. The
toring points). graphs are divided into 3 groups – graphs with
10. Surveying by GRII 1966-1968 (6000 moni- settlement development from 20 to 90 mm,
toring points). graphs with settlement development from 90 to
11. Surveying by GRII 1982-1986 (8000 moni- 200 mm, and graphs with settlement develop-
toring points). ment in excess of 200 mm.
12. Surveying by GRII 2003-2009 (more than Looking at the settlement development
11000 monitoring points). graphs for historic buildings contained in the fig-
ures for the last 130 years, it becomes obvious
The collected catalogues were digitized and that most buildings had developed significant
arranged in chronological order. As a result the long-term settlements. The values of such set-
author obtained an upper triangular matrix of tlements appear commensurable with expected
large dimension comprising absolute levels of final settlement values and in many cases even
82 V.A. Vasenin / Analysis of Long-Term Settlement of Buildings in St. Petersburg Based on Monitoring Results
exceed them. Such significant settlement rates on singled out for which it was possible to draw
historic buildings indicates a long-term devel- conclusions about subsoil deformation rates at
opment of adverse natural and man-made pro- the stage of long-term settlement development,
cesses in the central areas of St. Petersburg. i.e. the creep stage. The buildings in that group
Similar settlement development graphs were had avoided negative influence of any sort that
plotted in the chronological order for the other might have been rendered by man-made factors
lines of the absolute city levels matrix (see Fig. (reconstruction/construction of added storeys,
15) for the periods of the main setting-out cam- underexcavation, influence of adjacent construc-
paigns 1892-2003, 1911-2003) which gave an tion, services installation, transport related vibra-
opportunity to carry out general numerical evalu- tion, etc.). In the entire data pool, for the purpos-
ation of results and to make the basic verification es of my research, settlement development
procedures. curves of historic buildings constructed on soft
According to the basic purposes of the re- subsoils with the minimum long-term settlement
search, from the entire setting-out network, a rates proved the most interesting.
group of "referential" historic buildings was
Time, years
1872 1892 1912 1932 1952 1972 1992 2012
0
20
40
Settlement, mm
60
80
100 Monitoring
point 1872 –
The Admiralty M1980
120
Reference point 4135
140
Fig. 1. A settlement development curve (mm) of the Admiralty Building - the north-east wall (former Respublikansky Proezd).
¢#! # '!# " ! £ ,(B$
" # ! £(H5* according to Savitsky.
Time, years
1872 1887 1902 1917 1932 1947 1962 1977 1992 2007
0
20
40
!"#
Monitoring point 1872.
$&' '*
Palace Quay 28 – M1169
Settlement, mm
60 *
Reference point 6953
80
100
120
140
160
180
Fig. 2. A settlement development curve (mm) of a building on Dvortsovaya Naberezhnaya (Palace Quay), 28. Comparison to
# '!# " ! £ /H$B
" # ! £((/H "" '
V.A. Vasenin / Analysis of Long-Term Settlement of Buildings in St. Petersburg Based on Monitoring Results 83
Time, years
1870 1885 1900 1915 1930 1945 1960 1975 1990 2005
0
20
40
Settlement, mm
60
80
Monitoring point 1872.
!"#
$&'*'
Palace Quay 36 M. 1174
100
Reference
*
point 6748
120
140
Fig. 3. A settlement development curve of the State Hermitage Building (Dvortsovaya Naberezhnaya, 36). Comparison to set-
# '!# " ! £ /9,5
" # ! £((/H "" '
Time, years
1872 1887 1902 1917 1932 1947 1962 1977 1992 2007
0
10
20
30
Settlement, mm
40
50
60
+',-
Monitoring point 1872. Ploshchad Isskusstv 1
70 Reference
point 5373
Polynomial (Monitoring point 1872. Ploshchad
80 +$:$;<
Isskusstv 1)
+',->
90
100
Fig. 4. A settlement development curve of Mikhailovsky Theatre Building (Ploshchad Isskusstv, 1). Comparison to settlement
'!# " ! £ $B9B
" # ! £ +(*H "" '
Time, years
1872 1887 1902 1917 1932 1947 1962 1977 1992 2007
0
20
Settlement, mm
40
60
Monitoring point 1872.
?&'@--C
Kutuzov Quay 14
80
Reference
point 4234
100
120
Fig. 5. A settlement development curve of a building located at Naberezhnaya Kutuzova (Kutuzov Quay),14. Comparison to
# '!# " ! £ ,+B,
" # ! £ (($+ "" '
84 V.A. Vasenin / Analysis of Long-Term Settlement of Buildings in St. Petersburg Based on Monitoring Results
Time, years
1872 1892 1912 1932 1952 1972 1992 2012
0
20
Settlement, mm
40
60
120
Fig. 6. A settlement development curve of a building located at Kanal Griboedova (Griboedov Canal), 42/29. Comparison to
# '!# " ! £ ,HB$
" # ! £ (+(, "" '
Time, years
1872 1887 1902 1917 1932 1947 1962 1977 1992 2007
0
20
40
Settlement, mm
60
80
100
120
Monitoring point 1872.
E& $'
$&'
Griboedov Canal 58
*
140 Reference point 6935
160
Fig. 7. A settlement development curve of a building located at Kanal Griboedova (Griboedov Canal), 58. Comparison to set-
# '!# " ! £ /HB$
located near monitoring poin £ (+(H "" '
Time, years
1872 1887 1902 1917 1932 1947 1962 1977 1992 2007
0
10
20
30
Settlement, mm
40
50
60
+',-
Monitoring point 1872. Ploshchad Isskusstv 1
70 Reference
point 5373
80
Polynomial (Monitoring point 1872. Ploshchad
+$:$;<
Isskusstv 1)
+',->
90
100
Fig. 8. A settlement development curve of Mikhailovsky Theatre Building (Ploshchad Isskusstv, 1). Comparison to settlement
'!# " ! £ $B9B
" # ! £ +(*H "" '
V.A. Vasenin / Analysis of Long-Term Settlement of Buildings in St. Petersburg Based on Monitoring Results 85
Time, years
1872 1887 1902 1917 1932 1947 1962 1977 1992 2007
0
10
20
Settlement, mm
30
40
50
80
Fig. 9. A settlement development curve of a building located at Naberezhnaya Fontanki (Fontanka Quay), 26. Comparison
# '!# " ! £ $9*5
" # ! £ (*(5 "" '
Time, years
1872 1887 1902 1917 1932 1947 1962 1977 1992 2007
0
20
40
Settlement, mm
60
80
100
120
Monitoring point 1872.
?&'
F$$
140 Fontanka Quay 92
Reference point 3992
160
Fig. 10. A settlement development curve of a building located at Naberezhnaya Fontanki (Fontanka Quay), 92. Comparison
# '!# " ! £ BHH+
" # ! £ (*B5 "" '
Time, years
1872 1887 1902 1917 1932 1947 1962 1977 1992 2007
0
50
Settlement, mm
100
150
'*G#-'
250
Fig. 11. A settlement development curve of a building located at 6-ya Sovetskaya St., 109. Comparison to settlement develop-
# " ! £ $H+$
" # ! £ /($ "" '
86 V.A. Vasenin / Analysis of Long-Term Settlement of Buildings in St. Petersburg Based on Monitoring Results
Time, years
1872 1892 1912 1932 1952 1972 1992 2012
0
10
20
Settlement, mm
30
40
50
60
70
Moshkov Pereulok 6/4
I'*J Kolokolnaya 18/14
@:$# J Serpuhovskaya Ulitsa 9
G-K#-'
80 Ploshchad Lomonosova 1
+'L$JN'L$ Sadovaya 21
G#<$OK'> Muchnoy Pereulok 2/36
-Q$'J*
Ploshchad Ostrovskogo-Zodchego
Ligovsky Prospekt 60
L'* Nikolaevsky 2/Sadovaya 62
?JG* 'T*JU'
90 Rossi 6/23
Sadovaya 42
G#-' Angliyskaya Naberezhnaya
$#$&' 30 ,:#$#
Italyanskaya 11
Alexandrinsky Theatre-
Liteyny 3/19
L$;J
$''T +#*J
Pestelya 6/27
Ploshchad Ostrovskogo
Torgovy Pereulok 1
V;' ,:#$#J
Italyanskaya 45 L
Ligovsky 37
Sadovaya 21 masonry wall
G#' @Q$#
Kirochnaya 31 @$$'&-'
Konnogvardeysky Boulevard 13
Konnogvardeysky Boulevard 17
@$$'&-'
#$*
Myasnikova 6
Fig. 12 # '!# "' ) " (59+ ¤ " # ' +* ## – 85 mm.
Time, years
W
1872 1892 1912 1932 1952 1972 1992 2012
0
20
40
60
80
Settlement, mm
100
120
140
160
Fig. 13 # '!# "' ) " (59+ ¤ " # ' 5$ ## – 200 mm.
V.A. Vasenin / Analysis of Long-Term Settlement of Buildings in St. Petersburg Based on Monitoring Results 87
Time, years
1872 1892 1912 1932 1952 1972 1992 2012
0
50
100
Settlement, mm
150
250 Shpalernaya 29
Z$#
Bolshaya Podyacheskaya 7
Y'+[#Q#
Nevsky
? 91
300
350
Fig. 14 # '!# "' ) " (59+ ¤ " # ' +** ## – 300 mm.
Benchmarks 1911–2003
Time, years
1911 1924 1937 1950 1963 1976 1989 2002
0
50
100
Settlement, mm
150
200
250
300
9 11 19 27 28 32
34 38 65 75 122 154 173
163 163 173 173 185 197 211
217 219 242 251 252 318 333
341 360 395 429 439 467 487
560 578 612 653 746 752 798
817 826 864 879 1116 1131 1176
1226 1242 1243 158 855
Fig. 15. Settlement development curves of buildings for the period 1911-2003 (entire selection of the setting-out network).
Based on results of my analysis, in each and on timber piles driven into quick clay depos-
considered time interval of the setting-out works its, and having long-term settlement rates in
(conducted from 1872 until 2003) a certain num- the range of 0.15-0.4 mm/year. It is necessary
ber of buildings (both ordinary and special) was to point out that the number of such buildings in
singled out, constructed both on natural subsoil the central districts of the city is rather small,
88 V.A. Vasenin / Analysis of Long-Term Settlement of Buildings in St. Petersburg Based on Monitoring Results
owing to the inherent congestion of central resi- that at the initial stages of settlement develop-
dential areas. In this way, Fig. 16 contains set- ment had low settlement rate values to be subse-
tlement graphs of such buildings for a selection quently accelerated towards gaining additional
of monitoring points and reference benchmarks settlements for the reasons of man-made interfer-
in the periods 1927-1932/2003-2005. On these ence.
graphs clearly distinguishable are such buildings
Selection “1927–1932”–“2003–2005”
Time, years
1927 1937 1947 1957 1967 1977 1987 1997 2007
0
10
20
30
Settlement, mm
40
50
60
70
835 689 835 837 856 885 927 1003 1058 1292 1721 1723 1734 1738 1742
1834 2279 3007 3243 3370 3775 3861 3904 3932 3956 3983 4014 4098 4258 4310
80 4320 4350 4402 4442 4475 4492 4492 4530 4539 4547 4549 4549 4573 4613 4667
4673 4673 4718 4718 4731 4736 4760 4760 4762 4771 4789 4812 4839 4844 4860
4880 4887 4897 4923 4950 4959 4959 4967 5000 5002 5026 5085 5085 5110 5132
90
5167 5182 5202 5204 5210 5221 5226 5227 5244 5262 5272 5278 5278 5302 5312
5315 5325 5329 5335 5354 5383 5394 5401 5415 5425 5432 5442 5450 5456 5472
5492 5501 5508 5524 5569 5569 5585 5622 5625 5653 5666 5683 5718 5719 5730
5745 5746 5752 5762 5775 5788 5794 5809 5819 5821 5822 5823 5826 5847 5851
5852 5853 5867 5917 5920 5939 5943 5944 5970 5989 5996 6020 6026 6033 6046
6095 6100 6103 6109 6109 6139 6177 6217 6264 6296 6300 6315 6368 6373 6388
6420 6438 6453 6473 6477 6482 6483 6540 6590 6625 6778 6803 6832 6832 6853
6859 6956 6973 6995 6996 7000 7017 7023 7169 7234 7287 7377 7463 7487 7488
6771 6510 6273 6186 5918 4577 4451 4278 3886 3281 840 528
Fig. 16. Settlement curves of the "referential" buildings for a selection of monitoring and reference points 1927-1932/2003-2005.
Traditionally rates of creep related settle- 0.6-1.3 mm/year. This value of average long-
ment of historic buildings’ subsoils were thought term settlement rates appears to be 2–5 times
to be in the order of 2–3 mm per year. Such val- greater than the average long-term settlement
ues of settlement rates appear on the average rates of referential subsoils for the considered
much higher than the obtained range of values class of buildings and soil conditions. In the giv-
and are a consequence of erroneous calculations en case the time for the final settlement to be
of settlement rates of historic buildings over the complete increases 2–3 times in comparison with
period 1911-1953 owing to overlooking the cor- the buildings having low rates of long-term set-
rection coefficient adopted in 1953 to account for tlement (the "referential" buildings). Even at av-
change of initial levels of the city’s monitoring erage rates of long-term settlement having the
network (47 mm – for the setting-out tie-in be- order of 1 mm/year, typical for central areas of
tween 1947 and 1953). As evidenced by the the city, prolonged development of differential
analysis, the term of reaching final settlement settlement occurs, and so does the corresponding
by the buildings of the referential group with long-term development of structural defects, as
low rates of long-term settlement is in the or- well as general decrease of rigidity in the main
der of 20-25 years. bearing structures.
Analysed rates of long-term settlement pro- For the purposes of the present research it
gress of the greater part of historic buildings in was interesting to assess average settlement val-
the central areas of the city (around 2300 build- ues of historic buildings for the entire period of
ings were investigated) show average long-term monitoring, as well as for its latest cycle. The
settlement rates of subsoil to be in the order of historic centre of St. Petersburg includes, first of
V.A. Vasenin / Analysis of Long-Term Settlement of Buildings in St. Petersburg Based on Monitoring Results 89
all, the central construction and development re- erage settlement rates of historic buildings it
striction zone (commonly referenced in Russian seemed quite reasonable to see the Admiralty
¥¦¥-1) and the cultural heritage restriction and the Central Districts as one. Such territorial
zone. The historic city centre comprises more association makes sense not only geographically,
than 11000 buildings, of which more than half but also from the point of view of uniformity of
are listed, as well as about 10000 engineering buildings, as more than 90% of the total number
infrastructure objects. The main scope of historic of all buildings in this area had been constructed
buildings is concentrated in the Admiralty, Vasi- prior to 1917. Let us analyse average settlement
lievsky Island, the Petrograd Side (Petrogradsky) rates for the districts under consideration. To in-
and the Central Districts of St. Petersburg. crease accuracy of all long-term settlement rate
The biggest part of the territory belonging to evaluation, I used results of all setting-out works
the Admiralty and Central Districts of the city, is carried out in the city until 1982/1987 – 2003, i.e.
located within the area bordered by the Greater two latest cycles. As it was pointed out earlier,
Neva and the Obvodny Canal. This, interestingly research of long-term settlement rates of historic
enough, is practically the same area where the buildings in a city territory is reasonable to con-
afore-mentioned setting-out works by captain duct both for the entire period of monitoring, and
Savitsky were conducted in 1872. For the sake of for its latest cycle.
increased convenience in the analysis of the av-
Table 1. Average settlement rates for the Central and the Admiralty city districts (entire selection).
Table 2. Average settlement rates for the Central and the Admiralty city districts (selection of buildings without significant man-
made influences)
Table 1 contains results of statistical analysis buildings in the selection decreased by 0.15
of settlement rates for all geodetic points in the mm/year and approached the 1 mm/year value.
Central and the Admiralty Districts. The selec- Table 3 contains values of average settle-
tion of geodetic points being rather extensive, the ment rates of the entire selection of geodetic sta-
observed settlement rates are also quite consider- tions on Vasilievsky Island. Significant long-
able, and were in excess of 1.4 mm/year for the term settlement rates reaching 2.2 mm/year are
latest monitoring cycle. Such significant rates are clearly visible. This average settlement values of
caused, first of all, by man-made influence, i.e. the entire selection are caused by the settlement
construction and reconstruction of underground of buildings belonging to the active phase of
facilities of the city metro. Excluding from the construction of the Soviet period during the late
selection an insignificant number of buildings 1960-s – the early 1980s mainly in the newly re-
which had been influenced by underground de- claimed dredged territories of the island. If one
velopments (values of settlement rates of certain excludes the buildings constructed during the
buildings sometimes reached 8-12 mm/year) a Soviet period in those dredged alluvial territories
statistical analysis of the thus truncated selection from the selection, then the average settlement
was performed. Results of the analysis are con- rates of city quarters of Vasilievsky Island will
tained in Table 2. The table shows that with the be in the order of 1.2 mm/year (Table 4). Inter-
exception of 65 buildings exhibiting extreme set- estingly enough, the average settlement rates in
tlement rates, the average settlement rate of the the Admiralty, Central and Vasilievsky Island
90 V.A. Vasenin / Analysis of Long-Term Settlement of Buildings in St. Petersburg Based on Monitoring Results
Table 4. Average settlement rates of buildings located on Vasilievsky Island (selection of buildings outside the quarters con-
structed on reclaimed dredged areas).
Table 5. Average settlement rates for the Petrogradsky city district (entire selection).
The average settlement rates situation is gradsky District. The areas of the Stone,
quite the opposite in Petrogradsky District (the Krestovsky and Yelagin Islands in general are
Petrograd Side). Here with a rather big selection classified as park zones and are exempt from the
of buildings, long-term settlement rates only just otherwise inevitable congestive construction.
exceed the value of 0.5 mm/year. The latest cycle If one excludes from the general selection
of settlement rate monitoring gave the average of the setting-out stations located on buildings of
0.8 mm/year (see Table 5). It is possible to as- Krestovsky, Yelagin and Stone Islands, the long-
sume that such insignificant rates are caused by term settlement rate of Petrogradsky District in-
an essentially smaller degree of construction creases to 0.6 mm/year. Long-term settlement
congestion. Indeed, the density of buildings is rate for the latest cycle of monitoring also in-
significantly lower in the northern parts of Petro- creases to 0.9 mm/year.
Table 6. Average settlement rates for the Petrogradsky city district (selection of buildings outside the quarters constructed on
Krestovsky, Stone and Yelagin Islands).
Considering the results contained in Tables age settlement rates in the central city districts
1-6, for the four city districts, where most histor- exceeded 1,3 mm/year.
ic buildings are located, it is possible to evaluate If one excludes the extreme high and ex-
settlement rates throughout the considered selec- treme low settlement rate values from the general
tion. Results of average settlement rate evalua- selection in each area, as it was done in Tables 2,
tion in the four considered city areas are con- 4 and 6, it is possible to obtain average settle-
tained in Table 7. The average long-term settle- ment rates for the historic centre. Results of av-
ment rate within the considered territory for the erage settlement rates of geodetic stations for
entire monitoring period exceeded 1,2 mm/year. such selection are contained in Table 8. Accord-
For the latest cycle of setting-out works the aver- ing to the results of statistical processing, the set-
V.A. Vasenin / Analysis of Long-Term Settlement of Buildings in St. Petersburg Based on Monitoring Results 91
tlement rate in the historic centre approaches the of setting-out works the average settlement rate
order of 1 mm/year. Whereas for the latest cycle was closer to 1,2 mm/year.
Table 7. Average settlement rates for the Central, the Admiralty, the Petrograd city districts and Vasilievsky Island (entire selec-
tion).
Table 8. Average settlement rates for the Central, the Admiralty, the Petrograd city districts and Vasilievsky Island (truncated
selection – mainly historic city quarters without significant man-made influences).
A general evaluation shows that average set- works in the given selection of buildings refers to
tlement rates of the overwhelming majority of the time interval between 1927/1932–2003. Thus,
historic buildings are within the range of 0,6-1,5 the calculated average rates of long-term settle-
mm/year and in this connection to build maps of ment for the considered buildings are statistically
contemporary local vertical displacement it is representative. According to the analysis, for a
possible to use the average settlement rate value group of buildings located near particular quays,
of 1 mm/year. However, to assess true creep de- average settlement rate development was defined.
formations, average evaluation will be unac- The results will be found in Table 9.
ceptable. Paying attention to distribution of long-term
A preliminary statistical analysis also will be settlements in Table 9, it is possible to notice the
useful to the purposes of mapping absolute set- highest rates along the Robespierre Quay, the
tlements within the considered territories. Moyka River, Kryukov Canal, the Fontanka Riv-
My study clearly demonstrates that signifi- er and the University Quay. Moreover, rates of
cant concentration of increase in rates of long- long-term settlements of historic buildings locat-
term settlement of historic buildings is ob- ed on practically all quays have increased con-
served when the buildings are located near siderably over the last 20 years. The only excep-
quays of rivers and canals (see Figures 17,18). tion to this trend is the quayside of the Obvodny
To evaluate possible changes of settlement rates, Canal where average settlement rates for the last
those for the latest time interval of 1985-1987/ 20 years have increased only very insignificantly.
2003 were calculated separately. In total, results However, this effect can be explained simply by
of monitoring more than 396 buildings located loss of monitoring points and reference bench-
near river and canal quays were used in the anal- marks at the latest interval of monitoring owing
ysis. To calculate settlement rates, results of re- to reconstruction of buildings along the Obvodny
peated setting-out cycles for the various time pe- Canal.
riods were used. Most results of setting-out
Table 9. Average long-term settlement rates of buildings located near quays in the central city districts.
Dvortsovaya Naberezhnaya
(The Palace Quay)
Moyka Quay
The
Greater
Neva
Dvortsovaya
Ploshchad (The
Palace Square)
Fig. 17. Contours of settlement progress rates (mm/year) along Palace and Moyka Quays adjacent to the Palace Square
and the buildings of the State Hermitage.
V.A. Vasenin / Analysis of Long-Term Settlement of Buildings in St. Petersburg Based on Monitoring Results 93
Moyka Quay
Gri-
boedo
v
Canal
Quay
Fig. 18. Contours of settlement progress rates (mm/year) along Griboedov Canal and Moyka Quays in the area limited
by Griboedov Canal Quay, Moyka Quay and the Nevsky Avenue.
The reasons for such distribution of settle- transport units of the city metro, in particular, by
ment rates are manifold. Apparently, the prevail- the inclined tunnels of stations’ vestibules. Set-
ing influence on such distribution is rendered by: tlement development of historic buildings after
long-term horizontal movements of artificially launch of the metro structures (the after-effect
created coastal areas, ground conditions, washing period) continues for 35-40 years and seems to
out of fines arising through natural unloading of be nowhere near stabilizing. The way how these
ground water, as well as construction and subse- rates of long-term settlement development
quent operation of services (concentrated primar- change suggests that the prolonged settlement
ily in the given areas). In cases of complex re- processes in buildings located under the influ-
construction of the overwhelming majority of the ence of the metro structures will continue devel-
buildings located in the historic city centre in oping for decades.
proximity to quays it is expedient to consider un- Importantly, it must be borne in mind that
derpinning options for subsoils of such buildings the presented conclusions refer mostly to the un-
(either directly with root piles or by means of derground works conducted from the middle of
subsoil improvement). the 1950-s until the middle of the 1970-s. In view
Also, significant concentration of long-term of the accumulated experience of underground
settlement rates of historic buildings is noticed in construction contemporary settlement prevention
zones of influence rendered by underground technologies have advanced since then.
94 V.A. Vasenin / Analysis of Long-Term Settlement of Buildings in St. Petersburg Based on Monitoring Results
The long-term settlement rates of historic retying-in of the old monitoring points of the
buildings contained in the analysis (in total, previous settings-out even if the initial setting
more than 2300 buildings were considered) are out stations and points have been irretrievably
necessary for the purposes of setting the parame- lost. It has opened an opportunity to continue
ters as well as for development and verification historical monitoring in order to obtain long-term
of numerical rheological subsoil models in the settlement graphs for time intervals of 77 years
ground conditions of St. Petersburg. (for example, for the monitoring activity con-
The conducted research of the setting-out ducted in the late 1930-s).
network has allowed to develop a technique for
Time, days
Vertical movement, mm
Time, years
Fig. 19. Settlement progress curve for a four-storeyed school building on Vasilievsky Island in St. Petersburg (1936-2012)
For instance, Fig. 19 contains a settlement In total, during the first stage of research, up
curve of a four-storeyed school-building over 76 to 80 buildings and structures were surveyed,
years of monitoring. The building was con- over which monitoring had begun a certain time
structed on a natural subgrade of soft soils with before. Most of the data (about 70%) was ob-
thickness of up to 25 m. Settlement monitoring tained by specialists of Soils and Foundations
commenced in 1936 by a working team of pro- Department of St. Petersburg Transport Universi-
fessor Tsytovich [6]. ty. Express condition surveys were carried out
Likewise, in the first precinct of Vasilievsky during the reconnaissance works, with the pur-
Island monitoring was conducted over eight pan- pose of trying to establish, amongst others, pos-
el buildings constructed over a powerful stratum sibilities of resuming the setting-out activities.
(more than 40 m) of soft clay deposits. Here, set- Based on that reconnaissance it appeared possi-
tlement of a 12-storeyed panel building con- ble to do so for 30 various buildings constructed
structed on a raft foundation reached 113 cm in on soft soil subgrades.
43 years and is currently continuing. Complex renewal of regular monitoring on
According to the curves of differential set- 30 sites made it possible to create a verification
tlement development in a building, represented basis to evaluate accuracy of both engineering,
in Fig. 20, the settlement is continuing rather in- and numerical calculations. Fig. 21 contains a
tensively in time without any trends of fading. map indicating the monitoring sites, from which
The differential settlement value between the ga- it is obvious that monitoring so far has covered a
ble and the central parts of the building for the significant part of the city’s territory (including
period of monitoring has totalled 47 cm. the southern districts), in other words monitoring
V.A. Vasenin / Analysis of Long-Term Settlement of Buildings in St. Petersburg Based on Monitoring Results 95
appears to be significantly representative. Set- tlement rates of historic buildings in the city’s
tlement curves for the buildings constructed on setting-out network has allowed to make the
soft soils were plotted based on the results of the complex analysis of settlement development of
conducted research works. The general period of the buildings constructed in conditions of soft
monitoring has been from 23 to 77 years. This soils of St. Petersburg. The present pare contains
monitoring in conjunction with research of set- only the primary results of this analysis.
Time, days
Vertical movement, mm
Time, years
Fig. 20. Differential settlement progress curves (mm) for a typical building of series BS-12 on a piled foundation,
Nalichnaya Street, 36/2 (Vasilievsky Island, quarter 1, building 12; object 1–4) (1970-2012)
Time, days
Vertical movement, mm
It is possible to point out that buildings were the so-called "final" settlement will not make any
constructed at various times (over a number of sense, except for the purposes of tentative evalu-
architectural periods – from the onset of Stalin’s ations. Indeed, what is the point in calculating
Classicism until the dawn of contemporary pan- "final" settlement if it cannot be measured in situ,
elled housing schemes) and have different num- and there will be no settlement stabilization over
bers of stories (from 4 up to 16 levels). Fig. 22 the actual prescribed term of the building’s life?
contains settlement plots for buildings construct- In these conditions calculation of buildings’ set-
ed on St.-Petersburg’s soft soils. It is clearly vis- tlement in time will be the only worthwhile ac-
ible from the figure that graphs quite faithfully tivity.
follow the logarithmic dependency. According to
the analysis it appears that the optimum approx- y = 2,4475x
imation of the settlement in time will be a de-
Total settlement as
R² = 0,77
measured, mm
Investigation conducted at the first stage of codes limiting ultimate subsoil deformation val-
research helped determine distribution of build- ues appear quite reasonable in light of the availa-
ings’ settlements and their differential in time, as ble long-term monitoring data.
well as provide some answers to practical ques-
tions addressed by designers, amongst which the
following were the most interesting:
- What will be the percentage of settlement
realization for a building constructed on soft
subgrade at the moment of construction com-
pletion?
- How effective are the criteria of limiting fi-
nal settlement and the differential thereof?
Fig. 23 contains a curve representing the
correlation between settlement observed at the
moment of construction completion and the mon-
itored settlement. Statistically, on average some-
thing like 30 % of the overall settlement (ex- Fig. 25. Correlation between settlements calculated as per SP
pected to occur throughout a normative life of a 22.13330.2011 and SNiP 2.02.01-83* codes (with account of
soil buoyancy within the limits of compressible stratum
building constructed on soft soils) is realized at throughout the entire depth).
the moment of construction completion. In terms
of real design it means that sequential construc-
tion by phases will not solve the problem of set-
tlement differential in existing buildings in case
of complex development of a city area. Problems
associated with settlement differential should be
solved through a set of complex measures.
Concerning the issues of restricting absolute
settlement and its differential, the currently ap-
!") "
" " ¨ ©ª
standards, limit absolute settlement to a value of
15-20 cm and relative settlement differential to
the range of 0.015–0.002. Fig. 24 contains the
curve correlating maximum settlement differen- Fig. 26. Correlation between observed settlement values and
tial and the absolute maximum settlement value those computed as per SNiP 2.02.01-83*
as obtained from long-term in situ monitoring. It
is indeed permissible to say that the dependency
is expected and predictable, inasmuch as the big-
ger the settlement, the higher the relative differ-
ential. A settlement of 15-20 cm has companion
differential in the order of 0.0015–0.002. Ac-
cording to monitoring results for buildings hav-
ing similar numbers of storeys, it can be con-
cluded that no observable defects were identified
should the relative differential not exceed 0.0025.
However, based on archive records several panel
buildings had cracks opening in joints between
panels even at the stage of handover (and that
with differential value being around 0.002).
These defects were removed after handover and
Fig. 27. Correlation between observed settlement values and
the joints subsequently displayed no further those computed as per SNiP 2.02.01-83* (with account of
problems in terms of either strength or integrity. soil buoyancy within the limits of compressible stratum
Thus, the restrictions contained in the Russian throughout the entire depth).
98 V.A. Vasenin / Analysis of Long-Term Settlement of Buildings in St. Petersburg Based on Monitoring Results
Abstract: This article presents a developed design method applied in an actual structure (St John´s Church, Tartu, Estonia)
during the process of extensive underpinning. The article focuses on a practical theory of how to implement soil-foundation-
pile interaction using pretested, end-jacked piles and how to design a strip foundation using only one method based on ultimate
limit state (ULS) and serviceability limit state (SLS).The method takes the following into account as geotechnical requirements:
contact pressure, total settlement and angular distortions. As far as structural requirements are concerned, the method considers
admissible plastic rotations, end movements due to displacement angle, as well as control of cracking.
Keywords. strengthening of foundations, plasticity theory of structures, soil-pile interaction, geotechnical limit state design
Figure 2. Total strip foundation and separated minimum With reference to the theories of plasticity and
parts I, II, and III. Forces acting in ULS (Fd) from the elasticity the design may further be simplified by
superstructure and opposing forces (Rd).
applying the lower and upper bound theorems for
the limit load. In doing so, the deformations of
substructure have to be estimated by means of
elastic theory and based either on a loading
specified particularly for this purpose, or on a
loading derived from the ultimate mechanism
including the partial safety factors. The plastic
mechanism of a strip foundation used in the
design calculations is demonstrated in Figure 5.
Figure 17. Geotechnical and structural design procedure in Using the well-known formula of Schleicher
limit states for strip foundation. Double lines around boxes (1926), and expanded by Gordunov-Posanov
represent serviceability state. (Tsytovich 1981), the settlement, SFI is 9.7mm.
6Fd The subgrade reaction (KFA) and spring coeffi-
p 0d L0 B cient (kF) are:
Ssq d 0.70 (3)
q 0d 6R d K FA
qd
(6)
L0 B SF
When Ssq is less or equal 0.70, then a settle- kF K FA A F 49 685 kN/m (7)
ment examination required for the feasibility of
the strip foundation is carried out. As the contact AF is the effective sub-area of the foundation
pressure (q) is used, the design value is (slab). The total spring coefficient (kPT) for the
the design value is three piles is:
qu
qd , (4) k PT 6k P 3 k P and kp is 15 000 kN/m (Figure
J 10),
where qu ! N overall
safety factor representing a combination of safety 6k P 3 15 000 kN/m 45 000 kN/m.
" Q " (/G X (/
taking into consideration that 1/1.6 = 0.625 < Compatibility:
0.7.
SF SP (8)
®
6. Design calculations in uls and sls in line b ¯Q F 6FdP VdP
6.1. Soil-foundation-pile interaction in ULS QF (9)
°SF k
Characteristic loads in SLS (Fk) are multiplied by ° F
a safety factor, =1.4, (1.35 G + 1.5 Q), to get ®
°S 6FdP
design loads in ULS (Fd). The design load (for
°
¯
dP
6k P
K.C. Avellan / Application of Soil-Foundation-Pile Interaction Design Method for Strengthening Foundations 109
uses separate parts of a strip foundation for Brinch Hansen J. 1961. A general formula for bearing
practical calculations. capacity. Copenhagen. Akademiet for de tekniske viden-
For the soil-foundation interaction the spring skaber Geoteknisk institute.Bulletin no. 11:38-46
coefficient of pile must be tested or predicted. DIN 1054 1976. Baugrund – Zulässige Belastnung des
Baugrunds. German code, November 1976. Berlin,
8. Conclusions Beuth Verlag.
EN 1992. Eurocode 2. Design of concrete structures.
The author of this article is convinced that in EN 1997. Eurocode 7. Geotechnical design.
order to leverage the technical advantage found FIP recommendations (1984) Practial design of reinforced
in computerized systems and Eurocodes, it is and prestressed concrete structures based on the CEB-
important that manual design calculations be FIP model code (MC78). London, Thomas Telford Lim-
taken into consideration and implemented ited.
throughout the design process. For example, Frank, R. 2006. Some aspects of soil structure interaction
manual design calculations could be used as according to Eurocode 7 – Geotechnical design. Engen-
complementary tests to computerized design haria Civil. UM 2006/25, pp. 5-16.
models to verify particular functionalities of Horn, A. 1970. Sohlreibung und räumlicher Erdwiderstand
software applications. In addition, they could bei massive Gründungen in nichtbindingen Böden. Stras-
also be used to compare the results with comput- senbau und Strassenverkehrstechnik. Heft 110. Bonn.
erized calculations. Bundesminister für Verkehr. Block 110. Federal Minis-
This consolidation would essentially contrib- try of Transport, Building and Urban Environment.
ute to integral quality of design methods and Neal, B. 1970. The plastic methods of structural analysis.
enhance the application and applicability of Great Britain, Norwich. Barnes & Noble Inc.
Eurocodes. Schultze, E. 1961. Distribution of stress beneath a rigid
foundation. Proc. 5th International Conference on Soil
9. Acknowledgement Mechanics. Paris, pp. 807-813.
Schleicher, F. 1926. Zur Theorie des Baugrundes. Der
The author would like to thank engineer Erika
Bauingenieur. Heft 48: 931-935, Heft 49: 949-952. (The
Belopotocanova M.B.A. for finalizing the text.
theory of the ground. The Civil Engineer. Block 48, pp.
10. References 931-935, Block 49, 949-952.)
Tsytovich, N. 1981. Soil Mechanics. Moscow. English
Avellan, K. 1992. Geotechnical and Structural Ultimate translation Mir Publishers.
Limit State Design of Foundations Resting on Soil (in
Finnish). Licensiate thesis. Helsinki University of Tech-
nology, Faculty of Civil and Environmental Engineering.
Avellan K. 2008. Ultimate Limit State Design of a Strip
Foundation. In Development of Urban Areas and
Getechnical Engineering. St. Petersburg 16-19 June
2008.
Avellan, K. 2011. Limit state design for strengthening
foundations of historic buildings using pretested drilled
spiral piles with special reference to St. John’s Church
in Tartu. Doctoral thesis. University of Oulu, Faculty of
Technology. Acta Universitatis Ouluensis, C Technica
392.
Avellan, K & Maanas, M. Strengthening the foundations of
St. John’s Church in Tartu, Estonia. Proceedings of the
Fifteenth International Conference on Soil Mechanics
and Geotechnical Engineering. Istanbul, 27-31 August
2001,pp. 1687-1690.
Bergdahl, U. & Eriksson, U. 1983. Bestämning av jordegen-
skaper med sondering – en literturstudie. Linköping.
Statens Geotekniska Institut. Rapport 22. In Swedish..
Swedish Geotechnical Institute. Report 22.
Soil-Structure Interaction, Underground Structures and Retaining Walls 111
V.M. Ulitsky et al. (Eds.)
IOS Press, 2014
© 2014 The authors and IOS Press. All rights reserved.
doi:10.3233/978-1-61499-464-0-111
Abstract: The surrounding conditions of tunnel lining may be changed during the investment period, which lead to a redistribu-
tion of forces acting on the tunnel lining, and perhaps generate new forces that are not taken into account during the design. In
this paper, a case study of Alsafkon tunnel is regarded, which is one of the Syrian Railway deep tunnels. The lining of this tunnel
suffers from cracks, leakage, and other defects. The results of numerical analysis are conformed to the current situation of the
tunnel in terms of distributed cracks places in the concrete lining. The study showed also the importance of improving the lens
properties to avoid the total fall down of the tunnel.
600
400
200
1600
1400
1200
1000
18
16 n [5, 12]
Modulus
concrete lining
14
12 of elastici- 23E3 13E3 5E3
10
ty E(Mpa)
8
6 Poisson
0.3 0.35 0.45
4 ratio v
2
0
1qqx 2qqx 3qqx 4qqx 5qqx By re-analyzing, it is noted that the maximum
value of horizontal displacement at point A has
study case àâ! BB "#
!"
Figure 18. The number of plastic elements in the lining of
¨! H+9 #
the tunnel according to case study in figure (19).
10. References
1. Hamza, Mohammed Marwan, soil and rock mechanics,
Directorate of books and publications, publications of
the University of Aleppo, 2002.
2. Detailed report of the experts of the Bulgarian
KOEHNE to RVP , Syria- Aleppo-Lattakia, Amhan
Center for engineering consultation, Aleppo, 2004.
3. P. Yankov, K. Jiponov, G.Georgiev Measures for
reconstructing of a part of tunelNo.8 Aleppo-Lattakia,
Sofia 2003
Figure 19. Horizontal displacement in the lining of the 4. A.P. Ledyaev,Yu.S. Frolov, Abdulfattah Mansur
tunnel, soil after processing the lens Design forUpgradingRailwayTunnelsNo.3 And No.8 In
Note from (figure 20) that no cracks have seen in the lining Aleppo-Lattakia LineIn Syrian Arab Republic,
after injection. St. Petersburg2007
5. Sukkar, Naseem, Analysis for Syrian railways tunnels
and rehabilitation, master, Aleppo, 2006.
6. RocksienceInc, Phase2 User's Guide, Toronto, Ontario,
Canada 2010.
7. Leca, E. Clough, W. Preliminary design for NATM
tunnel support in soil. Journal of Geotechnical Engi-
neering 118 (4), 558–575, 1992.
8. Matsuoka, S. M. Astudy on simulation of tunnel lining
which involves cracks. Proceedings of Japan society
civil engineers No.554/III-37. Tokyo, 1996.
9. Idris J. Al Heib M. Numerical modeling and mechanical
behavior analysis of ancient tunnel masonry structures -
Tunneling and Underground Space Technology, 2006.
10. Moeller, S.C., Vermeer, P.A. On numerical simulation
Figure 20. The non-appearance of cracks in the lining after of tunnel installation. Tunneling and Underground
injection. Space Technology, Volume 23, Issue 4, Pages 461-475,
July 2008.
11. Hoek, E. Practical Rock engineering. Rock Mass
9. Results and recommendations Properties. Rocscience.com on line, pp. 161–203
(Chapter 11), 2000.
1. The results showed the existence of a consen- 12. Mollon, G. Dias, D. and Soubra A. Probabilistic
sus between the real condition of the Alsaf- analyses of tunneling-induced ground movements.
kon tunnel and the model proposed, which ActaGeotechn. 8: 181-199, 2013.
confirms the validity of the proposed model- 13. Bobet, A. Analytical Solutions for Shallow Tunnels in
Saturated Ground. Journal of Engineering Mechanics,
ing method. This method can be applied to 127(12), p 1258 – 1266, (2001).
analyze other tunnels with different condi- 14. Shin, H. Youn, D. Chae, S. and Shin, J. Effective
tions, in order to verify other factors that may Control of Pore Water Pressures on Tunnel Linings us-
cause cracks developing in lining of tunnels ing Pin-Hole Drain Method. Tunnelling and Under-
through the stages of investment. ground Space Technology, 24, pp 555 – 561, (2009).
15. Cheehan, L. An Investigation of the Effect of Erosion
2. Groundwater level plays a negative role in Voids on Existing Tunnels. Department of Civil Engi-
cracks developing in concrete lining; there- neering & Applied Mechanics McGill University, Mon-
fore it is recommended to clean the drainage treal, Quebec, Canada Sept, 2009.
system inside the tunnel periodically to pre- 16. Thomas, K. Günther, M. A numerical study of the
vent this rise of water level. effect of soil and grout material properties and cover
depth in shield tunneling, Institute for Structural Me-
3. Developing of weak lens near the tunnel has chanics, Ruhr University Bochum, Denmark, 2006.
an important role in increasing cracks in lin- 17. David, H. Fundamentals of finite element analysis,
ing. If these lenses remain without treatment, McGraw Co, New York, 2004.
the wall lining will deteriorate and the tunnel 18. David, R. Finite Element Analysis, Department of
will becomes out of investment. Materials Science and Engineering, Massachusetts In-
stitute of Technology, Cambridge, MA 02139, 2001.
4. The properties of the lens can be improved by
injecting mixture of cement, sand and benton-
Soil-Structure Interaction, Underground Structures and Retaining Walls 119
V.M. Ulitsky et al. (Eds.)
IOS Press, 2014
© 2014 The authors and IOS Press. All rights reserved.
doi:10.3233/978-1-61499-464-0-119
Abstract. Three techniques for the study of the soil structure interface properties will be presented. The first technique involves
the development of a 3D surface roughness parameter to evaluate the skin friction. The second technique is used to map contact
stress concentrations at the soil structure interface using sensitive films developed for medical profession. The third technique is
manufacturing artificial crushable sand at the desired properties to study static and dynamic interface behavior of soil and
concrete. All of the techniques presented have a potential to understand the interface behavior in more detail. The techniques
presented are implementable and practical.
The quantification of the surface roughness in ] = number of sections along the profile
terms of relative interface roughness ] as L = mean grain size D50
reported by Uesugi and Kishida, 1986b, is
defined by measuring @ Q'" " Actually, the above equation represents a two
) ! G dimensional approach where the other dimension
along a profile length L equal to the mean grain is neglected. Since the data recorded in the
size D50 and then normalizing it by D50 as shown scanning process are in three dimensions, there is
in Figure 1Q +â algorithmG. The surface a need of a three dimensional approach which
roughness of each concrete block investigated is can take into account the other direction. The
measured and digitized with a commercially addition of y dimension to the equation above,
available compact size 3D scanning and milling turns the profile into a three dimensional surface
device Roland Modela MDX-+* The MDX-+* which is more representative. The illustration of
3D scanner is connected to the computer with an the three dimensional approach and the three
RS-+B+¢ ") X " "!) "n- dimensional surface roughness parameter ]
ning objects at four to 15 mm per second with a ' J ( Q Bâ #G
! **$ ## ) ""
#-
point height-sensing. The scanning pitch was
" *$ ## ' u-
tion data which results in accurate information
about the concrete surface topography. The 3D
scanning software Dr. Picza is used in the scan-
ning process. The scanning area is selected and
" "" )" Q¢(
¢+
¢B ¢,G are
placed on the table of the scanner and are
scanned one by one. The scanning process takes
approximately five hours for each block. Eleva-
tions on the surfaces are recorded in three dimen-
sions which provides accurate information
differently from other approaches where rough-
ness characteristics are calculated from a set of
" ! +â e segments. The
obtained data are recorded as point cloud format
" "#) " Q@
&G # !"
An algorithm is developed to measure the
normalized surface roughness parameter ] of
each concrete block in three dimensions. The
# " Q]G s identical with the one
reported by æ
(H59 ç]. ] can
be defined by measuring @ Q'" "
between the highest peak and t G
along a profile length L equal to the mean grain
size D50 and then normalizing it by D as men-
tioned before. The average of normalized surface
roughness parameter ] ! +
dimensions can be given as;
max ]
¦
]
1
] L
]
where Rn = average of normalized surface Figure 2. Concrete block surfaces with their 3D roughness
roughness along a line, coefficients and corresponding direct shear interface test
Rmax= vertical distance between the highest results.
peak and the lowest trough,
G. Baykal / Three Techniques for the Study of Soil Structure Interface Properties; 3D Roughness Parameter 121
adequate. The high capacity films are composed correct calibration chart for the color intensity-
of one sheet only. The films are available in rolls pressure relationship. The recommended meas-
with protected cover and must be cut to the # #! ' ) +*
desired size. Two sheet films are supplied with B$ í¢ Q -+* (5* í¢ !) ) "a-
two rolls which are cut and fit on top of each G ' #ty
other ' ) B$ 5* è á !"
calibration sheet is provided for the software. A
Table 2. Properties of the Prescale Films scanner cover is provided for the specified
scanner. The calibration sheet must be scanned
Pressure Low for each session. The films must be scanned
High MPa within half hour. Once the scanning process is
Range MPa
completed the stored image can be analyzed
Extreme Low **$ *+
using several tools provided. The scanned image
Ultra Super & +â X ! '
*+ */
Low ! +** ! "
Super Low *$ +$ Q*(+$##G
X $* ## @ $* ## @ +* ## # )"
Low Pressure +$ (**
are used to sandwich the sensitive film under
investigation. Odometer loading setup is used to
X " # *+ (, load the mortar block to approximately 1MPa.
The contact pressure versus color intensity +,$B ª (*Ü( #"" '
calibration c' ! J , for !! #! +,$B ª #! X
two temperature and relative humidity ranges. pressure applied on the sample is 981 kPa. The
+ # " !!"a-
tion. The samples are then scanned and pressure
maps are obtained.
The stress map of the sample studied is pre-
sented in Figure 5. The LLW pressure sensitive
film has a measuring range *$ +$ ê
The green color represents stress values below
the operating range of this film and the yellow
color shows the stress levels above the measura-
ble range for the pressure film used. In certain
cases two or three different range films can be
used, provided that they are analyzed individual-
ly at their specified ranges. The concrete to
Figure 4. The color intensity range and corressponding stress concrete interface shows that the interface
levels for the prescale films Q " " G. surface topography is not smooth causing stress
concentrations.
Due to their sensitivity to pressure the films
must be handled gently. The suitable range of the
pressure measuring film is selected for the
proposed study. Each sheet is cut to the size of
the area to be measured. Two sheets are matched
together and placed between the surfaces to be
measured. The load is applied. There are two
measurement types; momentary and sustained.
For the momentary contact load application the
l # ) !! + "
removed. For the sustained measurement the
duration of contact load application is for two
minutes. The temperature and the relative humid-
ity of the measurement environment is recorded. Figure 5 ¢" #! ) $* ## ) $* ##
This data is needed for the selection of the concrete blocks.
G. Baykal / Three Techniques for the Study of Soil Structure Interface Properties; 3D Roughness Parameter 123
The companion software has several func- contact stress mapping. Pressure sensitive films
tions which help to analyze the stress distribution are good and satisfactory tools to investigate the
at the interface in a detailed manner. The relative stress distribution at the soil structure interface.
magnitudes of the stress levels may easily be Comprehensive calibration techniques must be
detected using the software QJ /G. The real developed for geotechnical applications. Even if
stress values however require very thorough the calibration of the films may not be adequate,
calibration and must be used with caution. The the information about the distribution of the
real stresses may be measured with other tactile stresses, location of high stress concentrations
sensors or pressure transducers but still they can and data about the contact areas are invaluable
only validate stress values on the points they are for researchers. The practicality of the films and
measured. There are calibration tools available to
ease of analysis with the provided software
adjust the load level to the applied load. The
larger the efficient contact area the more reliable makes this method a good candidate for many
are the results. For rough surfaces stress concen- geotechnical applications. The range of operating
trations are easily detected by the software, temperatures and relative humidity allows the
however the magnitude of these high stresses are use of these films in the laboratory and field
mostly out of range of the used film. A higher conditions. The films must be scanned within
capacity film may be used together to determine half hour for further processing. If the films will
the magnitude of these stress concentrations be used in the field, the special scanner must be
when needed. present in the field.
The measurement data display shows the ef-
fective measurement rate which is an important
tool to understand the reliability of the measure- Manufactured sand
ment. A high value is the sign of good contact
and reliable measurement. Generally a rate of Artificial sand may be produced by crushing and
# 5* ! " î' screening of rock fragments or can be produced
uneven topographies and rigid materials this is by pelletization of powder fines. Powder materi-
not an easy task to achieve. Maximum pressure als may be pelletized to form sand, gravel and
values, average pressure values, calculated load aggregates. The geotechnical parameters of the
and contact area are presented in the measure- pellets are production controllable parameters
ment data. All of the recorded data may be during the pelletization process such as round-
exported to Excel for further analysis. ness, sphericity, unit weight, porosity, water
absorption, internal friction angle and crushing
strength. Fly ash, fly ash with lime and fly ash
with cement addition are pelletized into sand
size and their engineering properties are meas-
ured. The addition of lime and cement helps to
obtain any desired crushing strength. A typical
collection of manufactured pellets is presented in
J 9
The manufactured pellets behave like The grain diameter of the powder material
calcareous sands found in the nature. The source influences the magnitude of the surface tension
and shape difference of the natural calcareous force; small grain diameter is necessary to create
sands do not exist in the manufactured pellets enough pulling force to initiate agglomeration.
having nearly perfect sphericity and roundness. Agglomeration can be achieved by drum or disc
The crushing behavior of the manufactured soil pelletizers QJ HG
is studied in detail. For potential applications
like backfill for retaining walls, fill under the
footings, pile installation in existing
manufactured soil embankment, anchor
installation in manufactured fills, the interface
behavior and the influence of crushability on the
interface behavior is also studied. Finally
odometer tests, direct shear tests are conducted.
By controlling the parameters during the pelleti-
zation process sands with desired crushability
also can be produced. This type of manufactured
sand may be used to study calcerous sand behav-
ior at the soil structure interface.
Pelletization process is the agglomeration of
moisturized fines in a rotating drum or disc. The Figure 9. The disc pelletizer.
product at the end of the process is called the
“fresh pellet”. The crushing strength of the fresh The revolution speed of the disc can be
pellet must be enough for hauling and stockpil- " ) * 9* !#
ing purposes. The pelletization technology is of the disc plane to the normal can be adjusted
widely used in powder metallurgy engineering, ) * H* X #
and medicine industry. disc used in this study *,* # "p-
The pelletization theory was developed in ing blades are placed from center to one edge at
(H,* X !#" !& **/ # ' â '
!" " N G disc the grains pulled by surface tension are
!! # !&N G the compacted further. The agglomerated grains hit
# # ##N G to the scraping blades, falling free to the bottom
mechanical process parameters such as the angle section of the disc. This free fall action compacts
of balling drum or disc to the normal and the the agglomerated product more. This repeated
revolution speed. Observations and analysis revolving and free fall action densifies and
performed on these parameters with respect to makes the agglomerated product stronger for
mechanic and kinetic laws formed the theory of handling. To achieve the most suitable pelletiza-
!& !" Q â' +***G tion process; the revolution speed and the angle
The formation of capillary force between two of disc plane to the normal should be set in a
grains is presented in Figure 8. manner to avoid the dominancy of gravitational
or centrifugal forces. For various disc diameters
the effect of operating angle and revolution
speed on centrifugal and gravitational forces are
! J (*.
References
Abstract: The structure of the Villa Méditerranée is in an unstable position, which gives to four of its foundations a permanent
traction load. Each of these foundations is stabilised by 4 to 6 active permanent anchors (149 to 349 Tonnes of load service
each). In order to limit deformations due to elasticity of prestressing cables, a specific level of prestressing has been defined. The
paper presents the design, the achievement and the monitoring of the prestressing anchors, in order to insure the stability of the
structure.
2. Geological context
A 2 to 3m thick layer of relatively permeable 3.3. Procedure for injecting the anchors
gravely sand, was identified in the marls, be-
tween -20 to -25m NGF and is a captive aquifer. The anchors are performed according french
Fills and alluvial formations constitute a free recommendations TA95 (1) procedure for global
surface aquifer. Both aquifers have nearly the injection type of grouting.
same static water level. First, drilling is carried out until the de-
signed length; then the armature and the tube for
grouting with cuffs are put in the hole. A first
grout fills the hole from the bottom to the top of
the whole length. After several hours, the injec-
tion of the fixing part is operated from the cuffs
tube to a low pressure of 1 MPa.
3.4. Distortions
Considering that the stability of the whole
structure is carried out on the anchors, it was
decided to be more conservative than the Euro-
code 7 for service situations load cases: The
specific service limit state being considered
gives up to 10 to 40% more load than usual
characteristic service load.
The design of such foundations needs to es-
timate accuratly the distortion to be expected in
Figure 3. Foundations of the metallic structure. accordance with elasticity of the cables, and with
different load cases, and to tight enough the
cables, in order to limit the deformations to
acceptable values for the long term behavior of
3. General design of foundations the structure.
A1 B1 C1 D1 SDLqsu , (1)
QELS
QELU (kN) 4173 1990 1830 4512 2
(x6) (x4) (x4) (x6)
QELS CAR 2892 988 855 3148
(kN) (x6) (x4) (x4) (x4) Where D= 0,18m is the diameter of the drilling,
QELS SPE 3148 1368 1190 3588 L=6m the grouted length, and qsu= 600 kPa, the
(kN) (x6) (x4) (x4) (x6)
ultimate friction.
Free lenght (m) 11 11 11 11
Grouted lenght 20 9 9 20
(m) The tests were carried out in conformity with
french norma NF P94-153 (2) up to test traction
Te= 2232 kN. This value is three times the
theoretical load service value. Table 2 gives the
4.2. General stability main results of these tests. The ultimate value
qsu= 600 kPa, defined out of these tests, is 20%
To verify the uploading general stability, we less than the lower measured value, in order to
calculated a reduced volume of soil, associated be conservative.
to one anchor, in conformity with TA95 recom-
mendations (see fig. 5). This condition gives the
Table 2. Main results of preliminary tests
minimum length of the free length of the an-
chors. Anchor T'c Tc Tu (kN) Ts qsu
(kN) (kN) (kN) (kPa)
TR5 1688 1519 > 2232 1215 716
4.3. Preliminary tests TR6 1825 1662 > 2232 1329 783
TR7 1925 1733 > 2232 1386 817
At the beginning of works, three preliminary
tests were carried out in order to define properly 4.4. Creeping test
the ultimate friction to be considered in marls for
In order to check the risk of creeping in marls,
global injection type of grouting.
one of the test (TR7) was carried out with a long
In order to test the whole zone of embedding,
time stage of 24 hours under the load of 1360
the anchor tests were vertically drilled in 0.18m
kN. On fig.6, it can be seen the results of the
diameter, had a free length of respectively
test, in term of a values of creeping, calculated
26.5m, 31.5m, and 36.5m, and a fixing length of
according to formula 2, with gives the speed of
six meters.
the displacement of the anchor under the load.
This speed is decreasing with time. Extrapolat-
ing these measures, it can be assumed that
creeping becomes null after around 50 hours.
130 C. Jacquard / Foundations by Prestressing Anchors of the “Villa Méditerranée” in Marseille
Time (minutes)
Figure 7. PAF of the anchor TR7 24 hour test
'l (2)
D
't
ES'l (3)
PAF
Te T 0
Tension (kN)
7. Aknowledgements
8. References
Abstract. Many river dikes and embankments in the coastal region of Tohoku area were damaged due to strong ground motions
of the 2011 Off the Pacific Coast of Tohoku Earthquake and subsequent attack by the tsunami. This paper analyzes the cause of
damage to the Yoshihama river dike located in Ofunato city of Iwate prefecture based on past field investigations and numerical
simulations. Analyses revealed that the dike body had low liquefaction resistance and the volume change after liquefaction was
rather large. Due to main shock and aftershocks there is likelihood of liquefaction in dike body and the re-liquefaction
possibilities are high even under small ground motions. Based on the series of analyses it could be confirmed that even though
the dike body was covered with concrete panels, due to liquefaction induced settlement of the embankment, gaps developed
between the embankment crown and the concrete covering. These gaps ultimately reduced the tsunami resistant character- istics
of the dike, and thus resulting in the collapse of the dike.
Keywords. ground motions, tsunami impact, damaged dikes, numerical modeling, dike protection
areas under the jurisdiction of the Tohoku Re- wall and road embankment, situated perpendicu-
gional Bureau of MILIT, extensive investiga- lar to the dike were also collapsed, however,
tions were carried out to determine the damage there was no sign of damage to the nearby build-
characteristics. Based on those damage patterns ings. In the upstream of the river, several con-
and characteristics as well as analyses of the necting parts of the vertical buttress wall were
dikes due to the earthquake and tsunami, appro- moved away by about 30 m due to damage to the
priate reconstruction measures were suggested crown and protective cover of the embankment
(MILIT, 2011b). Many researchers conducted as shown in the Figure. Close to the damaged
case study analysis of river dikes damaged due to parts of the dike, the crown of dike body (re-
liquefaction, and reported the damage patterns ferred hereafter as embankment) subsided by
and the extent of deformation (Imai et al. 2013, about 0.6 m to 1 m. Near the drainage ditch of
Wakinaka et al. 2013). the sloping side, sand boils were observed (these
The purpose of this research is to analyze the resemble sand with fine contents that may have
performance of a river dike under the compound formed from granite) in several places. No sign
effects of earthquake and tsunami due to the of scouring related damage was observed in
2011 Off the Pacific Coast of Tohoku earthquake. those areas, and also there were no damage to the
The dike of Yoshihama river (referred hereafter drainage ditches.
as Yoshihama dike), located in the southern cen-
tral part of Iwate prefecture, was picked up as the Completely
target of the study. Several field investigations, collapsed sea wall
in-situ testing, and laboratory testing were con-
ducted in the past (Hazarika et al. 2013a, Yama-
naka et al. 2012) for investigating the cause of
damage to the dike. In this research, the perfor-
mance of the Yoshihama dike was analyzed by
numerically simulating the conditions of the dike
before and after the arrival of tsunami by taking
into the account the histories of the earthquakes
(main shock and aftershocks) and tsunami (in-
coming tsunami and backrush).
survey to determine the elevations after the tacked the dike. In addition, stability of the em-
earthquake (2) survey using total station (Laica bankment was evaluated based on the tsunami
Geosystems 㸸 TCR407S 㸧 to determine the impact force experienced by the damaged cross
cross section of the dike (3) Sounding (Portable section due to overtopping tsunami.
Dynamic Cone Penetration (PDCP) test) and (4)
Surface wave exploration method. The locations 3.1. Numerical model
of the surveys are shown in Fig. 2. The surveys
were conducted in three stages The first stage In order to simulate the influence of ground mo-
was after two months (May 2011), the second tions on the damaged embankment, two dimen-
stage was after six months (September 2011) and sional effective stress analyses were performed
the third stage was after almost two years and that can take into the account the post liquefac-
four months (July 2013). tion settlement due to drainage (Oka et al. 1994).
Figure 3 shows the cross section of the Yo- Here, attention was paid to the influence of long
shihama dike along the AA’ line shown in Fig. 2. continuation time of the main quake and several
The dike wall consists of a vertical buttress with aftershocks that followed within a short interval.
2 m wide crown with back slope (inclination of In the analyses, timing of the concrete cover
1:1.2), which were covered with concrete. As of peeling off in the crown and back slope of the
July, 2013, the crown height of the embankment dike, subsidence of the embankment and the de-
was T.P. + 5.10 m. Wide area around the epicen- veloped gap at the crown of the embankment,
ter of the 2011 Off the Pacific Coast of Tohoku which were observed during the field surveys
Earthquake suffered huge tectonic subsidence. were evaluated in the time scale.
According to the standard cross section used in The numerical model is shown in Fig. 4. The
design (1968) of Yoshihama dike (TDL 2011), target structure is the damaged part of the Yoshi-
the design crown height was T.P. + 6.15 m. hama dike shown in Fig. 3. The bottom of analy-
Therefore, it can be inferred that in the concerned sis model was set where measured shear wave
location, the crown subsided by about 1 m as velocity (Vs) in the field survey (Hazarika et al,
compared to the original height at the time of 2013a) was over 350 m/s. In the analyses, the
completion of the embankment. bottom was taken as viscous boundary and the
sides were set as equal displacement boundaries.
The contact and separation between the structure 3.2. Simulation results
and the embankment were carefully considered
during the simulations. The ground water level Figure 6 shows the deformation of the dike due
was taken to be at T.P. + 1.78 m (Based on the to the ground motions. During the main shock
field survey results). and two aftershocks, the embankment subsided
Earthquake records of KiKnet recorded near by about 0.67 m and the back slope was highly
the surveyed areas were adopted as the input deformed resulting in the peeling off of the con-
ground motions. Ground motions observed at crete covers as observed in the in-situ survey.
KiKnet Kamaishi (IWTH23), which has almost However, near the toe of the embankment no
the same Vs as the base of analysis model, was large deformations were observed due to the
used in the analyses. Figure 5 shows the synthe- ground motions.
sized input acceleration record used in the anal-
yses. It is to be noted that the target area experi-
enced main shock at 14:46 hours and two after-
shocks at 15:06 hours and 15:09 hours (JMA
2011).
overflow and the backrush, dikes need to pre- collect data on more similar sites in order to
serve its water retention capacity and retain its develop adequate relationship amongst structural
function to prevent erosion and flow. elements (such as dike materials, dike structures),
In order to adopt adequate seismic resistant topographical elements (such as nearby
and tsunami resistant measures for dikes, and topography, rived shapes) and elements that
proper design standard for the expected tsunami contributes to compound disaster. It is also
in the future, it is necessary that the tsunami run necessary to conduct simulations of undamaged
up does not cause much damage to the landside flexible structures (Hazarika et al. 2013b), which
area beyond the dike (MILIT, 2011c). In princi- may through light on the damage mechanism of
ple, levee works materials used should be of conventional rigid structures.
proper standard, so that they do not undergo liq-
uefaction. Such measure could maintain height
of the design tsunami water level (Nakane et al. ACKNOWLEDGEMENT
2013). Especially, the filled soil used in the foun-
dation needs to be compacted and stabilized well The authors greatly appreciate the contributions
to prevent liquefaction. of the following individuals in the in-situ testing,
It may be difficult to adopt the above men- field survey and data analysis of this research: Dr.
tioned techniques for existing structures. How- Minoru Yamanaka, Kagawa University; Mr.
ever, liquefaction of foundation soils still can be Nozomu Kosaka, JAFEC USA, U.S.A.; Mr. Hi-
prevented by installing preventive piles or sheet roshi Takenaka and Masaaki Fujii of Japan
piles in the riverside, landside as well as crown Foundation Engineering Co., Ltd., Sendai
of the dike (Otsushi et al. 2011). Such structural Branch, Japan.
measures would be able to preserve the water re-
tention capacity of the dike by minimizing the
vertical and horizontal deformation within the REFERENCES
dike (Kochi Prefecture, 2013). Use of flexible
structure behind the back of sea wall is another Hara, T., Okamura, M., Uzuoka, R., Ishihara Y. and
low cost approach which could be adopted to Ueno, K. 2012: Damages to river dikes due to
prevent scouring and erosion due to overtopping tsunami in southcentral coastal area of Iwate
tsunami and its backrush (Hazarika et al, 2013c). Prefecture in 2011 off the Pacific Coast of
Tohoku Earthquake, Japanese Geotechnical
Journal, Japanese Geotechnical Society, 7(1), pp.
5. CONCLUDING REMARKS 2536 (In Japanese).
Hara, T., Kokusho, T. and Hiraoka, R. 2004:
(1) Embankment material in the dike body had Undrained strength of gravelly soils with
low liquefaction resistance similar to those of al- different particle gradations, Proc. of the13th
luvial soils and decomposed granite. World Conference on Earthquake Engineering,
(2) The volume change due to dissipation of the Paper No. 144, pp.19.
excess pore water pressure at the end of the cy- Hazarika, H., Kasama, K., Suetsugu, D., Kataoka, S
clic loading is larger than those of alluvial sand and Yasufuku, N. 2012: Damage to geotechnical
and decomposed granite with 50% relative den- structures in waterfront areas of northern Tohoku
sity (Hara et al. 2004). due to the March 11, 2011 tsunami disaster,
(3) When several ground motions from the main Indian Geotechnical Journal, Invited paper,
shock and the aftershocks act on the embank- Indian Geotechnical Society, 43(2), pp.137152.
ment, in the saturated zone of the foundation Hazarika, H., Hara, T., and Furuichi, H. 2013a:
soils a large change of the excess pore water SoilStructure Interaction during Earthquake and
pressure develops. As a result even under small Tsunami Two Case Studies from the Latest
aftershocks, the foundation soils undergo re-liq- Disaster in Japan , Proc. of the special workshop
uefaction. on challenges of soilstructure interaction and
(4) The subsidence of the embankment due to retaining walls, 18th ICSMGE, Paris, pp.131142.
dissipation of the excess pore water pressure af- Hazarika, H., Hara, T., Ohsumi, T., and Furuichi, H.
ter liquefaction increases due to repeated loads 2013b: Time to Deviate from Wrong Perception
under main shock and aftershocks. – A Lesson Relearned from the 2011 Great East
In this research attention was focused only on Japan Disaster , Theme Lecture, 5th International
a particular site. It is, however, necessary to
H. Hazarika et al. / Effect of Ground Motions and Tsunami Impact Force 139
Geotechnical Symposium (IGS Incheon 2013), JGS Shikoku Branch Annual Conference, pp.
Incheon, Korea, CDROM. 5556, 2013. (In Japanese)
Hazarika, H., Etoh, I., Pradhan, K.H., Yasufuku, N., Oka, F., Yashima, A., Shibata, T., Kato, M., and
Ishikura, R., & Fukumoto, Y. 2013c: Fundamen- Uzuoka, R. 1994: FEMFDM coupled
tal Research on Tsunami Resistant Sea Wall using liquefaction analysis of a porous soil using an
Low Carbon Material, 10th National symposium elastoplastic model, Applied Scientific Research,
on Environmental Geotechnology, Tokyo, pp. 52, pp. 209245.
351354 (In Japanese). Otsushi, K., Koseki, J., Kaneko, M., Tanaka, H.,
Imai, Y., Suzuki, A., and Tobita, T. 2013: Case Study Nagao, H., Experimental Study of Reinforcement
Analysis of river embankment caused by the of Dike by Steel Sheet Pile, Japanese
Great East Japan Earthquake, 48th Annual Geotechnical Journal㸪Vol. 6㸪No. 1㸪pp. 114
Meeting of the Japanese Geotechnical Society, (In Japanese)..
Toyama, pp. 11291130 (In Japanese). Shikoku Regional Development Bureau, Ministry of
Iwate Nippo 2011: Roads to Recovery Verification Land, Infrastructure, Transport and Tourism
and Testimony of the Great East Japan (MILIT), 2013: Pamphlets for Earthquake
Earthquake in Iwate, Part 6, Moving to Highland Resistance and liquefaction countermeasure,
No.4, Ohfunato and Yoshihama, Lessons of our Report of the Kochi Coastal Projects Run by
ancestors defended life, Morning newspaper National Government (In Japanese)..
article, Dec. 1, 2011 (In Japanese). Tsunami Digital Library (TDL) 2011:
Japanese Geotechnical Society (JGS) 2011: Joint http://tdl.civil.tohoku.ac.jp/TSUNAMI/TDL_top
Survey of Tohoku Branch, and Shikoku Branch _e.html,http://tsunamidl.jp/docimg/080/chile80_
Investigation Team: Damages in Southcentral 125.jpg/medium.
Coastal Area of Iwate Prefecture in 2011 off the Ueno, M., Hara, T., Hazarika, H., Ohsumi, T.,
Pacific Coast of Tohoku Earthquake, Yamanaka, M and Furuichi, T. 2012: Shear
Geotechnical Engineering Magazine, 59(6), pp. strength and deformation properties before and
3035, (In Japanese). after liquefaction of the tsunamiaffected river
Japan Meteorological Agency (JMA) 2011: 2011 dike, Proc. of the 67th JSCE Annual Meeting, III,
Great East Japan Earthquake JMA Seismic pp. 3940 (In Japanese).
IntensityMap,http://www.jma.go.jp/jma/en/2011_ Wakinaka, Y., Ishihara, M., and Sasaki, T. 2013: Case
Earthquake/Information_on_2011_Earthquake.ht study analysis of River dike damage caused by
ml liquefaction during the Great East Japan
Kochi Prefecture 2013: Report of The Nankai Earthquake, 48th Annual Meeting of the
earthquake longterm flood Study Measures, Japanese Geotechnical Society, Toyama, pp.
Chapter 5, Water sealing method and Wastewater 17011702 (In Japanese).
treatment method, pp. 5.15.54 (In Japanese). Yamanaka, M., Hara, T., Hazarika, H., Ohsumi, T.,
Ministry of Land, Infrastructure, Transport and Furuichi, H., Ueno, M., Yamazaki, T., Okada, H.
Tourism (MILIT) 2011a: Restoration and 2012: Stability of Earth Structures against
Damage of Rivers and Coastal facilities in the Overflow of Tsunami during the 2011 off the
Great East Japan Earthquake, Report of the River Pacific Coast of Tohoku Earthquake, Japan
section of the Tohoku Regional Bureau (In Association for Earthquake Engineering㸪12(5)㸪
Japanese). pp. 89101 (In Japanese).
Ministry of Land, Infrastructure, Transport and .
Tourism (MILIT) 2011b: Kitakami River Dike
Restoration, Technical Review Group Report,
The Tohoku Regional Bureau (In Japanese).
Ministry of Land, Infrastructure, Transport and
Tourism (MILIT), 2011c: Proposal for the
measures against tsunami in river (main part),
The 1st river tsunami measure investigation
commission, Document No.6 (In Japanese)..
Nakane, H., Shimamoto, E., Tsunekawa, Y., Kuroda,
S., Nakajima, N., Ishikawa,Y., Hara, T. and
Okabayashi, K.: Consideration of river dike
materials damaged by the 2011 Off the Pacific
Coast of Tohoku Earthquake, Proc. of the 2013
140 Soil-Structure Interaction, Underground Structures and Retaining Walls
V.M. Ulitsky et al. (Eds.)
IOS Press, 2014
© 2014 The authors and IOS Press. All rights reserved.
doi:10.3233/978-1-61499-464-0-140
Abstract. Cases of deformation in reinforced soil retaining walls have been reported in Hokkaido, Japan. Based on a survey of
the deformed cases, the deformation was attributed to poor winter construction in which portions of the embankment soil froze
and then subside during the thawing period, which resulted in bulging of the wall.
In this study, to clarify how the temperature acting on the groundaffects the pulling resistance of the strip, pull- out tests
were conducted under various embankment materials and curing conditions.The pulling resistance under the various construction
conditions is discussed.
Keywords. reinforced soil retaining walls, influence of thermal conditions, strip pull-out test
Start of experiment
The soil zone from 100mm above to
100mm below the strip was measured to
The specimens are left for 24 hours (Room determine whether the freezing speed of
temperature conditions : 20 䉝; low-temperature and U=1.0 㹼 2.0mm/h was achieved. It was
freeze-thaw conditions : 2䉝䠅 confirmed that the freezing sp eed was
within the desired range for each test
The soil specimen is placed in the tank to a depth of case(Fig. 5).
150mm (Dc=90%) After confirming that the temperature
of the soil 100mm below the strip was
The strip is placed lower than 0 °C, the tank was filled with
Room-temp condition Freeze-thaw condition additional soil and the density was
100mm of soil is placed over the strip adjusted to attain adegree of compaction
Low-temp condition
(Dc=90%㸧
of Dc=90%.
The temperature inside the container
The soil frozen from the strip upward was set as 25 °C and the 100-mm-thick
100mm and downward 100mm
layer above the strip and the 100-mm-
thick layer blow the strip were
The experiment tank is filled with the soil (Dc=90%)
thawed.After the temperature of the soil at
Room-temp condition Freeze-thaw condition
100mm below from the strip was
The soil is thawed from the strip
Low-temp condition confirmed to be exceeding 0 °C, the same
upward 100mm and downward 100mm
procedure as that done for the room
temperature and low temperature tests was
The lid is placed on the top of the experiment tank
taken before the lid was closed.
Pull-out test
For the freeze-thaw conditions, the pull-out maximum pulling resistance(kN/m2), Tmax:
test was done ± 6 hours after 1, 3and 14 days of maximum pulling force(kN)
thawing.
The overburden stress was applied in two 2.2 Behavior of soil materials in the pull-out
patterns using vertical stress of v=60 test
and140kPa on the strip.The vertical stress was 2.2.1 Examination focusing on the differences
confirmed by the earth pressure gauge installed between the thermal conditions
at the same location as that of the strip.
The displacement velocity of the strip was set
The curves in Figs. 6 a) and b) show the
as v=1mm/min.
relationship between the pul-lout force T and
the length of the strip pulled out for each soil
2. EXPERIMENT RESULTS AND
material.
DISCUSSION
In Fig. 6 b), the curves are for overburden
stress v of 60kPa (Cases 11, 13, and 15) and
2.1 Interpretation on the pull-out test results overburden stress v of 140kPa (Cases 12, 14,
and 16).The curves all have a similar, gently
The pulling force per unit area acting on the convex shape. No clear peaks in pulling force
stripwas defined as the pulling resistance. Tmax are observed.
Equation (1) below expresses the area of the Fig. 6 a) shows the curves for the two values
strip in contact with the soil, which was of overburden v, the group of Cases 1, 3and 5,
obtained by considering the reduction in area and the group of Case 2, 4 and 6.The shapes of
corresponding to the increase in the length of the relationship curves are similar to those in
strip pulled out L.As expressed in Equation (2), Fig. 6 b). Tmax of Case 5 is about 40% to 50%
the maximum pulling resistance max was lower than those of Cases 1 and 3, and that of
determined by dividing A, which was obtained Case 6 is also 40% to 50% lower than those of
by Equation (1), by the maximum pullingforce Cases 2 and 4.
Tmax obtained in the pullout test. The above examination reveals that, for the
volcanic ash soil,Tmax differs from the other
temperature conditions only in the freeze-thaw
§ L ǼL · B
A ¨ 0 ¸u u 2ࠉ (1) condition with overburden stress v=60kPa.
© 1000 ¹ 1000 For the sandy soil, however, it was found that
Tmax the values of Tmax for the freeze-thaw condition
ȫmax ࠉ (2) were much lower than those for the room
A temperature and the low-temperature conditions.
where, A: surface area that is determined from Fig. 7 a) and b) show the curves for v versus
the length of strip pulled out (m2), L0: length of max for each soil material obtained from the
the strip laid (mm), L: length of strip pulled pull-out test.
out (mm) B: width of the strip (mm), max:
144 H. Hashimoto et al. / Influence of Thermal Condition on the Extraction Characteristics
㻤㻜 䃣max
㻹㼍㼤㻚㻌㼟㼔㼑㼍㼞㻌㼟㼠㼞㼑㼚㼓㼠㼔 max (kPa) 㻤㻜
䃣max
㻢㻜 㻢㻜
a) b)
㻠㻜 㻠㻜
㻿㼍㼚㼐㼥 㼟㼛㼕㼘 㼂㼛㼘㼏㼍㼚㼕㼏 㼍㼟㼔㻌㻌㼟㼛㼕㼘
㻾㼛㼛㼙㻌㼠㼑㼙㼜㻚䠄㼟㻓㻩㻠㻠㼗㻺㻛㼙㻞㻘㻌䃥㻩㻞㻜㼻䠅 㻾㼛㼛㼙㻌㼠㼑㼙㼜㻚䠄㼟㻓㻩㻝㻥㼗㻺㻛㼙㻞㻘㻌䃥㻩㻞㻜㼻䠅
㻞㻜 㻞㻜
㻸㼛㼣㻌㼠㼑㼙㼜㻚䠄㼟㻓㻩㻡㻝㼗㻺㻛㼙㻞㻘㻌䃥㻩㻝㻠㼻䠅 㻸㼛㼣㻌㼠㼑㼙㼜㻚䠄㼟㻓㻩㻞㻢㼗㻺㻛㼙㻞㻘㻌䃥㻩㻝㻟㼻䠅
㻲㼞㼑㼑㼦㼑㻙㼠㼔㼍㼣㻌㻝㼐㼍㼥䠄㼟㻓㻩㻠㻟㼗㻺㻛㼙㻞㻘㻌䃥㻩㻡㼻䠅 㻲㼞㼑㼑㼦㼑㻙㼠㼔㼍㼣㻌㻝㼐㼍㼥䠄㼟㻓㻩㻤㼗㻺㻛㼙㻞㻘㻌䃥㻩㻞㻟㼻䠅
㻜 㻜
㻜 㻞㻜 㻠㻜 㻢㻜 㻤㻜 㻝㻜㻜 㻝㻞㻜 㻝㻠㻜 㻝㻢㻜 㻝㻤㻜 㻞㻜㻜 㻜 㻞㻜 㻠㻜 㻢㻜 㻤㻜 㻝㻜㻜 㻝㻞㻜 㻝㻠㻜 㻝㻢㻜 㻝㻤㻜 㻞㻜㻜
㻻㼢㼑㼞㼎㼡㼞㼐㼑㼚㻌㼟㼠㼞㼑㼟㼟㻌㼍㼏㼠㼕㼚㼓㻌㼛㼚㻌㼠㼔㼑㻌㼟㼠㼞㼕㼜㻌㻌v (kPa) 㻻㼢㼑㼞㼎㼡㼞㼐㼑㼚㻌㼟㼠㼞㼑㼟㼟㻌㼍㼏㼠㼕㼚㼓㻌㼛㼚㻌㼠㼔㼑㻌㼟㼠㼞㼕㼜 v (kPa)
Fig.7 Relationship between overburden stress acting on the strip and maximum shear strength
Previous studies showed that the Mohr- Fig. 8 shows the result of the pull-out
Coulomb yield criterion (max = s’+vtan) holds test,i.e.,the relationship between T and the
between v, which is generated between the length of the strip pulled out, for Cases 5 to 10.
strip and the ground,and max. In the above The shapes of the curves for Cases 5, 7 and 9, in
equation, s' is the adherence and is the friction which the overburden v is 60kPa and those for
angle between the strip and the ground. Cases 6, 8 and 10, in which the overburden v is
From Fig.7 a), it is found that the relationship 140kPa are of similar convex shapes.However,
between v and max tends to be similar to those the decrease in Tmax of Cases 5 and 6 (1 day
for the room temperature and low-temperature after thawing), were greater than those for Case
conditions, even though there are differences in 7 (3 days after thawing) and Case 9 (14 days
the values of s' and . after thawing) with overburden of 60kPa and for
However, in the freeze-thaw condition, Case 8 (3 days after thawing) and Case 10 (14
maxis very low, regardless of the magnitude of days after thawing) with overburden of 140kPa.
v. In Fig. 7 b), the curves of v versus max tend Fig. 9 shows the results of the pull-out test
to be similar for the room temperature and the according to thermal conditions expressed as
low-temperature conditions, even though there the relationship between v and max.
are differences in the magnitude of s' and . max for the freeze-thaw condition (1 day after
For the freeze-thaw condition, it is found thawing) is about 30% lower than those for the
that max becomes smaller than those for the room temperature and low-temperature
other two conditions with decreases in v. conditions regardless of the value of v.
However, the curves for the freeze-thaw
condition (3 days and 14 days after thawing)
2.2.2 Examination of sandy soil according to show that max at v=140kPa for the freeze-thaw
temperature conditions condition (14 days after thawing) is only
slightly lower than those for the room
temperature and the low-temperature
H. Hashimoto et al. / Influence of Thermal Condition on the Extraction Characteristics 145
㻝㻜㻜 䃢v
㻢㻜
㻠㻜
㻾㼛㼛㼙㻌㼠㼑㼙㼜㻚䠄㼟㻓㻩㻠㻠㼗㻺㻛㼙㻞㻘㻌䃥㻩㻞㻜㼻䠅 㻿㼍㼚㼐㼥㻌㼟㼛㼕㼘
㻸㼛㼣㻌㼠㼑㼙㼜㻚䠄㼟㻓㻩㻡㻝㼗㻺㻛㼙㻞㻘㻌䃥㻩㻝㻠㼻䠅
㻞㻜 㻲㼞㼑㼑㼦㼑㻙㼠㼔㼍㼣㻌㻝㼐㼍㼥䠄㼟㻓㻩㻠㻟㼗㻺㻛㼙㻞㻘㻌䃥㻩㻡㼻䠅
㻲㼞㼑㼑㼦㼑㻙㼠㼔㼍㼣㻌㻟㼐㼍㼥䠄㼟㻓㻩㻡㻟㼗㻺㻛㼙㻞㻘㻌䃥㻩㻝㻠㼻䠅
㻲㼞㼑㼑㼦㼑㻙㼠㼔㼍㼣㻌㻝㻠㼐㼍㼥䠄㼟㻓㻩㻡㻞㼗㻺㻛㼙㻞㻘㻌䃥㻩㻝㻝㼻䠅
㻜
㻜 㻞㻜 㻠㻜 㻢㻜 㻤㻜 㻝㻜㻜 㻝㻞㻜 㻝㻠㻜 㻝㻢㻜 㻝㻤㻜 㻞㻜㻜
㻻㼢㼑㼞㼎㼡㼞㼐㼑㼚㻌㼟㼠㼞㼑㼟㼟㻌㼍㼏㼠㼕㼚㼓㻌㼛㼚㻌㼠㼔㼑㻌㼟㼠㼞㼕㼜 v (kPa)
Fig 8. Relationship between pulling force Fig 9. Relationship between v and max
and length pulled out(Sandy soil) (Sandy soil)
conditions.Based on the above examination, it found.Table 3 shows the natural water content
was found that max recovers with time after the wn of the samples collected from the three layers
third day after thawing. as in the above. The values of wn for Case 11
In the pull-out test using sandy soil, the and 13 do not differ considerably by the
maximum shear strength max in the freeze-thaw location of sampling; however, for Case 15, wn
condition at 1 day after thawing is much lower for the top layer of the zone is greater than wn
than those in other cases.The decrease in is for the bottom layer.This is because the water
particularly remarkable.The low value of is moved toward the upper part of the soil-filled
attributed to the thawing of pore water between tank where the cold air was applied.
the soil particles, which disturbed the Fig. 11 shows the soil surface where the
arrangement of the particles. strip was placedafter the pullout test for Case 15
It is thought that the interlocking effect was (volcanic ash soil, 1 day after thawing in the
temporarily reduced because of the freeze-thaw condition). A trace of strip
disturbance.It was also found that max recovered placement is seen.There is no water visible on
with time to the value equivalent to those for the
room temperature and low-temperature
conditions.
Volcanic ash soil
Percent passing (%)
2.2.3 Physical characteristics volcanic ash soil
Case 11 Top
Case 11 Middle
Case 11 Bottom
Case 13 Top
Case 13 Middle
Many types of volcanic ash soil in Hokkaido Case 13 Bottom
exhibit clear fracture characteristics under Case 15 Top
Case 15 Middle
compaction, consolidationand shear forces, Case 15 Bottom
because of the influence of freeze-thaw
action.Therefore, use of this type of material in Grain size (mm)
4. REFERENCES
Abstract: Based on long-term research on soil frost heave conducted in laboratory and in situ including both experimental and
expert investigation sites the influence of groundwater level on frost heave-induced deformations has been identified. Nature of
change of moisture during frost penetration has been studied, parameters of changes of water accumulation coefficient for the
main types of soil in Minusinsk cavity have been obtained. Options of anti-heave activities during construction for driven and
end-bearing piles have been proposed.
Keywords. soil frost heave, cryogenic processes, buildings’ deformations, interaction of foundations and freezing soil
tween adjacent foundation cushions all basement up, in summer – down, residents made a mound
area was filled with 300-mm layer of claydite of earth. Nowadays there is thermal insulation of
gravel. In spring the building was completed, in plinth and paving which prevents freezing and
summer after complete soil thawing claydite frost heave under the foundation base. However
gravel was taken from the basement to attic the most efficient method is not thermal insula-
floor. tion of plinth and paving which employs energy
Thermal insulating material filling provided of a house and prevents soil freezing under the
even thawing of soil from the bottom and a fro- foundation base but replacement of a heaving
zen soil layer on the top eliminated the danger of soil by a thermal insulating material which
heave of oversaturated thawed lower soil. Soil would also exclude heat escape under the build-
heave from foundation bases of unconstructed ing area [2].
facilities in Khakasia in the thawing period has
been described [1]. However it is advisory to 2. Research methods
make thermal insulation before frost penetration
to avoid migration of water accumulation. The first laboratory tests for studying frost heave
To conclude discussion on thermal protection compared with field data resulted in necessity to
anti-heave solutions one should mention recon- conduct experiments in test hods to reproduce
struction of individual houses built in 1960–70s. conditions of natural soil cooling in a tempera-
Most of them were constructed on shallow foun- ture regime and atmospheric pressure alternating
dations with foundation depth of 30–40 cm. To in time. It was found out that in case of negative
reduce frost heave strains – a house constantly temperatures fluctuations from –1 to –5qÊ there
moves in relation to a chimney: in winter it goes is more intensive migration and soil frost heave.
Stationary soil pan pavilion Scheme of the soil pan pavilion
Pavilion
Gym
Soil pan
Post Gym
Sandwich panels
panels
Thrust system
The need to study cryogenic processes in by running water in plastic tubes with holes fed
natural freezing environment led to construction from a 0.5 m3 tank which was filled with water
a stationary pavilion with a soil pan in the after each cycle. Soil for the pan was taken from
courtyard of Khakassia Technical Institute – 27, the third field site (4th district of Abakan),
Schetinkina str. Methods of research in the pa- where water content distribution from 21% to
vilion are based on natural freezing of soil in the 74% was recorded (fig. 1). The same pan was
pan when the pavilion in opened at negative used to develop a technology on punching
temperatures and artificial thawing of soil when holes, making foundations and studying their
the pavilion is closed and electrical heaters are behavior during freezing-thawing.
switched on. Groundwater modeling was made
O.Z. Khalimov / Methods of Research on Frost Heave of Soil and Foundation 149
Observations of soil strains in field sites ture after freezing. It was layer by layer sam-
could lead to loss of results due to poor infra- pling from each geological element to define
structure. The most valuable field results re- soil water content before and after freezing.
ferred to results of research on changes of mois-
Fragment 2 to figure 3
Therefore, deformation of a cryogenic tex-
ture with vertical and tilted ice lenses is ex-
Improved soil Layer contact Unimproved plained by interaction of shrinkage and volu-
boundary soil metric deformations of improved and unim-
Figure 3. Cryogenic texture, 2012. proved soils as well as corresponding fractures
150 O.Z. Khalimov / Methods of Research on Frost Heave of Soil and Foundation
of the contact area where pressured water flows In terms of power efficiency one can-
(as there is some release in fractures). Since ice not reduce foundation depth counting on heat
in formed vertical and tilted lenses is not char- release from a building, therefore, a foundation
acterized by long-term strength it is easy to im- should be put at the regulatory freezing depth or
agine that the described weakened contact area a smaller one, however one should rely on
later develops and becomes a zone of active properties of soil after a freezing-thawing pro-
movement of various soils [3]”. Only recently cess. Construction Code 2.02.01.83* admits
one has managed to conduct more detailed in- foundation construction in a seasonally frozen
vestigation on causes of forming such lenses soil but there is a need to comply with require-
and to find explanation of considerable reduc- ments of heave deformation and additional set-
tion of heave strains for piles with anti-heaving tlements occurring after thawing. However vast
coating (fig. 3). The factor of accelerated cool- majority of design and construction companies
ing of improved soil and horizontal migration of in Eastern Siberia, in particular in Krasnoyarsky
water from unimproved soil to the improved krai and Republic of Khakasia do not risk to
one plays the main role. construct foundations at the depth smaller than
the calculated freezing depth. The coefficient of
4. Soil properties change during frost pene- thermal influence is taken 0.7–0.8. The prospec-
tration tive development is associated with coefficient
of thermal influence to be equal to 1 (the calcu-
The issues of foundation engineering lated freezing depth is equal to the regulatory
are the most difficult. It is know that a depth of one).
foundation depends on a depth of frost penetra- In order to save energy it is advisory to
tion. If the groundwater level is 2–3 m and exclude heat release into the ground. At the
freezing depth sometimes exceeds 3 m founda- same time it is not expedient to deepen under a
tion depth is always smaller than the freezing foundation base where frost penetration can
depth, i.e. a foundation base is located in frozen occur. It is difficult to predict how soil proper-
soil. There is a question on regulated freezing ties will change during frost penetration. There
depth calculated with account of thermal impact is a table below which shows results of expert
and freezing depth calculated without account investigations of soils deformed by frost heave
of thermal impact (for non-heated buildings or under buildings and structures.
buildings under construction).
Table 1 Results of expert investigations of soils deformed by frost heave under buildings and structures
The most serious deformations occurred in first winter an outhouse containing a three-floor
the kindergarten, the settlement of Zeleny, Vil- staircase collapsed, a court looking façade was
ner’s house in Minusinsk and State Technical deformed since it served a location where water
College -59 in the city of Abakan which partly flows from adjacent areas went to. A western
collapsed. Analysis of stresses under the base and a central facades were less damaged. Only 2
showed that, for example, under State Technical hair-shape cracks were recorded. It can be ex-
College -59 stresses reached 0.55 P«. Howev- plained by the fact there were many transversal
er, the building was standing unconstructed for walls for the staircase which reduced the load
3 years, only a part without concrete floors in applied to thawing soil. Works on foundation
the basement collapsed (fig. 4). In a part of the underpinning included installation of a monolith
building with concrete floors they rose due to reinforced concrete disc of floor above the
soil bulk thrusting along the slipping surface but basement which prevented successive develop-
saved the basement from the heave (fig. 5). ment of cracks in the building. Cracks contin-
Silty sand was located under Vilner’s house ued to develop only in a partition along the
foundation base, the ground water level com- middle and longitudinal axes where a two-floor
prised 4.5 ». After the building had been set for part of the building had been originally located,
reconstruction and heating had been cut frost then the third floor was added [6].
penetration began from the basement. After the
Concrete floor
No concrete
floor
Figure 4. A fragment of the cross-section of State Technical College -59 building including soil profile.
152 O.Z. Khalimov / Methods of Research on Frost Heave of Soil and Foundation
imental facilities
Sand
Clayey sand Laboratory investigation in a thermal cham-
ber, test hods and the stationary pavilion with
the soil pan as well as in situ investigations in
6 sites with different groundwater levels pro-
vided sufficient data on nature of frost heave
in Minusinsk cavity, allowed to construct ex-
perimental facilities on end-bearing piles and
to ensure reliable anti-heave stabilization of
Location, soils, conditions of frost penetration the constructed structures.
Figure 6. A plot of moisture redistribution (coefficient of
In order to make anti-heave envelope from
water accumulation) at frost penetration in different con- improved soil a technology of a bearing struc-
ditions. ture installation has been updated and author
1) In the freezer with a constant negative "" ' £ ((*(,H5 )
temperature; approved. According to the technology for
2) In a soil hod in the open air; making anti-heave envelope below the
3) In the stationary soil pan during natural groundwater level the following operations
freezing and artificial thawing in the period of have been proposed: to drill a 1-m-diameter
negative temperatures; borehole by a hole driller down to the
4) Under the foundation base in Vilner’s groundwater level; to put a calculated amount
house, Minusinsk; of potash fertilizer on the bottom, to fill the
5) Under the foundation base in 27, Vyat- borehole with gravel-sand mixture and to in-
kina str. in Abakan during freezing from the stall an end-bearing pile. Salt poured before
basement. the groundwater level is needed for a con-
6) In a field site during freezing below the struction period in case of possible lack of
groundwater level; loading of the end-bearing pile and lack of
7) A site of geo-engineering investigations thermal influence.
for a residential house, 2, Katernaya str. in Two end-bearing piles did not show the
Abakan; clayey sand lies at 0.4–0.8 m from calculated failure, the actual failure comprised
the ground surface, silty sand -at the depth of 76 mm, thus, it was necessary to reach the
0.8–1.8 m underlain by gravel, the groundwa- designed benchmark by an idle blow with
ter level is 3 m. lifting a hammer part to the middle of a dis-
O.Z. Khalimov / Methods of Research on Frost Heave of Soil and Foundation 153
Floor beam
Floor slab of 18 m
Wall panel
End-bearing
Rigid joint 2 pile
Liquid-plastic clayey
sand
Doubling pile
Natural gravel
Figure 7. Construction of warehouses for medical
equipment in the settlement of Askiz.
Envelope
of gravel
Groundwater
level
End-bearing
pile
Natural gravel
Liquid-plastic clayey sand Doubling
pile
Figure 8. Driving of doubling piles to the lens of soft soils in axes “”, “5–6”.
The first negative situation in the ware- did not make an anti-heave gap structure un-
house of medical equipment in the settlement der the wall panel (fig. 9).
of Askiz was caused by the fact that builders Making a paving concrete was filled under
failed to implement the design solution: they the base panel instead of putting glass wool.
154 O.Z. Khalimov / Methods of Research on Frost Heave of Soil and Foundation
law is used to determine the translational and computational cost. Accordingly, the gravity
rotational motion of each particle arising from acceleration used in the simulations was scaled
the contact forces, applied forces and body up in such way that the scaled model provided
forces acting upon it, while the contact model is the same gravity-induced stress field as the
used to update the contact forces arising from prototype. Considering that water was not
relative motion at each contact. physically modeled in the simulations, the
The contact model of the conventional DEM gravity acceleration was justified to account for
is only suitable for un-cemented granular the reduction of effective stresses due to sub-
materials, since adhesive forces and moments mergence. Hence, the gravity acceleration was
can not be transmitted at the contacts. In view
set as 12.31 g here.
of this limit, efforts have been made to invoke
cohesion arising from chemical bonds in order 2.3. Bond contact model of MHBS
to extend DEM to cemented geomaterials (e.g.,
Potyondy & Cundall 2004; Jiang et al. 2007; The contact model proposed by Jiang et al.
Obermayr et al. 2013; Jiang et al. 2014). (2014) was used to simulate the mechanical
behavior of soil particles cemented by MHs.
2.2. Problem description This model is only recapped here for complete-
Figure 1 illustrates a hypothetical marine ness, and details have been given by Jiang et al.
ground with a single drilled pile under plane- (2014).
strain conditions. This model is configured to In this model, two adjacent sandy grains are
match the conditions of a site in South China idealized as two disks with radii R1 and R2
Sea, where MHs have been truly detected. The cemented by MH, which forms a symmetric
truncated seabed is 60 m wide and 24 m deep, at bond in between with a finite width (B) and the
a depth of 800 m below the sea level (i.e., pore minimum thickness (t0). The bond can resist
water pressure of 8 MPa). It is assumed that the normal force (Fn), shear force (Fs), and moment
temperature remains constant as 5ºC throughout (M), which can be computed as follows:
the entire calculation domain, provided that, in Fn K n un t0 ; Fs ¦ K 'u
s s ;M K rT (1)
South China Sea, the water temperature is about
5.6°C at water depth of 800 m with the aver- where un, 'us, and T is the overlap, the relative
aged geothermal gradient equal to 37.6 °C/km tangential displacement increment, and the
(Wang et al. 2006). Two situations were consid- relative rotation angle of particles, respectively;
and Kn, Ks and Kr are normal, tangential, and
ered in this study: (1) the clean ground without
rolling stiffnesses, respectively. The parameters
MHs (clean ground for short), and (2) the
associated with contact stiffnesses can be
ground cemented by MHs with the hydrate computed from the bond thickness and Young’s
saturation degree SH equal to 25% (MH ground modulus of pure MHs, which correlates the
for short). Note that SH is defined as the ratio of temperature and pressure.
the volume occupied by MHs to the total Once any component of contact forces ex-
volume of voids. ceeds a certain threshold, the bond will break in
Sea level an irreversible manner and the bond contact
Velocity
model will reduce to a rolling resistance model
800 m T (Jiang et al. 2005) that characterizes the in-
terparticle rolling resistance by introducing a
Pile particle
12 m dimensionless shape parameter Gr. The thresh-
olds (i.e., bond resistances) can be determined
24 m
Marine ground
based on a bond failure criterion which was
derived from a series of micro mechanical tests
Pile (¶=80cm)
(Jiang et al. 2012). This failure criterion de-
pends on the bond width, the tensile strength
60 m (qt,max) and the compressive strength (qc,max) of
Figure 1. Dimensions of the problem in prototype. pure MHs, which correlate the temperature and
pressure. Note that the geometry features of
The DEM model was scaled down to 1/20 in each bond (i.e., the bond width and thickness) is
size of the prototype to largely reduce the
F. Liu and M. Jiang / Oblique Pullout Capacity of a Single Drilled Pile in Sandy Grains Cemented by MH 157
determined from a pre-defined SH for a given 2.4. Contact model for pile particle
packing conditions of soil particles.
Table 1 summarizes the parameters of the The close-up in Figure 1 shows the detail of the
model used for MHBS particles with diameters pile, which is formed by regularly arranging
ranging from 6 mm to 9 mm. MH bonds are 960 identical disks with diameter of 5 mm in the
present between two adjacent particles as long loosest packing. The particle density was
as their separation is less than 5% of the average chosen in such way that the pile was compara-
particle diameter d50. The width of a MH bond ble to a concrete pile. The contact behavior of
is characterized by a shape parameter G defined pile particles was modeled by the parallel bond
as the ratio of the width to the common radius contact model proposed by Potyondy & Cundall
R0=R1R2/(R1+R2), where R1 and R2 are the radii (2004) with parameters summarized in Table 2.
of the two adjacent particles. For simplification, Extremely large values were assigned to the
Gis identical throughout the entire model, and strength parameters of the parallel bonds in
is computed from SH = 25%. The parameters of order to avoid failure in pile itself. The inter-
the bonded particles were computed according particle friction coefficient of pile particles was
to the temperature-pressure conditions (i.e., 5°C set to be 5 to eliminate unwanted particle
and 8 MPa) using the approaches presented by slippage. The parameters associated with
Jiang et al. (2014). The parameters of unbonded particle stiffness and parallel bond stiffness
particles were calibrated by comparing DEM were selected in such way that the pile resem-
simulation of biaxial compression tests with
bled a pile made of C30 concrete with the
experimental observation from triaxial com-
Young’s modulus of 30 GPa.
pression tests on a clean sand sample (Masui et
al. 2005). The stress-strain response of a MHBS Table 1. Parameters of soil particles
sample using the parameters listed in Table 1 Parameter value
was examined from DEM simulation of a series Particle size 6 mm ~9 mm
Initial planar void ratio 0.21
of biaxial compression tests in comparison to Particle density 2.6 g/cm3
triaxial test conducted by Masui et al. (2005). Local damping coefficient 0.7
Figure 2 shows results of one of the simula- Viscous damping coefficient 0.0
tions under confining pressure of 1 MPa. Bonded particles
Different from the clean sample, the MHBS Tensile strength of MH qt,max 4.73 MPa
sample exhibits higer stiffness, and more Compressive strength of MH qc,max 8.57 MPa
Young’s modulus of MH 290 MPa
remarkable strain softening and volumetric Minimum bond thickness t0 5% d50
dilation. Nevertheless, the two stress-strain Shape parameter G (for SH = 25%) 0.568
curves coincide at large axial strains as a result Un-bonded particles
of progressive bond breakage, because the Particle normal stiffness 6×108 N/m
residual strength is mainly controlled by the Particle tangential stiffness 4×108 N/m
void ratio regardless of the presence of MHs. Interparticle friction coefficient 0.75
Shape parameter Gr 0.8
DEM simulations under different confining
pressures (not shown in this paper) infer that the Table 2. Parameters of pile particles
cohesion and the peak internal friction angle of Parameter value
MHBS are approximately 50 kPa and 30°. Particle number 960
Particle diameter 5 mm
3 Effective confining pressure 1 MPa -4 Initial planar void ratio 0.27
Temperature 5qC Particle density 3.0 g/cm3
Pore pressure 8 MPa -3 Interparticle frictional coefficient 5
Deviator stress (MPa)
64,000 particles was generated using the multi- 2002). Softening-like response is also noted
layer with the under-compaction method (Jiang from the pile in MH ground subjected to
et al. 2003). The ground was consolidated under oblique loads at loading angles of 45 and 60°.
the gravity acceleration of 12.31 g until the The shaft resistance rapidly reaches the peak at
force equilibrium was reached. In the case of the vertical displacement less than 3% D regard-
the clean ground, the ground generation was less of the loading angle, and then drops to a
terminated. In the MH ground case, hydrate relatively steady value. The peak shaft re-
bonds were activated where the inter-particle sistance decreases with the decrease of the
distance was less than 5% of the average parti- loading angle, because the increased lateral
cle diameter. Second, soil particles were re- component of forces cause de-bonding of MH
moved and replaced by pile particles where the cementation and therefore reduce the shaft
pile was intended. This pile can be regarded as a resistance at small displacement of the pile. On
non-displacement pile (e.g., drilled shaft), since the contrary, the residual shaft resistance in-
it was virtually installed in the space left by creases with the decrease of the loading angle,
soils previously removed without disturbing the because the increased lateral earth pressure
surrounding soils or changing the stress state at against the pile mobilizes the shaft resistance at
any point of the surrounding soils. Third, large displacement of the pile.
external loads upon the pile head were simulat- 200
o o
90 (MHBS) 90 (Clean)
ed by imposing a uniform velocity to the parti- o
60 (MHBS)
o
45 (MHBS)
cles located at the upper 10-mm segment of the 150
Net uplift load Py(kN)
nent of an oblique force has detrimental effect Several semi-empirical relationships were
on the lateral capacity of the pile. proposed based on model tests to link the
400
o
oblique capacity to the lateral and uplift capaci-
60 (MHBS)
o
ty of piles. The method given by Chattopadhyay
45 (MHBS)
o
0 (MHBS)
and Pise (1986) is extended here to predict the
300 o oblique load capacity:
Lateral load Px (kN)
0 (Clean)
PT ª1 D § T ·º
200 sin 2 T exp « ¨1 ¸ »
PU ¬ 1 D © 90 ¹¼
(2)
ª1 D T º
100 D cos T exp «
¬1 D 90 »¼
0
0 5 10 15 20 25 30 35
where D = PL/PU is the ratio of the lateral
Normalized horizontal displacement dx/D (%) capacity to the uplift capacity. Note that T is the
inclination of the pile head displacement to
Figure 4. Lateral load against horizontal displacement
at different loading angles.
horizontal, while it is defined in the original
formula (Chattopadhyay and Pise 1986) as the
3.2. Oblique capacity inclination angle of the applied load to horizon-
tal. As shown in Figure 6, Eq. (2) fits the data
Figure 5 presents the applied oblique load (PT) well particularly for oblique capacities defined
against the resultant displacement of the pile (d) in terms of service limits. For comparison, the
in order to facilitate a possible comparison method proposed by Meyerhof and Satry (1985)
between different inclined loads. Curves ob- is also superimposed in the same figure, which
tained under different loading angles intersect. provides better prediction for PT,p while worse
Nevertheless, the oblique load increases with for PT,10 and PT,15 than Eq. (2).
the decrease of the loading angle when the 90
PT,p (DEM)
resultant displacement is larger than 5% D. For 5 75 PT,10 (DEM)
the sake of simplicity, the oblique load capacity 60 PT,15 (DEM)
4 Eq. (2)
of the pile is defined in terms of serviceability, 45 Meyerhof & Sastry
i.e., PTp, PT,10, and PT,15, which are oblique 3 (1985)
capacity corresponding to the first peak, and 30
Ratio PT/PU
300 0 1 2 3 4 5
T R Ratio PT/PU
PTp PT10
200 T R Figure 6. Oblique capacity of the pile in MH ground at
d different loading directions
T
100 T R
4. Conclusions
D
This study investigates the response of a drilled
0 pile subjected to oblique forces and in a marine
0 10 20 30 40 50
Normalized displacement d/D (%) ground with MHs by means of DEM analyses.
Figure 5. Oblique load vs. resultant displacement
The major conclusions are as follows:
(1) The presence of MHs significantly en-
Figure 6 presents the oblique load capacities
hances the uplift and lateral capacities of the
of different definitions normalized by the uplift
load capacity of the pile subjected to a pure pile, of which the oblique capacity correlates
uplift load (i.e., T= 90°) defined in the same the loading angle.
manner. The ratio corresponding to the peak (2) The peak shaft resistance of the pile de-
capacity decreases with the decrease of the creases with decreasing loading angle due to de-
loading angle, while the ratios corresponding to bonding, while the residual shaft resistance
the capacities at service limits increase. increases because of increasing lateral earth
pressure mobilized at large displacement.
160 F. Liu and M. Jiang / Oblique Pullout Capacity of a Single Drilled Pile in Sandy Grains Cemented by MH
(3) The uplift component of an oblique force Jiang MJ, Zhu FY, Liu F, Utili S. 2014. A bond contact
has adverse effect on lateral capacity of the pile. model for methane hydrate bearing sediments with in-
ter-particle cementation. Int. J. Numer. Anal. Met. (in
Acknowledgements press).
Jung J, Santamarina JC, Soga K. 2012. Stress-strain
This study was supported National Natural response of hydrate-bearing sands: Numerical study
Science Foundation of China (51025932, using discrete element method simulations. J. Geophys.
51239010), the Fundamental Research Funds Res. – Sol. Ea., Vol. 117. doi: 10.1029/2011JB009040
for the Central Universities, and Guangxi Key Lee J, Kyung D, Hong J, Kim D. 2011. Experimental
Laboratory of Geomechanics and Geotechnical investigation of lateral loaded piles in sand under multi-
Engineering (12-KF-02). layered conditions. Soils and Foundations Vol. 51, pp.
915-927.
Liu F, Jiang MJ, Zhu FY. 2014. Discrete element analysis
References
of uplift and lateral capacity of a single pile in methane
Brugada J, Cheng YP, Soga K, Santamarina JC. 2010. hydrate bearing sediments. Int. J. Numer. Anal. Met.
Discrete element modelling of geomechanical behav- (under review).
iour of methane hydrate soils with pore-filling hydrate Maslin M, Owen M, Betts R, Day S, Jones TD, Ridgwell A.
distribution, Granul. Matter. Vol. 12, No. 5, pp. 517- 2010. Gas hydrates: past and future geohazard? Phil.
525. Trans. R. Soc. A, Vol. 368, pp. 2369-2393.
Chattopadhyay BC, Pise PJ. 1986. Ultimate resistance of Masui A, Haneda H, Ogata Y, Aoki K. 2005. Effects of
vertical piles to oblique pulling loads. The Proceedings methane hydrate formation on shear strength of synthet-
of 1st East Asian Conference on Structural Engineering ic methane hydrate sediments. The 15th International
and Construction, Bangkok, pp. 1632-1641. Offshore and Polar Engineering Conference. Seoul,
Cundall PA, Strack ODL. 1979. A discrete numerical model Korea; pp. 364-369.
for granular assemblies, Géotechnique. Vol. 29. No. 1, Meyerhof GG, Sastry VVRN. 1985. Bearing capacity of
pp. 47-65. rigid piles under eccentric and inclined loads. Can. Ge-
Cundall PA. 1988. Formulation of a three-dimensional otech. J. Vol. 22, pp. 267-276.
distinct element model—part I. A scheme to detect and Miyazaki K, Tenma N, Aoki K, Yamaguchi T. 2012. A
represent contacts in a system composed of many poly- nonlinear elastic model for triaxial compressive proper-
ties of artificial methane-hydrate-bearing sediment
hedral blocks, Int. J. Rock Mech. Min. Sci. Geomech.
samples. Energies, Vol. 5, pp. 4057-4075.
Abstr. Vol. 25, No. 3, pp. 107-116.
Hart R, Cundall PA, Lemos J. 1988. Formulation of a three Obermayr M, Dressler K, Vrettos C, Eberhard P. 2013. A
dimensional distinct element model—part II. Mechani- bonded-particle model for cemented sand, Comput.
Geotech. Vol. 49, pp. 299-313.
cal calculations for motion and interaction of a system
composed of many polyhedral blocks, Int. J. Rock Potyondy DO, Cundall PA. 2004. A bonded-particle model
Mech. Min. Sci. Geomech. Abstr. Vol. 25, No. 3, pp. for rock, Int. J. Rock Mech. Min. Vol. 41, No. 8, pp.
117-125. 1329-1364.
Ho CE, Lim CH, Tan CG. 2002. Characteristics of bored Uchida S, Soga K, Yamamoto K. 2012. Critical state soil
piles installed through jet grout layer. Journal of Per- constitutive model for methane hydrate soil. J. Geophys.
formance of Constructed Facilities. Vol. 16, pp. 160- Res. – Sol. Ea., Vol. 117. doi: 10.1029/2011JB008661
Wang S, Yan W, Song H. 2006. Mapping the thickness of
168.
the gas hydrate stability zone in the South China Sea.
Jiang MJ, Konrad JM, Leroueil S. 2003. An efficient
Terr. Atmos. Ocean Sci. Vol. 17, No. 4, pp. 815-828.
technique for generating homogeneous specimens for
DEM studies. Comput. Geotech. Vol. 30, pp. 579-597.
Jiang MJ, Yu HS, Harris D. 2005. A novel discrete model
for granular material incorporating rolling resistance.
Comput. Geotech. Vol. 32, pp. 340-357.
Jiang MJ, Yu HS, Leroueil S. 2007. A simple and efficient
approach to capturing bonding effect in naturally mi-
crostructured sands by discrete element method, Int. J.
Numer. Meth. Eng. Vol. 69, No. 6, pp. 1158-1193.
Jiang MJ, Sun YG, Li LQ, Zhu HH. 2012. Contact behavior
of idealized granules bonded in two different interparti-
cle distances: An experimental investigation. Mech.
Mater. Vol. 55, pp. 1-15.
Soil-Structure Interaction, Underground Structures and Retaining Walls 161
V.M. Ulitsky et al. (Eds.)
IOS Press, 2014
© 2014 The authors and IOS Press. All rights reserved.
doi:10.3233/978-1-61499-464-0-161
W
'X 'w
'w = H'X
Figure 1.Schematic of the analysis of extensible geosynthet-
ic reinforcement layer during pull-out: (a) soil-reinforcement
system; (b) forces acting on the reinforcement; Figure2.Problem Definition
(c) soil-reinforcement shear stress-displacement curve;
(d) forces on a differential reinforcement segment during The interface shear stress along the geosyn-
pull-out. (Madhav et al. 1998) thetic sheet depends on the relative displacement
between the sheet and the soil interface based on
and ensures stress or strain equilibrium inside the built in model of Itasca (2006) shown in Fig.
the elements. The problem is modeled for 4 (b). The shear stress along the interface
different sizes of grid such as 120 x 120, 90 x
Fs
90, 60 x 60 and 30 x 30. A grid size of 60 x 60 is W kb uc um
found adequate to obtain reliable results not L (1)
affected by the boundary conditions and at an
optimum time required for solving for each set where Fs is the shear force along the reinforce-
of parameters. ment, L is the length of sheet, kb is the shear
The subgrade beneath the reinforcement is stiffness, uc and um are the axial displacement of
represented as an linear elastic material. The sheet and the adjacent soil respectively.
input parameters required are density, deforma- The mobilized shear stress is limited based on
tion modulus and Poisson’s ratio of the sub- the adhesion and the interface friction along the
grade. The density of fill, , deformation mod-
P.V.S.N. Pavan Kumar and M.R. Madhav / Modeling of Interaction of Polymeric Sheet Reinforcement and Backfill 163
soil – sheet interface. The maximum interface should converge to zero when the model reaches
shear stress from Fig. 4 (a) is the equilibrium state. The other equilibrium
max criterion is stress ratio defined as maximum
Fs
W max sbond V c u tan I r u perimeter unbalanced force to the representative internal
L (2) force. In the present work the equilibrium of the
where sbond is the adhesion, Vc is the normal model is established based on the following
stress on the reinforcement, Mr is the interface criteria – number of iterations are restricted to
friction angle. The shear stress - displacement one hundred thousand, the minimum stress ratio
relation in the present continuum model follows of 0.001 or the maximum unbalanced force of
the initial slope of shear stress - displacement zero Newton. In most of the cases equilibrium
relation presented by Madhav et al. (1998) (Fig. state based on minimum stress ratio of 0.001
1 (c)). governs the solution.
max
Compressive Fs
Force
L
ycomp
EArea sfriction
rebuilt to attain equilibrium. The axial displace- and 8 respectively. The tension profiles for
ment of reinforcement, w, and tension developed different shear stiffnesses, kb are close to each
at each node, T are recorded from the output of other and nearly independent of shear stiffness,
the FLAC. kb. The increase of shear stiffness, kb increases
The node B of sheet reinforcement is sub- the shear stress acting on the reinforcement for
jected to same pullout load and the displacement the same amount of displacement compared with
and tension profiles are obtained. Accurate that for lower shear stiffness. The axial pullout
predictions of deformation and tensile force in force, Tmax = 24 kN/m displaces the end B of
the reinforcement are obtained by adopting sheet reinforcement by 113 mm and free end of
“large strain” mode of FLAC program. reinforcement by 85 mm for shear stiffness, kb =
25 kN/m3 while the node B displaces marginally
3. Results &discussion by 27 mm and the free end of reinforcement by 6
mm for a higher shear stiffness, kb = 200 kN/m3
The parametric study is carried for the following (Fig. 8). Hence the displacement at node B and
ranges of parameters: Depth of reinforcement, D the slip of reinforcement decrease drastically
= 1 – 5 m, length of reinforcement, L = 3 – 7 m, with increase of shear stiffness, kb from 25
shear stiffness of interface, kb =25 – 200 kN/m3, kN/m3 to 200 kN/m3 and the elongation of
axial stiffness of reinforcement, J = 0.1 – 5 reinforcement decreases marginally from 28 mm
MN/m. Tensile yield strength of sheet rein- to 21 mm.
forcement = 200 kN/m. Interface friction angle,
Mr = 30q.
The tension and displacement profiles for
different pullout forces, Tmax at node B are
presented in Figs. 5 and 6 respectively for an
axial stiffness of reinforcement, J = 1 MN/m and
shear stiffness, kb = 100 kN/m3. The application
of axial pullout force or horizontal displacement
at node B develops shear stress at the surface of
sheet reinforcement. This shear stress decreases
towards the free end of reinforcement depending
on the displacement of reinforcement at different
nodes. The sheet reinforcement subjected to a
pullout force, T of order 7 kN/m displaces the
node B of reinforcement by 13 mm and the free
end of reinforcement by 5 mm. Thus reinforce- Figure5.Tension profiles for different pullout loads
ment slips by 5 mm and elongates by 8 mm. The
increase of pullout force, T by about three times
to 22 kN/m results in a linear increase of dis- 50 M adhav et al. (1998)
placement of node B by 40 mm and free end J = 1 M N/m
displacement by 16 mm. These results compare Tmax = 23 kN/m
kb = 100 kN/m3
closely with the results of mechanical model Mr = 30°
presented by Madhav et al. (1998). The mechan- 15 kN/m
w (mm)
L = 5.0 m
ical model based on discrete approach predicts 25 D = 1.0 m 7 kN/m
an increase of displacement at node B and a
subsequent decrease of displacement at free end
of reinforcement for the same order of pullout
force at node B in comparison with continuum
model. 0
The node B of sheet reinforcement with an 0.0 2.5 5.0
axial stiffness, J = 1 MN/m is subjected to a Distance along reinforcement, x (m)
pullout force, Tmax = 24 kN/m at node B and the
shear stiffness, kb is varied from 25 kN/m3 to
Figure6.Displacement profiles for different pullout loads at
200 kN/m3. The corresponding tension and
node B
displacement profiles are presented in Figs. 7
P.V.S.N. Pavan Kumar and M.R. Madhav / Modeling of Interaction of Polymeric Sheet Reinforcement and Backfill 165
24
M adhav et al. (1998)
J = 1 M N/m
Mr = 30° kb = 200 kN/m3
L = 5.0 m
T (kN/m)
0
0.0 2.5 5.0
Distance along reinforcement, x (m)
75 L = 5.0 m, D = 1.0 m
50 kN/m3 1 M N/m
T (kN/m)
12 0.25 M N/m
100 kN/m3
3
200 kN/m 0.1 M N/m
0
0.0 2.5 5.0
Distance along reinforcement, x (m) 0
60 1 M N/m 11
J = 1 M N/m, L = 5.0 m
5 M N/m Mr = 30°, L = 3.0 m
L = 7.0 m
kb = 100 kN/m3
D = 1.0 m
T (kN/m)
0 6
0.0 2.5 5.0
Distance along reinforcement, x (m)
The shorter length of reinforcement results in of reinforcement do not have any influence on
slip or rigid body displacement while increase of the response of reinforcement to the pullout
length reinforcement results in elongation of force.
reinforcement for the same pullout load. The
displacement and tension profiles presented by
mechanical model compare closely with the 4. Conclusion
results of the present model for different lengths
of reinforcement, L ranging from 3.0 m to 7.0 m. The response of polymeric sheet reinforcement
26
subjected to axial force at its free end is studied
M adhav et al. (1998) J = 1 M N/m, using FLAC for a range of interface shear
Mr = 30°, D = 1.0 m stiffnesses, axial stiffnesses of reinforcement and
kb = 100 kN/m3 for different lengths of reinforcement.
L = 3.0 m x The results of axial pullout of sheet
w (mm)
5.0 m
loaded end of reinforcement is large
D = 1.0 m while it is negligible at the free end.
28
3.0 m
x For the same axial pullout force, the
shorter length of reinforcement under-
goes rigid body displacement. The
loaded end of longer length of rein-
forcement elongates and the displace-
0
ment at free end is small compared to
0 45 90 that for shorter length of reinforcement.
wmax (mm)
The present study predicts a linear increase
Figure15.Variation of maximum tension with horizontal of displacement with pullout force. The pullout
displacement at end B – Effect of length of reinforcement force to be applied for a given displacement
increases with increase of shear stiffness, axial
The applied pullout force is less than the stiffness of reinforcement and marginally with
maximum pullout resistance of reinforcement length of reinforcement.
and hence the interface friction angle and depth
168 P.V.S.N. Pavan Kumar and M.R. Madhav / Modeling of Interaction of Polymeric Sheet Reinforcement and Backfill
Abstract: The paper introduces an original approach to face seismic hazard on existing buildings by treating with a soft grout a
thin layer of soil at a certain depth. The paper presents a parametric analysis with reference to the case of a typical clayey soil,
considering the ideal cases of an horizontal or vertical layer. The information gained from these two extreme cases are essential
to understand the true behavior of a real soft caisson, whatever its shape. Surprisingly, the seismic isolation capacity of the
vertical layer depends not only on shear stiffness but also on bulk stiffness, and this has to be taken into account in the design of
a seismic barrier composed of both vertical and horizontal layers.
Keywords: seismic isolation, soil grouting, parametric analyses
2. Numerical analyses
The adoption of soft barriers has been already 2.2. Testing program
considered in literature, for instance to reduce the
vibratory effects of trains as well as traffic As previously mentioned, before analyzing more
induced vibrations (eg. Chouw 1992, Kellezi realistic geometrical configurations, simplified
2011). The isolation mechanism which this paper schemes with a single vertical or horizontal layer
will analyse takes advantage of some experiences
D. Lombardi et al. / Soft Grouting for the Seismic Protection of Existing Buildings 171
of grouted material have been studied, assuming peak acceleration (amax) was kept constant and
a thickness of the treated layer of 1 m. equal to 0.5 g.
All the analyses consider a linear elastic be-
haviour for the involved materials, assuming a
given value of stiffness and damping ratio D and
shear deformations.
The mesh used for these analyses has a width
of 120 m in the horizontal direction (x) and a
thickness 60 m in the vertical direction (y).
In the scheme with the vertical layer, the ver-
tical layer was placed along the vertical sym-
metry axis, and the dynamic motion was propa-
gated only on the right side of the soil deposit, to
check the isolating effectiveness of the vertical
layer.
The geometrical configuration adopted was
reported in Fig.4.
Figure 5. Model adopted for the propagation through
horizontal layer.
Soil
Table 1. Main properties of the materials.
D=3% þ D
Material Vs [m/s]
þ+*,*Í# 3 [kN/m3] [kg/m3] [%]
Vs=300m/s Natural
20 2040 300 3
soil
Bedrock 20 2040 1000 0
10, 20,
Grouted
20 1020 50, 100, 0
layer
200,300
amax,g /amax,s
Vs=300m/s
waves. From this point of view, being the prob-
lem mostly related to the transmission of S 1
waves, it is therefore much more interesting to
check the relevance of K on the performance of
the vertical grouted layer. 0.5
0
Table 2. Summary of the numerical analyses. All tests were 0 20 40 60 80 100 120
carried out for the three frequencies f=1, 3, 5 Hz.
x (m)
Grouted Vs,g G K K/G
layer [m/s] [kPa] [kPa]
Figure 6. Vertical barrier: results in terms of the ratio
10 102 221 between peak ground acceleration with and without treatment
50 2550 5525 at the ground surface under 1 Hz Ricker input motion.
100 10200 22100 2.17
200 40800 88400
vertical
300 91800 198900 However, the first experimental tests carried
K1=441 1.08 out in laboratory have indicated that, even with
20 408 K2=884 2.17 very soft grouts, the value Vs=10 m/s is difficult
K3=8733 21.41 to obtain. For this reason, a value of shear wave
K1=441 1.08 velocity equal to 20 m/s will be considered in the
horizontal 20 408 K2=884 2.17 following: such a value, in fact , is still effective
K3=8733 21.41 in mitigating the dynamic effects, and more
realistic.
Therefore, considering a single value of the
shear wave velocity for the vertical treated layer,
2.3. Results of the analyses for the scheme with a the results for different input frequencies are
vertical layer shown in Fig. 7. At a frequency of 1 Hz, the high
A series of analyses with no treatment has been values of acceleration are due to the closeness of
previously performed for the soil, thus obtaining the fundamental frequency of the signal to the
a reference behavior of the soil deposit. The natural frequency of the soil layer without any
signals, which are Ricker wavelets transmitted as treatment (assuming H=60m,
SV waves, have been propagated starting from fn=VS/(4H)=1.25Hz).
the base of the right side of the model. The left Then, keeping constant G, the value of volu-
side of the model is thought as the part to be metric stiffness of the soft layer has been varied,
protected from the dynamic input. in order to evaluate the influence of this latter in
The results were evaluated in terms of maxi- the system response. This variation has been
mum acceleration at the ground surface. assigned imposing values of the Poisson ratio
For the sake of brevity, only the results per- included in a range consistent with the elastic
taining to f=1 Hz, which is the most critical constitutive model (0dQd0.5).
frequency for the natural deposit, are shown. Surprisingly, looking at the results plotted in
Figure 6 demonstrates that the lower is the Vs Fig. 8, it is clear that the volumetric stiffness of
of the grouted layer, the larger is the reduction of the grouted layer has an influence the wave
acceleration on the protected (left) side of the propagation.
mesh. The same results have been obtained for
the other Ricker wavelets.
D. Lombardi et al. / Soft Grouting for the Seismic Protection of Existing Buildings 173
1
It can be noted that, for the volumetric stiff-
1 Hz ness K1 (the lowest one), the results are almost
3 Hz identical to the ones of the K2 scheme, while for
volumetric stiffness K3 (largest value), the
5 Hz
accelerations are close to those of the natural
amax (g)
0.5
soil.
These numerical evidences point out that it is
convenient to have, on the lateral side of a soft
caisson (vertical barriers) a grout as soft as
possible not only in terms of shear stiffness but
0 also in terms of volumetric stiffness.
0 20 40 60 80 100 120 The first possible reason for this evidence
x (m) may be a modal conversion of the waves trans-
mitted through the barrier, with the transfor-
Figure 7. Results in terms of ratio between peak ground mation of energy into compressional waves.
acceleration with and without treatment at the ground surface
by varying the fundamental frequency of the Ricker wavelet However, no modal conversion of the propagated
and assuming a shear wave velocity VS of the thin layer equal waves was observed on the other side of the
to 20m/s . isolating barrier.
Then, the only possible explanation of the
influence of K of a vertical layer on the transmis-
0.5
natural soil
sion of S waves is its effect on the mechanism of
deformation of the barrier itself: since vertically
0.4 K1 propagating S waves induce shear displacements
K2 in the vertical grouted barrier, its volumetric
0.3
stiffness rules the relationship between the
amax (g)
K3
displacements of the two sides of the barrier.
0.2
For an extremely stiff grouted material (high
values of K), it is expected that the two sides
0.1
display remaining parallel, while for low stiff-
ness a distorsion is expected. Fig. 9 sketches the
0
0 20 40 60 80 100 120 expected mechanism on a single volume ele-
ment.
x (m) (a)
2.4. Results of the analyses for the scheme with
natural soil an horizontal layer
0.5
K1 In this section, the effects of the insertion of a
0.4 horizontal layer of grouted material on the wave
K2
propagation are shown.
0.3
amax (g)
amax(g)
0 0.2 0.4 0.6 0.8 1
0
10
High K
2@1
2 2#1 1 20
Vs,2 Vs,1
z (m)
30
40
natural soil (1 Hz)
(a) natural soil (3 Hz)
50
natural soil (5 Hz)
horizontal layer (1 Hz)
60 horizontal layer (3 Hz)
horizontal layer (5 Hz)
amax(g)
0 0.2 0.4 0.6 0.8 1
Protected Grouted Free 0
side barrier propagating
side 10
(b)
20
Figure 9. Sketch of the effects of K of the grouted soil on the
z (m)
of different values of the volumetric stiffness K tigation. Soil Dynamics and Earthquake Engineering, 13,
of the soft grouted layer has been investigated. In 3, 328-344.
particular, two further values of volumetric Kirtas E., Rovithis E., Pitilakis K. (2009). Subsoil Interven-
stiffness have been considered. Surprisingly, K tions Effect on Structural Seismic Response. Part I: Vali-
may have a relevant dynamic effect on the dation of Numerical Simulations. Journal of Earthquake
behaviour of vertically placed soft barriers. A Engineering, Vol.13(2), pp.155-169.
high value of volumetric stiffness makes the Ryan, H. 1994. Ricker, Ormsby, Klauder, Butterworth- A
barrier useless, while a low value of volumetric choice of wavelets. CSEG.
stiffness is able to sharply reduce the lateral
transmission of vertically propagating waves.
By analysing the scheme with an horizontal
layer, it was observed that for each seismic input
considered, the peak acceleration above the
barrier sharply decreases, keeping reduced values
up to ground surface. The volumetric stiffness,
instead, does not influence the results. Clearly, a
high volumetric stiffness may be necessary in the
horizontal layer for static reasons.
In conclusions, the proposed innovative ap-
proach to the seismic protection of existing
buildings seems to be a valid alternative to other
more conventional and invasive solutions, such
as the structural reinforcement and base isola-
tion, and can result suited for the historical
constructions for which integrity has to be
preserved.
Since the effectiveness of grouting depends
on the dynamic coupling of the input motion
with the subsoil and the building response,
further analyses considering 2D more realistic
geometrical configurations (Fig. 1) need to be
carried out to get more conclusive considerations
on the feasibility of such interventions.
4. References
Abstract. The paper presents some results of a joint research of three universities from Rio the Janeiro, Brazil, aiming at the
observation of concrete creep and shrinkage effect on soil structure interaction. In this paper an 18 story building resting on
shallow and deep foundations was submitted to a site instrumentation to monitor settlements and loads of some columns close to
the ground floor during the construction phase. Particular attention is given to the effect of concrete creep and shrinkage in soil
structure interaction effects. Due to the extent of the experimental results, only part of the database is presented.
Keywords. soil-structure interaction, creep and shrinkage of concrete, structural modeling, building settlements
ARGILA ARENOSA
NA
ARGILA ARENOSA
NA
ARGILA ARENOSA
NA
remaining instrumented columns. As long as the
AREIA FINA, MÉDIA E GROSSA
AREIA FINA, MÉDIA E GROSSA
POUCO ARGILOSA
isolated caps with single piles would have
AREIA FINA, MÉDIA E GROSSA
ARGILOSA
ARGILOSA
2.4. Site instrumentation For each phase a structural model has been
prepared. Table 1 summarizes the construction
The site instrumentation had the main objective phases with settlement instrumentation as well as
of monitoring the settlement of some main its time from the beginning of the construction.
columns with time in different construction The first monitoring operation occurred with the
phases and also the loads reaching some columns structure constructed up to the fourth floor. .
during the various construction phases. Figure 5 shows the equal settlement curves
Settlement measurements involve the per- mesured for the 6th (last) instrumented phase.
178 L.d.M.P. da Rosa et al. / Soil Structure Interaction: An Analysis of the Effect of Creep and Shrinkage of Concrete
The greatest settlements occurred for the col- process. In such process the loads that a set of
umns founded in spread footings at the central footings and/or piles transmit to the soil mass are
part of the construction whereas the slower ones decomposed in an equivalent set of point loads
on the columns founded on deep steel piles. with superposed effects calculated at the desired
point (selected columns base). In this way
settlements for the whole columns of the struc-
ture have been estimated from the design loads.
25
20
15
10
5 10 15 20 25 30 35 40
25 25
20 20
15 15
10 10
5 5
0 0
5 10 15 20 25 30 35 40 5 10 15 20 25 30 35 40
Figure 6. Equal settlement curves (predicted values), in Figure 7. Equal settlement curves for the 6th phase, consider-
centimeters. ing a fixed support analysis (in centimeters).
made for the 1, 3 and 6, spring coefficient of the Figure 8. Equal settlement curves for the 6th phase, consider-
construction phase up to the convergence of the ing the first iteration calculation (in centimeters).
spring coefficients. The convergence has been
attained at the 3rd iteration. 25
20
4.2. Analysis Results
15
4.2.1. Measured and estimated settlements
From the numerical models elaborated for each 10
structure iteration analyzed. Figure 9. Equal settlement curves for the 6th phase, consider-
It can be observed that the greatest settle- ing the second iteration calculation (in centimeters).
ments have been estimated with the loads ob-
tained by the analysis with fixed support. The In fact, it can be observed in the 6th phase that
location where the greatest settlement occurs is from the fixed support condition to the first
modified if a soil structure interaction is intro- analysis with spring supports (from figure 7 to
duced in the analysis. If a higher refinement is figure 8) 26 out of 43 columns presented loading
introduced in the interaction analysis, the change differences higher than 5%.
in the settlement pattern is less significant (a In the same 6th phase, if one starts the analy-
large number of iteration steps), up to the con- sis with the first estimation of spring supports to
vergence of the results (figures 8, 9 and 10). the second one (from figure 8 to figure 9) only 3
out of 43 columns presented loading differences
higher than 5%. From the second estimation of
180 L.d.M.P. da Rosa et al. / Soil Structure Interaction: An Analysis of the Effect of Creep and Shrinkage of Concrete
spring supports to the third (from figure 9 to The similarity observed in the settlement val-
figure 10) only 2 out of 43 columns presented ues indicates the adequacy of the compressibility
loading differences higher than 5%. of the sandy soil mass, already observed by
Conde de Freitas et al. (2012) in distinct case
histories. The applied correlation for estimation
25
of the elastic soil moduli from N60, proposed by
20
Conde de Freitas (2012), has been found to be
very promising. The similarity in the settlement
15
behavior pattern indicates the adequacy of the
numerical model to reproduce the soil x structure
10
interaction.
5 25
0 20
5 10 15 20 25 30 35 40
th
Figure 10. Equal settlement curves for the 6 phase,
15
considering the third iteration calculation (in centimeters).
is changed from figure 7 (fixed support) to figure Figure 11. Equal settlement curves for the 6th phase,
8 (spring support, first estimation). However, for considering the measured values (in centimeters).
the remaining calculation, from figure 8 (spring
support, first estimation) to figure 9 (spring
25
support, second estimation) and from figure 9
(spring support, second estimation) to figure 10
20
(spring support, third estimation) no sensible
difference has been observed in loading redistri- 15
bution or settlement equalization.
In the following analysis the equal settlement 10
curves predicted and measured have been com-
pared in order to better infer the compressibility 5
of the soil mass, the capability of the Aoki Lopes
(1975) Method to properly estimate the settle- 0
Reação (kN)
without creep and shrinkage effects;
(3) Structure resting on elastic supports
(like springs) and structural material
with creep effects;
(4) Structure resting on elastic supports
(like springs) and structural material 3 5 Tempo
4 (dias)2 6 1
with shrinkage effects;
(5) Structure resting on elastic supports Figure 13. Normal loading with time for column M5,
(like springs) and structural material considering the 6 enumerated conditions.
with creep and shrinkage effects;
(6) Structure resting on fixed supports and
structural material with creep and
shrinkage effects; Reação (kN)
25 10
20 5
15
5 10 15 20 25 30 35 40
(a)
10
25
20
0
5 10 15 20 25 30 35 40
15
Figure 16. Equal settlement curves predicted for the 6th
phase, with soil structure interaction with due consideration
10
of creep and shrinkage of the concrete.
5
25
0
5 10 15 20 25 30 35 40
20
(b)
25
15
20
10
15
5
10
0
5 10 15 20 25 30 35 40
(c)
It can be concluded that for the case in study Figure 18. Equal settlement curves for the 6th construction
the combined effect of creep and shrinkage of the phase: (a) measured, (b) estimated considering soil structure
concrete has not contributed to a better equaliza- interaction and (c) estimated considering soil structure
tion of settlement when compared to the really interaction and also the effect of creep and shrinkage of the
most relevant effect of soil structure interaction. concrete (in centimeter).
Figure 18 shows the equal settlement curves
for the 6th construction phase, reducing for the It can be observed that there has been a major
estimated settlement the part that occurred before consistency between the measured settlement
the beginning of the instrumentation. Figure 18
curve and that including only the soil structure
includes: (a) the measured settlements, (b) the
interaction. When the effect of creep and shrink-
settlements estimated considering soil structure
age of the concrete is also considered, a change
interaction and (c) the settlements considering
can be observed in the central part of the curve,
also the effects of creep and shrinkage of the
with a more clear equalization of the settlement.
concrete.
It can be then concluded that the considera-
tion of creep and shrinkage of the concrete
indicates a prediction in the numerical analysis in
L.d.M.P. da Rosa et al. / Soil Structure Interaction: An Analysis of the Effect of Creep and Shrinkage of Concrete 183
excess of its real significance when compared to Estimating Young Module in Sands from the Normalized
the sole effect of soil structure interaction, N60 Blow Count, Soils and Rocks, Vol. 35, N.1.
according to the instrumentation. Danziger, F. A. B.; Barata, F. E.; Santa Maria, P. E. L.;
Danziger, B. R. e Crispel, F. A. 1997. Measurement of
Settlements and Strains on Buildings from the Beginning
5. Conclusion of Construction, Vol. II, Proc., XIV International Con-
ference on Soil Mechanics and Foundation Engineering,
The article presented an analysis of a construc- Hamburgo, pp. 787-788.
tion founded on a shallow foundation, in the Gonçalves, J.C. 2010. A Influência dos Recalques das
central part, and a deep foundation in the periph- Fundações no Comportamento de Edificações ao Longo
ery. An instrumentation has been performed do Tempo, Tese de Doutorado, Engenharia Civil, Uni-
with settlement and loading monitoring in some versidade Federal do Rio de Janeiro, 301 p.
columns. Gusmão, A. D.; Gusmão Filho, J. A. 1994. Avaliação da
In view of the results the authors consider Influência da Interação Solo-Estrutura em Edificações, X
that for current foundation design, the analysis Congresso Brasileiro de Mecânica dos Solos e
including soil structure interaction, recommend- Engenharia de Fundações, Rio de Janeiro, RJ.
ed in the new version of the Brazilian Code, is Rosa, L.M.P. 2005. Interação Solo-Estrutura – Análise de um
justifiable for a sole iteration. Caso de Obra Envolvendo Danos Estruturais,
The consideration of creep and shrinkage of Dissertação de M.Sc., Faculdade de Engenharia Civil,
the concrete in the numerical analysis indicates a Universidade Federal Fluminense, Niterói, RJ, 117p.
prediction in excess of its real significance when Russo Neto, L. 2005. Interpretação de Deformação e
compared to the effect of soil structure interac- Recalque na Fase de Montagem de Estrutura de Concreto
tion alone, according to the instrumentation. com Fundações em Estaca Cravada, Tese de Doutorado
em Geotecnia, Escola de Engenharia de São Carlos, Uni-
versidade de São Paulo, 279 p.
6. Acknowledgement
7. References
Abstract. The paper describes principles of solving problems to calculate changes in stressed-strained conditions of soils during
their freezing and thawing, necessary to predict behaviour of structures contiguous with these soils. Approaches are suggested
whereby to choose mathematical soil models necessary to predict deformations of frost heave during freezing, including cases of
water migration in soils and formation of frost related cracks, as well as models to figure out potential soil deformations during
thawing.
2. Finite element modelling of freezing and fields are defined irrespective of the stressed-
thawing problems strained condition of soil. Thus, the solution of
this problem can be presented in two stages. At
A system of equations of a finite element method the first stage we solve the system of equations
for a deformable environment in conditions of based on an equation of non-stationary heat
static (quasistatic) loading is reduced to a linear conductivity:
system of equilibrium equations satisfying the
following boundary conditions: wT
Cth ( f ) U d
>K @^Gt ` ^F t ` , (1) wt , (3)
§ w 2T w 2T w 2T ·
Where [K] – a rigidity matrix of system of finite O th ( f ) ¨¨ 2 2 2 ¸¸ qV
elements (RMS), describing geometry of the © wx wy wz ¹
calculated area and mechanical properties of the
environment, ^G t ` – a vector of node move- where Cth ( f ) – specific thermal capacity of
ments, ^F t ` – a vector of node forces. thawed or frozen soil; Ud – density of dry soil;
The vector of node forces is represented in – temperature; t – time; Oth ( f ) – thermal conduc-
the following way:
tivity of a thawed or frozen soil; x, y, z – coor-
^F t n ` ^Fe t n1 ` ^'Fe t n ` dinates; qV – capacity of internal heat sources.
^FJ t n1 ` ^'FJ t n ` ^Fp t n1 ` , (2) Function of thermal capacity consists of two
^'Fp tn ` ^F f tn1 ` ^'F f tn ` parts. The first part is a volumetric thermal
capacity of soil (thawed or frozen), the second
part is the latent heat of phase transitions within
where {Fe (tn-1)} – the vector of forces caused the below zero temperatures interval, absorbed or
by external load factors (external loading) at the released by soil due to changes of ground water
previous step of the solution (the previous phase, presented in the following way:
moment in time);
wWW
{'Fe (tn)} – an increment of loading on the C( f ) C( f ) L0 (4)
current solution step (either a time step and/or a wT
loading step); where L0 – heat of phase transition water-to-ice;
{FJ(tn-1)} – a vector of forces from mass loads; Ww – moisture of frozen soil due to presence of
{'FJ(tn)} – an increment of the vector of mass unfrozen water.
loads; “Termoground” utilizes the so-called Kronik
{Fp (tn-1)} – a fictitious vector of accumulated model (Kronik Y.A. (1982)), in which latent heat
forces defined by nonlinear mathematical model of phase transitions within below-zero tempera-
of environment and including accumulated ture range from “just beginning to freeze”
deformations (pressure) in the environment. This temperature to “completely frozen” tempera-
vector is defined by the law of deformation ture is accounted for. Heat conductivity and
accumulation (plastic, viscoplastic, viscoelastic, thermal capacity of soil are defined depending on
non-linear elastic, etc). Its calculation by means its condition (thawed or frozen). Within the
of the finite element method normally uses the interval from to thermophysical characteris-
so-called methods of initial stress or initial tics are taken as variable.
deformations; Thermophysical characteristics are defined,
also, by moisture of soil defined by the water
{'Fp (tn)} – an increment of the fictitious vector content, initially contained in pores, and also
of node forces; migrating towards freezing front at a high level
{Ff (tn-1)} – additional load factors defined by of ground water.
temperature influences (frost heave or thaw);
{'Ff (tn)} – an increment of additional forces.
The problem of freezing and thawing of
soils is realized as unconnected, i.e. temperature
186 V.N. Paramonov and A.V. Nabokov / Specifics of Finite Element Modelling of Freezing and Thawing Deformations
wT wT wT
wx l wy wz , (9)
l sx l
wT sy wT sz wT
wS wS wS
wT ª§ wT · 2 § wT · 2 § wT · 2 º
«¨ ¸ ¨¨ ¸¸ ¨ ¸ » (10)
wS wx wy wz
¬«© ¹ © ¹ © ¹ ¼»
To “cause” such frost heave deformations, it
is necessary to subject a finite element to addi-
Figure 2. Dependence of relative settlement of frozen soils
tional loads {'Ff (tn)} in the vector of the right on vertical pressure during thawing.
part of the finite elements equations system,
equal to In this case relative deformations of thawing
^'Ff t n ` ³ >B @ >D@^dH fh `dV soils are defined from the following expression:
T
Abstract. A comparative analysis of a deep foundation of a highrise building in Bangkok subsoil shows that adoption of the
piled raft design concept could result in a significant cost saving of the foundation work. When the raft was placed in the first
stiff clay layer, the load share by piles was reduced to 77% from 100% generally considered in the traditional practice. The
required number of piles could be significantly reduced without scarifying stability of the foundation and settlement requirement
for the building. The 3D FEM analysis yielded a more realistic settlement condition of raft and load distribution of piles than the
conventional plate on spring analysis owing to its direct and complete consideration of soil structure interaction.
Keywords. high-rise building, raft settlement, load distribution of piles, barrettes, FEM analysis
addition, the raft size needed to extend beyond Therefore, the adopted piled foundation concept
the building footprint to accommodate the large for the actual design was reasonable. However
number of piles. For factor of safety of 2.50 the for the purpose of a comparative study, two
ultimate capacity of the pile and barrette were additional levels of basement were hypothetical-
30,000 kN and 45,000 kN, respectively. ly included in analysis to have the raft reach the
stiff clay layer at 15m depth. Thus the potential
benefit of piled raft foundation design concept
could be explored.
3. Methods of analysis
model the raft. Soils above the raft depth were distributed pile load condition underneath the raft
idealized as a surcharge. given by the conventional combined stress
In this study, all column loads were modeled method, which is commonly adopted in the piled
as point loads and only a half of foundation was foundation design in Bangkok, is un-realistic for
modelled owing to the symmetrical shape of the this case study.
foundation. Even though the stability of the whole pile
group under the raft is satisfactory, the designed
pile foundation in this case does not seem eco-
nomical. When considering interaction effect
among raft, pile and soil, some of the piles in this
designed layout have to bear very high axial
loads while others bear very small loads. There
is big contrast in factor of safety of individual
piles in the group. A better and more rational
design would involve rearrangement of the pile
layout as well as a change in amount of pile.
Figure 4. 3D FEM Mesh and elements
Table 1 Results of analyses
4. Analysis results
M1 M2 M4 M5
more uniform. In that case, the piles would not done. There were no time- or floor progress-
be placed in the uniformed pattern as they were settlement records. It was merely reported at the
actually made following the conventional design end that the settlement was in the order of 20-
method. 30mm. Nonetheless, it can be seen that the two
plate-on-spring analyses (M2 and M4) yielded
smaller maximum settlement than the observed
value. The rigorous 3-D FEM analysis shows the
value that seems to be more reasonable.
Bending moments in the raft from the three FEM The results are compared with those of the initial
methods are shown in Fig.8. The maximum design as summarized in Table 2. The load
positive moment from M5 method is about 1.4 distribution on piles for the adjusted design case
times those given by M2 and M4, while the is shown in Fig. 10. As can be expected, the load
magnitudes of maximum negative moment from shared by piles is reduced to 77%. The mini-
the three methods are not so different. The mum safety factor of individual piles decreased
smaller magnitude of bending moment given by from 2.23 and 2.89 to 2.08 and 2.11 for barrette
the plate on spring analyses was due to incom- and bored piles, respectively. The maximum
plete consideration of interaction effect between settlement only slightly increases from 37.2 to
piles underneath the raft. 41.4mm (Fig. 11). The bending moment in the
raft was practically unaffected (Fig. 12).
The construction cost of the building (including By means of numerical analyses, a comparative
superstructure, basement and barrette piles study on the design of pile foundation for a
excluding raft and bored piles) was 30.25 million highrise building in Bangkok subsoils indicated
US dollars. Cost of the project is significantly that the current practice of the design could be
reduced by 1.2 million Baht or about 3.5% of the improved for valued engineering. If the piled raft
cost of the building. It can be said that by allow- designed concept was adopted for the case
ing the foundation raft of the building to reach building, the load share by pile would be only
the first Bangkok stiff clay layer, a significant 77% of the total building load when the raft was
reduction in the number of pile and raft size and placed down to the first stiff clay layer. The 3D
thickness can be made without scarifying the FEM analysis yielded larger magnitude of raft
safety and stability requirement on the pile settlements and consequently larger bending
capacity, raft settlement and stresses in the raft. moment than the FEM Plate on spring method of
In such the condition the stiff clay will share analysis. The latter also gives unrealistic results
approximately 23% of the vertical superstructure on load distribution among piles because it can
load. The adjustment would result in a significant not simulate interaction between piles.
cost reduction. The adjustment and improve-
ment in the foundation design requires an in- 7. References
depth soil-structure interaction analysis made
possible by the 3D finite element analysis. Amornfa, K. (2012) Analysis of piled raft foundations with
their application to Bangkok subsoil condition. Doctoral
dissertation, Asian Institute of Technology, Bangkok
Brinkgreve, R.B.J. and Broere, W. (2004). PLAXIS 3D
Foundation manual - version1. Balkema Publishers,
Netherlands.
Katzenbach, R., Arslan, U. and Moormann, C. (2000). Piled
raft foundation in Germany. In H. J.A. (Ed.), Design ap-
plications of raft foundations. London, Thomas Telford
Ltd.
Phienwej, N., Giao, P.H. and Nutalaya, P. (2006). Land
subsidence in Bangkok, Thailand. Engineering Geology,
82: 187-201.
Poulos, H. G. (2001). Piled raft foundations: design and
application. Geotechnique, 51(No.2): 95-113.
Randolph, M. F. (1994). Design methods for pile groups and
Figure 11. Raft settlement along longitudinal axis of initial piled raft. Proc. of the 13th International Conference on
and adjusted designs Soil Mechanics and Foundation Engineering, New Del-
hi:
Tan, Y.C., Cheah, S.W. and Tahal, M. R. (2006). Methodol-
ogy for design of piled raft for five-storey buildings on
very soft clay. Foundation analysis and design: innova-
tive methods, Geotech. Spec. Publ. (ASCE),153: 226-
233.
Yamashita, K., Kakurai, M. and Yamada, T. (1994). Investi-
gation of a piled raft foundation on stiff clay. Proc. of the
13th International Conference on Soil Mechanics and
Foundation Engineering, New Delhi: 543-546.
Abstract. A non-classical alternative was adopted for the foundation system of a thirty-storey building in Beirut, Lebanon. The
subsurface strata included a thick layer of highly weathered marls and/marly limestone overlying the karstic limestone base
stratum. Ground improvement by rigid inclusions/ mortar columns was proposed, analyzed, designed and successfully executed.
2D & 3D Finite element simulations were conducted to assess the degree of improvement, and then used to perform/finalize the
design of the foundation system. The paper discusses the conducted analyses & presents some parametric studies & recommen-
dations.
Keywords. mortar-column-improved ground, deep foundations, high rise building, differential settlement
2. Substratum characterization
3. Foundation system
3.1. Methodology
The methodology adopted for the founda-tion
system design considered identifying the con-
cerns for every option, and weighing them in
reference to cost and practicality. The classical
options to consider first were the two extremes:
raft foundation on one end, and drilled cast – in –
place shafts on the other. The discussion below
presents the advantages and concerns of each
Figure 2. UCS Values alternative
Another equally important concern is the Moreover, such an alternative is in line with the
karstification, and potential presence of cavities. growing trend to more sustainable designs by
Thus, a cavity probing campaign needs to be eliminating the reinforcing steel from the mortar
conducted adding to the cost, with little effect in pile elements.
filling the cavities beyond the probe hole due to It is worth noting that the proposed mortar
the nature of the material. piles are not directly in contact with the raft, and
The excessive settlement concern was con- have a separation layer consisting of a blinding
firmed by a settlement estimation trial that layer with a thickness of ~50 cm. This layer acts
yielded ~7 cm. Although such a value may be as a “softer” transfer cushion and prevents large
acceptable for certain structures, the structural concentrated stresses from affecting the top of
design team for this project required a maxi- the mortar piles. In addi-tion, this layer facilitates
mum total value within 5 cm and strict differen- the load transfer process and the subsequent
tial settlement and distortion requirements. arching instigated by the relative movement
Moreover, the presence of large cavities close to around the mortar piles (ASIRI 2012).
the raft elevation would render this option as In what follows, the methodology and cal-
risky even if settlements were tolerable. culations which verify the adequacy of the
proposed system are presented.
3.3. Piles
4. 2D Finite element analysis
The classical foundation piles option would have
reduced the settlements at considerable cost. The proposed system requires a three di-
Moreover, it is the authors’ opinion that an mensional analysis to capture the prominent
intermediate option between the two extremes interactions between the raft, the separation
could be sought, especially that the bearing layer, and the mortar pile elements. Albeit the 3D
material is not soft/weak and could/should be nature, and knowing that 2D analysis tools are
utilized in resisting part of the applied load. much simpler and more readily available; two
simplified approaches are presented herein for
3.4. Ground Improvement Using Mortar Piles the analysis. The advantages and limitations of
After identifying the two concerns, a solution each approach are discussed, and then compared
that targets them both at a reasonable cost ap- to a full 3D analysis.
pears viable, especially if this option was also The aim is to provide a raft foundation design
cost effective thus satisfying value engineering that accounts for the serviceability criterion
requirements. Ground improvement by rigid (settlement).
inclusions has been successfully implemented for
soft soil deposits under different names (Con- 4.1. Unit Cell Idealization - Single Mortar Pile
trolled Modulus Columns, Vibro-Concrete Model
Columns, deep mixed columns, etc...), resulting A single mortar pile along with the surrounding
in the improvement of the load bearing charac- in-situ material in its “tributary zone” is consid-
teristics of the reinforced ground . ered in axi-symmetric conditions and modeled
For the current project, adding rigid inclu- using a Finite Element Analysis approach (ref.
sions would improve the load-bearing charac- Figure 4 for schematic). The boundaries allow
teristics of Unit 1 because of the higher strength for vertical movement, while no lateral defor-
and stiffness properties of the inclusions than the mation is possible. This model is considered
surrounding soft rock matrix. Moreover, these representative for a loaded area in the middle of
inclusions/elements, if executed on a regular grid the raft (away from the edges). This statement is
(on the order of 2m to 3m), will intercept and fill supported by the fact that the mortar pile grid is
any large cavities that may be present in the first uniform, and the spacing is relatively close and is
20m below the foundation level. comparable to the raft thickness.
The adopted option comprises 20 m long 60 This model would be useful for the evalua-
cm diameter mortar piles, and is to be construed tion of the average resulting settlement after
as an effective, practical and economical option installing the mortar piles, and is expected to
which could meet the design require-ments. It is result in an upper bound settlement value (as the
a relatively easy to execute option with readily model boundary conditions allow for free
available local equipment and skill levels.
198 S. Sadek and S. Hasan / Foundation System Design for a Tall Structure
movement in the vertical direction). An equiva- being more suitable than a rectangular grid). But,
lent vertical subgrade modulus could thus be if the mortar pile elements are not laid in a grid,
evaluated. However, this model lacks the possi- such an approach would not be suitable.
bility of determining the differential settlements,
and does not account for the flexural rigidity of
the raft in affecting the applied pressures.
Raft 2 m
Unit 1
Mortar Pile
MPs
Raft Edge
Figure 10. 3D vs 3D FEM Comparison – 2.5 m Rectangular Figure 11. Vertical Stress Distribution along Section A-A –
Grid Effect of Separation Layer Parameters
6. Parametric Study
full 3D and unit cell averaged modulus of sub- The resulting settlement shadings are shown
grade reaction values. Due to the uniform grid in Figure 15 below, with a maximum value of
adopted, a single value is estimated, and com- 2.7 cm.
pared with that inferred from the unit cell model.
The subgrade modulus value from the 3D
FEM case is determined along two lines, one in-
between the mortar piles “3D”, and another
passing through them “3D @ MP”. The results
are plotted in Figure 13. It is noticed that the
average subgrade modulus value in-between the
mortar piles is lower than that of the unit cell
model, while that at the mortar piles is larger, so
that the arithmetic average between the “3D” &
“3D @ MP” values is reasonably close to that
inferred from the unit cell model.
10. References
Abstract. Herein the situation is considered when a thawing subsoil of the piled foundation of a frame building is accompanied
by significant settlements of separate pilecaps, damaging steel frame elements of the superstructure and rendering the building
potentially unsafe. Numerical simulation of the soil-structure complex was performed, where movements of the pilecaps were
traced during freezing and thawing of the subsoil in the basement over the period of 36 months. Evolution of the bending
moments in rigid connections of columns with beams of the steel frame was established.
Keywords. settlement of pile foundation, interaction of structure with freezing and thawing soils
Later, one year after beginning of operation, It proved to be true upon inspection of weld-
the building started exhibiting deformations. ed joints between columns and beams as well as
Their progress was revealed by appearance of of some gusset plates, where cracks had been
cracks in floors and partition walls accompanied identified (Fig. 3-$G
by development of significant tilts in certain It is necessary to point out, that only some of
locations (Fig.1). the foundations had been affected by settlement,
During a detailed condition survey of the as shown in from Fig. 2. In this connection, the
school’s structures large settlement differential primary objective was to establish the causes
of some contiguous separate foundations was leading to settlements. Condition survey revealed
revealed (Fig. 2). Maximum settlement of one of the soil floor in the basement to have settled so
the pilecaps exceeded 19 cm, the relative settle- extensively as to lead to some piles becoming
ment differential with the adjacent foundations exposed (Fig. 6,7), which primarily would have
exceeding **(5 been explained by thawing deformations. How-
ever, in view of such local settlement develop-
ment, it was more logical to relate its character to
a greater thickness of the frozen soil in the areas
under consideration, particularly in view of the
fact that in basements of unheated buildings frost
penetration is usually greater in the northern part.
The initial conditions were of uniform The shaded area in the northeast corner of the
distribution of temperatures throughout the building led to formation of inactive layers
depth, equal to (
$°Ê
as had been measured in (layers remaining frozen during the summer
+**+ site prior to construction time) under the basement of the building as it
commencement. As it was already pointed out still remained unheated (Fig. 11, 12). Subsequent
the basement of the building was unheated for 2 thawing of the frozen soils in the basement led to
years. The air trapped in the basement, was taken settlements of the soil floor and the pile clusters
as an insulating layer for the soil floor and additionally loaded by forces of negative friction.
modelled by volumetric final elements. In the
calculation, as boundary conditions for the
problem, Class One conditions were adopted. In
this connection monthly average temperatures
were applied to the ground surface (outside the
building) and to the surface of the floor located
immediately above the basement. The third year
assumed constant air temperature inside the
basement, equal to (*°Ê
Data on air temperatures were obtained from
the meteorological station nearest to the site,
located in the village of Muzhi of Yakut National
Autonomous Region (Table 1). The number of
steps in the solution was equal to 36 at a 3-year Fig. 11. Formation of an inactive layer at the east wall (to the
period of calculation and the monthly break- left of the corner) and the north wall (to the right of the
down. The solution time for the problem using a corner) for the month of June +**$.
computer of average capacity exceeded 7 days.
Table 1.
Temp Month
March
June
May
Sept
Nov
Aug
July
Dec
Apr
Feb
Oct
Jan
References
Fig. 14. Change of the bending moments in beam-to-column ( Ê«½«º¸¹ ÐÐ
Ë«º«»¸®¸¹ ; ¯¬¿®®«Ç
joints along gridline 3 (y-axis - moment, kN/m; x-axis – time,
¸¶®·« Áϸº»«¶¯Ä ·«º·«¬®¸¤¸ ÅÁ«®¯Ç ƺ¯
months)
ƺ¸»ºÅ«®¯¯ ¯ ¸¾¾«¯¹«®¯¯ ¬¹«Ä®¸¤¸
¸¬®¸¹«®¯ÇÍʸ¹º»®®¼ ¯®®¸¹«¶¯¸®®¼
Conclusions ¾½®¸¿¸¤¯¯ ¯Å¼¬·«®¯Ä
ƺ¸·¾¯º¸¹«®¯Ç ¯
¬¾º¸¯¾¿À¬¾¹« ¹ Ȭ¿¸¹¯Ç½ º«Ä®¤¸ ʹº«
1. Principal reasons causing settlement in the «¾-¿¼ »ÂÁÈ®«º ·¸®Ï ·È¾¬·
+*(+ Ê(++ –
pilecaps are the processes of freezing and thaw- (+5
ing, developing within the limits of the northern
and the southern wings of the building.
2. A considerable contribution towards the
settlement of the pilecaps of the northern and the
eastern wings of the school building may also be
added by:
a) possible displacements of load-free piles
! +**+Í+**B r-
mation of cavities under pile toes;
b) excessive humidifying of soils in the
basement due to drainage of precipitation water
from the roof;
c) greater deformability of soils around pile-
toes in comparison with deformability of soils in
the southwest part of the building.
208 Soil-Structure Interaction, Underground Structures and Retaining Walls
V.M. Ulitsky et al. (Eds.)
IOS Press, 2014
© 2014 The authors and IOS Press. All rights reserved.
doi:10.3233/978-1-61499-464-0-208
Abstract: The paper describes the design concept of pile group system for strengthening of ‘’Most Sava – Zeleni’’ railway
bridge. It gives detail description of soil profile, geometry of the system, design methodology and results of static pile load test.
The pile group system consists of 24 bored piles, with 1000 mm in diameter and 24.5 m long. Soil profile consists of stiff soils,
typical for alluvium of River Sava in the area of Zagreb. The basic design concern was to take into account interaction between
the old pier footing, the new pile system and the surrounding soil. The system was designed with Ensoft-GROUP software and it
was additionally analyzed by Plaxis 3D software. Static pile load test was performed to validate the performance of vertically
loaded pile.
Keywords. pile group system, bridge strengthening, interaction old pier footing, piles and soil
Figure 2. Characteristic soil profile with corresponding mechanical and physical properties of the soil
(SPT – standard penetration test; qu – uniaxial compressive strength; PP - pocket penetrometer; PSV – pocket shear vane; Ic –
index of consistency.) . .
210 I. Sokolić and B. Vukadinović / Design of Pile Group System for “Most Sava – Zeleni” Railway Bridge
Principal strain corresponding to the 50% of the Due to complexity of the calculation model
strength of ‘p-: used in Ensoft-GROUP software, it was neces-
H 50 0.002 Ln(c >
# @) 0.0162 (3) sary to check the performance of the system with
some independent methodology. The basic idea
Initial stiffness of ‘p-: was to model the same structure in numerical
>
K
$ / 2 @ 1810 c >
# @ (4)
program Plaxis, and to use the comparable
experience with modelling retaining walls to
determine the soil parameters. Full 3D soil-
COHESIONLESS SOILS: structure interaction model was calculated using
the PLAXIS 3D Foundation software
Shaft resistance: (Brinkgrawe & Swolfs, 2007). The old footing
was modeled with soil elements using the linear-
q E V '
(5) elastic soil model and corresponding parameters
0.25 E 2 0.15( y>@) 0.75 1.8 for the concrete. The same parameters were used
for foundation slab modeled with the 3.3 m thick
Base resistance: floor elements. The piles were modeled with
q 60 N 60 (6) embedded pile elements available in Plaxis 3D
software.
Initial stiffness of ‘p-: The strategy for deriving the parameters for
>
$ / 3 @ 822 N 60 (7) HS soil model was the same as proposed by A.
Szavits-Nossan (2008) for numerical modelling
The calculation model was analyzed as 2D of anchored retain structure, for the soil profile
problem, taking into account the inclination of typical for the area of Zagreb city. The back
the piles in the plane parallel to the railway axes. analysis of horizontal wall movement, per-
The schematic view of the deformed model is formed on several case histories (A. Szavits-
shown on the Figure 4. Detailed description of Nossan et al. 2010) show that the reference
!# " ) M
at stiffness of the HS soil model E50ref correlates
al. 2010. According to the results, almost all the
well with the NSPT values corrected with depth:
moment action on the pile cap is taken by the
vertical reactions in the piles (94 to 110 %) ref
E 50 5 N1 60 (8)
while the rest of the moment is taken by the
reaction moment of the piles and the footing. The strength parameter used for the gravel
Total amount of the applied load transmitted to # ¨Rê # Q
the old footing reach maximum 7% of H, 2% of = 35o), while the undrained strength of the clay
V and only 2% of M action. layer (cu = 120 kPa) was modeled with effective
strength parameters and by performing the
undrained type of analysis.
6. References
Keywords. nuclear power plant foundations, soil-structure interaction for static and seismic conditions
contact pressure at the foundation bottom or and the following options were examined:
values of acceptable total and differential settle- - Subsoil improvement with the help of sandy-
ments. gravel cushion, where part of the less appro-
For closer study 2 typical examples of the priate soil (loess) is substituted by sandy-
nuclear reactors were selected, one with so called gravel;
lower energy output and the second one with - Subsoil improvement with the help of piles,
higher energy output which also had different which are embedded into gravel layer and are
ground plans, the first one with circular founda- in upper part interconnected by reinforced
tion slab and the second one with rectangular concrete slab.
foundation slab. Both options have preferred sub-options,
when sand-gravel cushion is reinforced by
2. Soil structure interaction for static loading geosynthetic layers. Similarly the piles are also
reinforced around outer perimeter by geotextiles
Analytical calculation model was used for the or geogrids with very high initial tensile stiff-
basic limit states – ultimate and serviceability ness. The proposal comes from the positive
limit states. As expected more important is the evaluation of the sensitivity of reinforced earth
limit state of serviceability, namely value of the structures during seismic loading when very high
total settlement. Calculated values of the total initial tensile strength is increasing stiffness for
settlement significantly exceeded acceptable short time seismic loading.
values (usually around 200 mm), even when for M ) " *$ # " ' "#!"t-
the settlement the model using so called principle ed gravel layer is proposed between the zone of
of structural strength was used, Van M improvement and foundation slab as this layer
Van Q+**5G
M
has positive impact on the interaction of building
M. (2013). This model is giving generally lower of reactor with subsoil for seismic loading.
values of settlement than settlement calculation X " " " )e-
based on the theory of elastic half space. There- tween 4.5 to 6.0 m for different NPP. So it means
fore it was necessary to propose subsoil im- that still about 2.0 to 3.0 m of fine soil remained
provement. in place, which is able to protect ground water in
Subsoil improvement was proposed in two lower gravels from potential contamination from
different alternatives, specified in the Fig. 1. the surface. This fact can play very positive role
Proposals for the subsoil improvement are during the EIA evaluation process.
based on the preference of shallow foundation The length of the piles is about 10 m and they
are embedded into gravel layer on the length of 3. Analysis of soil structure interaction for
about 3 m. reactor building during the seismic load-
Subsoil improvement has very positive im- ing
pact on the limit state of serviceability for both
cases of subsoil improvement. Maximum value This analysis is the main part of the feasibility
of foundation slab settlement is now about 160 study as it should prove that the behaviour of
mm what is now acceptable value. ground and realized structure during the seismic
Software PLAXIS was used for numerical loading is within acceptable limits.
modelling, namely in connection to the soil
structure interaction problems. This software 3.1. Evaluation of subsoil from the view of
based on FEM has the possibility to simulate the seismic wave spreading
real soil behaviour using different constitutive
models. During this phase of study the planar The values of seismic wave spreading were
solution (2D) was applied for geometrical model discovered during the ground investigation,
with width 2×200 m and depth 100 m. 3 D namely geophysical investigation. Seismic wave
solution will be applied in the future as it is more S (secondary/shearing) is generating significant
sensitive to the arrangement of individual parts amplitude of the horizontal movement of the
of NPP and to the specification of the technology earth surface and is evaluated as the main reason
of construction (with respect to the different of seismic damage. The velocity of the seismic
depth of foundation and different contact pres- wave spreading S (Vs) is increasing with depth.
sure for different buildings). For upper layer composed from fine grained soil
The subdivision of ground respected different this velocity is increasing from the value of 150
layers in subsoil together with ground water m/s up to 400 m/s. For gravel and sandy soils is
table. Finite element spacing was adapted to the increasing up to the 1000 m/s for the base of this
boundary conditions with detailed spacing layer with another small increase with depth (in
around places of concentration of stresses. With the depth of 100 m – for the model base it is
the help of parametric study it was approved that roughly 1200 m/s). The velocity of the prima-
the constitutive model “Hardening soil small ry/longitudinal seismic wave P (Vp) were discov-
strain model” is the closest to the analytical ered roughly twice higher. For the purposes of
solution for settlement using principle of struc- the seismic analyses the observed locality was
tural strength. This specified model can define classified into type 2 according to the IAEA,
different stiffness of soil; for virgin loading, for Safety Standard NS-G-3.6, (2006), for which:
repeated loading or for the phase of unloading. 1100 m/s > Vs > 300 m/s.
The soil structure interaction for static load-
ing was described in more detail by the authors 3.2. Design seismic level
previously – Van M
Q+*(,G
The observed locality is falling into Seismic
with the following main outputs, which are in Level 2 – SL 2, which corresponds to a level
very good agreement with theoretical assump- with a probability of 10-4 per year of being
tions: @"
" ""& ) ! ""l-
- Just before the top of tertiary clays the eration values (PGA) – for the horizontal direc-
deformation values are close to zero. tion PGAhor and for the vertical direction PGAvert.
- Total settlement is roughly 160 mm for Seismic load ! ""
sandy-gravel cushion even for the cushion values of PGAhor = 0,195 g and PGAvert = 0,101 g
" only 4 m. for the model base resp. PGAhor = 0,349 g a
- Lower total settlement of about 60 mm was
PGAvert = 0,214 g for the upper layer of gravels
calculated for the subsoil improved by piles.
was used. This loading respects new directive of
- The contact pressure in the footing base is
relatively constant with excesses at the end of IAEA SSG-9 (2010) and also the information
slab. obtained on the locality. After that 3 artificially
- Small plastic zones are confined and there are created accelero-graphs with total length of 20
# !! seconds and with step of 0.002 second were
specified independently for each direction, it
means 2 for horizontal directions (x and y) and 1
for vertical direction (z).
M. Vaníček and I. Vaníček / Soil-Structure Interaction Analyses for the Study of NPP Foundation Alternatives 217
3.3. Modelling of seismic waves spreading During the first verification phase 2D model
through ground was applied which was at the base loaded by
accelero-graph from the free field. Numerical
For specified accelero-graph, see for x direction modelling was very comprehensive, not only for
Fig. 2, its change was modelled from the level of both representatives of the energy output NPP
the gravel layer surface (-20 m) to the base of but also for both cases of ground improvement
numerical model (-100 m). At the base the and for basic combination of accelero-graphs x +
accelero-graph was slightly rearranged in such z resp. y + z. The results of these analyses are
manner that after recalculation to the gravel interpreted in the graphical forms as e.g. residual
surface the same maximum value was obtained deformation after 20 seconds of the seismic
as initially defined. Software PLAXIS recalcu- loading, as maximal values of vertical or hori-
lated the given accelero-graph for new boundary zontal deformation during the phase of seismic
conditions at the modelled base – for defor- loading, maximal values of acceleration (in g)
mations which are used as input data for the during seismic loading, resp. in the form of
actual numerical modelling – see Fig. 3. isolines of the total deformation in individual
directions after 20 seconds.
3.4. Results of the soil structure interaction for During the second final phase for the actual-
seismic loading ized accelero-graph defined to the base of the
The modelling with the same software and numerical model and recalculated for the new
modelled area was performed step by step, boundary conditions as described earlier, the
however the outputs are interpreted only for the numerical modelling of the seismic loading
final phase of seismic loading. From the view of started, similarly as during the first phase of
boundary conditions the software is using an verification. The outputs are presented in Figures
adsorption boundary for which the model is 4, 5 and 6. However the most illustrative form of
using viscose boundary (dashpot). the results presentation is the video showing the
movement of the reactor building during the
phase of the seismic loading.
4. Conclusion
Figure 4. Maximal horizontal deformation during the seismic Figure 5. Maximal acceleration [g] during the seismic event -
event - Subsoil improvement by geogrid reinforced sandy- Subsoil improvement by geogrid reinforced sandy-gravel
gravel cushion. cushion.
M. Vaníček and I. Vaníček / Soil-Structure Interaction Analyses for the Study of NPP Foundation Alternatives 219
Figure 6. Residual horizontal deformation after the seismic event – Subsoil improvement by piles.
5. References
Abstract: The pressuremeter is one of the leading tools that allow direct measurement of the soil structure interaction. Its
almost full scale loading of soils during testing, its applicability to most soil types, and recently, to soft rocks has allowed
major progress in the understanding, design and control of soil improvement, where treated soils will have to meet preset
criteria. Structures can be designed based on Menard’s rules ,this with a tested volume allowing the 98 % confidence in
predicting the soil structure behavior, which is by far more than current practice. Indeed the approach of Louis Menard directly
uses the limit pressure, PLM, as the parameter governing bearing capacity, without relying on failure criteria and hypotheses,
and the pressuremeter modulus, EM, is directly related to deviatoric deformation and adopted for volumetric deformations as
well. The selected case histories illustrate the essential input of the Menard Pressuremeter in various applications.
2 Pressuremeter and self bearing in fills Table 1. Acceptance criteria for villa areas
Where w is settlement in cm, h is fill thickness in Dynamic compaction was implemented in the
cm the structure coefficient variable according project using pounders weighing up to 25 tons to treat
to the nature of the soil and the ratio of Menard more than 90% of the loose fill. The excessive
modulus, EM, to Pl. The unit for Pl in Equation 1 is treatment depth of the remaining 10% of the project
bars. required the implementation of heavier pounders;
Al Quo’a New Township was a new development hence the MARS (Menard Accelerated Release
in the deserts of UAE that was to be constructed on System) that is shown in Figure 3 was developed to
levelled dune sands. While some areas of this 3.8 drop a 35 ton pounder in cable-less free fall. This
million m2 site was on competent ground, release system was able to self-attach itself to the
approximately 1.13 million m2 of the project was pounder at the end of each drop cycle.
located on loose fill with thicknesses similar to Figure
2, and sometimes up to 28 m thick (Hamidi et al.,
2010). The project’s developer had first-hand
experience of creep, excessive total and differential
settlements and building cracking in the first phase of
the development, and was seeking a means to ensure
that the same problems would not be repeated.
potential risk for the project that was already bound by As the investigation was not able to clearly define
a tight schedule. the soil profiles and consequently the ground
The settlement criteria specified by the tank improvement method to be applied, a dynamic
fabricators limited total long term settlement at the reconnaissance phase was adopted by dropping a 20
tanks’ shells to 150 mm and the long term differential ton pounder from 20 m and visual observation.
settlement around the shell perimeters and across the
diameters to 1:360 and 1:80, respectively. Settlement
limits related to the design life of 50 year design life of
the tanks, including hydrotest under a full tank
loading.
Thus, an alternative construction method using
dynamic consolidation was proposed by the
contractor. In the new scheme the fill mass was
improved by dropping pounders weighing up to 40
from as high as 40 m. The prints were backfilled with
rock and further compacted with the same equipment
to create a stiffened raft.
Pressuremeter tests were performed in the rock
columns and the sand fill and settlements were
predicted using finite element analysis based on a
medium of varying stiffness with depth over a rigid Figure 7. Schematic cross section of the KAUST ground
base. The D-60 rules (Menard, 1975) were also used conditions (Hamidi et al., 2010b)
and tank settlements of 4 to 7 cm were predicted.
Hydrotest settlement data showed comparable results
Pressuremeter tests provided the parameters for the
with the analytical estimates. The most critical tank
design of footings on sand or silt. As the buildings and
was No. 1311 which was located on a fill with a
thus footing locations were not defined at that phase, a
variable thickness of 20 m from one side to the other
2 m thick sand platform was allowed on top of the silt
side and a maximum fill thickness of 40 m. Hendy and
to ensure the distribution of loads to dynamic
Muir (1997) report the settlement of this 46 m
replacement columns by arching.
diameter tank during the hydrotest to have been from
The design and construct ground improvement
55 to 77 mm. The pressuremeter showed to be the
proposal that met the project manager’s technical
ideal tool for measuring the required parameters in the
requirements, schedule and budget was based on the
rock columns and the compacted sand fill.
below design criteria:
o Footing location: Any place within the
4 Pressuremeter for Pre-Test, Design and treatment area
o Maximum footing load: 1,500 kN
Simplified criteria o Allowable bearing capacity: 200 kPa
The 5.6 million m2 King Abdulla University of o Maximum total settlement: 25 mm
Science and Technology (KAUST) is located in o Maximum differential settlement between two
Rabigh on the coast of the Red Sea and near the city of adjacent footings: 1/500
Jeddah in Saudi Arabia. KAUST, originally o Liquefaction mitigation for an earthquake with
anticipated to have buildings with at most two to three peak ground acceleration equal to 0.07g
storeys. The project was fast track and master o Level: 0.8 m below final ground level, but in
planning, architectural and structural design and any case at least 2 m above sabkah level
construction all had to be completed in less than three A pilot test was realised with pressuremeter testing
years. Thus there was great need for flexibility, and SPT (with grain size) to define boundaries of
coordination and overlapping of tasks (Hamidi et al., application of the dynamic compaction and dynamic
2010b). replacement techniques as a function of grain size,
The preliminary geotechnical investigation that was limit pressure and applied energy (see Figure 8).
carried out rather sparingly indicated that the ground Furthermore, a spread sheet based on D60 rules
was very heterogeneous loose or soft soils with rapid (Menard, 1975) was prepared for the quick estimation
variations of ground conditions within short distances of the bearing capacity and settlement by the site
of even 10 m. This investigation and further testing engineer.
during the works indicated that more than 2,600,000
m2 of the construction area was to be built on soil
consisting of up to 9 m of loose silty sand or soft
sandy silt that is locally called sabkah. A schematic
cross section of the site is shown in Figure 7).
224 S. Varaksin and B. Hamidi / Analysis of Soil-Structure Interaction by Ménard Pressuremeter Tests
through the pounder with the least resistance. It was Po & ! at the test level at
also with dual side functionality; i.e. it was 1.7 m by the time of the test.
1.7 m on one side and used for driving rock The internal friction angle
for sands can be
dynamically into the clay and 2.3 m by 2.3 m on the estimated in sands from the pressuremeter test by
side to dynamically compact the rock blanket. Figure (Menard, 1970):
10 shows this marine pounder.
The self-bored slotted tube or Staf technique Y'Z[ (7)
(Arsonnet et al., 2005) was utilised from a jack up 67 W = 2.5 × 2 [
barge, to perform the pressuremeter tests down to a
depth of more than 30 m. The technique consists of
sealing a casing to the sea floor and driving a BX size
slotted casing with advanced drilling and by utilising
an eccentric bit. The slotted casing is advanced to the
required depth, the bit is removed and the
pressuremeter probe is inserted to depth. After the test,
the slotted casing is jacked up one meter and the next
test is performed. Figure 11 shows the Staf drag bits
that can either have blades or buttons.
Since the stability analysis was performed using the
classical the Mohr Coulomb failure criteria, the
friction angle and cohesion were necessary for the
stability analysis.
Shear strength, c, can be estimated from the
pressuremeter test by (Menard, 1970):
6V W
U= (5) Figure 11. Staf drag bit
5.5
Y'[ (8)
67 W = 4 × 2 ]
embankment
surcharge p
load redistribution by
geogrid reinforcement
load on membrane p´
Figure 2. Model of a reinforced zone on piles by BELL et al. (1994), based on findings of GUIDO, KNUEPPEL & SWEENY
(1987) for punching effects, reduced by lateral reinforcement (right)
Height of embankment H in [m]
grain redistribution
modulus and stress-strain characteristic of the this method, it is possible to consider not only
reinforcement used. the deformation below the geosynthetics but also
The construction method has been imple- the pile settlements. This method allows all three
mented within the wide range of construction load approaches as illustrated in Fig. 5.
methods used for special geotechnical works and ZAESKE (2001) has mainly analyzed the load
can be seen as standard. A majority of projects approach 3 and developed an analytical solution
shows also the economic efficiency of this method for this one. He also modified the design
construction method. The economic savings can into a simplified graphical form, which is the
be expected in the range of 10 to 20 % against basis for the EBGEO (2010) calculation method,
conventional solutions, and shortening the where the load figure is assumed as “Load
construction period. The options combining the Approach 2” and the piles will be considered as
reinforcements and piling systems are nearly rigid elements in comparison to the soft bedding
without limits. behavior of the subsoil. In other words, after
As piling elements driven piles, piles from fulfilling the mentioned bedding requirement,
wood, steel, pre-cased reinforced concrete or the deformation of piles will be ignored and the
ductile cased metal piles can be used as well as design will be focused on the deformation
soil improvement techniques as gravel columns, behavior of the geogrids between the piles,
treated or non-treated, CMC-columns, mixed-in- which enables a design method on the safe side.
place columns etc. Nevertheless, the load trans- This simplified graphical calculation method can
fer on the piles has to be ensured by e.g. pile be used according to EBGEO after fulfilling a
caps or widening of columns in the top-area to a bedding requirement that the bedding modulus
typical diameter of approx. 0.6m … 0.9m. of the piles has to be 75 times greater than the
subsoil bedding modulus.
1.2. Overview of the calculation methods After ZAESKE (2001), GEDUHN &
VOLLMERT (2005) developed and simplified
The first calculation method was developed by the Zaeske method (ZAESKE 2001) into a new
KEMPFERT et al. (1997). In this method, the form, called “Zaeske, simplified”, where the
bedding of the subgrade was considered through piles are considered as rigid elements analogous
its bearing capacity. In this method, a uniformly to EBGEO calculation method. In addition, the
distributed load was taken into consideration as “Zaeske, simplified” calculation method enables
shown in Fig. 5 (Load approach 3). to consider all three load approaches (see Fig. 5),
Following the mentioned Kempfert calcula- which were already described by ZAESKE
tion method, ZAESKE (2001) developed a (2001) and are commented on by VAN
design method based on the publication of EEKELEN (2011). A brief comparison of the
EMDE (1995) where the subsoil bedding could mentioned calculation methods is summarized
be considered via catenary differential equations and illustrated in Table 1.
for the geosynthetics between the piles. With
2 x Secugrid®
Secugrid® 400/40
HEITZ (2006) extended the calculation method firmed by the parameter study of VOLLMERT
of ZAESKE (2001) in terms of cyclic loading. et al. (2012).
Based on the results of the performed model
tests, the soil arching effect of the geogrids tends 2. Worked example: Hafen City Hamburg
to be reduced due to cyclic loading. This reduc- (weihrauch et al. 2013)
tion can be considered by a calculation method
which was developed in the mentioned publica- At the HafenCity in Hamburg, previously part of
tion of HEITZ (2006). the harbour area, the current inner-city area
The effect of the chosen calculation method south of the historic warehouse district is being
(Table 1) has a big influence on the design extended by some 40 % to a total of 157 hec-
results. The research of VAN EEKELEN et al. tares. 5,500 dwellings for 12,000 residents are to
(2012a,b) indicates that the shape of a deformed be constructed, along with office space for
geosynthetic reinforcement between the piles 40,000 employees.
does not follow a power law function lower than In the course of these infrastructure projects,
the third order in any of the performed tests. The the trafficked areas – with the exception of the
inverse triangular load distribution (load ap- quay and embankment promenades – are being
proach 1) also gives a third order power law raised from the current MSL + 5 m to MSL
function, which demonstrates the most similarity +7.5 m to + 8.0 m, to make them safe for flood
(among the other load approaches) to the real events. The soil conditions around today's
deformation form of the geosynthetic reinforce- Honkongstrasse (formerly Magdeburger Strasse)
ment according to the test results. For the same are typical for the HafenCity Hamburg. Fill
geometry considered in the publication of VAN material of low bearing capacity overlies soft
EEKELEN (2012b), the same total load with the organic layers of clay and peat which in turn
inverse triangular load distribution (load ap- overlie firm sands. Raising the level of the road
proach 1) results in a tensile force in the geosyn- embankment by approximately 3.0 m would
thetics that is 75% of the tensile forces calculat- have resulted in long-term settlements of be-
ed by a triangular load distribution (load ap- tween 300 and 400 mm, and significant differen-
proach 2), which shows the big importance of tial settlements would have been expected.
the load distribution choice on the design. To put The system is characterised by vertical col-
in a nutshell, load approach 1 leads to a mini- umns (lime-cement treated gravel, unreinforced)
mum stress on the geosynthetics and a maximum and an overlying sand layer horizontally rein-
stress on the piles and the load approach 2 shows forced with geogrids. The placement grid of the
exactly the opposite tendency, whereas load supporting elements should be designed to
approach 3 ends up somewhere in the middle. transfer the geogrid loads in an orthogonal
These mentioned tendencies were also con- manner. For Hongkongstrasse, this resulted in a
rectangular grid with a spacing of 2.3 m normal
L. Vollmert and J. Schlee / Piled Embankments in Soft Estuary Clay 231
to the embankment axis and 2.5 m in the axial al. (2014). The measurement results figure out
direction, the diameter of the elements was the ductility of the system and robustness against
0.6 m. Reinforced concrete columns with con- subsoil influences gained by an excavation for
tinuous steel reinforcement were used at the the new Greenpeace building, right to the con-
edge of the structure to cope with a bending struction.
moment (e.g. should any excavation be required
at a later date) as a result of lateral pressure. The 3. Worked example: SCHLEUSE NEUER
geogrid-reinforcement is installed 150 mm Hafen, BREMERHAVEN (Vollmert et al.
above the columns in order to guarantee ade- 2006)
quate safety against shear during the construc-
tion phase, and in case of large settlements. The An essential part of the project New Harbour in
design required a short-term tensile strength of Bremerhaven, Germany, in the years 2003 to
400 kN/m. 2005 was the new development of a sluice as
To avoid the risk of subsequent construction connection between the New Harbour and the
activity of investors endangering or destroying Weser River.
sections which have already been built, an area For the crossover of a road across the outer
of 1.0 m of the traffic section must be able to be head of the sluice fillings with thicknesses of up
removed. Fill which intrudes into investor areas to 5 m were required in the ramp area. Due to
must also be removable. In order to prevent any deep clay layers with a low bearing capacity
damage to the embankment support system, the different measures for the reduction of the
geogrid-reinforced fill layer was built with expected large settlements were required.
sufficient overlap. The outer section can thus be Amongst others, due to the earlier installed sheet
removed in the course of normal earthworks. piles a dam filling on cement stabilised columns
The geogrid-reinforced layer was designed was planned. Above the columns a geogrid
according to a verification concept which has reinforced load distribution platform was in-
already been used and proven itself several times stalled to support the bearing arch and to transfer
at HafenCity Hamburg. The limit bearing ca- the dam loads into the pile group. The sequence
pacity was first verified in accordance with a of strata is particularly affected by highly com-
suggestion from KEMPFERT et al. (1997). A pressible, holocene organic soft layers in depths
conservative value for subgrade reaction was of up to approx. -5 m below sea level or in a
used. This verification procedure does not enable total thickness of between 15 m and 20 m. The
any deformations to be inferred. However, sea silt is characterised by a predominant weak
verification of serviceability and of deformation consistency, a medium oedometric modulus of
limitation is compulsory for all construction ES = 1.5 MN/m² and an undrained shear strength
projects in the HafenCity Hamburg. A comple- of cu 15 KN/m².
mentary design procedure was therefore adopt- On an approx. 0.3 m thick sand layer a dou-
ed, using a method developed and extended from ble-layered biaxial geogrid reinforcement
EBGEO, which at the time of planning was only together with an intermediate 0.3 m thick sand
available in its 2004 draft stage. This extended layer is installed, followed by the embankment
design method was verified for similar subsoil fill. The geogrids with a biaxial short term
and loading conditions (VOLLMERT et al. strength of 80 kN/m (Secugrid 80/80/Q6) are
2006). The anticipated further settlements at the characterised by a very high stiffness of J2%
level of the reinforcement were estimated at less 2000 kN/m.
than 50 mm after termination of construction. To prove the serviceability of the load distri-
Comparisons with the current EBGEO (2010), bution platform a comprehensive instrumenta-
available in its final form now that construction tion of the geogrids with strain gauges is carried
is complete, show that the design and verifica- out. The instrumentation has been complement-
tion of the system is sufficiently robust to cater ed by vertical pressure gauges above the cement
even for the special case "Loss of Subgrade stabilised piles and between the piles as well as
Reaction" in Load Case 3. by settlement measurements. To register the arch
Extensive monitoring using horizontal incli- effects earth pressure cells have additionally
nometer, strain gauges and stress transducers has been installed in the middle of the field and on
been worked out in this project, reported by one pile cap. The earth pressure cells in the
WEIHRAUCH et al. (2013) and SCHÄFER et middle of the field are thereby graded in the
232 L. Vollmert and J. Schlee / Piled Embankments in Soft Estuary Clay
altitude in such a way that reductions of vertical (DMS). The instrumentation could already be
stress due to arching effects can be comprehend- prepared in-house and thus in-situ be installed
ed, Fig. 6. The strains occurring within the together with the geogrids in only a few hours.
reinforcement are registered by strain gauges
OKroad
Planum
surface
H = 4…5m 3
0,3 0,3
2
1 *
Auff.
sand
Sand
h = 15…20m Klei
sea silt
sand
Sand
d=0,6m
2,0m 2,0m 2,0m
1 1. Lage lastenverteilende
1. load Sandschicht
distribution layer (sand)
2 2. Lage lastenverteilende
2. load Sandschicht
distribution layer (sand)
3 Sandschüttung
embankment fill (sand)
2 Lagen Geogitterbewehrung
Secugrid 80/80 Q6 (two layers)
* Herstellungsebene vermörtelte
working plane for column Säulen
installation
pile cap
top
pile cap middle
top bottom
middle
bottom
vertical stress
vertical stress
Abstract. To investigate the effect of soil-structure dynamic interaction (SSI effect) on the dynamic response of base-isolated
structure, some shaking table tests have been completed before. Based on the test results and the existed studies, an efficient
simplified model and corresponding calculation method to estimate the dynamic characteristic of base-isolated structure with SSI
effect is developed and verified. Compared with the shaking table test and with the standard specification method suggested in
China code without SSI effect, the new developed method can efficiently reflect the influence law and degree of SSI effect on
the dynamic responses of base-isolated structure.
Keywords. base-isolated structures, earthquake impact, shaking table tests, soil-structure interaction effect
Accordingly, the influence laws of SSI effect the additional weight is added in each floor of
on the dynamic characteristic and response of model structure with 0.736t. The total weight of
soil-isolated structure interaction system has the model structure and the additional weigh is
been studied by four shaking table tests before. about 4.0t.
In this paper, an efficient simplified method to
calculate the dynamic characteristic of isolated Four lead rubber bearings are used as the
structure with SSI effect is developed and isolators in the isolation layer of structure.
verified by tests, which can be used easily in According to the stress similitude ratio of
seismic design of isolated structure. isolator and the upper weight on them, the model
isolators are designed with diameter of
2. Shaking table test transection 100mm and the average compressive
stress designed 1.3N/mm2, which geometry
shape and the physical and mechanical quantities
A four-layer steel frame structure is made by are given in Figure 1(b) and Figure 1 (c).
square steel tube used to make the columns and Four shaking table tests are designed to model
H-shaped steel used as the beams. The main sizes the dynamic response of a four-floors structure in
of model steel structure are shown in Figure 1(a). four different conditions such as the non-isolated
The size of model structure’s cross section is structure on rigid foundation (Test I), the
0.8m (vibration direction) × 0.6m (vertical to the base-isolated structure on rigid foundation
vibration direction), which height is 2.1m. Steel (Test II), the non-isolated structure on soil
plate with thickness 6mm are used as the floor foundation (Test III) and the base-isolated
slab. Accordingly, the aspect ratio of model structure on soil foundation (Test IV). The tests
structure in vibration direction is 2.625. The are explained in Figure 2, and some photos in
weight of model structure is 0.32t. To consider test are shown in Figure 3.
the similitude ratio of gravity as soon as possible,
(a) Size of model structure (c) Structure of the lead rubber bearing
Figure 1. Model structure and lead rubber bearing designed for test (Unit of size: mm)
Z. Haiyang et al. / An Efficient Method for Estimating the Dynamic Response of Base-Isolated Structure 237
Lead rubber
isolator
Rigid steel Lead rubber Rigid steel
isolator
Structure
Isolation Bearing
(a) Fixed on rigid foundation (b) Fixed on soil foundation (c) Lead rubber isolator
Figure 3. Photos of tests for isolated structure on different foundations
3. Efficient method to estimate ssi effect the model shown as Figure 4(a) can be
simplified to be a more simply model shown as
Figure 4(b). In Figure 4, k0 and c0 are the
3.1. Simplified analysis model
horizontal impedances of isolation layer, ku and
In theory, with SSI effect, the inertia force of cu are the horizontal impedances of soil
upper structure will be transited to foundation, foundation, k © and c © are the rotation
which causes additional translation and rotation impedances of soil foundation, h is the distance
of foundation. By analyzing on the existed from the gravity center of foundation to the
earthquake damages, Sivanovic (2000) also whole inertia force’s action spot of isolated
found that the SSI effect is obvious, which structure for the first-order vibration mode, u is
causes the swing of foundation. This the horizontal displacement of isolated structure
phenomenon has also been found during the relative to the foundation, uf and T are the
shaking table tests in this paper. Accordingly, an horizontal displacement and rotation of pile
efficient model should be advanced to estimate foundation to the lateral soil foundation,
the influence of SSI effect on the dynamic respectively, ug is the horizontal displacement of
response of isolated structure. What’s more, soil foundation at ground surface.
according to the study by Wenguang Liu (2003),
238 Z. Haiyang et al. / An Efficient Method for Estimating the Dynamic Response of Base-Isolated Structure
mn
1^ u
cn k n m
hi?^ i ui mi
ci k i
m2 uf
h
k0
m1 c2 k 2 hi
ku c 0
Isolator m0 c1 k 1
c0 k 0
h0
cu
uf ^ k^ c^
?
Soil foundation ug
[ [0 ¦k [ i i
(6)
isolated structure, Equation (9) and Equation (10)
can be expressed as
k
~ k § k h2 ·
According to the Equation (1) ~ Equation (3), T T 1 ¨¨1 u ¸¸ (11)
~ ku © kT ¹
the frequency Z and the damping factor ~
[
can be divided by ~ ~ 0.05
[ [0 ~
T T 3 (12)
1 1 1 1
~ ~ ~2 ~2 (7)
Z2 Z Z Z
2
s u T
Z. Haiyang et al. / An Efficient Method for Estimating the Dynamic Response of Base-Isolated Structure 239
~ ~
Where, T and [ are the first-order vibration Vki
period and damping factor of interaction system, D im K rD i Kr n
(16)
respectively, [~0 is the damping factor of
foundation which can be decided by the values ¦G
j 1
j
developed method (Methodĉ: with effect of [ are all largest by Method III, and which are
~ ~
isolation layer’s rotation, Method Ċ: without all smallest by Method IV. However, T and [ of
effect of isolation layer’s rotation) are compared isolated structure calculated by Method II are
240 Z. Haiyang et al. / An Efficient Method for Estimating the Dynamic Response of Base-Isolated Structure
very closer to those by shaking table test, vertical coordinate axis denotes the isolation
especially for the damping factor. It means that layer and other numbers denotes the floor
the developed method is efficient to estimate the number of isolated structure.
influence of SSI effect on the dynamic By Figure 5, the ISFs by ERDC of China are
characteristic of isolated structure. To rigid all smallest compared with other three methods,
~ ~ and the different between them become larger
foundation, T and [ calculated by ERDC is with the PGA of input motions increasing.
very close to the test results. However, with SSI effect, the ISFs by developed
Based on the acceleration responses at each method without corrected by r are very closer
floor of isolated structure, the displacement to the test results, especially for PGA=0.1g and
responses can be calculated by the integration of PGA=0.2g. However, the ISFs by r are bigger
acceleration responses. Then, the ISFs of than those by shaking table test, and this effect
isolated structure can be calculated by the becomes stronger with the height of structure
interlayer relative displacements and the increasing. It means that the rotation effect of
interlayer lateral stiffness. The ISFs of isolated isolation layer is excessively valued by using r
structure by different methods are shown in to correct the ISF. To PGA=0.3g, the ISFs at the
Figure 5 with EI wave inputted. To other waves, upper floor of isolated structure can be corrected
the change laws of ISF are also same to those by r roughly.
with EI wave inputted. In Figure 5, zero at the
Floor number
3 3
2 2
1 1
(a) PGA=0.1g (b) PGA=0.2g
0 0
Interlayer shear force (KN) Interlayer shear force (KN)
Abstract. This paper presents the recommendations of determination of the bearing capacity of driven piles according to
dynamic cone penetration test (DCPT). As well as the sufficiently close correlation between the main parameters of dynamic
cone penetration test (DCPT) and static cone penetration test (CPT) was identified during the work and presented in this article.
Existing regulations prescribe the DCPT test- However, during the juxtaposition of the re-
ing regardless of ground conditions of the sults obtained by aforementioned equations and
construction site in accordance with GOST the dynamic tests of piles, with the results of
19912-2001 to estimate the possible depth of parallel conducted CPT, the discrepancy of
piling and the piles’ bearing capacity, as well as bearing capacities of the piles of all three tests in
to define the mechanical and deformation approximately the same ground conditions is
characteristics of investigated soils. detected. It should be noted that the bearing
However, to determine the bearing capacity capacities of the piles, determined by CPT near
of driven piles in the existing regulations are the test piles with those obtained by dynamic
given recommendations only for CPT. Currently tests are approximately equal.
DCPT is used in clayey soils only for the indi- In this situation, it was clear that it is neces-
rect determination of possible depth of piling sary to clarify abovementioned equations, by
and clarify the lithological section (soil strata), comparing the values of the bearing capacities of
in sandy soils, there to evaluate the mechanical real piles with those of CPT and DCPT.
and deformation characteristics. 163 DCPT were performed on 16 construc-
During the detailed study of the sectoral lit- tion sites, in the places, assigned for the dynamic
erature (Mariupol,1989) we have found the tests of the piles according to the pile outlines of
recommendations to determine the bearing the project. CPTs also are carried out at the same
capacity of driven piles according to DCPT places. In other cases, when the depth of DCPT
results. Mariupol proposes the use for sandy exceeds a depth of CPT for the comparison of
soils the following equations to determine the bearing capacities of the piles the data of CPT to
resistance of the soil under the toe of pile R, probing depth was taken and for lower depth the
MPa, and the lateral surface soil resistivity of bearing capacity of pile capacity defined by
pile f, MPa: dynamic test has been taken.
The following equations were found accord-
0 ,13 ing to the results of performed works:
ª § p ·º § p · - for the determination of the resistance of
R p d «0.91 0.57 lg¨¨ d0 ¸¸» ¨¨ d0 ¸¸ (1)
the soil under the toe of pile R, MPa, and the
¬« © p d ¹»¼ © p d ¹
lateral surface soil resistivity of pile f, MPa, in
sandy soils using the following equations:
0 ,13
pd ª § p ·º § p ·
f «0.58 0.32 lg¨¨ d0 ¸¸» ¨¨ d0 ¸¸ (2)
100 «¬ © p d ¹»¼ © p d ¹ R 0.4 105 pd3 0.8 103 pd2 0.0948 pd
0.0177
where pd- dynamic soil resistivity, MPa; pd0 = 1
(5)
MPa.
8 5 3
To determine the resistance of the soil under f 0.9 10 p 0.3 10 p 0.6 10 p d
3
d
2
d
the toe of pile R, MPa, and the lateral surface
soil resistivity of pile f, MPa, in clayey soils it is 0.2 10 3
proposed to use the following equations: (6)
where pd- dynamic soil resistivity.
0 ,17 - for the determination of the resistance of
ª § p ·º § p · the soil under the toe of pile R, MPa, and the
R p d «0.54 0.35 lg¨¨ d0 ¸¸» ¨¨ d0 ¸¸ (3)
«¬ lateral surface soil resistivity of pile f, MPa, in
© p d ¹»¼ © p d ¹
clay soils using the following equations:
R 0.2 10 5 p d3 0.5 10 3 p d2 0.0568 p d
0.32
pd ª § p ·º § p · 0.0056
f «2.82 2.11 lg¨¨ d0 ¸¸» ¨¨ d0 ¸¸ (4)
100 ¬« © p d ¹»¼ © p d ¹ (7)
R 0.0091 pd0, 4456 (8)
where pd- dynamic soil resistivity, MPa; pd0 = 1
where pd- dynamic soil resistivity.
MPa.
244 A.Zh. Zhussupbekov et al. / Determination of Bearing Capacity of the Precast Piles by DCPT
It should be noted that to determine the re- A characteristic feature of sandy soils is the
sistance of the soil under the toe of the pile R, existence of the relatively clean (free of impuri-
MPa, it is necessary to use an average value pd, ties clay particles) and dense layers of different
MPa, obtained from the test, in a portion located thicknesses. A representative case is shown in
one diameter d above and four diameters d Fig. 2.CPT test was terminated at a depth of 7.2
below the pile toe (where d - diameter or square meters, due to reaching the limit values of
side, or large rectangular side of a driven pile, frontal resistance of the probe tip. However,
m). during DCPT, the layer described above was
As an example for determining the bearing traversed. The layer of sandy soil with worse
capacity of the piles in clayey soils, we present strength and deformation characteristics under
the lithological column of the borehole with the abovementioned layer remained hidden; during
executed near the CPT and DCPT. Near this the CPT. The same fact was confirmed during
borehole the driven test pile was penetrated and the dynamic tests of the real piles.
was subjected to the dynamic load test for
determination it’s bearing capacity. The results
of experimental works are presented in Fig. 1.
Figure 1. Comparative analysis of determining the bearing capacity of driven piles in clayey soils by DCPT, CPT and dynamic
load test of real pile.
A.Zh. Zhussupbekov et al. / Determination of Bearing Capacity of the Precast Piles by DCPT 245
Figure 2. Comparative analysis of determining the bearing capacity of a driven pile in sand soils by DCPT, CPT and dynamic
load test of real pile.
Table 1 Comparative results of CPT, DCPT and dynamic tests of real piles
The following relations from a comparison of of probe fs, MPa, depending on the dynamic soil
data in parallel conducted experimental work by resistivity pd, MPa, in sandy soils (Fig. 3 and 4);
DCPT and CPT were found: - for the determination the resistance of soil
- for the determination the resistance of soil under the cone probe qc, MPa and sleeve friction
under the cone probe qc, MPa and sleeve friction of probe fs, MPa, depending on the dynamic soil
resistivity pd, MPa, in clayey soils (Fig. 5 and 6).
Figure 3. Diagrams - ratio between dynamic soil resistivity pd and soil resistance under the cone probe qc in sandy soils.
Figure 4. Diagrams - ratio between dynamic soil resistivity pd and sleeve friction of probe fs in sandy soils.
Figure 5. Diagrams - ratio between dynamic soil resistivity pd and soil resistance under the cone probe qc in clayey soils.
Figure 6. Diagrams - ratio between dynamic soil resistivity pd and sleeve friction of probe fs in clayey soils.
A.Zh. Zhussupbekov et al. / Determination of Bearing Capacity of the Precast Piles by DCPT 247
4. Conclusions
5. References
Abstract: Secant pile walls are increasingly becoming popular groundwater cut-off systems of choice on underground construc-
tion sites, as they offer a number of advantages over traditional sheet piling and diaphragm walls. However, some elements of
uncertainty surround the design and construction of secant pile walls. Vertical and horizontal installation tolerances are major
issues, especially in the case of multi-level basements, as required tolerances are difficult to achieve with deep piles associated
with deep basements. In addition, while conventional foundation piles are routinely statically or dynamically load-tested, hard
piles in secant pile walls are rarely axially tested and this does not allow for the routine verification of existing design methods
and assumptions.
This paper presents the design, construction and testing of the secant pile wall on a 3-4 level basement construction scheme
in Central London. Simple empirical correlations for estimating the geometry of secant pile walls for effective groundwater cut-
off are derived and validated. An alternative approach for estimating the vertical capacities of secant pile walls is also presented
and compared with a common routine method in the industry. The alternative method is shown to be less conservative than the
common industry approach, while predictions are also in good agreement with results of a recent series of model tests on
perimeter pile groups.
This paper focuses on the prediction of verti- 18 storey-residential block of luxury apartments.
cal axial capacities of secant pile walls, as well The development included a 3-4 level basement;
as the fundamental principles of design of secant a 3 level-basement over the entire footprint of the
pile walls to provide effective groundwater cut- site and a 4th level of basement in the centre of
off. For both subjects, direct references are made the site. Due to the proximity of the site to the
to a recently completed multi-level basement Thames River, groundwater level was relatively
construction project in Central London, United high and subject to tidal fluctuation. Therefore,
Kingdom. in addition to soil retention, groundwater cut-off
was also a necessity, to enable basement con-
2. Site geology & ground model struction and a hard-firm secant pile wall was
considered to be a potentially effective and
The site is located in Vauxhall, Central London. economical solution for the first 3 levels of
Site stratigraphy comprises of up to 3m thick basement across the site. The secant pile wall
made ground underlain by a layer of very soft to system comprises of interlocking male and
soft alluvial clay and peat to 9m depth, which in female piles. The female piles are only required
turn overlies a 1m thick layer of medium dense to provide groundwater cut-off and they do not
to dense sandy gravel. Firm to stiff to very stiff provide any structural support in respect of wall
to hard London clay lies below the sandy gravel overall stability. An illustration of the wall
stratum at 10m depth and this is underlain by a geometry is shown in figure 1.
layer of very dense Thanet Sand at 48m depth. Male piles were made of grade 35 structural
The London clay is an overconsolidated material concrete and reinforced down to depth required
of high plasticity, with plasticity indices ranging for overall wall lateral stability, while the female
between 35%-45%. As the site was immediately piles were made of 10 N/mm2 concrete and
adjacent to the River Thames on the southeastern unreinforced. Both male and female piles were
boundary, groundwater level was subject to tidal provided with a minimum embedment of 1m in
fluctuation and standpipe monitoring results London clay below basement formation level, for
show that water level could be as high as 1m groundwater cut-off. Male piles were of 880mm
below ground level (bgl). Generalised site diameter, while female piles were of 900mm
stratigraphy and associated geotechnical design diameter. A photo of the secant pile wall is
parameters are presented in table 1. shown in figure 2. However, as the 4th level of
Coefficient of earth pressure Ko was general- basement wholly lies within the London clay
ly limited to 1.0 for the overconsolidated materi- stratum, groundwater control was not considered
als (after CIRIA Report No. C580, 2003). This to be a major issue and therefore, a contiguous
was to allow for remoulding and stress relief in pile wall was considered to be the most econom-
the soils; conditions associated with embedded ical and effective solution. Contiguous pile wall
retaining wall installation and excavation. In the comprised of 750mm diameter bored piles
temporary condition, effective wall friction spaced at 900mm centre to centre (c/c).
angles G on the active and passive sides were Maintaining a perfect interlock between adja-
taken to be 0.67Ic and 0.5Ic respectively (where cent male and female piles is often difficult to
Ic is effective angle of shearing resistance). realize when dealing with multi-level basements,
However, in the permanent condition, angle of as required horizontal and vertical tolerances are
wall friction was assumed to be zero; this was to difficult to achieve with many conventional
allow for the effect of vertical compressive piling rigs, for deep bored piles associated with
loading on the wall from the superstructure in the deep basements. To mitigate this risk, prior to
long term. commencement of piling, a guide wall was put in
place along the wall alignment, for improved
3. Proposed development & construction horizontal positional tolerance. In addition, for
procedure improved verticality tolerance, male piles were
installed with a Cased Secant Piling (CSP) rig,
The site was adjacent to the River Thames and which is a customized version of the traditional
the Vauxhall Bridge in Central London. The Continuous Flight Auger (CFA) drilling rig,
project was a large scale high rise development equipped with a temporary heavy duty steel
comprising of two separate structures on the casing over the full depth of the pile.
same site; a 28 storey-commercial tower and an
A. Adekunte / An Investigation into the Vertical Axial Capacities and Groundwater Cut-Off Capabilities 253
The CSP technique improves the achievable The 4th level of basement at the centre of the
vertical tolerance to 1:150, when compared with site, which had to be constructed beneath the
1:75 achievable with a conventional uncased excavation level of the perimeter 3 level-secant
CFA bore (after ICE SPERW, 2007). In addition, pile wall, was supported with a contiguous pile
the customized technique results in better fin- wall. This was installed by a Bachy Soletanche
ished wall appearance and reduced overbreak Large Diameter Auger (LDA) rotary boring
when compared with traditional CFA drilling, technique. Above the London clay stratum, the
while it is also faster and more economical when LDA rig advanced the drillhole with the combi-
compared with rotary bored piling. nation of an auger and a thick-walled temporary
Economy is another important factor in foun- casing, rotating in opposite directions. The
dation engineering and construction (especially temporary casing was required to provide lateral
from a contractor’s perspective) and this has to support to the hole in the unstable made ground
be accounted for in every aspect of wall design and alluvial deposits above the London clay. In
and installation. For economical reason, female the London clay, hole was drilled with the auger
piles were designed to be installed by traditional only. This rotary boring technique allowed each
CFA drilling; the reduced verticality tolerance hole to be partially concreted up to the formation
was off-set by the commercial benefit derived level of the general 3-level basement (approxi-
from the lower higher rate of the CFA rig, when mately 11m bgl), while the rest of the hole was
compared with the CSP rig. backfilled with carefully selected gravel peas up
254 A. Adekunte / An Investigation into the Vertical Axial Capacities and Groundwater Cut-Off Capabilities
to ground/piling platform level. This approach depth required for cut-off; this is a function of
allowed for the installation of both the secant and horizontal positional tolerance, verticality toler-
contiguous pile walls from ground/piling plat- ance and length of pile L required for cut-off
form level, such that excavation for the top 3 (equivalent to length of female pile). D is mini-
levels of basement was done without having to mum allowable bite thickness in mm between
deal with concrete piles projecting from the 4th male and female piles at cut-off depth. This may
level contiguous pile wall in the middle of the be taken as 25mm (after ICE SPERW, 2007).
site; this also resulted in savings in concrete.
Overall, the basement works package in- G in Equation (1) can be expressed as;
volved the use of three different bored piling
techniques for improved efficiency and econo- G = _ [( X W a) + 1] (2)
my. Figure 3 shows wall installation with piling
rigs on the site. where X is average verticality tolerance of
both male and female piles in mm/m depth of
4. Groundwater cut-off installation, which is dependent on rig type and
piling technique. h is horizontal positional
Many cases of flooded and/or abandoned se- tolerance in mm at piling commencement level;
cant wall-supported basements have been rec- this is dependent on availability of guide wall
orded. Such problems arise because many prior to the commencement of piling. A con-
engineers are unaware of the fundamental princi- servative value of 25mm may be assumed (after
ples underlying the accurate estimation of opti- ICE SPERW, 2007). L is pile length required for
mum interlock thickness between male and cut-off; this is equivalent to depth of female pile
female piles (and ultimately pile spacing) re- toe. k is a constant, which accounts for the
quired to ensure an effective groundwater cut-off potential horizontal and vertical deviation of
solution. To provide effective groundwater cut- immediately adjacent male and female piles in
off with a secant pile wall, many factors would opposite directions. Generally, the value of k
need to be accounted for. These include; should be taken to be a minimum of 2. There-
fore, Equation (1) may be rewritten as;
x rig type;
x rig capability; E = _ [( X W a) + 1] + D (3)
x installation technique;
x pile size; The value of E can be used to estimate the
x verticality tolerance; required centre to centre spacing of the male
x horizontal positional tolerance; piles, provided male and female pile diameters
x depth of sealing stratum; are known. This is illustrated in figure 4. In
x minimum required pile embedment figure 4, Dm is male pile diameter, Df is female
in sealing layer. pile diameter and E is thickness of interlock
between overlapping male and female piles.
For effective groundwater cut-off to be guar- From the figure, male pile centre to centre
anteed, all of the above factors must be account- spacing S can be expressed as;
ed for. Inability to account for only one of the
factors could result in significant consequences. eO ei ei eO
While publications with comprehensive specific S=d f + dh j 4 Ef + d f + dh j 4 Ef
2 2 2 2
information on this subject are very rare, field
experience has shown that the minimum required (4)
interlock thickness between male and female
piles may be expressed by the equation;
E=G + D (1)
? k = eO + ei 4 2E (5)
? k = eO + ei 4 2_ [(X W a) + 1] 4 2D (6)
Adopting a value of 2 for k as suggested
above, Equation (6) becomes;
k = eO + ei 4 4 [Xa + 1] 4 2D (7)
Constant k accounting for potential deviation the superstructure. The RC capping beam on the
of overlapping piles in opposite directions can be piles was designed to spread the point loads from
taken to be 2 as suggested above. the columns over a minimum of 3 No. hard piles,
By substituting the above values in Equation such that the maximum estimated service com-
(3), minimum required bite thickness E between pressive load on any hard pile was 2450 KN.
overlapping male and female piles is estimated to As stated in earlier sections, the estimation of
be; axial capacities of pile retaining walls is an area
E = 2[(8mm/m * 12m) + 25mm] + 25mm # shrouded in uncertainty because unlike tradition-
270mm al bearing piles, hard piles in bored pile walls are
Male pile diameter Dm = 880mm, Female pile rarely tested. It is common practice among
diameter Df = 900mm designers to routinely approach this problem by
Substituting the above values in Equation (8), estimating the capacity of a singly acting pile in
required male pile centre to centre spacing is similar ground conditions and thereafter, factor-
estimated to be; ing down the estimated capacity to 50%-60% of
S = [880 + 900 – 2[2(8 * 12 + 25) + 25]]mm estimated figure, to account for pile group
S = 1246mm # 1.25m efficiency in the wall. This approach is common
Therefore, required centre to centre spacing for walls in clay. However, Rose & Taylor
of male piles is 1.25m c/c, which is approximate- (2010, 2012 & 2013) carried out some experi-
ly equivalent to 1.4D (where D is pile diameter). mental work and parallel finite element model-
Installation of the secant pile wall on the site was ling on the behaviour of pile groups arranged in
based on the above figures and the end result was grids in clay at the City University London. They
a perfectly dry excavation as can be seen in discovered that piles on the perimeter are subject
figure 5 below, despite the wall having to deal to higher loads when compared with the internal
with relatively high water level that was subject piles; this they have attributed to increased soil-
to tidal fluctuation and more than 10m of cohe- pile interaction.
sionless made ground, very soft to soft alluvial Further to this, they carried out a series of
deposits and sandy gravel. tests on piles arranged in single row, to model
the behaviour of perimeter pile groups with no
internal piles, whilst also considering several
variables that influence pile group behaviour;
number of piles, pile spacing, length/diameter
ratio and soil strength. From their experimental
results, Rose & Taylor (2013) discovered that
pile group efficiency of perimeter piles could be
as high as 100%, while perimeter pile groups
with pile centre to centre spacing of less than 2D
(where D is pile diameter) tend to exhibit block
type failure. Hard piles in secant pile walls are
typically spaced at d 1.5D centre to centre.
Based on Rose & Taylor’s (2010, 2012 &
2013) test results, at design stage, the wall on the
current site was considered to act as a continuous
deep strip footing below basement formation
Figure 5 – Perfectly dry excavation was achieved with secant
level, surrounded by a block of soil, with the
pile wall on the site
assumption of a block type failure mechanism in
the ultimate state. In addition, the bearing capaci-
5. Prediction of wall axial capacity ty factor Nc in the London clay was reduced by
multiplying with a reduction factor g, to account
In addition to earth and water retention, the for the existence of non-structural female piles in
hard piles in the secant wall were required to the wall. The reduction factor g is expressed as;
support vertical compressive column loads from
A. Adekunte / An Investigation into the Vertical Axial Capacities and Groundwater Cut-Off Capabilities 257
been shown to be considerably less conservative Rose, A.V 2012. Behaviour and efficiency of perimeter pile
groups. PhD thesis, School of Engineering and Mathe-
when compared with a common routine design matical Sciences, City University London.
approach, while predictions of the method are Sowers, G.I, Wilson, L, Martin, B and Fausold, M 1961.
shown to be in reasonably good agreement with Model tests of friction pile groups in homogenous clay.
the results of a recent research-based small scale Proceedings of the 5th International Conference on Soil
Mechanics and Foundation Engineering, Paris, 1961.
model tests on perimeter piles.
Taylor, R.N, Rose, A.V and Gorasia, R.J 2013. Pile and pile
group capacity; some findings from centrifuge tests. In-
7. References ternational Journal of Geo-Engineering, volume 5, issue
2, pp. 5-15.
Adekunte, A 2013. Alternative deep basement solutions for Taylor, R.N and Rose, A.V 2010. Modelling the performance
built-up areas. Proceedings of the Deep Foundations of micropiles at close spacing. 10th International Work-
Institute’s 38th Annual Conference on Deep Foundations, shop on Micropiles, Washington DC, September 2010.
Phoenix Arizona, USA, September 2013.
Adekunte, A, O’Hara, P and Dennany, A 2010. Novel 8. Acknowledgements
methods of restraining embedded retaining walls. Pro-
ceedings of the Deep Foundations Institute’s 35th Annual
Conference on Deep Foundations, Hollywood Ca., Oc- I would like to specially dedicate this paper to my
tober 2010. newborn baby; Princess Alicia Omojadesola
Adekunte, A 2008. An innovative use of bored tension piles Adebimpe and her mum Adedoyin. They both
in embedded retaining wall design and construction. make my world so colourful and exciting. The pa-
Proceedings of the Deep Foundations Institute’s 33rd per is also dedicated to every child in Yewaland,
Annual Conference on Deep Foundations and 11th Inter-
national Conference on Piling and Deep Foundations, Southwest Nigeria and every member of the
pp. 167-176, New York, October 2008. Young Yewa Professional Foundation; a non-
Adekunte, A 2007. Predicted and observed performance of an profit youth-led organization committed to the im-
anchored retaining wall in granite. Proceedings of the provement of the standard of education and the
Institution of Civil Engineers’ International Conference overall standard of living in the community.
on Ground Anchorages and Anchored Structures in Ser-
vice, London, November 2007.
The author is grateful to the management of Bachy
Broms, B.B 2007. Foundation engineering. Online Text Book, Soletanche UK, the geotechnical specialist con-
webforum europe, tractors for the study site (Riverwalk House,
www.geoforum.com/knowledge/texts/broms/index.asp Vauxhall, Central London) and the main contrac-
Construction Industry Research and Information Association, tor; Sir Robert McAlpines UK. The contributions
2003. Report No. C580 – Embedded Retaining Walls –
Guidance for Economic Design, CIRIA, London.
of Barry Osborn, David Hard, Rob Cunningham,
Finno, R.J and Bryson, L.S 2002. Response of a building Alistair Mitchel and Dimitrios Leventakis towards
adjacent to stiff excavation support system in soft clay. the successful completion of the deep basement
Journal on the Performance of Constructed Facilities, scheme are acknowledged.
American Society of Civil Engineers, volume 16, issue 1, The contribution and support of Tine Gretlund (Bachy
pp. 10-20.
Soletanche UK) during the preparation of this pa-
Institution of Civil Engineers, (2nd ed.) 2007. ICE specifica-
tion for piling and embedded retaining walls. London: per are also well appreciated.
Institution of Civil Engineers, the Crown and Thomas
Telford Publishing Limited.
Long, M, Brangan, C, Menkiti C, Looby M and Casey, P
2012. Retaining walls in Dublin boulder clay, Ireland.
Proceedings of the Institution of Civil Engineers, Ge-
otechnical Engineering, volume 165, issue GE4, pp. 247-
266.
Looby, M and Long, M 2007. Deep excavations in Dublin –
recent developments. Paper First Presented to a Meeting
of the Geotechnical Society of Ireland, Dublin, 11th De-
cember 2007.
Paikowsky, S.G and Tolosko, T.A, 1999. Extrapolation of
pile capacity from non-failed load tests – Report No.
FHWA-RD-99-170. Virginia: US Department of Trans-
portation, Federal Highway Administration.
Rose, A.V, Taylor, R.N and El Naggar, M.H 2013. Numeri-
cal modelling of perimeter pile groups in clay. Canadian
Geotechnical Journal, volume 50, issue 3, pp. 250-258.
Soil-Structure Interaction, Underground Structures and Retaining Walls 259
V.M. Ulitsky et al. (Eds.)
IOS Press, 2014
© 2014 The authors and IOS Press. All rights reserved.
doi:10.3233/978-1-61499-464-0-259
Abstract. In 2D modelling computer programs which calculate in plain strain conditions, it is considered that modelled section
contiunes infinitely with the same geometry through the axis perpandicular to the section. However, any variation in soil type,
surcharge loads, topography or supporting conditions in this parpendicular axis effects the behaviour of the stress distribution in
the soil. Therefore 3D modelling computer programs provide more realistic solution for the sections which contain variety of
above mentioned conditions in parpendicular axis. In this paper, calculations in 2D and 3D modelling studies for two different
shoring works are presented. Calculated plain strain ratio (PSR) values are compared with previously published studies.
Surface settlement behind the shoring wall is Calculated lateral diaphragm wall displace-
calculated as 102mm and 12 mm for sections ments and surface settlements in two and three
neighboring building and the street sides respec- dimensional model analysis and also for both of
tively. two cases are summarized in the Table 3.
Lateral displacements of the diaphragm walls Table 3. Lateral Displacements and Settlements
with 3D model are estimated as follows;
For the sides neighboring the buildings, lat- Case/ Exc. Depth Surc. Load Lateral Disp. Settlements
eral displacements are 57mm and 54mm for the Model m kPa mm mm
short and long sides respectively. The lateral 1. / 2D 11.35 15 64 68
displacement of the short side is smaller than the 1. / 3D 11.35 15 38/30* 34/30*
other because the building load effects to only 1. / 2D 8.30 15 33 34
the half of this side. 1. / 3D 8.30 15 27/18* 25/18*
For the sides neighbouring to the streets, lat- 2. / 2D 11.30 100 147 102
eral displacements are 33 mm and 37 mm for the 2. / 3D 11.30 100 54/57* 31/31*
short and long sides respectively. 2. / 2D 11.30 15 10 12
2. / 3D 11.30 15 37/33* 22/16*
* First value is for the long, second value is for the
short side of the excavation.
3. Plain strai
PSR values are calculated for the sections results of comprehensive monitoring data in both
both considering plaxis results and the method cases the reasons of the incompatibility of PSR
submitted by Finno et al. (2007) and summarized values could be evaluated.
in the table below; These studies showed that with 3D calcula-
Table 4. Calculated PSR values
tions a more realistic approach could be provided
for estimating displacements in problems that
Lat. necessitate 3-D modeling.
Exc. Surc. Lat. PSR, PSR, calc.
Disp.
Case Depth Load Disp. Plaxis with
m kPa mm-2D
mm-
Models formula (1)
5.
3D
1 11.35 15 64 38 0.593 0.854 Beadman, D., Cheng, A., 2002. 2002-06. Arup
International Document
1 8.30 15 33 27 0.818 0.990
Burland, J.B., Simpson, B. & St. John, H.D., 1979. Move-
2 11.30 100 147 57 0.388 0.748 ments around excavations in London Clay. #{ E
2 11.30 100 147 54 0.367 0.921 Q
*!| ;
/
'
!ZBritish
Geotechnical Society, London U.K. 13-30
Finno, R.J. & Blackburn, J.T. & Roboski, J.F., 2007. Three-
dimensional effects for supported excavations in clay,
4.
'
, 133 (1) 30-36
Fuentes, R., Devriendt, M., 2010.Ground Movements around
Numerical studies of two different deep ex- corners of excavations: Emprical Calculation Method
cavation projects are summarized above. Two
'
, 136 (10) 1414-1424
cases which have similar excavation depth but Lee, F.H., Yong, K.Y. & Liu, K.X., 1995. Tree-dimensional
analysis of excavation in soft clay. # ZZ ;
with different dimensions, surcharge loads, soil
!
`_Egyptian Geotechnical Socie-
types and supporting systems have been studied. ty, Cairo, Egypt
Consequences of the calculated displace- Lee, F.H., Yong, K.Y., Quan, K.C.N. & Chee, K.T. 1998.
ments and PSR calculations can be listed as; Effect of corners in strutted excavations: Field monitor-
x In both cases 3D model calculations re- ing and case histories.
'
, 124
(4) 339-349
sulted with less lateral displacements and settle- Liu, K.X. 1995. Tree-dimensional analysis of excavation in
ments. soft clay, MEng thesis, National University of Singapore,
x Lateral displacements and settlements Singapore
increased with the increase of excavation depth, Ou, C.Y. & Chiou, D.C., 1993. Three-dimensional finite
surcharge loads and the length of the excavation element analysis of deep excavations. # ZZ *
;
The Institution of Engineers,
side where displacements are calculated. Malaysia, Kuala Lumpur, Malaysia, 769-774
x As an exception, in the second case Ou, C.Y., Chiou, D.C. & Wu, T.S., 1996. Tree-dimensional
since building load does not affect to the whole finite element analysis of deep excavations,
of the long wall, smaller displacements are '
122(5), 337-345
calculated when compared to the short side. Ou, C.Y., Shiau, B.Y. & Wang, I.W., 2000, Three-
dimensional deformation behavior of the Taipei National
x PSR values which were calculated with Enterprise Center (TNEC) excavation case history.
Plaxis results are compared with the ones calcu- 37, 438-448
lated with considering method submitted by Simpson, B., 1992. Retaining Structure: displacement and
Finno et al. (2007). It is seen that PSR values design,
, London, U.K., 42(7), 541-576
calculated with Plaxis models are below the St. John, H.D. 1975. Field and theoretical studies of the
behavior of ground around deep excavations in London
values calculated with this method. Clay. Phd thesis in Cambridge University, Cambridge,
x In the first case excavation pit is not U.K.
rectangular and has variable excavation depths. Wong, L.W. & Patron, B.C., 1993. Settlements induced by
In the second there are heavy surcharge loads deep excavations in Taipei, # ZZ * ;
which are not a direct parameter of the calcula-
The Institution of Engineers, Malaysia,
Kuala Lumpur, Malaysia, 787-791
tion method submitted by Finno et al. (2007).
Also surcharge load is not symmetric and there
are barrettes which effects basal heave. All these
reasons can be related with the incompatibility of
calculated PSR values.
x In both cases since excavation phase
will initiate shortly, no monitoring data could be
provided for comparison, it is clear that with the
Soil-Structure Interaction, Underground Structures and Retaining Walls 265
V.M. Ulitsky et al. (Eds.)
IOS Press, 2014
© 2014 The authors and IOS Press. All rights reserved.
doi:10.3233/978-1-61499-464-0-265
Abstract. The design of flexible retaining structures is very frequently conducted using simplified soil-structure interaction
schemes such as the well known non-linear spring method. With this method, several distinguishing aspects of flexible wall
behaviour can be usually captured with reasonable accuracy with regards to construction stages and static conditions. Seismic
design, however, still represents a challenging aspect. In this paper some common issues of seismic analysis with the spring
method are addressed and a proposal to include seismic effects within a pseudo-static framework is presented. Some examples
analyzed with the proposed approach are finally discussed.
KA·V'v
In essence such geometrical parameters are
'
set equal to the average mobilized soil wedge
widths in active conditions (driving side) or
Figure 1. Constitutive law for a non-linear spring (cohesion-
passive conditions (resisting side).
less soil) The (secant) elastic modulus E' can be deter-
mined by means of usual correlations adopted for
For granular soils, yield limits linearly de- other geotechnical problems.
pend on the effective overburden stress V'v via For sandy soils, a reasonable estimate can ob-
the usual active and passive thrust coefficients tained by the one of the following equations:
KA and KP, usually determined according to E ' | 2 y 3 N SPT (MPa) , (3a)
classical Coulomb equations for KA and values
by Caquot et al. (1973) for KP, accounting for or
wall-soil friction G and dredge line slope.
The spring stiffness depends on the elastic E ' | 0.8 y 1 G 0 , (3b)
properties of the soil and on the excavation where G0 is the small strain shear modulus.
geometry. According to Becci & Nova (1985), In Becci & Nova (1987) a simplified cap log-
at each step, the spring stiffness K is computed ic is also included in order to differentiate load-
by the following equation: ing modulus Evc from unloading one EUR. Includ-
E' t ing such model in a general nonlinear finite
K , (1) element code, several modeling features can be
L easily considered, such as stage analysis by
where E' is the soil Young modulus, t is the adding and/or removing parts of the soil as well
spring spacing and L is a geometrical parameter as ground anchor installation and also the inter-
accounting for actual wall geometry. This value action with water. At each analysis stage, bal-
is different between up- (LM) and down-hill (LV) anced conditions are computed by means of
side: Newton-Raphson iterations. The straight-forward
model definition as well as the simplicity in soil
2 § S I' · parameters selection makes such kind of proce-
LM H' tan¨ ¸
3 ©4 2¹ dure a quite attractive option, at least in prelimi-
, (2) nary design phases, albeit some important design
LV
2
H'D tan§¨ S I' ·¸ aspects, such as settlements of nearby founda-
3 ©4 2¹ tions or global stability checks are not directly
available.
D is the excavation depth, H is the overall wall
height and H’ =min (2D,H) (Figure 2). 3. Pseudo-static seismic analysis
the practice. In essence a uniform seismic accel- kinds of retaining structures, such as cantilevered
eration with horizontal and vertical components or propped bulkheads. Since the performance
is statically applied to the wall and to the soil design method requires the calculation of a
mass interacting with the wall. The M-O method desired ultimate capacity, just rigid-plastic
and its extensions offer a closed form solution of methods, such as the Blum method or similar
the total soil thrust on the wall, assuming that the extensions (e.g. Conti & Viggiani, (2013),
soil mass behind yields during seismic event, Callisto (2014)), are essentially applicable. In
thus considering a wall subjected to relevant other words, reduced design accelerations ac
deformations. Following the approximation given by a performance design approach must be
proposed by Seed & Whitman (1970), the M-O applied only if seismic increments are deter-
increment to static active pressures can be mined by a method like Eq. 4, but not by Eq. 5 or
applied as a constant pressure distribution given any other elastic method. In this respect, it is
by clear that, for many kinds of walls, such as multi-
propped bulkheads, the performance (or capaci-
a
'p E ,M O | 0.375 J H , (4) ty) seismic analysis with pseudo-static approach
g requires particular care and further research.
Finally, within the performance design
where H is the wall height, J the average backfill
framework:
unit weight and a/g the normalized seismic
acceleration. Eq. 4 strictly holds for a dry granu- 'p E , Wood a
lar fill with a friction angle I'=35° and a soil- | 2.67 max , (6)
wall angle G=½ I'. 'p E , M O ac
Other methods, including the widely used Therefore active seismic thrusts may drop down
Wood method (Wood, 1973)) provide the seis- to a quite low percentage of elastic thrusts: for
mic increment of the soil thrust on a wall which example, taking ac/amax | 0.70, according to Eq.
is essentially rigidly restrained so as to prohibit 6, active thrust is just 26% of elastic thrust.
remarkable wall deformations. Such increment is
represented by a constant pressure distribution 4. Pseudo-static seismic analysis by the non-
given by: linear spring method
a
'p E ,Wood JH, (5) 4.1. General
g
Once the construction process has been com-
Comparing the formulations above, it can be
pleted, seismic conditions are usually modelled
realized that the seismic thrust increment for
according to one the two following alternatives:
rigid walls (Eq. 5) may exceed the M-O incre-
ment (Eq. 4) by a factor greater than 2.5. There-
A. the active and passive coefficients are modi-
fore it is important to select the appropriate
fied to account for seismic effects or;
approach, in the light of wall behaviour under
consideration.
B. just passive coefficients are modified where-
Beyond the method adopted in computing
as, on the driving side, an appropriate distri-
seismic pressures, a most important issue arose
bution of external pressures is applied,
even since early '70s: i.e. the selection of an
whose resultant corresponds with the ex-
appropriate design acceleration ac with respect to
pected seismic thrust increment: for example
the prescribed maximum seismic acceleration
Eqs. 4 or 5 may be used, depending on the
amax. Thanks to the fundamental works by Rich-
expected wall behaviour.
ard & Elms (1979), Whitman & Liao (1985) and
later works, the concept of design acceleration
Approach A is usually not recommended since it
ac<amax related to selected wall performance
is just applicable to very flexible walls, in which
(acceptable deformation) rather than to maxi-
active conditions are reached at the end of
mum site intensity was clearly established. Since
excavation process and in general may not
then, most design standards such as Eurocode
sufficiently reproduce expected seismic actions.
comply with this approach. Research efforts have
Approach B, on the contrary, allows the designer
been devoted in extending this method that was
to keep under his/her control the seismic actions
originally proposed for gravity walls, to other
268 B. Becci and M. Carni / Pseudo-Static Analysis of Flexible Retaining Structures
4.3. Discussion
According to this procedure, the computed
seismic pressures on the driving side fall be-
tween a minimum (active) and a maximum
B. Becci and M. Carni / Pseudo-Static Analysis of Flexible Retaining Structures 269
As most pseudo-static methods, this approach Conti, R. 2010. Modellazione fisica e numerica del
has a reasonable ability to capture internal forces comportamento di opere di sostegno flessibili in
condizioni sismiche. PhD thesis, Università di Roma
in soil, as well as in wall and props, but fails in Tor Vergata, Rome, Italy (in Italian).
computing physical deformations when relevant Conti R., Madabhushi G. S. P, Viggiani G. M. B. 2012. On
plastic deformations are progressively accumu- the behaviour of flexible retainingwalls under seismic
lated during seismic events: this circumstance is actions. Géotechnique, Vol. 62, No. 12, pp 1081–94.
particularly evident in case of cantilevered or Conti, R. & Viggiani, G. 2013. A new limit equilibrium
method for the pseudostatic design of embedded
singly propped walls and must be clearly taken cantilevered retaining walls, Soil Dynamics and
into account in final design. Earthquake Engineering, Vol.50, pp. 143–150.
However, on authors' best knowledge, this Itasca Consulting Group, 2011. Installation of a Triple-
limitation is also present in any kind of pseudo- Anchored Excavation Wall in Sand, FLAC version 7.0,
static analysis, even using more advanced nu- Example Applications Manual, section 17.
Lancellotta, R. 2007. Lower-bound approach for seismic
merical models. passive earth resistance, Géotechnique, Vol. 57, No. 3,
Based on the examples studied with this ap- pp. 319–321.
proach, the following concluding remarks can be Mononobe, N., and Matsuo, H. 1929. On the determination of
proposed: earth pressure during earthquakes. Proceedings of
World Engineering Conference, Vol 9.
Okabe, S. 1926. General theory of earth pressure. Journal,
- Multi-propped walls with deformable interme- Japanese Society of Civil Engineers, Vol. 12, No. 1.
diate supports such as ground anchors are nor- Pappin, J. W., Simpson, B., Felton, P. J., and Raison, C.
mally so flexible to reach active limit state at 1985. Numerical analysis of flexible retaining walls.
driving side at the end of excavation as well as Proc., Numerical Methods in Engineering Theory and
during seismic events. In this case both the Applications (NUMETA’85), Balkema, Rotterdam, pp
789–802.
proposed approach and the classical M-O pres- Psarropoulos P.N., Klonaris G., Gazetas G. 2005. Seismic
sure distribution can be used in practical de- earth pressures on rigid and flexible retaining walls,
signs. Soil Dynamics and Earthquake Engineering, vol. 25,
pp. 795–809
- For multi-propped walls with rigid supports, Richards, R. & Elms, D. 1979. Seismic behaviour of gravity
retaining walls, Journal of geotechnical Engineering
active seismic pressures may be inappropriate,
Division, ASCE, Vol. 105, No. GT4, pp. 449-464.
whereas fully elastic solution such as the one Schweiger, H. F. 2002. Results from numerical benchmark
by Wood (1973) may be too conservative. The exercises in geotechnics, Proc. 5th European Conf.
proposed algorithm can provide a more reason- Numerical Methods in Geotechnical Engineering,
able estimate of seismic conditions falling Presses Ponts et chaussees, Paris, 2002, pp 305–314.
between lower and upper bound solutions. Seed, H. B. & Whitman, R. V. 1970. Design of earth
retaining structures for dynamic loads. Proceedings of
ASCE Special Conference on Lateral Stresses, Ground
7. Acknowledgment Displacement and Earth Retaining Structures, Ithaca,
N.Y., pp. 103–147.
The authors are grateful to Prof. Giulia Viggiani Soubra, A. H. 2000. Static and seismic passive earth pressure
and Dr. Riccardo Conti of Università Tor Verga- coefficients on rigid retaining structures - Can. Ge-
otech. Journal. Vol. 37, pp. 463–478.
ta (Roma, Italy) for their kind assistance in Whitman, R. V. & Liao, S. 1985. Seismic design of retaining
providing detailed information on their research- walls, Miscellaneous Paper GL-85-1, U.S. Army Engi-
es, as well as for their valuable suggestions. neer Waterways Experiment Station, Vicksburg, Mis-
sissippi.
8. References Wood, J. H. 1973. Earthquake Induced Soil Pressures on
Structures, Doctoral Dissertation, EERL 73-05, Califor-
nia Institute of Technology, Pasadena, CA.
Becci, B. & Nova, R. 1987. Un metodo di calcolo automatico
per il progetto di paratie, Rivista Italiana di
Geotecnica,1,33–47 (in Italian).
Brinkgreve, R.B.J., Engin E., Engin H.K., 2010. Validation
of empirical formulas to derive model parameters for
sands. Proc. Numerical Methods in Geotechnical Engi-
neering NUMGE2010, Trondheim, Norway, June 2-4 ,
pp. 137–142
Callisto, L. 2014. Capacity design of embedded retaining
structures, Géotechnique, Vol. 64, No. 3, pp 204–214.
Caquot A., Kerisel J., Absi E. 1973. Tables de butée et de
poussée, Gautiers-Villars, Paris.
Soil-Structure Interaction, Underground Structures and Retaining Walls 273
V.M. Ulitsky et al. (Eds.)
IOS Press, 2014
© 2014 The authors and IOS Press. All rights reserved.
doi:10.3233/978-1-61499-464-0-273
Abstract. The current design practice of sheet pile walls are based on horizontal force and moment equilibrium using active and
passive earth pressures. The methods consider soils only in direct contact with the wall to determine the pressures, and do not
take into account the effect of existing ground surface profile and conditions below the wall. However, the initial ground and
sub-wall soil conditions affect the wall behavior and deformations. The effect of ground conditions and varying sub-wall soil
conditions on the behavior of sheet pile walls are presented in this paper.
Keywords. sheet pile walls, constitutive soil models, soil-structure interaction analyses, anchor forces
cut and fill operations for varying slope angles presented in the following. Parametric study was
and cut to fill ratios on wall behaviour were performed for varying soft soil thicknesses and
investigated using numerical methods (Bilgin, depths below the wall tip. The finite element
2009 and 2010). method was used to perform the numerical
The objective of this study was to analyze analyses by Plaxis.
and provide an overview of the effects of ground
surface and sub-wall soil conditions on the sheet 3.1. Geometry and material properties
pile wall behaviour. The conditions presented
include the presence of soft soil layers below the The floodwall height, H, selected for parametric
wall tip and the sloping ground surface where study was 5 m with a depth of wall penetration,
varying amounts of cut and fill is required during D, also equal to 5 m. The baseline case assumed
the construction. The finite element method was to have medium stiff clays at the site. Then a soft
used to perform numerical analyses and to clay layer with varying thickness and depth
conduct the study. below the wall tip was introduced for the para-
metric study cases. Thickness of the soft layer
ranged from zero (baseline case) to 25 m, and the
depth below the tip of sheet pile to the top of soft
2. NumericAl and constitutive soil models
layer ranged from 0 to 15 m. There were total of
The finite element analyses were performed by 29 cases analyzed for various combinations of
using Plaxis finite element code (Brinkgreve et soft layer thickness and depth to soft layer below
al., 2006). The finite element modeling com- the wall tip. A schematic of a typical wall section
prised two-dimensional plane strain analysis. analyzed along with the parameters selected is
Soil layers and sheet pile walls were modeled shown in Figure 1. Soil model parameters used
using 6-node triangular elements and 3-node in the analyses for medium stiff and soft clay
plate elements, respectively. A finer mesh was soils are given in Table 1. Steel sheet pile was
used around the wall because of the stress con- modeled using elastic plate elements, with axial
centration. and bending stiffnesses of EA = 2.68u107 kN/m
The Mohr-Coulomb constitutive model, a and EI = 40,000 kNm2/m, respectively.
linearly elastic and perfectly plastic model, was
used to model soil behaviour. The Mohr-
Coulomb model has been commonly used in the
finite element modelling of retaining walls. The
yield function, f, is defined by the Mohr-
Coulomb criterion using soil stresses (normal
stress, V, and shear stress, W) and strength param-
eters (cohesion, c, and friction angle, I) as
f W V tan I c (1)
The analyses conducted and the results are Figure 2. Final wall displacements (Case A: No soft soil
given in details by Bilgin (2009). The overview layer; Case B: Depth to soft soil is 5 m and soft layer
of the results is provided in the following. thickness is 15 m; and Case C: All soft soil below wall tip).
(unloading) and fill (loading) operations. The Table 2. Soil and wall properties
conventional design methods commonly used in Property Value
sheet pile wall design do not consider whether
Soil:
the walls are constructed by cutting, filling, or
Unit weight, J 17 kN/m3
both. Numerical analyses, on the other hand,
Friction angle, I 36°
allow modelling of the construction processes.
Dilatancy angle, \ 6°
Modulus of elasticity, E50 35,000 kPa
Poisson’s ratio, Q 0.28
Wall:
Axial stiffness, EA 4.36u106 kN/m
Bending stiffness, EI 9.86u104 kNm2/m
Soil/Wall interface strengtha, Rint 0.65
a
Rint = tanG/tanI (G=interface and I=soil friction angle)
and fill ratios, has an effect on the bending along the slope, and therefore the cut and fill
moments. Figure 7 shows that as the amount of ratios, affects the anchor forces significantly. As
cut decreases and fill amount increases, i.e. as the cut amount decreases and the fill amount
the wall location moves from the top of the slope increases, the anchor forces drop. There is
towards the toe of the slope, the maximum approximately 30% difference between only cut
moments increase. For the cases studied the and only fill cases. The reason of the decrease in
increase in bending moment was approximately anchor forces is related to the wall deformations.
14%. In addition, increased fill amounts result in Since the anchor is installed at the beginning of
moment reversal close to the tip of the wall. The the fill process, the movements are restricted at
results also show much higher moments, as much the anchor level. The stresses developed by the
as three times, at the anchor level in cut case backfill soil cause wall tip to move outwards and
compared to fill case (Figure 7). rotate the wall about the anchor location. This
rotation also causes wall top to move backwards
4.4. Anchor forces as it was discussed previously and shown in
Figure 6.
The anchor forces obtained are shown in Figure
8. The figure shows that the location of the wall
278 Ö. Bilgin / Soil-Structure Interaction for Sheet Pile Walls
6. References
Abstract. Construction of deep excavations in congested housing conditions in St. Petersburg without damaging adjacent
buildings is rather a pressing challenge. The paper presents a case of underground construction and associated monitoring of
deformations on existing buildings and structures. The obtained data on deformations are studied and collated with calculation
results.
rigid monolith reinforced concrete internal walls Cantilevered cofferdam protection in the giv-
and columns, rigid horizontal disks of monolith en case was unacceptable. A rigid system of
reinforced concrete intermediate floors and rigid struts was necessary to minimize cofferdam
interface joints between intermediate floors, movements. Anchorage solutions were unequiv-
walls and columns. The external walls of the ocally rejected in view of their high ductility,
building have a bearing function and are con- and, besides, placement of anchor roots under the
structed of hollow bricks. existing buildings could provoke additional
The new building extends its zone of influ- deformations.
ence onto 6 buildings (Fig. 1) Amongst those In this connection three basic options for the
) £( ## ì"" cofferdam protection were considered:
site. B £B is listed and protected by the - diaphragm wall with one level of struts
National Committee for the Preservation of shaped as sections of monolithic intermediate
Historic Monuments. All of the existing build- floors (“top-down”);
ings
@"! ) £$ £/
are of the 2nd - diaphragm wall with two levels of struts
technical condition category. B £$ shaped as sections of monolithic intermediate
the front elevation of ) £/ are of the 3rd floors (“top-down”);
technical condition category. - short sheet pile wall with one level of steel
struts and the preliminary construction of a jet-
grouted layer below the bottom of the pit.
To make the most relevant choice of coffer-
dam stabilization a series of calculations was
conducted in both analytical and numerical
settings. As was shown by the calculations, the
diaphragm wall option even with two levels of
struts would not have been able to hold addition-
al settlements of the adjacent territory within the
limits of permissible values. In spite of the fact
that in the latter case total additional defor-
mations did not exceed the permissible levels,
reaching 2.9 cm, there was no "reserve" neces-
The building to be Houses affected sary to accommodate possible additional settle-
constructed by construction ments during the works, which, should they have
appeared, would have pushed the settlements
30-restriction zone boundary
beyond the maximum tolerances for buildings of
Figure 1. The situational plan of the construction site. the 2nd technical condition category (3 cm).
The option safe for the surrounding buildings
The new project required construction of a pit and existing communications was that of short
extending down to ~8.7 m from the ground level sheet pile wall with one level of steel struts and
in immediate adjacency to existing buildings, the preliminary jet grouted layer below the
services and communications. bottom of the excavation.
Thus, considering the geotechnical situation As the Designer concluded, observance of the
it is possible to define three basic factors which time intervals for the cofferdam construction
can lead to development of deformations of both prescribed by them would ensure total settlement
the new and the existing buildings: of the adjoining territory occasioned by move-
- soft clay deposits in the subsoil of the new ments of the cofferdam walls and static loading
building (total thickness of the soft clay strata in of the subgrade by weight of the new building
the subsoil ~ 14 m); not to be in excess of 2.2 cm, which does not
- presence of existing buildings in the zone of exceed maximum permissible value for buildings
influence rendered by the new project; of the 2nd technical condition category (3.0 cm).
- necessity to construct the complex under- This would additionally provide the “reserve” of
ground space. 0.8 cm to accommodate possible additional
settlements as the works on site are carried out.
S.G. Bogov et al. / Monitoring Results for Construction of a Building with a Complex Underground Space 281
Depth, m
vented even greater deformations of the building.
Settlements after the end of piling and jet
grouting were registered at 20-22 mm.
Length 3 of the curve corresponds to the bulk
excavation and the actual underground construc-
tion. For the given period settlement rate reached
6,7mm/month, whereat the settlement value grew
by 40 mm, totalling 60 mm.
Ground level
Intermediate floor
level
Intermediate floor
level
Foundaiton raft
level
that the maximum motion of soil around building This idea, however, was never fully realized.
£( " +B mm, hence, it is possible This happened because drilling for piles began
to assume that the vertical deformations related before the columns had had time to set and gain
to bulk excavation and the construction of the strength properly. As a result overdrilling could
underground scopes had reached a similar value. not be completely ruled out. The deformations
Thus, the difference between the accumulated continued, although for some time their rates
settlement and that occasioned by static loading were held at a minimum.
by weight of the structures and the cofferdam It solicits itself as the most probable reason,
yield will comprise the technological settlement therefore, that both during piles construction and
and total 37 mm. for a short while after, the most probable reason
Conducting works without infringement of for the deformations was the overdrilling. If a
working regimes and method statements would CPT investigation had been carried out during
have caused a settlement of the residential house the construction of piles, a quantitative character-
with values well within the permissible 23 mm, istic of piling influence on natural soil composi-
which is in total agreement with the calculations, tion and on soil strength would have been
Moreover, this would have generated the “re- known.
serve" necessary to accommodate possible Construction experience on the given site, in
additional settlements during construction of the complex soil conditions and in congested city
project. environment has shown that observance of piling
regimes is always extremely important, as is
5. Conclusions monitoring; the latter allowing to obtain realistic
data on deformations and adjust working regimes
The most probable reason for development of to minimize adverse influence on existing build-
deformations of the building was infringement of ings.
piling regimes, which led to deterioration of
properties of the surrounding soils and provoked 6. References
additional settlements. When bored piles are
constructed in soft saturated soils, overdrilling is ÊË +,(BBB*+*(( ʹ«Ä®¼ ÏÈ®Á«»®¾¼ ·¾È«¿¯Å¯º¸Ã
a common enough mistake, and the correspond- ¹«®®«Ç ºÁ«·¶¯Ç ÍÍ : – ¯®¯¬¾º¬¾¹¸ º¤¯¸®«¿À®¸Ã
¤¸ º«Å¹¯¾¯Ç ¦¸¬¬¯Ä¬·¸Ä Áº«¶¯¯
+*((
ing settlements come as no surprise.
¸¤¸¹ ÊÒ Á«Æ¾«¶¯Ç ¬¾ºÈÄ®¸Ä ¾½®¸¿¸¤¯¯ Á¿Ç Ȭ¾º¸ÄÃ
One of the measures whereby to reduce nega- ¬¾¹« ƸÁÅ»®¸¤¸ ¸Ì»« ÅÁ«®¯Ç ¹ ¯¬¾¸º¯É¬·¸Ä É«¬¾¯
tive influence on the existing buildings could Ê«®·¾-˾ºÌȺ¤« ©½®¸¿¸¤¯É¬·«Ç ¸¬«Á·« ¯ ·¸®Ã
have been construction of casing protected bored ¾º¸¿À ¹¿¯É¯®¼ ʸ¹º»®®¼ ¤¸¾½®¸¿¸¤¯¯ ¹
¬¾º¸¯¾¿À¬¾¹ ¯ ¯½ ®«ÈÉ®¸-¾½®¯É¬·¸ ¬¸Æº¸¹¸ÂÁÃ
piles under water pressure. However the piling ®¯
ÊËÌÜ ÊËÌÒÊ
+*(,
contractor lacked the technical means to see it ¦·¸®¬¾ºÈ·¶¯Ç ÅÁ«®¯Ç ƸÁ »®¸¤¸·¹«º¾¯º®¼Ä Á¸» ¬¸
through. Therefore, another way of more sparing ¹¬¾º¸®®¼»¯ Ƹ»®¯Ç»¯ ¯ ƸÁÅ»®¼» ¤«º«Â¸»
construction of piles in proximity to the building QÆ«º·¯®¤¸»G Ò¸¾½®¯É¬·¸ ¸Ì¸¬®¸¹«®¯ ӯϺ 5$
– 10 – ËÐÍÔ(*/ – Ò² ²²² ËÐ Ò¸º·¸®¬¾ºÈ·¶¯Ç
was chosen, namely, preventive grouting of soft ÊËÌ
+*(*¤
soil along the pile axis. As a result soil columns ¿¯¶·¯Ä ;
ӫη¯® Ò Ò¸¾½®¯É¬·¸ ¬¸Æº¸¹¸ÂÃ
were formed in the ground around locations of Á®¯ º·¸®¬¾ºÈ·¶¯¯ ¤¸º¸Á¸¹ Q¸Ì¬¿Á¸¹«®¯
º«¬ÉÃ
the piles (depth ~20 m and diameter exceeding ¾¼
¹Á®¯ º«Ì¸¾
»¸®¯¾¸º¯®¤G – Ü ÐÅÁ«¾¿À¬¾¹¸
Ê;
(HHH-B+9 ¬Ü ¯¿
that of the piles), passing through the entire
thickness of soft strata. This measure is generally
subsumed under the idea of replacing soft soil
with a stronger material. This was followed by
drilling a borehole in the thus strengthened soil
to the design toe level in proper pile locations.
Thus, drilling in soft soil per se was never done,
which excluded overdrilling and made one
hopeful that the technological settlement during
construction of the project would be minimized.
284 Soil-Structure Interaction, Underground Structures and Retaining Walls
V.M. Ulitsky et al. (Eds.)
IOS Press, 2014
© 2014 The authors and IOS Press. All rights reserved.
doi:10.3233/978-1-61499-464-0-284
Abstract. The design of containment walls suffering seismic loads traditionally has been realized with methods based on
pseudoanalitic procedures such as Mononobe-Okabe's method, which it has led in certain occasions to insecure designs, that they
have produced the ruin of many containment walls suffering the action of an earthquake. A method is proposed in this paper for
the design of containment walls in different soils, suffering the action of an earthquake, based on Performance-Based Seismic
Design.
method, considering his action with agreement back line (i=0º), granular landfill (Mononobe-
to Richards-Elms[15] of 0.66 H over the foot of Okabe does not consider cohesion), friction
the wall, being H the height of it. angle of landfill-wall δ 0 and angle g of
Later it is calculated the maximum friction inclination of the surface of the area ß 0, the
force that is opposed to the sliding and the formulation used by the program calculates the
stabilizing moment, obtaining finally the safety active and passive pressures with the point of
coefficients for sliding and overturning, in static application defined by Prakash[12], Richards-
and also in seismic conditions. Elms[15] and Whitman-Liao[18].
STATIC ACTIVE FORCE ACCORDING TO
COULOMB´S THEORY
The static pressure according to Coulomb´s
theory on a wall of vertical back line has the
expression (figure 1):
Active force:
1
EA KA H 2
2
cos 2 1
KA 2
Figure 2. Dynamic pressure against a wall according
sen ( 1 ) · sen ( 1 ) Mononobe-Okabe´s theory
cos 1
cos · cos
K Ah K A cos K AV K A sen Active force:
1 1
E Ah K Ah H 2 E AV K AV H 2
2 2
Fe F R N ( P E AV ) tg 2 Fd E AAh
H
Me P · x E AV · B1 MV E AAh ·
3
1
E AE H 2 ( K AE KA)
2
Where kh y kv are the horizontal and vertical
seismic coefficient accelerations relative to g,
that is the horizontal and vertical pseudostatic
accelerations are: ah=kh g y av=kv g.
Figure 1. Static pressure against a wall according
Coulomb´s theory. The coefficient kv can have positive or nega-
tive sign, if the vertical acceleration due to the
The total active force according to earthquake is ascending or descending.
Coulomb´s theory forms an angle δ with the
normal to the wall (figure 1). The total active force PAE has his point of
application at a height on the base of the wall
DYNAMIC ACTIVE FORCE ACCORDING of:
MONONOBE-OKABE´S THEORY
With the conditions established in the
program MUROSIS for a wall with vertical
286 C. Jurado Cabañes / Seismic Soil-Interaction on Earth Retaining Structures
H 2H
EA E AAE
h 3 3
PAE
Once calculated KAE, it can be obtained:
K AEH K AE ·cos K AEV K AE ·sen
E AEH 1 · ·K ·H 2
2 AEH
E AEV 1 · ·K AEV ·H 2
2
Fe FR · N ( P E AV
·N E AEV ) ·tgg 2
Fd E AH E EAH
Me P · x E AV ·B1 E AEV ·B1 Figure 3. Seismic forces on the wall by
H 2H MONONOBE-OKABE
Mv E AH · E AEH ·
3 3 Substituting in these equations:
The safety coefficients with the action of
earthquake are: T N tg
g b
Fe Me
FSD 1.1 and FSV 1.1
Fd Mv
PERFORMANCE-BASED SEISMIC DESIGN OF
( PAE ) h PAE cos(( )
A WALL ( PAE )v PAEsen( )
In a containment wall with the action of an Where ι is the inclination of the back of the
earthquake, when the horizontal acceleration is wall with the vertical one and δ is the friction
lower than a critical or limit value acrit, the wall angle between soil and wall, obtaining the value
does not suffer movement. of the limit or critic acceleration that it puts in
But, when the acceleration of the soil equals situation of imminent sliding the wall:
the critical or limit acceleration, the wall is in
situation of imminent sliding. In this instant the W
N tg b a crit
cr PAAE cos ( )
total active forces due to the static and seismic g
pressures equal the stabilizing force due to the
friction between the wall and the soil at the N W PAE sen ( )
base. Then:
Considering only horizontal acceleration W
W PAE sen ( ) tg b acrit PAE
A cos ( )
and naming PAE the total static and seismic g
force calculated by Mononobe-Okabe´s method
and by (PAE)h and (PAE)v the horizontal and Wherefrom:
vertical components of this, the horizontal and
vertical balance of forces on the wall (figure 3.) PAE cos PAE sen tg b
acrit t
tg b g
gives: W
T = Fh + (PAE)h Being Фb the friction angle between the
N = W + (PAE)v base of the wall and the underlying soil. In this
equation PAE's value must be obtained with
Mononobe-Okabe's method.
The solution of this equation must be solved
by iterative form, since before to calculate PAE's
value by Mononobe-Okabe's method is needed
C. Jurado Cabañes / Seismic Soil-Interaction on Earth Retaining Structures 287
to know the value of acrit, and it is necessary to ing with the process, until acrit and
proceed by trial and error. amax fulfill this condition.
6. The value of dperm finally calculated
Richards y Elms[15] (1979) have calculated
must be less than the prearranged limit
the permanent displacement of the wall when
value and in contrary the dimensions of
amax>acrit obtaining the value:
the wall must be increased until the
2
Vmax 3
a max mentioned condition is fulfilled.
Dperm = 0,087
4 STUDIED CASES
a crit
[18]
Whitman y Liao (1983) correcting It has been realized the seismic calculation
Richards-Elms's simplifications have proposed of 16 cases of walls by means of the program
the following expression for the permanent named MUROSIS, using the next two types of
seismic displacement: walls:
1. Wall with 4,00 m of height
2. Wall with 6,00 m of height
acrit
2 9.4 In both cases the static pressures have been
37Vmax amax
d perm ·e calculated by Coulomb´s theory. The dynamic
a max pressures have been obtained by Mononobe-
Okabe and the permanent seismic possible
This value must not overcome the displacements by Whitman-Liao. All the exits
admissible limit for the permanent displacement have been verified manually and with the
of the wall. If this value is exceeded, it is bibliography to check the kindness of the
necessary to increase the dimensions of the results. In all cases the surface of the embank-
wall. ment is horizontal = 0º and the wall has the
back vertical α = 90º. The thicknesses of the
ORDER FOLLOWED BY THE PROGRAM wall in the base and the on the top has been
The order followed by the program (Jurado, selected to the specific situations of every case.
C.[6] is: The program verifies that the safety coeffi-
cients of sliding and overturning in seismic
1. First it calculates the weight W of the situation fulfill the following conditions:
wall.
2. A previous value of ah = kh·g, will be FSD 1.10 FSV 1.50
supposed, calculating for the value of The following situations have been studied
av = kv · g, the total active force PAE by for these two types of walls:
- Wall with 4.00 m /8.00 m of height
Mononobe-Okabe’s method. 1. Without earthquake, without cohesion
3. For this value of PAE it calculates the and without friction soil-wall
critical or limit acceleration. 2. Without earthquake, without cohesion
and with friction soil-wall
PAB co
cos
os l PAE sen l tg b 3. Without earthquake, with cohesion and
acrit tg
tg b g
W without friction soil-wall
4. Without earthquake, with cohesion and
4. If the value of acrit differs of ah = amax with friction soil-wall
less than one prearranged value 5. Without earthquake, without cohesion
(0,001), it calculates the permanent and without friction soil-wall
displacement by Whitman-Liao’s 6. With earthquake, without cohesion and
method according with the equation: without friction soil-wall
7. With earthquake, with cohesion and
acrit
without friction soil-wall
2 c t
9 ,4
37Vmax
37Vm amax 8. With earthquake, with cohesion and
d perm e m
Table 1. Exit of results for wall of height H = 4.00 m. seismic design of a wall. This is the line
followed nowadays for the “Performance Based
Wall of 4.00 m Seismic Design”.
kh REFERENCES
c1=c2 δ B1 B2 dperm
Wall (kv= kcrit [1] Brinkgreve, R. B. J, 2004. PLAXIS Reference
(kPa) (º) (m) (m) (mm)
1/2kh) Manual Delf.
[2] Coulomb, C.A., 1776, Essai sur une application
1 0 0 0 1.80 1.30 0.120g 0
des regles des maximis et minimis a quelques
2 0 0 15 1.60 1.20 0.161g 0 problèmes de statique relatives à l’architecture.
Mémoires de l’Académie Royale prés Divers Sa-
3 0 10 0 1.50 1.20 0.083g 0 vants, Vol. 7,
4 0 10 15 1.30 1.10 0.126g 0
[3] Fajfar P, Krawinker H., 1977, Seismic Design
Methodologies for the next generation Codes. Ed.
5 0.10g 0 0 1.85 1.50 0.1497g 0.19 Balkema.
[4] International Building Code (IBC 2006).
6 0.10g 0 15 1.80 1.50 0.202g 0 [5] Itasca Consulting Group. FLAC3D Manual.
Advance, Three-Dimensional Continuum Model-
7 0.10g 10 0 1.60 1.30 0.1135g 1.84
ling for Geotechnical Analysis for Rock, Soil and
8 0.10g 10 15 1.50 1.00 0.1435g 0.28
Structural Support.
[6] Jurado C., 2012, Problems of Soil-Structure
Interaction on Footings, and Walls of Contain-
ment. Influence of Lift-up and Sliding Phenome-
Table 2 Exit of results for wall of height H= 6.00 m. na. Ph. D. Thesis. Polytechnic University of Ma-
drid.
Wall of 6.00 m [7] Lysmer, John, Ostadan, Farhang, Tabataie,
Manour, Tajirian, Frederick and Vahdani, Shah-
kh riar, 2000, SASSI 2000- A system for Analysis of
c1=c2 δ B1 B2 dperm
Wall kcrit Soil-Structure Interaction.
(kv= [8] Mononobe, N. and Matsuo H., 1929, On the
(kPa) (º) (m) (m) (mm)
1/2kh) Determination of Earth Pressures during Earth-
quakes. Proceedings 2nd World Engineering
1 0 0 0 2.70 2.40 0.1439g 0
Conference, Vol. 9, paper nº 388, pages 177-185.
2 0 0 15 2.30 2.10 0.1712g 0 [9] Newmark, N.M., 1965, Effects of earthquakes on
dams and embankment. Geotechnique, v. 15, n2,
3 0 10 0 1.90 1.70 0.050g 0 pages 139-160.
[10]Prakash, S. and Basvanna, B. M., 1969, Earth
4 0 10 15 1.80 1.50 0.104g 0
Pressure Distribution Behind Retaining Walls
5 0.10g 0 0 2.80 2.50 0.163g 0
during Earthquakes. Proceedings Fourth World
Conference on Earthquake Engineering, Chile,
6 0.10g 0 15 2.70 2.50 0.212g 0 Vol. 3, pages 133-148.
[11]Prakash, S. and Rafnsson, E. A., 1995, On
7 0.10g 10 0 2.20 1.90 0.0985g 4.72
Seismic Design displacements of Rigid Retaining
8 0.10g 10 15 2.10 1.90 0.1575g 0 Walls. Proceedings 3rd International Conference
on Recent Advances on Geotechnical Earthquake
Engineering and Soil Dynamics, Vol.3, pag.
CONCLUSIONS 1183-1192, St. Luis, Missouri.
[12]Prakash, S. and Wu, Y.,1996, Displacement of
It has been analyzed 16 cases of walls of rigid walls during earthquakes. Eleventh World
4.00 m and 6.00 m of height for two values of Conference on Earthquake Engineering.
friction angle soil-wall and for two values of the [13]Priestley MJN, 2000. Performance Based
cohesion of landfill, one null and other different Seismic Design. New Zealand. National Society
to zero. for Earthquake Engineering.
The analysis of the permanent seismic [14]Rankine, W., 1857, On the Stability of loose
displacement leads to the conclusion that walls Earth Philosophical Transactions of the Royal
that fulfill to stability in seismic situation can Society of London, Vol. 147, 1857.
have unacceptable displacements, being [15]Richards, R. and Elms, D. G., 1979, Seismic
advisable the determination of this value in the Behavior of Gravity Retaining Walls. Journal of
C. Jurado Cabañes / Seismic Soil-Interaction on Earth Retaining Structures 289
Abstract. This paper deals with the assumptions and analyses used to stabilize the excavation of holy Alavi courtyard located in
Iraq. Numerical modeling of the problem is performed using PLAXIS software. Due to the connection of the main structure to
the retaining system which consist of multi strand cable anchored concrete piles, the interaction between the structure inside the
excavation and the retaining structure has been taken into account and its amounts has been investigated based on different
scenarios for retaining system connection to the structure. Accuracy of the analysis is evaluated through comparing with actual
displacements derived from monitoring of the excavation face during construction and useful conclusions are drawn.
Existing
Historic
Development Plan Building
Figure 1. (top) Layout plan of the excavation in project site, (bottom) Layout plan of the blocks in the courtyard of Hazrat
Fatima [1]
3. Numerical modeling
4 to 8 Sand 30 30 19
Some results included horizontal deformation
Weak counters for static and pseudo-static cases are
8 to 14 20 2000 20 presented in Figures 5 and 6.
Sandstone
Loose
14 to 17 saturated 30 5 19
4. Comparison with monitoring results
Sand
Figure 7 shows the monitoring results of one
Weak of the piles in block 10 which derived from
Below 17 20 2000 20
Sandstone surveying methods. Comparing Figures 7 and 5,
illustrates that results obtained from monitoring
H. Elahi et al. / Shoring-Structure Interaction in Stabilization of Excavation Adjacent to Historic Buildings 293
Stage 1
Stage 2
Stage 3
Stage 4
Stage 6
Date -5 -5 -5
4
Anchor Level
3
A B C
2
Figure 8. Scenarios considered for Shoring-Structure free
Interaction analyses flexible
1 rigid
Based on abovementioned scenarios, numeri-
cal analyses carried out and profile of shoring
0 50 100 150 200 250 300
pile displacements, shear and moment forces
after pseudo-static loading are drawn in figure 9. Force (kN/m)
Also anchors’ force and transferred horizontal
force to main structure from shoring structure are Figure 10. Anchors’ forces after pseudo-static loading for
presented in figure 10 and table 3, respectively. different interaction scenarios
As seen in above results it is demonstrated
that considering the interaction effects between
retaining system and main structure in excava-
tion area affect on design of both of them. In the
other hand connecting the shoring system to
main structure reduced displacements of piles via
H. Elahi et al. / Shoring-Structure Interaction in Stabilization of Excavation Adjacent to Historic Buildings 295
6. Conclusion
7. References
Abstract. This paper discusses a typical soil-structure interaction analysis related to a quay wall design. The quay wall consists
of a deck on piles combined with a closed frontwall. The quay wall has been verified for both operational and seismic load
conditions. The earthquake conditions have been verified both pseudo-statically and dynamically. The analysis has been per-
formed using PLAXIS FEM software. A new type of embedded pile element has been used to model the piles in a 2D model.
The verification is based on Eurocode and PIANC. This paper presents the general findings regarding the performance of inte-
grated retaining wall and piled deck structure behavior and the comparison between the load conditions.
The quay wall is described as an integrated re- The verification is based on the Eurocode and
taining wall and piled deck structure. It consists PIANC. Eurocode 7 (EN 1997-1) design ap-
of a deck on piles with a front wall and an anchor proach DA1 is selected for operational and seis-
wall. The area behind the front wall and below mic load cases. Only DA1-2 is presented in this
the deck is completely backfilled. The combina- paper. DA1-2 is in general more decisive than
tion of an anchored retaining structure connected DA1-1 for quay wall structures.
to a deck supported by piles implies a complex The quay walls are assigned to Eurocode
soil-structure interaction mechanism. As a result importance class 3, which is equivalent to PI-
of this, the retaining structure is analyzed with ANC performance grade A. For performance
Plaxis 2D finite element software and not by grade A the following performance requirements
conventional Winkler analysis. apply: For Level 1 seismic loading the structure
Seismic conditions are calculated with pseu- should remain serviceable. The perfor-
do-static analysis and dynamic analysis. Dynam- mance/damage criteria categorized by PIANC as
ic analysis are performed in order to be able to Degree I are adopted. For Level 2 seismic load-
assess the structure performance more accurately ing the structures should be repair able. This
at higher range seismic loading. implies that irreparable damage (like e.g. below
ground plastic hinging of piles) is not allowed.
H.J. Lengkeek and F. Besseling / Verification of Deck on Piles Quay Wall Structures 297
1.3. Soil conditions and geotechnical parame- piles 1420 mm with secondary intermediate
ters sheet piling.
The anchorage is provided by tie-rods and
The site investigations consists geophysical anchor wall consisting of a row of tubular piles.
profiling and approximately 30 m deep bore- Behind the retaining wall foundation piles are
holes at a distance of typically 100 m along the installed in the backfill for foundation of con-
quays and breakwaters. No SPT’s or CPT’s crete decks and crane rails. The first row of
have been performed. The laboratory investiga- piles is placed close behind the front wall.
tion consists of Index tests, Atterberg limits, These piles will act as a shielding row for the
Oedometer tests, Direct Shear tests and UU front wall as it reduces the active pressures. The
tests. complex soil-structure interaction requires
The parameters as provided with the labora- either 3-D modeling or 2-D with embedded
tory tests tend to have relative high cohesion piles (Sluis et al., 2013).
and low friction angles for clay layers. This is Retaining heights measure about 20 and
likely related to the laboratory tests and meth- 22.5 m for section I and II respectively. The
ods. These parameters can be classified as deck width for section I is about 33 m, with 5
intermediate effective stress - total stress pa- rows of piles. Section II structure deck width is
rameters, or intermediate drained – undrained about 19 m, with 3 rows of piles. Figure 1 and 2
parameters. For the verification analysis the present cross sections of the quay walls.
parameters have been adjusted based on engi-
neering judgment to get effective stress parame-
ters for an effective stress analysis.
The main parameters of the geotechnical
units are presented in table 1. The two soil
profiles are presented in table 2 relative to
seabed level.
Table 1. Main parameters of geotechnical units
sat, ’, c’, Su, E,
Unit
kN/m3 deg kPa kPa MPa
2a, Sandy clay 19.5 26 4 45 25
Figure 1. Cross section quay wall I
2b, Clay 19.5 25 6 48 11
2c, Sand / silty 40
18.7 27 0 -
sand
4a, Clay 18.7 26.5 25 173 41
Table 2. Soil profile I and II
Soil profile I Soil profile II
Top of layer Top of layer
Unit Unit
depth depth
2a 0 2a 0
2c 1.3 4a 5.6
2a, 6.1
2b 10.9 Figure 2. Cross section quay wall II
2c 16.5
4a 26.3 The design of front wall with shielding pile
row is typical and more frequently used in
Russia. It is a practical solution in case of
1.4. Quay wall geometry
limited embedment or limited bending moment
The quay wall structures consist of an an- capacity of the front wall.
chored front wall with backfill. The front wall is The other typical aspect of the design is the
a combined wall, consisting of primary tubular anchor wall. The location of the anchor wall is
close to the deck with tie rod connection above
298 H.J. Lengkeek and F. Besseling / Verification of Deck on Piles Quay Wall Structures
the water line. This has practical benefits during Table 4. Load combinations combination factors
construction. However the effectiveness is
Load \ LC LC1 LC2 LC3 LC4 LC5
influenced too. The passive wedge overlaps the
foundation piles. The passive capacity and Extr. Extr. Extr. Seismic Seismic
stiffness of the anchor wall will be reduced. Vert. Low Crane L1 L2
water
1.5. Construction sequence Surcharge deck 1.0 0.7 - 0.3 0.3
Surcharge 1.0 0.7 0.7 0.3 0.3
The construction sequence can be summarized behind deck
as follows:
- Access dam construction at anchor wall Bollard 0.7 0.7 0.7 - -
- Installation of piles and walls Crane - - 1.0 - -
- Tie rod installation Water level Oper. Extr. Oper. MSL MSL
- Backfill low Low Low
- In situ construction of platform Seismic - - - 1.0 1.0
- Dredging to design depth. Hydrodynamic - - - 0.7 0.7
water pressures
1.6. Loads
A basic set of dominant load cases has been 1.7. Seismic conditions
taken into account for the verification. The
The earthquake conditions have been verified
surcharge load is divided in zones according to
both pseudo static and dynamically. The origi-
SNiP (SNiP2.06.01-86). The bollard load is 80
nal seismic loads are defined according to
tonnes and has been distributed over the plat-
Russian Standards (SNiP II-7-81) in MSK-64
form width and bollard spacing. The crane load
intensity zones point system. Additionally a
has been distributed over the crane length and is
project specific PSHA has been used. Table 5
equal for both rails. The maximum horizontal
presents the results of the PSHA.
load due to storm winds is 15% of the vertical
load. Table 3 presents the loads per section. Table 5. PSHA results
Load combinations including the specified load
cases follow Eurocode combination factors (EN Scenario Unity abedrock,ref asurface,ref abedrock,d asurface,d
1997-1) and are summarized in table 4. L1 (MRP
g 0.015 0.084 0.018 0.101
The differential water level is set to 0.5 m, 95 yr)
with specific levels for each load case. For L2 (MRP
g 0.115 0.278 0.138 0.333
seismic conditions the equivalent water level 475 yr)
has been calculated using the Westergaard
method, to account for hydrodynamic effects The structures are assigned to importance
according to EN1998-5 (EN1998-5). class 3, and consequently the reference seismic
Table 3. Loads per quay wall section
conditions are multiplied by an importance
factor 1.2 according to EN1998-1 in order to
Load case Unit Quay 1 Quay 2 obtain design conditions (EN1998-1). Table 6
20 / 40 / 60 7.5 / 15 / 20 / presents the seismic loads for pseudo-static
Surcharge kPa calculation.
/ 100 * 20 *
Bollard pull Table 6. Seismic loads for pseudo-static calculation
kN/m1 27 54
horizontal
Lateral
Crane load Design Lateral
kN/m1 416 655 Height seismic
vertical surface seismic
Scenario correction coeff. for
Crane load acceleration coeff.
kN/m1 63 104 factor calculation
horizontal (wind) asurface,d kh
kh
* surcharge load variation over pre-defined apron L1 (MRP
and non-apron areas (increasing distance from 0.101 0.101 0.81 0.082
95 yr)
berthing line) L2 (MRP
0.333 0.231 0.81 0.187
475 yr)
H.J. Lengkeek and F. Besseling / Verification of Deck on Piles Quay Wall Structures 299
The selected representative seismic records Table 7. Summary of ULS results quay wall I
for the dynamic analysis are presented by means
of their Fourier spectra in Figure 3. Anchor
Section_I Front wall Piles Tierod wall
500 yr. MEd
0.6 Load MEd,field MEd,head FEd [kNm/m
Spectral acceleration (g)
0.5
case [kNm/m] [kNm] [kN] ]
Coalinga
Dredging 42% 26% 99% 89%
0.4
Northridge
LC1 Op. 100% 57% 100% 95%
0.3
N. Palm Springs LC2 Op. 91% 60% 93% 100%
0.2
LC3 Op. 95% 100% 68% 72%
0.1
PS L1 >100% >100% >100% >100%
0.0
0.01 0.10 1.00 10.00 100.00 Dyn. L2 168% 142% 104% 153%
T (s)
- Global stability is critical for section II. The complex interaction between soil and
- Bending moment capacities of pile structure has been successfully modeled with
rows supporting the deck (except from the Plaxis embedded pile module. The ad-
the shielding pile row) are critical. vantage of the embedded pile is that a typical
3D configuration can still be effectively mod-
Russian standards typically require a lower eled in 2D.
reliability (SNiP CN&R 33-01-2003, SNiP CN The review was performed according to
23-13330.2011). Sensitivity analysis were Eurocode. In order to testify if the quay wall
performed in order to verify if above mentioned performance according to the Russian approach
main aspects would also be critical following to structural reliability would be sufficient, a
Russian partial factors on loads, resistance and sensitivity analysis was done. It was concluded
overall effects. In this sensitivity analysis that results from this sensitivity analysis were in
operational load cases and seismic load case L1 general slightly more favorable but with respect
are included. Dynamic calculations for seismic to quay wall global stability the results coincide.
load case L2 are excluded. The main conclu-
sions from comparison of the initial results with 5. REFERENCES
the sensitivity analysis results are:
- With respect to global stability the EN 1997-1, Eurocode 7: Geotechnical design – Part
analysis results coincide. 1: General rules, ICS91.080.01;93..020, March
- The conclusions regarding the too low 2005
anchorage effectiveness also holds EN1998-1, Eurocode 8: Design of structures for
based on sensitivity analysis results. earthquake resistance – Part 1: General rules,
- On average about 30% lower structure seismic actions and rules for buildings
internal forces (wall and pile bending EN1998-5, Eurocode 8: Design of structures for
moments) are obtained from the sensi- earthquake resistance – Part 5: Foundations, re-
tivity analysis calculations. taining structures and geotechnical aspects
Gazetas, G., Dakoulas, P. & Dennehy, K. 1990.
4. CONCLUSIONS Empirical seismic design method for waterfront
anchored sheetpile walls, Proc. ASCE Specialty
In the presented study an integrated retaining Conference on Design and Performance of Earth
wall and piled deck structure design has been Retaining Structures, ASCE Geotechnical Special
analyzed for operational and seismic load Publication No. 25, pp. 232-250.
combinations. Operational load combinations PIANC, Seismic design guidelines for port structures
are dominated by high surcharge loads of bulk / Permanent International Association of Naviga-
terminal areas and high horizontal crane loads tion Congresses, Working group nr. 34, 2001,
from wind in storm conditions. Review calcula- Leiden: Balkema
tions for seismic loads were performed pseudo- Sluis et al., 2013, Validation and Application of the
statically for L1 and by dynamic time-history Embedded Pile Row Feature in Plaxis 2D, Plaxis
analysis for L2. Limited application of pseudo- Bulletin, Autumn issue 2013 pp. 10-13.
static methods for higher range seismic loads is SNiP II-7-81: Construction in Seismic Regions.
often obtained and the present study shows that Building Codes and Rules, Russia, English trans-
this holds for this typical quay wall as well. lation, 93 pp.
The main critical aspect of the reviewed SNiP 2.06.01-86: Design loads on berthing facilities
quay wall designs was the low effectiveness of SNiP CN&R 33-01-2003: Construction Standards
the anchorage, which is related to the relatively and Rules of the Russian Federation, Hydraulic
short tie-rod length which causes a strong Engineering Constructions - Basic principles of
interaction between deck displacement and designing, 01-01-2004
anchor wall passive mobilization. The perfor- SNiP CN 23-13330.2011: Construction Standards
mance of the quay walls for seismic loads and and Rules of the Russian Federation, Foundation
operational load combinations with significant of hydraulic structures, 2011
lateral loads can easily be improved by increas-
ing the anchor wall distance, as is already
demonstrated by the difference between section
I and II.
302 Soil-Structure Interaction, Underground Structures and Retaining Walls
V.M. Ulitsky et al. (Eds.)
IOS Press, 2014
© 2014 The authors and IOS Press. All rights reserved.
doi:10.3233/978-1-61499-464-0-302
Abstract. This work presents a heady constructive method used for the retaining structures of the Underground Parking Clinics.
The project of the parking includes 315m long for 15,5m wide and 4 undergrounds which work was concluded in the end of
1998. The retaining structures of the excavation was executed through screw piles of 80cm diameter, with 23m of length, spaced
at every 2,50 m. The slabs of structure of the parking had been set up with precast beams of prestressed-concrete and the bottom
slab with cast in-in-place reinforced concrete. The excavation was accomplished in two phases, in the first one the walls worked
as cantilever; and the second one, after the installation of the upper slab working as struts, it was made the final excavation.
1. Introduction
structures it would be necessary the execution of layer of 2 centimetres of shotcrete among the
a “chalice”, whose constructive process was piles was just applied.
shown very difficult. Temporary berms had been foreseen during
the phase of final excavation, whose removal
would depend on the instrumentation results that
were constituted in measures of tension of the
steel of the reinforcement of the pile and hori-
zontal displacements of the piles. The obtained
results checked the viability of execution of the
excavation without the use of the berms.
After the excavation and execution of the bot-
tom slab, the installation of the intermediary
levels slabs was proceeded and the piles re-
strained.
The screw piles are constituted, generally
speaking, an economic solution for foundations
and as is not common to use them, in Brazil, at
that time, as retaining wall due to the construc-
Figure 2 – Aerial view of the parking after conclusion tive difficulty of reinforcing this type of pile,
tests were executed previous to the beginning of
So, it was decided to modify the project, the the work to confirm the executive viability of the
structure should be executed without intermedi- piles with 19 meters of reinforcement.
ary columns and precast beams of prestressed-
concrete that won the empty space of 15,5 meters 2. Geological-geotechnical profile
should be used. There were studied new support
alternatives, being adopted screw piles of 80 The geological-geotechnical profile of the
centimetres of diameter, that were used so much area, shown in the figure 3 presents a layer of
as retaining structures of the excavation as well approximately 1 to 1,5 meters of embankment of
as for foundation of the structure of the parking. silty clay, followed by a layer of 10 to 12 meters
The piles were arranged at every 2,5 meters in of silty clay with brown and red fine sand (po-
the two faces of the excavation, and they had the rous) that presents index SPT variable among 10
length increased for 23 meters and the steel to 20. Below this layer it comes a layer of 6 to 10
reinforcement introduced in the piles presented meters of silty clay with a lot of variegated fine
19 meters. The structure of the parking was sand, SPT among 8 to 16, over a layer of silty-
conceived with the upper slab and the intermedi- clayed sand of varied granulation. The level of
ary levels beam, in precast beams of prestressed- the water was found at 17 meters deep, therefore
concrete and the bottom slab with cast-in-place below the final excavation.
reinforced concrete. The bottom slab works as
complement of the foundation of the structure,
because the piles just have load capacity compat-
ible in relation to temporary phase, that is to say,
without being considered the overloads in the
floors of the underground.
The work began with the remotion of the in-
terferences and execution of the piles. The
placement of the upper precast beams was
accomplished with the aid of a portico of 30 tons
of capacity. The constructive method was just
constituted basically in two excavation phases.
The first excavation phase presented 5 meters of
depth and the walls worked as cantilever; after
the installation of the upper slab that worked as
struts, it was executed the final excavation. Due
to the excellent characteristics of the soil a fine Figure 3 - Geological-geotechnical profile
304 C.E.M. Maffei et al. / A Bold Constructive Method for the Retaining Structures
4. Constructive method
Figure 4 - Diagram of minimum pressure The outlines of the constructive method are
presented bellow. After the remotion of the
To determine the efforts and the displace- interferences the screw piles were executed in
ments due to the excavation the evolutive and the whole contour of the parking and metallic
no-evolutive models were used. piles were installed with pre-hole in the extremi-
In function of the geological-geotechnical ty of the parking where the ramp was located that
profile, a compartment was accomplished; in served as service road during the works and exit
which were determined three types of piles: of the parking after the conclusion of the work
Type “light”–axes 0 to 17–trench constituted (Figure 5).
of porous silty clay red, level of water of 3 In other to turn it possible the excavation till
meters above the end level of the embedded the installation level of the upper slab, with the
length, or below it. walls presenting 5 meters of cantilever, it was
C.E.M. Maffei et al. / A Bold Constructive Method for the Retaining Structures 305
Figure 11b
Figure 12a
Figure 12e
Figure 12b
Figure 12c
Abstract. Soil nailing is a world-wide frequently used technique for temporary and permanent slope stabili- zation applications and
excavations with vertical or nearly vertical cuts. A case study is analyzed applying both limit equilibrium (LEM) and finite
element method (FEM), including comparison with internationally used codes and standards. In this paper different approaches are
used in FEM for simulation of soil nails by different structural type in order to study its influence on results, e.g. (a) only axial
stiffness in nails is taken into account or (b) bending and shear stiffness beside the axial stiffness.
FRQWULEXWLRQ RI EHQGLQJ VWLIIQHVV LV VWLOO TXHV + FRQVLGHULQJ WKH ORZHU YDOXH
WLRQDEOH 6KORVVHU -HZHOO 3HGOH\ DSSURSULDWH IRU QDLOV FRQVWUXFWHG LQ ZHDWKHUHG
%HQGLQJ VWLIIQHVV FRQWULEX URFNVVWLIIVRLOVDQGWKHKLJKHUYDOXH IRUQDLOV
WLRQ FRPHV LQWR SOD\ ZKHQ WKH UHLQIRUFHG PDVV LQ FOD\H\ VRLOV 7KH KRUL]RQWDO VXUIDFH GLV
LV QHDU IDLOXUH DQG LWV PDMRU VLJQLILFDQFH LV LQ SODFHPHQWEHKLQGWKHZDOOįJHQHUDOO\UDQJHV
VRIWHU VRLOV &KHDQJ HW DO $FFRUGLQJ WR IURP + 7KHVH GLVSODFHPHQWV
)+:$ WKH EHQGLQJ DQG VKHDU VWLIIQHVV OHDGV WR WKH GLVWRUWLRQ RI WKH VRLO QDLO PDVV
DUH FRQVHUYDWLYHO\ RYHUORRNHG GXH WR WKHLU UHVXOWLQJLQVPDOOEHQGLQJPRPHQWVWKDWVKRXOG
FRQWULEXWLRQ RI QRW PRUH WKDQ RQ WKH EH GHYHORSHG QHDU WKH IDLOXUH VXUIDFH 7KH
RYHUDOOVWDELOLW\RIWKHZDOO OHQJWK RYHU ZKLFK GHIRUPDWLRQV DUH GDPSHQHG
ȜEHKLQGWKHVRLOQDLOZDOOLVGHILQHGDV
&2'(6$1'67$1'$5'6
λ = + ⋅ ( − WJη ) ⋅ N
$V LW ZDV VWDWHG SUHYLRXVO\ WKH VWDQGDUGV DQG ZKHUH N LV D FRHIILFLHQW LQ IXQFWLRQ RI WKH VRLO
FRGHVIRUWKHDQDO\VLVRIVRLO QDLO ZDOOVXVXDOO\ W\SH DFFRUGLQJ WR WKH 7DEOH + LV WKH VORSH
DSSOLHG DUH &ORXWHUUH ;3 3 KHLJKW DQG Ș LV WKH LQFOLQDWLRQ RI WKH IDFLQJ
)HGHUDO +LJKZD\ $GPLQLVWUDWLRQ UHJDUGLQJWKHYHUWLFDOD[LV
)+:$&,5,$&(1
%6HWF 7DEOH'LVSODFHPHQWSDUDPHWHUV&ORXWHUUH
$FFRUGLQJ WR )+:$ WKH FRQYHQWLRQDO GH
6RLOW\SH
VLJQ LQFOXGHV WKH SUHOLPLQDU\ GHVLJQ XVLQJ
VLPSOLILHG FKDUWV DQG ILQDO GHVLJQ 7KH ILQDO :HDWKHUHGURFNV
3DUDPHWHU 6DQG\ &OD\H\
6WLIIVRLOV
GHVLJQ RI VRLO QDLO ZDOOV LQFOXGHV DQDO\VLV RI
H[WHUQDO IDLOXUH PRGHV JOREDO VWDELOLW\ VOLGLQJ įK+§įY+
VWDELOLW\ EHDULQJ FDSDFLW\ LQWHUQDO IDLOXUH N
PRGHVSXOORXWIDLOXUHDQGQDLOWHQVLOHVWUHQJWK
IDLOXUH DQG GHVLJQ RI IDFLQJ DQG LWV IDLOXUH 'XULQJFRQVWUXFWLRQWKHDOORZDEOHKRUL]RQWDO
PRGHVIOH[XUHDQGSXQFKLQJVKHDUIDLOXUH GHIOHFWLRQV DUH XS WR + DFFRUGLQJ WR
7KH UHTXLUHG IDFWRUV RI VDIHW\ DSSOLFDEOH )+:$)XUWKHUPRUHDSRVWFRQVWUXFWLRQ
ZKHQ DOORZDEOH VWUHVV GHVLJQ LV DSSOLHG ZKHQ GHIRUPDWLRQ WHQG WR RFFXU GHSHQGLQJ RQ WKH
ORDGV DUH XQIDFWRUHG IRU WHPSRUDU\ VRLO QDLO JURXQG W\SH OHDGLQJ WR WKH LQFUHDVH RI
ZDOOVDUHJLYHQLQ7DEOH DGGLWLRQDOWHQVLRQLQQDLOV)+:$
7DEOH0LQLPXP)6IRUWHPSRUDU\VWUXFWXUHV %\ GHILQLQJ IROORZLQJ IRXU SDUDPHWHUV
&,5,$ & WKH FRPSDULVRQ EHWZHHQ
)DLOXUH 0LQLPXPUHFRPPHQGHG)6 GHVLJQV RI GLIIHUHQW SURMHFWV LV DOORZHG DQG
PRGH 5HVLVWLQJFRPSRQHQW 6WDWLF HYDOXDWLRQ RI D JOREDO SHUIRUPDQFH RI VRLO QDLO
([WHUQDO *OREDOVWDELOLW\ ZDOOV
VWDELOLW\ 6OLGLQJ / =/ +
U PD[
,QWHUQDO 3XOORXWUHVLVWDQFH
VWDELOLW\ 1DLOEDUWHQVLOHVWUHQJWK % ='⋅/ V
)DFLQJIOH[XUH U U K
)DFLQJ
VWUHQJWK )DFLQJSXQFKLQJIDLOXUH
6 U = ' V
K
2QHRIWKHPRVWLPSRUWDQWSDUDPHWHUIRUWKH
HYDOXDWLRQRIVRLO QDLO ZDOOSHUIRUPDQFHGXULQJ
3 =δ +
LWV GHVLJQ OLIH LV WKH GHIRUPDWLRQ RI WKH VDPH U K
ZDOODQGLWVLQIOXHQFHRQWKHDGMDFHQWVWUXFWXUHV ZKHUH /PD[ ± WKH PD[LPXP QDLO OHQJWK /U ±
7KHUH DUH HPSLULFDO FRUUHODWLRQV IRU LWV HVWLPD OHQJWK UDWLRQ %U ± ERQG UDWLR VK ± KRUL]RQWDO
WLRQPRVWO\EDVHGRQ&ORXWHUUH$FFRUG VSDFLQJ ' ± GULOO KROH GLDPHWHU 6U ± VWUHQJWK
LQJ WR &ORXWHUUH EDVHG RQ H[WHQVLYH UDWLR 3U ± SHUIRUPDQFH UDWLR įK ± KRUL]RQWDO
H[SHULPHQWDOVWXG\WKHPDJQLWXGHRIKRUL]RQWDO GLVSODFHPHQW
DQG YHUWLFDO GLVSODFHPHQW DW WKH WRS RI WKH 2QH RI WKH DSSOLFDWLRQV RI DQDO\VLV RI SHU
IDFLQJLVSUDFWLFDOO\HTXDOEHLQJLQWKHUDQJHRI IRUPDQFH RI VRLO QDLOHG ZDOOV EDVHG RQ FDVH
312 S. Melentijevic and J.L. Fernández / Comparison of Finite Element and Limit Equilibrium Methods
RUGHU WR FRPSDUH LW ZLWK UHVXOWV RI /(0 6RLO QDLOV DQG IDFLQJV DUH PRGHOHG DV HODVWLF
DQDO\VLV DQG WR VWXG\ WKH LQIOXHQFH RI GLIIHUHQW PDWHULDOV
SDUDPHWHUV RQ UHVXOWV )RU WKDW SXUSRVH WKH 7KH FRPPRQ SUDFWLFH LV WR XVH JHRJULG
FRPPHUFLDOFRGH3OD[LVYHUVLRQZDVXVHG HOHPHQWV IRU VLPXODWLQJ VRLO QDLOV ZKLFK
7KHSODQHVWUDLQWZRGLPHQVLRQDOPRGHOLVXVHG FRQVLGHUV RQO\ WKH D[LDO VWLIIQHVV RI QDLOV %\
7KH QRGH WULDQJXODU HOHPHQWV ZHUH XVHG WKDWZD\WKHEHQGLQJDQGVKHDULQJUHVLVWDQFHRI
WR JHQHUDWH WKH ILQLWH HOHPHQW PHVK 7KH VRLOQDLOVLVFRQVHUYDWLYHO\LJQRUHG,QWKLVVWXG\
ERXQGDU\ FRQGLWLRQV ZHUH VHW DW P DZD\ WKH XVH RI ERWK JHRJULG DQG SODWH HOHPHQWV LV
IURP WKH VRLO QDLO ZDOO DQG P EHORZ WKH VWXGLHG LQ RUGHU WR TXDQWLI\ WKH LQIOXHQFH RI
VRLO QDLO ZDOO ZKLFK LV ODUJH HQRXJK WR DYRLG EHQGLQJDQGVKHDUUHVLVWDQFHRQUHVXOWV
ERXQGDU\HIIHFWVLH PLQLPL]HLWVLQIOXHQFHRQ 0RUHRYHU WKH VLPXODWLRQ RI VRLO QDLO ZDOOV
QXPHULFDO VLPXODWLRQ UHVXOWV 7KH JOREDO XVXDOO\ GRHV QRW WDNH LQWR DFFRXQW ZDOO IDFLQJ
FRDUVHQHVV PHVK ZDV VHW WR PHGLXP W\SH 7R EHLQJ E\ WKDW ZD\ RQ WKH VDIH VLGH LQ WKH
HQVXUHWKHDFFXUDF\LQWKHDQDO\VLVWKHPHVKLV DQDO\VLV,QWKLVVWXG\WKHLQIOXHQFHRIWKHIDFLQJ
IXUWKHUUHILQHGLQWKHYLFLQLW\RIVRLOQDLOZDOODW HOHPHQW RQ WKH JOREDO VWDELOLW\ LV VWXGLHG 7KH
P ODWHUDOO\ DQG P EHORZ WKH VRLO ZDOO IDFLQJ LV PRGHOHG DV D SODWH VWUXFWXUDO
QDLOZDOO HOHPHQWLQRUGHUWRVLPXODWHWKHULJLGQDLOIDFLQJ
)LJXUHVKRZVWKHPHVKDQGJHRPHWU\XVHG FRQQHFWLRQ
LQ)(0 IRUWKHDQDO\VLVRIWKHVRLOQDLO ZDOOLQ 7KHLQSXWSDUDPHWHUVUHTXLUHGIRUPRGHOLQJ
WKHFDVHRIJHRJULGHOHPHQWVXVHGIRUPRGHOLQJ SODWH HOHPHQWV DUH EHQGLQJ (, DQG D[LDO ($
RI QDLOV 7KH VDPH PHVK W\SH DQG JHRPHWU\ LV VWLIIQHVV ZKLOH IRU JHRJULG PRGHOLQJ RQO\ WKH
XVHG IRU DOO GLIIHUHQW ILQLWH HOHPHQW DQDO\VLV D[LDOVWLIIQHVVLVUHTXLUHG&RQVLGHULQJWKDWVRLO
7KH JHQHUDO UXOH LQ )(0 LV WKH ILQHU WKH PHVK QDLOV DUH FLUFXODU DQG GLVWULEXWHG DW FHUWDLQ
WKH PRUH DFFXUDWH DQDO\VLV DOWKRXJK WKH KRUL]RQWDOGLVWDQFHLWLVQHFHVVDU\WRGHWHUPLQH
LQFUHDVH LQ PHVK GHQVLW\ UHVXOWV LQ JUHDWHU WKH HTXLYDOHQW VWLIIQHVV DV UHFWDQJXODU HOHPHQWV
RYHUDOOFDOFXODWLRQWLPH7KHLQIOXHQFHRIPHVK SHUOLQHDUPHWHU
GHQVLW\RQILQLWHHOHPHQWVLPXODWLRQRIVRLOQDLO 7KHHTXLYDOHQWPRGXOXVRIHODVWLFLW\(HTRI
ZDOOV LV JLYHQ LQ %DEX 6LQJK EHLQJ WKH VRLO QDLO LV GHWHUPLQHG E\ WDNLQJ LQWR
PD[ IRU )6 DQG IRU GLVSODFHPHQW DFFRXQWERWKUHLQIRUFHPHQWEDUDQGJURXWFRYHU
DQDO\VLV IRU WKH UDQJH RI YHU\ FRDUVH DQG YHU\ 7KDW YDOXH LV ODWHU XVHG IRU FDOFXODWLRQ RI
ILQHPHVK EHQGLQJ(HT,DQGD[LDOVWLIIQHVV(HT$
HT Q Q
(J© J
)
( = ( $ $ + ( §¨ $ $ ·¸
¹
20.75 m
30 m ZKHUH (Q DQG (J DUH WKH PRGXOXV RI HODVWLFLW\
RI WKH QDLO EDU DQG JURXW PDWHULDO UHVSHFWLYHO\
$QDQG$JFURVVVHFWLRQDUHDRIWKHQDLOEDUDQG
70 m JURXW PDWHULDO DQG $ LV WKH WRWDO FURVV VHFWLRQ
40 m
RIWKHJURXWHGVRLOQDLO ʌ'
7KH LQWHUDFWLRQ EHWZHHQ VRLO DQG QDLO DQG
ZDOO IDFLQJ SODWH HOHPHQWV LV DFFRXQWHG E\
LQWHUIDFH HOHPHQWV %HDULQJ LQ PLQG WKDW WKH
90 m YDOXHRIVRLOUHLQIRUFHPHQWLQWHUDFWLRQREWDLQHG
E\SXOORXWWHVWVLVODUJHUWKDQXQLW\:DQJHWDO
WKHGHIDXOWLQJVHWRIULJLGLQWHUIDFHLQVRLO
PDWHULDO LV XVHG LQVWHDG RI WKH XVH RI LQWHUIDFH
)LJXUH0HVKDQGJHRPHWU\RIWKHVRLOQDLOZDOO
HOHPHQWV
$Q D[LDO YHUWLFDO GLVWULEXWHG ORDG N1P
7KH PDWHULDO PRGHODGRSWHGWRVLPXODWHWKH
ZDV DSSOLHG DW WKH VRLO VXUIDFH EHVLGH WKH VRLO
VRLO LV WKH KDUGHQLQJ VRLO PRGHO DV WKH PRVW
QDLOZDOO
DSSURSULDWH WR VLPXODWH WKH VRLO IRU H[FDYDWLRQ
7KH FRPSOHWH VRLO PDVV ZDV LQLWLDOO\ VLPX
DQDO\VLV E\ )(0 7KH DYHUDJH JHRWHFKQLFDO
ODWHG IRU JUDYLW\ ORDGLQJ EHIRUH H[FDYDWLRQ LQ
SDUDPHWHUV RI WKH VRLO SURILOH SUHVHQWHG LQ WKH
RUGHUWRJHQHUDWHLQLWLDOVWUHVVHV'LVSODFHPHQWV
/(0 DUH XVHG LQ WKH )(0 LQ RUGHU WR VLPSOLI\
GXH WR JUDYLW\ ORDGLQJ ZHUH VHW WR ]HUR DW WKH
WKHDQDO\VLVDQGWRUHGXFH WKHFDOFXODWLRQWLPH
314 S. Melentijevic and J.L. Fernández / Comparison of Finite Element and Limit Equilibrium Methods
EHJLQQLQJ RI WKH VWDJHG FRQVWUXFWLRQ XVHG IRU REWDLQHG E\ PRGHOLQJ VRLO QDLOV DV JHRJULG
JHQHUDO QXPHULFDO VLPXODWLRQ RI WKH VRLO QDLO HOHPHQWVLVORZHUWKDQ ZKHQ PRGHOLQJDVSODWH
ZDOO ,Q HDFK H[FDYDWLRQ VWDJH VRLO FOXVWHU HOHPHQWV ERWK IRU PRGHOLQJ ZLWK DQG ZLWKRXW
UHSUHVHQWLQJ H[FDYDWLRQ VLGH LV GHDFWLYDWHG DQG IDFLQJ E WKH UHVXOWV REWDLQHG E\ /(0 DUH
WKH FRUUHVSRQGLQJ VRLO QDLO DQG ZDOO IDFLQJ LV FORVHWRUHVXOWVRI)(0ZLWKVRLOQDLOVVLPXODWHG
DFWLYDWHG (DFK VWHS RI FDOFXODWLRQ FRPSULVHG DVSODWHHOHPHQWVGXULQJDOOFRQVWUXFWLRQVWDJHV
VXIILFLHQW QXPEHU RI FDOFXODWLRQ VWHSV LQ RUGHU F WKH LQIOXHQFH RI EHQGLQJ UHVLVWDQFH RI QDLOV
WRUHDFKWKHVWHDG\VWDWHFRQGLWLRQDIWHUH[FDYD RQ YDOXHV RI IDFWRU RI VDIHW\ LV FRQFOXGHG
WLRQ HVSHFLDOO\ GXULQJ WKH EHJLQQLQJ RI FRQVWUXFWLRQ
7DEOH VXPPDUL]HV PDWHULDO SURSHUWLHV GXH WR PXFK ORZHU VDIHW\ IDFWRUV REWDLQHG E\
DGRSWHG LQ QXPHULFDO VLPXODWLRQ XVLQJ 3OD[LV PRGHOLQJ RI QDLOV DV JHRJULG HOHPHQWV G WKH
7KHVRLOLVPRGHOOHGZLWKLWVDYHUDJHFKDUDFWHU GLIIHUHQFHEHWZHHQUHVXOWVREWDLQHGIRU)(0RI
LVWLFVDORQJZKROHGHSWK VRLOQDLOVDVSODWHDQGJHRJULGHOHPHQWVLVPRUH
SURQRXQFHG GXULQJ ILUVW FRQVWUXFWLRQ VWDJHV
7DEOH0DWHULDOPRGHOSDUDPHWHUV
ZKLOHEHLQJYHU\VLPLODUIRUILQDOVWDJHVHWKH
3DUDPHWHU FRQWULEXWLRQ RI ZDOO IDFLQJ HOHPHQW RQ WKH
1DLOV VWDELOLW\ GXULQJ H[FDYDWLRQ LV PRUH SURQRXQFHG
GXULQJILUVWFRQVWUXFWLRQVWDJHV
<LHOGVWUHQJWKUHLQIRUFHPHQW I\ 03D
&ĂĐƚŽƌŽĨƐĂĨĞƚLJ
(ODVWLFLW\PRGXOXVUHLQIRUFHPHQW (\ *3D ϭ Ϯ ϯ ϰ ϱ ϲ
(ODVWLFLW\PRGXOXVJURXW (J *3D Ϭ
6RLO Ϯ
ŽŶƐƚƌƵĐƚŝŽŶƐƚĂŐĞ
8QLWZHLJKW Ȗ N1P ϰ
2HGRPHWULFWDQJHQWVWLIIQHVV UHI
PRGXOXV (RHG 03D
)LJXUH*OREDOIDFWRURIVDIHW\GXULQJGLIIHUHQW
8QORDGLQJUHORDGLQJVWLIIQHVV UHI
PRGXOXV (XU 03D FRQVWUXFWLRQVWDJHV
3RZHURIVWUHVVGHSHQGHQF\ P
ŝƐƉůĂĐĞŵĞŶƚ;ŵŵͿ
&203$5,6212)5(68/76 Ϭ ϭϬ ϮϬ ϯϬ ϰϬ ϱϬ ϲϬ ϳϬ ϴϬ ϵϬ
Ϭ
hdž'
'XH WR OLPLWHG H[WHQVLRQ RI WKH SDSHU RQO\ hdž'н&
ŽŶƐƚƌƵĐƚŝŽŶƐƚĂŐĞ
Ϯ
hdžW
VRPHRIWKHUHVXOWVDUHSUHVHQWHG ϰ
hdžWн&
hLJ'
7KHH[WHUQDOIDLOXUHPRGHDVDJOREDOVWDELO hLJ'н&
LW\LVVWXGLHGXQGHUVWDWLFFRQGLWLRQVLQWKLVFDVH ϲ hLJW
hLJWн&
)LJXUH SUHVHQWV UHVXOWV RI IDFWRU RI VDIHW\ IRU ϴ
ERWK/(0DQG)(0GXULQJFRQVWUXFWLRQVWDJHV ϭϬ
7KH FRPSDULVRQ RI VLPXODWLRQ RI VRLO QDLOV E\
XVHRIJHRJULGDQGSODWHVWUXFWXUDOHOHPHQWVKDV
ϭϮ
VHH)LJXUH
'н&
'н&
'н&
ϰϬϬ EĂŝů ϴ
'н& EĂŝů ϭϬ
EĂŝů ϭϮ
Wн&
DĂdžĂdžŝĂůĨŽƌĐĞ;ŬEͿ
ϯϬϬ '
'
W '
ϮϬϬ
'н&
'н&
'н&
ϭϬϬ
)LJXUH9DULDWLRQRID[LDOIRUFHVDORQJVRLOQDLO
Ϭ OHQJWKVDWGLIIHUHQWOHYHOVIRU*DQG*)
ϭ Ϯ ϯ ϰ ϱ ϲ ϳ
^ŽŝůŶĂŝůůĞǀĞů
ϴ ϵ ϭϬ ϭϭ ϭϮ
ϯϬ
)LJXUH0D[LPXPYDOXHVRID[LDOIRUFHVDWGLIIHUHQW Y;Wн&Ϳ
DĂdžƐŚĞĂƌĨŽƌĐĞ;ŬEͿĂŶĚŵŽŵĞŶƚ
VRLOQDLOOHYHO Ϯϱ
D;Wн&Ϳ
Y;WͿ
D;WͿ
ϮϬ
ϭϱ
' ' '
ϭϬ
W
W W
ϱ
EĂŝůϭϭ
EĂŝůϳ EĂŝůϵ Ϭ
' ϭ Ϯ ϯ ϰ ϱ ϲ ϳ ϴ ϵ ϭϬ ϭϭ ϭϮ
' ' ^ŽŝůŶĂŝůůĞǀĞů
)LJXUH0D[LPXPYDOXHVRIVKHDUIRUFHVDQG
W W W
EHQGLQJPRPHQWVYVVRLOQDLOOHYHO
)LJXUH9DULDWLRQRID[LDOIRUFHVDORQJVRLOQDLO
OHQJWKVDWGLIIHUHQWOHYHOVIRUPRGHOLQJQDLOVE\* )LJXUH VKRZV WKH YDULDWLRQ RI EHQGLQJ
DQG3HOHPHQWV PRPHQWV DQG VKHDU IRUFHV DORQJ WKH VRLO QDLO
OHQJWK DW GLYHUVH OHYHOV IRU VRLO QDLOV PRGHOHG
7KH GLIIHUHQFH LQ D[LDO IRUFH GLVWULEXWLRQ DV SODWH HOHPHQWV 7KH H[SHFWHG FRQFHQWUDWLRQ
ZKHQPRGHOLQJZLWKDQGZLWKRXWIDFLQJHOHPHQW RIIRUFHVQHDUIDFHRIWKHZDOOLVREVHUYHG
316 S. Melentijevic and J.L. Fernández / Comparison of Finite Element and Limit Equilibrium Methods
)LJXUHSUHVHQWVWKHYDULDWLRQHYDOXDWHGLQ SUHVHQWVYDOXHVRIGLVSODFHPHQWVEHKLQGWKHVRLO
SHUFHQWDJH EHWZHHQ PD[LPXP WHQVLOH IRUFH LQ QDLO ZDOO DQG WKH OHQJWK RYHU ZKLFK WKH
WKHQDLOZKHQPRGHOHGE\*RU*)DVZHOODV GHIOHFWLRQVDUHGDPSHQHG
WKH GLIIHUHQFH LQ PD[LPXP WHQVLOH DQG VKHDU
IRUFHV DQG EHQGLQJ PRPHQWV GHYHORSHG ZKHQ E;ŬEͬŵůͿ Y;ŬEͬŵůͿ D;ŵŬEͬŵůͿ
VLPXODWLRQSHUIRUPHGDV3DQG3)HOHPHQWVDW
Ϭ ϯϬ ϲϬ ϵϬͲϯϬ ϭϬ ϱϬ Ͳϭϱ Ͳϱ ϱ
Ϭ Ϭ Ϭ
HDFKQDLOOHYHO Wн&
'н&
Wн&
'н&
Wн&
'н&
ĞƉƚŚŽĨĨĂĐŝŶŐŽĨƚŚĞǁĂůů;ŵͿ
Ͳϱ ϱ ϱ
EĂŝů ϰ EĂŝů ϴ EĂŝů ϭϬ
Y
Y Y
Wн&
Wн& Wн& ͲϭϬ Ϭ Ϭ
D D D
Ͳϭϱ ϱ ϱ
Y
Y Y
W W
W
ͲϮϬ Ϭ Ϭ
D D D
)LJXUH9DULDWLRQRID[LDOIRUFHVVKHDUIRUFHVDQG
EHQGLQJPRPHQWVLQIDFLQJHOHPHQWVIRUVLPXODWLRQ
)LJXUH9DULDWLRQRIVKHDUDQGPRPHQWIRUFHVDORQJ E\*)DQG3)HOHPHQWV
VRLOQDLOOHQJWKVDWGLIIHUHQWOHYHOVIRUPRGHOLQJQDLOV
DV3HOHPHQWV
λ
ϭϬϬ
ϵϬ
ϴϬ
ϳϬ
ŝĨĨĞƌĞŶĐĞ;йͿ
ϲϬ
ϱϬ
ϰϬ
ϯϬ vertical displacements
ϮϬ
ϭϬ
Ϭ
ϭ Ϯ ϯ ϰ ϱ ϲ ϳ ϴ ϵ ϭϬ ϭϭ ϭϮ λ
^ŽŝůŶĂŝůůĞǀĞů
E;'Θ'н&Ϳ E;WΘWн&Ϳ Y;WΘWн&Ϳ D;WΘWн&Ϳ
)LJXUH'LIIHUHQFHLQWHQVLOHDQGVKHDUIRUFHVDQG
EHQGLQJPRPHQWVDWHDFKVRLOQDLOOHYHOIRUVLPXOD
WLRQE\**)3DQG3)HOHPHQWV
horizontal displacements
)LJXUHVXPPDUL]HVWKH YDULDWLRQRID[LDO
DQG VKHDU IRUFHV DQG EHQGLQJ PRPHQWV DORQJ )LJXUH'HIOHFWLRQVEHKLQGWKHVRLOQDLOZDOO
WKHIDFLQJHOHPHQWVIRUVLPXODWLRQVE\*)DQG
3)HOHPHQWV 7DEOH'LVSODFHPHQWVDFFRUGLQJWR)(0DQG
)LJXUH LV DQ H[DPSOH RI REWDLQHG &ORXWHUUH
GHIOHFWLRQVEHKLQGWKHVRLOQDLOZDOOPDVVLQWKH
FDVH RI QDLOV PRGHOHG DV JHRJULG HOHPHQWV E\ įKPP įYPP įK+ įY+
)(0 )(0*
7KH FRPSDULVRQ RI GHIRUPDWLRQ SDUDPHWHUV )(03
DV GHVFULEHG LQ 6HFWLRQ E\ HPSLULFDO
IRUPXODWLRQV JLYHQ LQ &ORXWHUUH DQG LWV )(0*)
YDOXHVGHGXFHGIURP)(0DQDO\VLVLVSUHVHQWHG )(03)
LQ7DEOHVDQG7DEOHVKRZVKRUL]RQWDODQG PD[ PD[
&ORXWHUUH PD[ PD[
YHUWLFDO GLVSODFHPHQWV DQG LWV FRUUHVSRQGLQJ
UDWLRV WR WKH KHLJKW RI VRLO QDLO ZDOO 7DEOH
S. Melentijevic and J.L. Fernández / Comparison of Finite Element and Limit Equilibrium Methods 317
7DEOH'LVSODFHPHQWVEHKLQGWKHZDOOQDLOPDVVDQG
OHQJWKRYHUZKLFKWKHGLVSODFHPHQWVDUHGDPSHQHG 5()(5(1&(6
DFFRUGLQJWR)(0DQG&ORXWHUUH
%DEX */ 6LQJK 93 6WDELOL]DWLRQ RI
įPP ȜP
YHUWLFDOFXWXVLQJVRLOQDLOLQJ3OD[LV%XOOHWLQ,VV
)(0* SS
)(03 %DEX */ 6LQJK 93 6LPXODWLRQ RI 6RLO
)(0*) 1DLO 6WUXFWXUHV XVLQJ 3OD[LV ' 3OD[LV %XOOHWLQ
)(03) ,VVSS
%6 &RGH RI SUDFWLFH IRU VWUHQJWK
&ORXWHUUH PD[ PD[
HQHGUHLQIRUFHGVRLOV3DUWVRLOQDLOGHVLJQ
&KHDQJ://XR647DQ6$ <RQJ.<
&21&/86,216 /DWHUDOEHQGLQJRIVRLOQDLOVLQH[FDYDWLRQ
,QW &RQI *HRWHFK DQG *HRO (QJ *HR(QJ
7KH FRPSDULVRQ RI D FDVH VWXG\ E\ )(0 DQG 0HOERXUQH$XVWUDOLD
/(0LVSUHVHQWHG &,5,$ & 6RLO QDLOLQJ ± EHVW SUDFWLFH
7KH )(0 LV SUHIHUUHG WR /(0 LQ RUGHU WR JXLGDQFH/RQGRQ
REWDLQ D JOREDO SHUIRUPDQFH RI D VRLO QDLO ZDOO 'XUJXQRJOX+7.HVNLQ+%.XODF+),NL]6
DQGDQHYDOXDWLRQRIGHIOHFWLRQVRIWKHVRLOQDLO .DUDGD\LODU 7 3HUIRUPDQFH RI VRLO
ZDOO QDLOHGZDOOVEDVHGRQFDVHVWXGLHV*HRWHFKQLFDO
7KH FRPSDULVRQ RI HPSLULFDO GHIOHFWLRQ (QJLQHHULQJ LQ 8UEDQ (QYLURQPHQWV 3URF WK
IRUPXODWLRQV DFFRUGLQJ &ORXWHUUH (XURSHDQ&RQIRQ6RLO0HFKDQLFVDQG*HRWHFK
JHQHUDOO\ DGRSWHG LQ RWKHU LQWHUQDWLRQDO (QJ0DGULG6SDLQ9ROSS
VWDQGDUGV ZLWK )(0 UHVXOWV LV SUHVHQWHG 7KH (XURSHDQ 6WDQGDUG (1 ([HFXWLRQ RI
HVWLPDWLRQ RI GGLVSODFHPHQWV E\ PHQWLRQHG VSHFLDOJHRWHFKQLFDOZRUNVVRLOQDLOLQJ
HPSLULFIRUPXODHLVVXIILFLHQWDWWKHSUHOLPLQDU\ )HGHUDO+LJKZD\$GPLQLVWUDWLRQ*HRWHFKQLFDO
VWDJHRIWKHVWXG\ HQJLQHHULQJFLUFXODU1R±VRLOQDLOZDOOV5H
7KH SODWH HOHPHQWV IRU VLPXODWLRQ RI VRLO SRUW )+:$$2,)26 :DVKLQJWRQ '&
QDLOVLVSUHIHUUHGWRJHRJULGHOHPHQWVLQRUGHUWR 86$
DFFRXQW IRU WKH EHQGLQJ VWLIIQHVV DQG VKHDU )UHQFK1DWLRQDO5HVHDUFK3URMHFW5HFRPPHQGDWLRQV
UHVLVWDQFHRIVRLOQDLOV &ORXWHUUH
7KH JOREDO VWDELOLW\ DQG GHIRUPDWLRQ -HZHOO 5$ 3HGOH\ 0- 6RLO QDLOLQJ
FRQVLGHUDWLRQV LQGLFDWH D VWDEOH V\VWHP RI WKH GHVLJQWKHUROHRIEHQGLQJVWLIIQHVV*URXQG(Q
FDVHVWXG\ JLQHHULQJSS
9DOXHV RI IDFWRU RI VDIHW\ REWDLQHG E\ -HZHOO 5$ 3HGOH\ 0- 'LVFXVVLRQ
PRGHOLQJVRLOQDLOVLQ)(0DVJHRJULGHOHPHQWV FORVXUH*URXQG(QJLQHHULQJSS
LVORZHUWKDQ ZKHQ PRGHOLQJDVSODWHHOHPHQWV -HZHOO 5$ 3HGOH\ 0- $QDO\VLV IRU VRLO
ERWK IRU FDVHV ZLWK DQG ZLWKRXW IDFLQJ 7KH UHLQIRUFHPHQW ZLWK EHQGLQJ VWLIIQHVV -RXUQDO RI
UHVXOWV RI )6 REWDLQHG E\ /(0 DUH FORVH WR *HRWHFK(QJ9RO1RSS
UHVXOWVRI)(0ZLWKVRLOQDLOVVLPXODWHGDVSODWH 3OD[LV 3OD[LV XVHU PDQXDO 8QLYHUVLW\ RI
HOHPHQWV 7HFKQRORJ\ 3OD[LVEY7KH1HWKHUODQGV
7KHLQIOXHQFHRIEHQGLQJUHVLVWDQFHRIQDLOV 6KORVVHU ) 'LVFXVVLRQ 7KH PXOWLFULWHULD
RQ YDOXHV RI IDFWRU RI VDIHW\ LV REVHUYHG DW WKHRU\ LQ VRLO QDLOLQJ *URXQG (QJLQHHULQJ SS
PD\RU VFDOH GXULQJ WKH EHJLQQLQJ RI
FRQVWUXFWLRQ RI VRLO QDLO ZDOO FRQVLGHULQJ :DQJ = 5LFKZLHQ : $ VWXG\ RI VRLO
VLJQLILFDQWO\ ORZHU VDIHW\ IDFWRUV REWDLQHG E\ UHLQIRUFHPHQW LQWHUIDFH IULFWLRQ -RXUQDO RI *HR
PRGHOLQJRIQDLOVDVJHRJULGHOHPHQWV WHFK *HRHQY(QJ9RO1RSS
7KH GLIIHUHQFH EHWZHHQ UHVXOWV REWDLQHG E\ ;3 3 6RLO UHLQIRUFHPHQW ± QDLO UHLQ
)(0 ZLWK VRLO QDLOV PRGHOHG DV SODWH DQG IRUFHG LQ VLWX UHWDLQLQJ DQG VORSH VWUXFWXUHV
JHRJULG HOHPHQWV LV PRUH SURQRXQFHG GXULQJ 9HULILFDWLRQRIGLPHQVLRQLQJ
ILUVWFRQVWUXFWLRQVWDJHVUHDFKLQJDOPRVWVLPLODU
YDOXHVDWILQDOVWDJHV
7KHFRQWULEXWLRQRIZDOOIDFLQJRQWKHVWDELO
LW\ GXULQJ H[FDYDWLRQ LV PRUH SURQRXQFHG
GXULQJILUVWFRQVWUXFWLRQVWDJHV
318 Soil-Structure Interaction, Underground Structures and Retaining Walls
V.M. Ulitsky et al. (Eds.)
IOS Press, 2014
© 2014 The authors and IOS Press. All rights reserved.
doi:10.3233/978-1-61499-464-0-318
Abstract. The increase in demand for underground tunnels in Singapore has added many challenges to tunnelling problems
beyond the capability of analytical or empirical methods. The prospect of using 3D numerical analysis in analyzing complex
practical applications is becoming more common. The paper presents the application of finite element analysis to assess the soil-
structure interactions in real tunneling problems. A benchmarking study was first carried out on a centrifuge test by Ong et al.
(2013) using a three-dimensional finite element method for the basic understanding of soil responses to tunneling. FE studies
were then extended to a practical application with the field data from a Singapore MRT project, Bendemeer Station, Downtown
Line 3 (DTL-3).
Keywords. tunneling, soft clay, centrifuge testing, modified Cam-Clay model, subsurface settlements, FE analyses
tunnel model (i.e, 6m diameter in prototype 10-noded linear strain tetrahedral solid ele-
scale) was carried out under 100g at National ments were used in the FE model. The displace-
University of Singapore. The tunnel was at 15m ment was completely fixed at the bottom bottom.
depth, and hence, tunnel depth to diameter, H/D Vertical sides were restrained against transverse
ratio of 2.5 was adopted in the centrifuge test. A movement for the displacement boundary condi-
centrifuge model set-up is shown in Figure 2. tion. The top surface mesh was assumed as the
drainage boundary where u=0. The two vertical
sides were assumed to be hydrostatic, recharging
"
ú* #!#e-
able boundaries were assigned at the bottom
plane and tunnel circumferences.
ling process was simulated in the procedures as Gaussian curve respectively. In other words,
follow: FEM underestimates the maximum surface
Initial step: Initial geostatic stress application. settlement results about 17% compared to Gauss-
Step 1 : The soil along the length of 20m ian method whereas only 3% by centrifuge test.
(tunnel length in the model) was excavated and
the support pressure was applied around the
excavated tunnel periphery in the same step.
If the support is equivalent to the initial all
round soil pressures, no volume loss is theoreti-
cally expected. Any reduction to this value will
introduce the volume loss as the end result.
Therefore, the support pressure in FE analysis is
a predetermined value which produced the target
volume loss of 3%.
In this FE analysis study with GeoFEA pro-
gram, the support pressure was applied in a
hydrostatic distribution with the values increas- Figure 5. Transverse settlement trough
ing from tunnel crown to invert level (Figure 4).
of tunneling induced settlement troughs in the greenfield soil response mainly owing to its
same trend especially at shallower depth. simplicity and reliability.
Figure 10. Soil displacement vectors (a) FEM , (b) centrifuge test
and results in the radial soil flow towards the Besar Station.At the monitoring array, the
tunnel. However, since the movements are tunnels were located at depth of about 32.5m
restrained at the boundaries, reverse soil below ground level.
movements occurred to maintain all the forces
in equilibrium state. This kind of effect may be 3.2. Geological and Soil Condition
reduced by using large sized FE mesh bounda-
ries. Unlike the centrifuge test, this problem is The DTL-3 tunnelling works encountered a
minimized due to the presence of drained layer high variation in geological condition along its
at the bottom and the soil flow directly toward alignment. Bendemeer Station project site also
the bottom or tunnel periphery. However, the involved the launching of TBM from good OA
amount of U turn movements are insignificant soil towards the Kallang formation, with the
(<5mm) compared to the movement at imme- challenging mixed-face soil condition at the
diate shear zone, major occurrence of soil transition. The field monitoring section under
displacements. Except that, the soil displace- this study was located in Old Alluvium (O)
ment pattern in FE results are comparable with formation, which was overlain by 2.5m thick
centrifuge test results. Fill and 3m thick Marine Clay (M) layers. The
water table from site investigation report
3. Understanding of soil responses due to ranged between 2 to 2.5m below ground
tunnelling –case study of bendemeer surface. The typical geological profile along
station (DTL-3), Singapore the proposed tunnel alignment is shown in
Figure 11.
3.1. Background and overview of the project
3.3. Ground Instrumentation
A 42km long Downtown Line (DTL) is the
longest and the fifth stage of underground Comprehensive instrumentation and monitor-
Mass Rapid Transit (MRT) line in Singapore. ing arrays were installed along the tunnel
The construction of DTL, which is proposed to alignment. Figure12 below illustrates the cross
serve 34 stations, was planned in 3 stages section of proposed monitoring array at the
(DTL-1, DTL-2, DTL-3). It will connect the section under this study, which includes:
North-Western and Central-Eastern regions to x series of ground settlement markers across
the new downtown of Singapore. The con- the section for the assessment of trans-
struction of the final stage of Downtown Line, verse surface settlement profile and along
DTL-3, comprises of 16 stations with three the tunnel alignment for the longitudinal
interchange stations. A 3-D finite element surface settlement;
analysis was carried out for Downtown Line x two magnetic extensometers placed at
(DTL-3), Bendemeer Station project to inves- each side of the two tunnels, for vertical
tigate the effect of tunnelling in Old Alluvium. subsoil movements at different depth (tips
The twin tunnels constructed by Earth Pressure at 3.5m, 12.5m and 32.5m respectively.)
Balance machines (EPBM) was passing
through from Bendemeer Station to Jalan
Figure 12. Ground instrumentation layout and section (Bukit Panjang Bound for this study)
3.4. Finite element mesh and boundary condi- were modelled using a 10-noded tetrahedral solid
tions elements. By taking into account of the effect of
symmetry, only half boundary sizes (105m x
Figure 13 shows the finite element mesh for the 60m x 72.5m) was used. Tunnel excavation was
simulation of single tunnel advancement. The simulated up to 75m out of 105m length.
soil, tunnel liner, overcut, shield and grouting
C.W. Ong et al. / Understanding of Soil Responses Due to Tunnelling 325
Displacement boundary conditions were tak- Table 2. Mohr-Coulomb model Soil Parameters
ing the same as the case history of tunnelling in Type of Soil E', MPa ø C', kPa
soft clay. For hydraulic boundary condition, the
Fill 8 0.3 0 30
vertical side parallel to the tunnel symmetry axis
was assumed to be hydrostatic, recharging 333cu
Clay (cu=15+1.5z) z 0.3 0 22
"
ú* #!#e- from +95
able boundaries were taken for symmetry plane,
Old Alluvium
bottom plane, tunnel circumferences and remain- (N<100)
125 0.3 10 32
ing vertical sides of the mesh.
Old Alluvium
250 0.3 20 34
(N<100)
Structural
E, MPa ø k, m/s
ªÍ#3
Element
Shield
200,000 0.25 10-12 80
Machine
Concrete
28,000 0.2 10-12 24
Lining
Overcut 1 0.2 0 0
Grout 2,800 0.2 10-12 24
Figure 17. Soil displacement versus depth (a) Soil Settlement (b) Transverse soil deflection (c) Longitudinal soil deflection
discrepancies are observed among the different strain characteristics in soil-structure interaction. Ge-
methods when it is further down the ground otechnique 36, No. 3, pp.377-396.
Lim K. C. 2003. Three-dimensional finite element analysis
level. Among them, FE prediction is closer to of earth pressure balance tunnelling. PhD thesis, Na-
empirical method while centrifuge test yield tional University of Singapore.
larger differences due to the boundary condi- Loganathan, N. and Poulos, H. G. 1998. Analytical
tion limit. However, the confrontations among prediction for tunneling-induced ground movements
the methods become insignificant when the in clays. Journal of Geotechnical and Geoenviromen-
tal Engineering, Vol. 124, No. 9, pp. 846-856.
distance x is more than a diameter of the LTA. 2010. Civil design criteria for road and rail transit
tunnel. systems, Land Transport Authority, Singapore
(3) The lateral soil deflections are promi- Mair, R. J., Taylor, R. N. and Bracegirdle, A. 1993.
nent at tunnel horizontal axis for the distance Subsurface settlement profiles above tunnels in clay.
less than 1D (D=diameter of tunnel). When Geotechnique, Vol. 43, No. 2, pp. 315-320.
New, B. M. and O’Reilly, M. P. 1991. Tunnelling induced
the distance is beyond 1D, the deflection at the ground movements: predicting their magnitude and
ground surface is more significant than that at effects. In Geddes (Ed) Proc. 4th Int. Conference on
the tunnel axis. ground movements and structures, Cardiff, invited
(4) The soil displacement vectors in FE re- review paper, Pentech Press, London, pp. 671-697.
sults give similar pattern with centrifuge test Ong, C. W., Leung, C. F., Yong, K. Y. and Chow, Y. K.
2007. Experimental study of tunnel-soil-pile interac-
results except that those vectors outside the tion. Underground Singapore 2007, pp. 55-66.
immediate shear zone flow occurs in reverse U Ong, C.W., Leung, C. F. & Yong, K.Y. 2013. Understand-
turn pattern due to the FE boundary con- ing Long-Term Effects of Tunnel-Soil Interaction.
straints. 18th International Conference in Soil Mechanics and
(5) Ground response in Old Alluvium is Geotechnical Engineering, Proceedings of the TC207
Workshop on Soil-Structure Interaction and Retaining
found to be fairly small. Walls 4 September 2013, Paris.
The results and findings have shown that Pang, C.H. 2006. The Effects Of Tunnel Construction On
with proper understanding of soil behavior and Nearby Pile Foundation PhD thesis, National Univer-
modeling technique, the complex soil structure sity of Singapore.
interactions can be predicted reasonably. Pang, C.H., Yong, K.Y., and Chow,Y.K. 2005. Three-
dimensional numerical simulation of tunnel advance-
These 3-D finite element studies will be
ment on adjacent pile foundation. Proceedings of the
extended further in the next paper for the 31st ITA-AITES World Tunnel Congress, Under-
condition of tunnelling in mixed faced soil ground Space Use: Analysis of the Past and Lessons
conditions with hard soil or rock at the lower for the Future, 7-12 May 2005, Istanbul, Turkey.
face of tunnel and soft soil at the upper face. Peck, R. B. 1969. Deep excavations and tunneling in soft
ground. Proc. 7th International Conference Soil Me-
chanics and Foundation Engineering, Mexico City,
4. Acknowledgement State of the Art Volume, pp. 225-290.
Potts, D. M. and Zdravkovic, L. 1999. Finite element
The authors would like to express their sincere analysis in geotechnical engineering: Theory. Vol. 1.
gratitude to Bendemeer Station project team, Thomas, London.
Land Transport Authority (LTA) of Singapore,
Tritech Consultants Pte. Ltd. and Tritech
Engineering Pte. Ltd. for providing the re-
quired tunnel information and field instrumen-
tation data of Bendemeer Station project.
5. References
Addenbrooke, T. I., Potts, D. M. and Puzrin, A. M. 1997.
The influence of pre-failure soil stiffness on the nu-
merical analysis of tunnel construction. Geotechnique,
Vol. 47, No. 3, pp.693-712.
Geotechnical Interpretative Baseline Report. Arup
Singapore Pte. Ltd.
Goh, T.L. 2003. Stabilisation of an excavation by an
embedded improved soil layer. PhD thesis, National
University of Singapore.
Jardine, R. J., Potts, D. M., Fourie, A. B. and Burland, J.
B.1986. Studies of the influence of non-linear stress-
Soil-Structure Interaction, Underground Structures and Retaining Walls 329
V.M. Ulitsky et al. (Eds.)
IOS Press, 2014
© 2014 The authors and IOS Press. All rights reserved.
doi:10.3233/978-1-61499-464-0-329
Abstract. Gate shaft is one of parts the power waterway system of dam structure and as an intake vertical tunnel or well
function for hydroelectric power plants in Jatigede area, Sumedang, West Java. Paper concerns some consideration in deter-
mining a constitutive model of soil and rock material. Comparison of existing lateral movement and the analysis results
identified that Hoek – Brown failure criterion were more suitable than Mohr – Coulomb. Study analysis was carried out using
without and with reinforcement system, the risk assessment to a rock shaft during excavation is greater than that to a soil shaft,
even though the two averages designed safety factors are the same. For the analysis, it should be considered to use the failure
criteria of intermediate material.
Keywords. fractured rock, gate shaft structure, reinforcement system, Hoek – Brown failure criterion
Figure 13 shows the typical of gate shaft de- wall (Elevation+ 295.500) was 1 (one) m depth
formation and the total displacement in element from the top of excavated gate shaft. Number of
modeling of 2D after excavation process to the installed inclinometer devices was 4 (four)
depth of 45 m. Figure 14 shows the deformation points with code number of BH 1, BH 2, BH 3,
in 3D analysis. Magnitude of total displacement and BH 4. The depth of installed inclinometer
was depended on the failure criterion of soil and devices were between 44.0 and 56.0 m depth.
or rock. From laboratory and field tests, it was Coordinates and depths of each inclinometer
difficult to select one constitutive model which devices are shown on Figure 15. Measurement
was suitable with the existing condition at the of inclinometer was performed for 8 (eight )
site. Mohr – Coulomb and Hoek – Brown months continually according to the time
materials were used to determine magnitude of interval and sequentially in accordance with the
lateral displacement. cycle excavation works from initial reading as
described by the flowchart in Figure 16.
existing maximum lateral movement measured loads. All steps analysis do not consider time -
inclinometer system during time observation. dependent phenomena such as volumetric
From the Figure 17, the lateral movement consolidation or pore pressure dissipation.
predicted by using Mohr – Coulomb material The layers of claystone stability condition
constitutive model is larger than Hoek – Brown. were more influenced by movements than
The results from inclinometer device show that volcanic breccia layers. Small movements of
the existing lateral movement measured is both layers could determine a large fall in the
smaller than assessment results applied by the available shear strength and therefore cause a
Hoek – Brown material failure criterion. From large safety reduction. Some reinforcement
observation at the field during the inclinometer systems (such as: concrete collar, shotcrete, and
measurement, the lateral movements were more rockbolt) could reduce the both soil movement
caused by the vibration during the excavation vertically and horizontally. From back analysis,
process using blasting method. The lateral safety factor without reinforcement system was
movement on volcanic breccia layers initiated existed between 0.98 - 1.03. And then, with
by the collapse of material composed of broken using reinforcement systems were generated
rock fragments varying in size. However, the between 2.32 - 2.97 analyzed by constitutive
lateral movements on claystone layers were model from Mohr – Coulomb and Hoek –
started by the emerged of the fault or crack Brown. However, the risk to a rock shaft during
zones. The fault zone of claystone layers can be excavation is greater than that to a soil shaft,
subjected to slumping and magnitudes of lateral even though the two average designed safety
movement were dominated by claystone layers. factors are the same. Consequently, the analysis,
it should be considered to use the failure criteria
of intermediate material.
References
Abstract. Since 1839, when the historical railway Napoli – Portici was inaugurated, Napoli has been characterized by a signifi-
cant system of urban and suburban railways. The idea of a fully integrated urban rail network dates back to the 1950s, but only in
1990s construction began. At present, the system includes 54 km of tracks and 69 stations belonging to 6 lines currently
operating; when the City Transport Plan will be completed, 10 lines with 93 km of track and a further 30 km of new light rail
linking 114 stations with 21 interchanges are foreseen.
Keywords. deep excavations, dewatering field tests, Napoli underground, San Pasquale site
Figure 2. Plan view and transverse vertical section of the San Pasquale Station
G. Russo et al. / San Pasquale Station of the Linea 6 in Napoli 339
0.5
f[log(k)]
0.3
0.1
N = 148 N = 34
0.0
-10 -9 -8 -7 -6 -5 -4 -3
log (k) (k in m/s)
S1 -31.97 1.3 8 5
S2 -38.94 1.3 9 4
SG1 -38.43 1.3 9
SG2 -42.23 1.3 9 3. Design issues
SG3 -45.23 1.3 9
SG4 -35.71 1.3 8 As already mentioned in the introduction, San
SG5 -33.97 1.3 8 Pasquale station is very close to some historical
SG6 -37.63 1.3 and valuable buildings belonging to the down-
SG7 -34.21 1.3 8 town. It is excavated in granular soils sitting
SG8 -43.49 1.3 9
SSP1 -39.87 1.3 0
above a rather homogenous and not largely
SSP2 -47.23 1.3 0 altered NYT layer with a rather superficial
groundwater table governed by the nearby sea
level.
Preliminarily, in the design stage approxi- Of course the design first and the construc-
mated permeability values were assigned to the tion later had to cope with several difficulties
different materials present in the subsoil. The arising by the above mentioned features. First of
permeability assigned to the NYT was obtained all the need to preserve the structure of historical
from 9 Lugeon tests whose results are summa- buildings some of these being enriched by
340 G. Russo et al. / San Pasquale Station of the Linea 6 in Napoli
valuable frescoes which could results as highly two different depths. The internal piezometers
vulnerable to even small displacement. were located respectively at -37.8 m a.s.l. and -
The method of construction and the stages of 40.8 m a.s.l.. The external ones were respectively
execution were revised several times in order to located at -12.80 m a.s.l. and -27.80 m a.s.l.
minimize such effects. In the end it was preferred below the ground surface.
the excavation with top-down technique using The trial field started on 18/05/2010 and fin-
the intermediate slabs as temporary supports and ished on 10/07/2010. Preliminary tests were
the association of deep dewatering wells to carried out on the single wells P1 and P2 sepa-
depress the ground water table inside the main rately. These initial tests were of the maintained
shaft (Russo et al., 2012). step type setting various outflows and waiting
It is clear that the problem is a very complex the time needed for stationary conditions to be
one with the settlement induced by subsidence reached. The large dewatering of the whole
increasing the already critical settlement induced station was carried out later since 26/05/2010 to
by the deformations of the diaphragm walls. 05/07/2010, for a total duration of 40 days.
With the main aim of correctly calibrating the Initially only wells P1 and P2 were activated
pumping stations needed to depress the inner while at a given time also P3 and P4 were addi-
groundwater table and of judging the entity of tionally activated reducing the amount of water
the external induced subsidence without having pumped out by the first two wells. The well P3
to sum the displacement induced by the excava- had some troubles and it was rapidly switched
tion a preliminary dewatering test was carried off relying upon the remaining three wells P1, P2
out in a trial field. The main steps of the dewater- and P4 which were however capable of lowering
ing field tests are first described allowing the the groundwater table in the whole station area
presentation of the main experimental results. (Mormone et al., 2013).
Subsequently the interpretation carried out via Table 3. Well’s features
numerical analyses is presented essentially to
discuss about permeability values attributed to
P1 P2 P3 P4
the different soil layers.
ground level
0.17 0.07 1.04 0.96
(m a.s.l)
4. Dewatering trial test depth
39.97 39.87 38.84 43.76
(m)
A dewatering trial test was designed with the well screen -27.8 -27.8 -25.8 -27.8
agreement of the construction company whose (m a.s.l.) -37.8 -37.8 -35.8 -37.8
main aim was simply to establish the minimum diameter
600 600 600 600
configurations of wells and pumps capable of (mm)
lowering down the groundwater table on the pump power
7.5 7.5 11 11
whole station area (see Figure 5 and 6). Prelimi- (kW)
nary calculations based on the design values
adopted for permeability were carried out.
A total of six pilot wells were installed in the 5. Test results
station area included within the existing dia-
phragm walls. Four pilot wells were installed in San Pasquale station was subjected to a careful
the western part of the station area. Diameters monitoring since the beginning of the construc-
and locations had to be discussed and agree also tion. The main civil works are now completed
with the Board of archaeological remains still but the present paper deals essentially with the
involved in surveying activities which usually interpretation of the dewatering trial field test
stop when the excavation proceed below the sea carried out before the main excavation took
level .The wells were 600 mm in diameter and place. The full result of the case history are still
were deep enough to get into the NYT deep under examination and will be published else-
layer. The main features of the pumping wells where. This work focuses only on the interpreta-
are summarised in Table 3. tion of few measurements for sake of simplicity
Moreover internal and external piezometers and for obvious space reasons.
were installed in couples on the same vertical at
G. Russo et al. / San Pasquale Station of the Linea 6 in Napoli 341
Figure 5. Plan view of station shaft with layout of monitoring system for the dewatering trial test.
T-T L-L
PZ13-14
PZ10-11
PZ7-14
PZ8-13
PZ9-12
PZ7-8
11-12
9-10
PZ5
PZ6
PZ4
PZ3
PZ1
PZ2
Figure 6. Cross and longitudinal section of station shaft with layout of monitoring system for the dewatering trial test.
Figure 7 shows the pumping rates and the line in the Figure 7. In the same figure the actual
measured groundwater hydraulic heads during outflow at each well and globally imposed is of
the dewatering trial test. Two different phases course represented at each time during the whole
can be clearly distinguished. During the first test. At the end of the test the total volume
phase only two wells (P1 and P2) were activated. pumped in 40 days was 72425 m3, 30657 m3
A flow rate of about 10 l/s was imposed at each coming from well P1, 27661 m3 from the well
well. P2, 1010 m3 from the well P3 and 13097 m3 from
The second phase started when the other two well P4.
wells (P3 and P4) were activated. After a few The elevations of the groundwater table in-
days pumping from the P3 well was stopped for side and outside the shaft are shown in Figure 7.
problems at the electric pump. Therefore the rest The diagram reported in Figure 7 was simplified
of the test was conducted using the wells P1, P2, and only the two envelopes coming out from the
and P4, with a flow rate measured at the end of internal and the external piezometers readings
the test approximately equal to 22 l/s. were represented as two different shaded areas. It
The average flow rate pumped during the can be inferred that the huge lowering of the
whole test was 21 l/s as it can be seen by the full groundwater table inside the excavation shaft
342 G. Russo et al. / San Pasquale Station of the Linea 6 in Napoli
Figure 7. Pumping rates and measured groundwater hydraulic heads during the dewatering trial test
was practically not causing any significant effect voir providing water without any significant
outside the excavation. The only internal device downwards flow. The average hydraulic gradient
showing some difference, PZ10, was left out of appreciated with several couples of external
the envelope for the group of internal piezome- piezometers is practically null and exactly
ters because no reasonable cause of the differ- i=-0.008 (negative is for downwards flow). On
ence could be detected. the other hand the internal piezometers allow to
It is worth noticing that the hydraulic head detect upwards flow with an average hydraulic
registered by the two piezometers composing gradient equal to +0.16.
each external couple disposed along the same
vertical are practically identical. It follows that
the monitoring system did not detect any signifi- 6. Back analysis
cant hydraulic head gradient in the vertical
direction across the soil layers outside the station The analysis and interpretation of the problem is
shaft; however it is very likely that the differ- highly complex due to three dimensionality, soil
ences in the groundwater head between the heterogeneity, filtration problems.
inside of the station shaft and the surrounding The main purpose of this study is to obtain
area were mainly dissipated inside the NYT the best interpretative model that allows to
layer. capture the induced effects inside and outside the
As a matter of fact the loose sandy deposits station due to the excavation and the process of
located above the tuff in the external areas pumping.
appears acting simply as a high capacity reser-
G. Russo et al. / San Pasquale Station of the Linea 6 in Napoli 343
T-T L-L
130 24 130 180 85 180 +1.3 m a.s.l.
+2.3 m a.s.l. +1.3 m a.s.l. +2.3 m a.s.l.
27
27
-14.7 m a.s.l. -14.7 m a.s.l.
-35.7 m a.s.l. -35.7 m a.s.l.
23
-41.7 m a.s.l. W1
23 -41.7 m a.s.l. W1 W2
W2
152,3
182,3
284
Pumping well
445
Figure 9. Transversal and longitudinal section of the shaft of San Pasquale Station: comparison between analyses and
measurements
The permeability values of the soil layers “stationary condition” of the dewatering test. A
were back-calculated by means of two bi- trial and error procedure was adopted to estimate
dimensional finite element models. However it the soil water permeability. Among many differ-
was indispensable to adopt a conventional ent simulations only a few best cases have been
criterion of equivalence to reduce a three- selected and summarised in Table 4. The results
dimensional problem to a simpler 2D one. First are plotted in the diagrams of figure 9 where also
of all it was necessary to estimate an equivalent the comparison with the measured values is
pumping rate to be assigned to the indefinitely found.
long “bi-dimensional” wells considered in the 2D Several attempts were carried out correspond-
model. Two fem models were implemented ing to different hypotheses. First of all the values
referring respectively to the longitudinal and to of permeabilities available from site and labora-
the transversal cross section of the station shaft. tory investigations were immediately checked as
Figure 8 shows the two alternative cross sections being too low. The equivalence on the pumping
adopted for the 2D parametric analyses. The rate obtained in such a case was considered as
geometry of the box meshed and used for the completely out of place and thus ignored. The
fem calculations is such to minimize the edge trial and error procedure started investigating
effects. The simplified stratigraphic sequence is higher values of permeability compared to those
the same for all calculations and consists of 4 obtained by laboratory tests and Lugeon tests. It
layers: the layers labelled with capital letters A, can be seen from the plot of figure 9 that the best
B, C and D correspond to those already identi- agreement between the results and the analyses is
fied in the simplified stratigraphy of figure 3. found with both longitudinal and transversal
The models permitted to simulate the filtration sections respectively in the case 1 and 2. In these
process that takes place in the station during the two cases for the NYT formation a significant
344 G. Russo et al. / San Pasquale Station of the Linea 6 in Napoli
(#) W1 W2 A B C D
x y x y
1 L 50 25 800 400 7.0 70 7.0 70
2 T 30 30 800 400 7.0 70 7.0 70
3 L 50 25 700 60 30 30 30 30
4 T 30 30 700 60 30 30 30 30
anisotropy was also presumed according to the among DICEA of Federico II, AnsaldoSTS Spa
widely diffused knowledge that vertical fissures and Metropolitana di Napoli spa.
are present in the deposit due to the rapid cooling
process subsequent its deposition. The letter y in 9. References
the table 4 represents of course the vertical
direction. After several attempts the best case Autuori S., Russo G., Nicotera M. V. (2013), Studio
among the several isotropic ones were individu- preliminare degli effetti indotti dallo scavo della stazione
ated in the case labelled 3 and 4. As it is shown San Pasquale, Incontro Annuale Ricercatori di
in the figure 9, in this case keeping the same Geotecnica, Perugia, ISBN 9788890642135
pumping rate of the case 1 and 2, the lowering of L’Amante D., Flora A., Russo G., Viggiani C. (2012).
the groundwater head is rather overestimated. In Displacements induced by the installation of diaphragm
conclusion none out of these two further cases is panels. Acta Geotechnica, 7:203-218.
as satisfactory as the cases 1 and 2. Mormone L., Falconio G., Mandolini A. (2013). Groundwa-
ter management during excavation of S. Pasquale station
7. Conclusions in Naples. Geotechnical Engineering for the preservation
of monuments and historic sites, ISBN 978-138-00055-1
Huge excavations are going on in the city of Russo G., Viggiani C., Viggiani G.M.B., 2012. Geotechnical
Napoli to build a new line of the underground design and construction iussues for Lines 1 and 6 of the
network. Some of these excavations present Naples Underground. Geomechanics and Tunnelling 5,
among the other difficulties the management of No. 3. Berlin.
significant dewatering processes in area which CENSIS, 2008. 42° Rapporto sulla situazione sociale del
are closely surrounded by important buildings. Paese. Roma. Franco Angeli.
As briefly shown in the present paper dewatering
tests are important in order to check in the
execution stage the possibility of obtaining the
whished results in the excavation area without
producing external effects on the groundwater
table. These last would of course be accompa-
nied by subsidence problems on the existing
buildings which must be carefully prevented.
The back-analyses of the executed tests allowed
also to deepen the knowledge of the permeability
of the geological formations involved at the field
scale.
8. Acknoledgements
Keywords. geotextile-reinforced soil walls, model walls, frost heave, depth of frost penetration, control of deformations
pushed outward. Once deformed, the wall measuring the deformation due to frost heave in
surface does not return to its original state even winter. The embankment material was volcanic
when the temperature rises, which leads to ash soil whose maximum dry density is not
cumulative deformation. When the cumulative measurable. Cone indexes set according to soil-
deformation reaches a certain level, the wall trafficability testswere3000kN/m2 or greater.
surface materials protrude outward due to frost This volcanic ash soil is a good material for the
heave and the wall collapses (Photo 1). construction of embankments, in that compac-
tion is relatively easy. But this soil is not totally
Wall surface insusceptible to frost heave, and frost heaving
may occur.
The model walls were built in an earth tank 5
m square and 2 m deep. The bottom of the earth
tank was covered with gravel 30cm thick for
drainage. Above the gravel, three tiers of em-
bankments were built. Each tier was 2.4m wide,
60cm high and 3.0m deep. Two sets of model
walls were built in the earth tank, with each set
separated by 10cm-thick Styrofoam. A total of
five sets of model walls were used for testing:
Cases 1, 2, 3, 4 and 5.
The five cases are outlined in Table 2, and
Photo 1 Example of deformation in geotextile-
side views of thefive models are shown in Figure
reinforced soil walls
Table 1 Basic physical property values of fill material
Physical property Volcanic ash
3. TESTING METHODS
Soil particle density Us (g/cm3) 2.470
The factors influencing frost heave are soil Natural water content wn (%) 46.98
quality, air temperature and water drainage; thus, 2 mm or larger (%) 22.8
soil replacement, maintaining temperatures Grain size 75 m to 2 mm (%) 51.4
above the freezing pointand insuring proper Less than 75 m(%) 25.8
drainage are the keys to preventing frost heave. Consistency limit N.P.
Because it is very difficult to improve drainage Classification symbol of soil material SVG
and replace soil in existing geotextile-reinforced Velocity of frost
soil walls, full-scale model walls were 0.27
heaving (mm/hr)
constructed for conducting tests focusing on Frost
Percentage of frost
temperature. heaving 18.1
heave (%)
Degree of frost heaving Medium level
3.1. Specifications of the full-scale model walls Permeability coefficient k (cm/sec) 2.34×10-5
The model walls were built in the City of Toma- Not
Maximum dry density Udmax (g/cm3)
komai, Hokkaido Prefecture, with the soil determined
materials shown in Table 1 for the purpose of Cone index qc (kN/m2) 3000 or larger
㻠㻜 㻞㻜㻜
㻠㻜 㻞㻜㻜
㼙㼑㼍㼚㻌㼠㼑㼙㼜㼑㼞㼍㼠㼡㼞㼑㻌㻔䉝䡡㼐㼍㼥㼟㻕㻌㻌
㻯㼡㼙㼡㼘㼍㼠㼕㼢㼑㻌㼠㼛㼠㼍㼘㻌㼛㼒㻌㼠㼔㼑㻌㼐㼍㼕㼘㼥㻌
㻯㻭㻿㻱㻝㻔㼟㼠㼍㼚㼐㼍㼞㼐㻌 㼙㼑㼠㼔㼛㼐㻕 㻯㼡㼙㼡㼘㼍㼠㼕㼢㼑㻌㼠㼛㼠㼍㼘㻌㼛㼒㻌㼠㼔㼑㻌㼐㼍㼕㼘㼥㻌㼙㼑㼍㼚㻌
㼙㼑㼍㼚㻌㼠㼑㼙㼜㼑㼞㼍㼠㼡㼞㼑㻌㻔䉝䡡㼐㼍㼥㼟㻕㻌㻌
㻯㼡㼙㼡㼘㼍㼠㼕㼢㼑㻌㼠㼛㼠㼍㼘㻌㼛㼒㻌㼠㼔㼑㻌㼐㼍㼕㼘㼥㻌
㻰㼑㼜㼠㼔㻌㼛㼒㻌㼒㼞㼑㼑㼦㼕㼚㼓 㼠㼑㼙㼜㼑㼞㼍㼠㼡㼞㼑
㻰㼑㼜㼠㼔㻌㼛㼒㻌㼒㼞㼑㼑㼦㼕㼚㼓㻌㼕㼚㻌㼠㼔㼑㻌
㻟㻜 㻜㻟㻜 㻜
㼑㼙㼎㼍㼚㼗㼙㼑㼚㼠㻔㼏㼙㻕
㻞㻜 㻙㻞㻜㻜
㻞㻜 㻙㻞㻜㻜
㻝㻜 㻝㻜
㻙㻠㻜㻜 㻙㻠㻜㻜
㻯㻭㻿㻱㻞
㻜 㻜
㻙㻢㻜㻜 㻙㻢㻜㻜
㻜㻤㻛㻝㻜 㻜㻥㻛㻝㻜 㻝㻜㻛㻝㻜 㻝㻝㻛㻝㻜 㻝㻞㻛㻝㻜 㻜㻤㻛㻝㻜 㻜㻥㻛㻝㻜 㻝㻜㻛㻝㻜 㻝㻝㻛㻝㻜 㻝㻞㻛㻝㻜
㼀㼕㼙㼑㻌㼛㼒㻌㼙㼑㼟㼡㼞㼑㼙㼑㼚㼠㻔㼥㼑㼍㼞㻌㻛㻌㼙㼛㼚㼠㼔㻕 㼀㼕㼙㼑㻌㼛㼒㻌㼙㼑㼟㼡㼞㼑㼙㼑㼚㼠㻔㼥㼑㼍㼞㻌㻛㻌㼙㼛㼚㼠㼔㻕
㻠㻜 㻞㻜㻜
㻠㻜 㻞㻜㻜
㻯㻭㻿㻱㻠
㼙㼑㼍㼚㻌㼠㼑㼙㼜㼑㼞㼍㼠㼡㼞㼑㻌㻔䉝䡡㼐㼍㼥㼟㻕㻌㻌
㻯㻭㻿㻱㻟㻔㻮㼛㼐㼥㻌㼛㼒㻌㼓㼛㼡㼞㼐㻌㼟㼠㼞㼡㼏㼠㼡㼞㼑㻕
㼙㼑㼍㼚㻌㼠㼑㼙㼜㼑㼞㼍㼠㼡㼞㼑㻌㻔䉝䡡㼐㼍㼥㼟㻕㻌㻌
㻯㼡㼙㼡㼘㼍㼠㼕㼢㼑㻌㼠㼛㼠㼍㼘㻌㼛㼒㻌㼠㼔㼑㻌㼐㼍㼕㼘㼥㻌
㻯㼡㼙㼡㼘㼍㼠㼕㼢㼑㻌㼠㼛㼠㼍㼘㻌㼛㼒㻌㼠㼔㼑㻌㼐㼍㼕㼘㼥㻌
㻰㼑㼜㼠㼔㻌㼛㼒㻌㼒㼞㼑㼑㼦㼕㼚㼓㻌㼕㼚㻌㼠㼔㼑㻌
㻟㻜 㻜㻟㻜 㻜
㼑㼙㼎㼍㼚㼗㼙㼑㼚㼠㻔㼏㼙㻕
㻞㻜 㻙㻞㻜㻜
㻞㻜 㻙㻞㻜㻜
㻝㻜 㻙㻠㻜㻜
㻝㻜 㻙㻠㻜㻜
㻜 㻙㻢㻜㻜
㻜 㻙㻢㻜㻜
㻜㻤㻛㻝㻜 㻜㻥㻛㻝㻜 㻝㻜㻛㻝㻜 㻝㻝㻛㻝㻜 㻝㻞㻛㻝㻜 㻜㻤㻛㻝㻜 㻜㻥㻛㻝㻜 㻝㻜㻛㻝㻜 㻝㻝㻛㻝㻜 㻝㻞㻛㻝㻜
㼀㼕㼙㼑㻌㼛㼒㻌㼙㼑㼟㼡㼞㼑㼙㼑㼚㼠㻔㼥㼑㼍㼞㻌㻛㻌㼙㼛㼚㼠㼔㻕 㼀㼕㼙㼑㻌㼛㼒㻌㼙㼑㼟㼡㼞㼑㼙㼑㼚㼠㻔㼥㼑㼍㼞㻌㻛㻌㼙㼛㼚㼠㼔㻕
㻠㻜 㻞㻜㻜
㼙㼑㼍㼚㻌㼠㼑㼙㼜㼑㼞㼍㼠㼡㼞㼑㻌㻔䉝䡡㼐㼍㼥㼟㻕㻌㻌
㻯㼡㼙㼡㼘㼍㼠㼕㼢㼑㻌㼠㼛㼠㼍㼘㻌㼛㼒㻌㼠㼔㼑㻌㼐㼍㼕㼘㼥㻌
㻯㻭㻿㻱㻡
㻰㼑㼜㼠㼔㻌㼛㼒㻌㼒㼞㼑㼑㼦㼕㼚㼓㻌㼕㼚㻌㼠㼔㼑㻌
㻟㻜 㻜
㼑㼙㼎㼍㼚㼗㼙㼑㼚㼠㻔㼏㼙㻕
㻞㻜 㻙㻞㻜㻜
㻝㻜 㻙㻠㻜㻜
㻜 㻙㻢㻜㻜
㻜㻤㻛㻝㻜 㻜㻥㻛㻝㻜 㻝㻜㻛㻝㻜 㻝㻝㻛㻝㻜 㻝㻞㻛㻝㻜
㼀㼕㼙㼑㻌㼛㼒㻌㼙㼑㼟㼡㼞㼑㼙㼑㼚㼠㻔㼥㼑㼍㼞㻌㻛㻌㼙㼛㼚㼠㼔㻕
4. TEST RESULTS
㻠㻜
㻹㼍㼤㼕㼙㼡㼙㻌㼐㼑㼜㼠㼔㻌㼛㼒㻌㼒㼞㼑㼑㼦㼕㼚㼓㻌㼕㼚㻌㼠㼔㼑㻌
㻯㻭㻿㻱㻝 㻯㻭㻿㻱㻞
㼑㼙㼎㼍㼚㼗㼙㼑㼚㼠㻌㼛㼒㻌㼙㼛㼐㼑㼘㻌㼣㼍㼘㼘㻌㻔㼏㼙㻕
collapsed.In and after 2009, the frost penetration construction.The data on maximum frost
depth was smaller than for standard geotextile- penetration depths in the embankments shown in
reinforced soil walls (i.e., Case 1). But the Figure 4 suggest that the wall surface material is
insulating materials used in Case 3 were proven displaced on a larger scale for greater frost
to be only marginally effective because the penetration depths.
difference from Case 1 in frost penetration depth The maximum annual displacements are
was only about 4cm. In Cases 4 and 5, a 7mm- shown in Figure 7. In Case 3, expanded metal as
thick drainage mat was placed in standard a wall surface material and insulating materials
geotextile-reinforced soil walls to control frost were applied in 2009. Except for Case 2, the
heave. In 2008, the frost penetration depths in largest value of the maximum annual horizontal
these two cases were much greater than the frost displacement was observed in the first freezing
penetration depth in Case 1. In and after 2009, season after the construction of the walls.
when an insulating material was applied to the Through the testing period, the value of the
reinforced soil walls, the frost penetration depths maximum annual horizontal displacement varied
were smaller than in Case 1. Over the four years depending on the frost penetration depth. In
after the application of insulating materials, frost Case 3, displacement was not effectively
penetration depths in Cases 3, 4 and 5 were controlled.Except for Case 4 in 2012, application
smaller in comparison with Case 1; thus, it can of insulating materials helped to control
be concluded that the insulating materials used displacement. Changes in displacement will be
are effective in preventing frost heave for at least monitored.
4 years after application.The effectiveness of the
insulating materials will be monitored over a 㻯㻭㻿㻱㻝 㻯㻭㻿㻱㻞 㻯㻭㻿㻱㻟
long period. 㻡 㻯㻭㻿㻱㻠 㻯㻭㻿㻱㻡 㻟㻜㻜
㻰㼕㼟㼜㼘㼍㼏㼑㼙㼑㼚㼠㻔㼏㼙㻕
㻞㻡㻜
㻯㼡㼙㼡㼘㼍㼠㼕㼢㼑㻌㼙㼑㼍㼚㻌㼐㼍㼕㼘㼥㻌
㼠㼑㼙㼜㼑㼞㼍㼠㼡㼞㼑㻔䉝䡡䡀䠽䡕䡏㻕
㻠
4.2. Horizontal displacement of the wall surface 㻯㼡㼙㼡㼘㼍㼠㼕㼢㼑㻌㼙㼑㼍㼚㻌 㻞㻜㻜
material 㻟 㼐㼍㼕㼘㼥㻌㼠㼑㼙㼜㼑㼞㼍㼠㼡㼞㼑
㻝㻡㻜
Figure 6 shows the horizontal displacement of 㻞
㻝㻜㻜
the wall surface material in each case.When the 㻝 㻡㻜
walls began to freeze, the wall surface material
began to be displaced outward. When the 㻜 㻜
embankment material completely thawed, the 㻞㻜㻜㻣 㻞㻜㻜㻥 㻞㻜㻝㻝 㻞㻜㻝㻟
㼀㼕㼙㼑㻌㼛㼒㻌㼙㼑㼍㼟㼡㼞㼑㼙㼑㼚㼠㻌㻔㼥㼑㼍㼞㻕
wall surface material was displaced slightly
inward. Outward displacement took place again Figure 7 The maximum annual displacements
in the following freezing season. The process of
outward and inward displacement was repeated
annually over the course of five years, resulting
in cumulative residual displacement outward. In 4.3. Strain in the reinforcing material
Case 1, the model walls were displaced outward
by 8 cm over the course of the five years after Strain in the reinforcing material (i.e.,
geotextile) in Case 1 is shown in Figure 8.
Except at the measurement point 140cm from
㻤 㻯㻭㻿㻱㻝
㻣 㻯㻭㻿㻱㻞 㻡㻌
㻯㻭㻿㻱㻟 㻰㼕㼍㼚㼏㼑㻌㼛㼒㻌㼜㼘㼍㼏㼑㻌㼟㼚㼐㻌㼠㼔㼑㻌
㻰㼕㼟㼜㼘㼍㼏㼑㼙㼑㼚㼠䠄㼏㼙䠅
㻢
㻯㻭㻿㻱㻠 㻠㻌 㼜㼛㼕㼚㼠㻌㼛㼒㻌㼙㼑㼍㼟㼡㼞㼑㼙㼑㼚㼠
㻡 㻯㻭㻿㻱㻡
㻟㻌 㻝㻜㼏㼙 㻟㻜㼏㼙 㻡㻜㼏㼙
㻠
㻿㼠㼞㼍㼕㼚㻌㻔㻑㻕
the reinforced soil wall surface, the strain geotextile-reinforced soil wallis as effective as
increased in the freezing season and slightly soil replacement with gravelly soil in
decreased in the thawing season, but thestrain embankments, because the strain isroughly the
never decreased to 0. Strain in the following same for these two methods.
freezing season added to the residual strain. In
Case 1, the greatest amount of strain was 5. SUMMARY
measured at a point 30cm from the reinforced
soil wall surface, and the second- and the third- This study focused on methods for controlling
greatest amounts of strain were measured at deformation due to frost heave in geotextile-
points 50cm and 70cm, respectively. At the reinforced soil walls in service in cold regions
measurement point 90cm from the reinforced such as Hokkaido. Specifically, materials that
soil wall surface, the strain remained almost provide insulation and are suitable to be planted
constant for the 3 years after the construction of on were attached to the outer surface of
the walls, but relatively large strain occurred in reinforced soil walls, and the effectiveness of
the 4th year. At the measurement point these materials in controlling deformation due to
only10cm from the wall surface, the strain frost heave were examined. Although the
decreased to 0 before the beginning of the effectiveness was found to varywith the
freezing season every year. insulating material used, it was determined that
A similar tendency was confirmed in the other the application of an insulating material to the
cases, although the strain varied and the distance outer surface of a reinforced soil wall helps to
from the wall for the greatest strain was control deformation caused by frost heave.
mutually different for each case. In Case 1, the Insulting material can be easily applied toa wall
strain at the point 30cm from the wall surface surface. Geotextile-reinforced soil walls are
was the greatest (3.9%). The greatest strain in designed to be planted on, and plants can grow
the other cases occurred at various distances on the attached insulating material. Over time,
from the wall surface. growing plants provide additional insulation8) 9),
Figure 9 shows the time-dependent change in thereby suppressing deformation further.
the strain at the point where the greatest strain
was measured in each Case. The largest strain 6. REFERENCES
was measured in Case 1 (i.e., standard
geotextile-reinforced soil walls) and the second- Hokkaido Regional Development Bureau, Road
largest strain was measured in Case 3 (i.e., Design Guidelines
geotextile-reinforced soil walls with an http://www.hkd.mlit.go.jp/zigyoka/z_doro/download/
insulating material of mesh structure). The strain pdf/01/1-7-1.pdf
was relatively small in Case 2, in which soil was Japan Road Association, 2009, Road Earthwork
replaced with gravelly soil, and in Cases 4 & 5 Guideline
in which nonwoven textile was used for Ono Takashi, 2012, Effects of Vegetation in the Frost
insulation. Heave Control, Collection of Lectures, the 47th
It is concluded that the application of Workshop of The Japanese Geotechnical
insulating materials to the outer surface of the Society,.7
Sato Atsuko, Nishimoto Satoshi, Suzuki Teruyuki,
and Nakamura Dai, 2010, Insulating Effect of
㻡㻌 Vegetation on Slopes, the 65th Annual Meeting,.9
㻯㻭㻿㻱㻝㻔㻟㻜㼏㼙㻕 Sato Atsuko, Nishimoto Satoshi, and Tatta Naoki,
㻠㻌 㻯㻭㻿㻱㻞㻔㻣㻜㼏㼙㻕 2012, Elucidation of and Measures against the
㻯㻭㻿㻱㻟㻔㻣㻜㼏㼙㻕
㻟㻌 Factors Involved in the Deformation of
㻿㼠㼞㼍㼕㼚㻌㻔㻑㻕
㻯㻭㻿㻱㻠㻔㻣㻜㼏㼙㻕
㻞㻌 㻯㻭㻿㻱㻡㻔㻡㻜㼏㼙㻕 Geotextile-reinforced Soil Walls Due to Frost
Heaving, the 55th Hokkaido Development
㻝㻌 Technology Conference, Hokkaido Regional
㻜㻌 Development Bureau, Ministry of Land,
Infrastructure, Transport and Tourism
㻙㻝㻌
㻜㻤㻛㻝㻜 㻜㻥㻛㻝㻜 㻝㻜㻛㻝㻜 㻝㻝㻛㻝㻜 㻝㻞㻛㻝㻜
The Japanese Geotechnical Society, 1994, Soil
Freezing - Its Mechanism and Actual Conditions
㼀㼕㼙㼑㻌㼛㼒㻌㼙㼑㼍㼟㼡㼞㼑㼙㼑㼚㼠㻔㼥㼑㼍㼞㻌㻛㻌㼙㼛㼚㼠㼔㻕
The Japanese Geotechnical Society, 2009, Frost Heave
Figure 9 Change in maximun Strain in respective cases Test Methods for Determining the Amount of
352 A. Sato et al. / Frost Protection for Geotextile-Reinforced Soil Walls in Service
Abstract. In the present paper, an attempt is made to monitor the behaviour of steel pipe pile on clayey slope while excavating
for a basement of a building. The basement excavation is supported by RCC diaphragm wall, which gets the lateral support from
temporary steel pipe piles to enable top-down construction in the peripheral area, which are connected to the diaphragm wall
with basement floor slabs. The horizontal displacement of the piles is measured based on inclinometer tubes embedded in the
diaphragm wall. A detailed finite element analysis has been carried out using PLAXIS 3D model and the theoretical results are
compared with the field observations.
Keywords. diaphragm wall, steel pipe pile, clayey slope, excavation, building basement, numerical soil-structure-interaction
analysis
of the basement was opted to overcome the slab. The soil was compacted, over which M10
instability problem during excavation. The PCC of 150 mm thickness was laid.
diaphragm wall along with two rows of The floor slab of 200 mm thick in the first
sacrificial pile is installed along the periphery of basement was cast and once the slab concreting
the construction area in order to support the wall was finished and cured the excavation of the
and to avoid any strut or anchors. The remaining basement below slab was done by mechanical
central area was constructed with open excava- means. Subsequently the entire slab is cast and
tion using the bottom-up method with an open the sacrificial piles supported the lateral loads
excavation. through the basement slab.
The first basement was cast only for a limited
width of 12 m covering the two sacrificial piles
and a construction joint was left to extend the
slab further later. The centeral portion without
floor slab was excavated to the next basement
floor level of - 5.5 m as an open excavation.
The excavation then progressed under the
slab already cast with support from sacrificial
pile as shown in Figure 2. In addition to the
vertical weight of the floor slab, the sacrificial
pile is also subjected to lateral load due to the
deformation of the diaphragm wall towards the
excavation. The behaviour of the piles and the
wall is analyzed using the FEM and the
deflection of the diaphragm wall obtained is
compared with the actual deflection measured
using an inclinometer. The input parameters of
soil have been varied till the theoretical
Figure 1. Typical Bore- Log Data deflection matches with the measured deflection.
At this stage, the shear in pile on slope and
Diaphragm wall of 800 mm thickness and corresponding deformations are noted.
length varying from 17 to 23 m is constructed
initially throughout the perimeter of the
construction area. Sacrificial piles of 860 mm
diameter were then driven in the specified
location adjacent to the wall. The displacement
in the diaphragm wall panel was measured at
selected location using inclinometer.
It is assumed that the displacement in the di-
aphragm wall is equal to that of sacrificial pile,
because of the rigidity of the floor slabs. The
typical reading for one panel was taken and Figure 2. During excavation
analysed. Figure 2 shows excavation beneath the
floor slab. 3. Geotechnnical and Structural Data
a. 3D view
356 S.V. Sivapriya and S.R. Gandhi / Evaluation of Pile Behaviour on Sloping Clayey Soil
a. Initial Stage
d.Stage V
c.Stage IV
S.V. Sivapriya and S.R. Gandhi / Evaluation of Pile Behaviour on Sloping Clayey Soil 357
Depth (m)
10 Stage I
Stage II
12 Stage III
14 Stage IV
Stage V
16
Figure 5. Comparison of measured and observed
displacement reading 18
a. Rear Pile
Bending Moment (kN.m)
-2 0 2 4
2
8
Depth (m)
10 Stage I
Stage II
12 Stage III
Stage IV
Figure 6. Load shared by the piles due to excavation 14 Stage V
The bending moment in the rear pile is more 16
due to slope effect. This is mainly due to the
fixity and the reduction in the resistance in the 18
rear pile. The bending moment of the pile was b. Front Pile
within the safe limit. The bending moment of the Figure 7. Bending moment in pile due to excavation
front and rear piles is shown in Figure 7.
6. Conclusion
7. References
Wong, I.H. and Poh, T.Y. 1999. Comparison of retaining
walls for basement construction in stiff clays. Tunelling
and Underground Space Technology,Vol. 14, No.4, pp.
461-468.
Muthukkumaran, K., Sundaravadivelu, R. and Gandhi, S.R.
2004 a Monitoring of lateral deflection in a berthing
structure during dredging- A case study. 5th
International Conference on Case histories in
geotechnical engineering, paper No.5.35
Muthukkumaran, K., Sundaravadivelu, R. and Gandhi, S.R.
2004 b Effect of slopping ground on single pile load de-
flection behaviour under lateral soil movement.13th
World conference on earth quake engineering, Canada,
1-6.
Ong, D. E. L., Leung, C. F. and Chow, Y. K. (2009) Behavior
of Pile Groups Subject to Excavation-Induced Soil
Movement in Very Soft Clay. Journal of Geotechnical
and Geoenvironmental Engineering,ASCE, Vol. 135,
No. 10, pp.1462-1474.
Stuedlein, A.W., Bailey, M., Lindquist, D., Sankey, J. and
Neely, W.J. 2010. Design and performance of a 46 m
high MSE wall. Journal of Geotechnical and
Geoenvironmental Engineering, ASCE,Vol. 136 , No.6,
pp 786 – 796.
Soil-Structure Interaction, Underground Structures and Retaining Walls 359
V.M. Ulitsky et al. (Eds.)
IOS Press, 2014
© 2014 The authors and IOS Press. All rights reserved.
doi:10.3233/978-1-61499-464-0-359
Abstract. The performance of soil-cement column (SCC) retaining wall system with top-down construction method was
observed and present in this paper. From the observation data, the 3D finite element model is used for back calculation to obtain
appropriate stiffness parameter for ground movement prediction during SCC installation stage and during excavation stage. The
suitable stiffness parameter for prediction of ground movement for soft clay is 50Su and 500Su at the SCC installation stage and
at excavation stage respectively. The influence of stiffness on lateral force is also reported.
Keywords. soil-cement column retaining wall, monitoring, instrumentation, back analyses, stiffness parameters
proposed. The SCC wall must be design as Figure 2 is the cross section of the site. The
gravity wall so it can resist earth pressure with- maximum excavation level is -7.9m close to
out any horizontal support. The gravity wall must public road. On the opposite side, the excavation
be thick enough to resist overturning, sliding, level is -6.3m. There are two basement, B1 and
bearing failure. However, at the detail design B2. The basement slab level is not equal, B1-A
stage, the space for thick gravity wall is found to slab is at level -2.90m and B1-B slab is at level -
be insufficient. The design has to be changed 4.50m. The excavation for B2 is unequal, for B1-
from SCC gravity wall to SCC wall with lateral A side the excavation level is -6.3 while the
support. The owner’s representative does not excavation level for B1-B side is -7.9m. As the
want to use temporary bracing as it spend extra construction is top-down method, the basement
time for installation and removal. Finally, the B1-A and B1-B which are the support to SCC
design team decide to use permanent basement wall must be constructed before foundation
as lateral support and construct basement from construction. Therefore, the slab must be sup-
top to bottom which is known as “top-down” ported with temporary stanchions which are pre-
construction technique. The combination be- installed and embedded into bored pile founda-
tween SCC wall and top-down construction has tion.
not been used before, therefore, the geotechnical Figure 3 is the subsoil profile from site inves-
design parameters (especially soil stiffness) are tigation. Subsoil consists of 2m thick fill soil
adopt from the observed performance of SCC overlying 11m thick very soft to softy clay.
gravity wall. Figure 1 is the layout of SCC wall Below this layer is 3m medium stiff clay. The
and foundation pile layout. Two sides of the soil below medium stiff clay is very stiff clay.
project are close to public road and another two
side close to two-storey and eight storey rein-
forced concrete building. The concrete block
fence with short pile foundation is on the bound-
ary line. The figure also indicate location of
temporary strut and inclinometers.
5.1. G und
Grou d mo
ovemennts
Th herre arre ttwo
o staagees of grrou und mo ovem mennt duri
d ing
coonsttrucctioon, firsst sttagee iss grroun
nd mo ovem mennt due
d
to
o SC CC installlatiion n annd ssecond d sttagee of grou
g und
m emeent due
move d to
o exca
e avaation n for bbaseemeent.
Fiigurre 7 is th he horrizo ontaal disp
d placcem mentt att SC
CC
in
nstaallattionn sttagee. The
T e m maxiimu um gro ounnd mov
m ve-
m t on
ment nA A-dirrection n whic
w ch perp p penndicculaar to
o wall
w
nmeent is -1
allign 1100mm m at 9m m dep
d pth. On O B-
diirecction, thee max
m imu um movem emeent is -7 70mmm.
Th he neegattivee sign
s n mea m ans th he gro ounnd mo ove
ouutwwardd thhe ex xcav vation arrea. Th his m moveemeent
caauseed by displacin ng of groun nd by 1.778 cub bic
m er of
mete o ccem mentt grroutt annd mix m xingg actioon of
o the
t
m ng toool.
mixin
Figu 6 In
ure 6. nstalllatio
on off SCC
C 20
00mm
m cllose to neigh
n hbor fencce.
4.. IInsttrum
menta
atioon pro
p gra
am
Inncliinommetterss were
w e in nstallled d onn fo
fourr sid de of thee
exxcaavattion n boun ndaary to mo onittor latteraal grou
g undd
m vem
mov mentt du ue to SC CC insstalllatiion an nd dur
d ingg
exxcaavattion n ass sh howwn in Figurre 1. To T meeasu uree
foorcee in
n temp poraary stru uts insstallled bettweeen B1 1-A
A
annd B B1--B, eleectrricaal sttrain n gaug ges aree attach hedd too
thhe neu utraal axis
a s off th he stru ut to avo oid beendingg
sttresss com
c mpo onen nt. Th he dumd mmy y sstraain gau ugee iss
usedd to eliimin natee teemp e ect. Ass thee sttrutt
peraaturre effe
iss suuppported witth w walle whiw ich is sim milaar to
t pin
pn
suuppportt; thhereeforre, thee beend dinggm mom mentt trranssferr Figure 7. Ground movement at SCC installation stage.
too strrut is min
m nim mum m.
T
The monim itorringg was w taaken n aat ever
e ry maajorr F ure 8 is the
Figu t seconnd stag ge gro ounnd mov
m ve-
chhannge in n sttagee oof con c nstru uctionn ass show
s wn inn mentts whi
m w ich come fro om
m th he exca
e avaationn of o soil
s
T le 1.
Tabl 1 TheT e sttarttingg day of moniitoriing g is att foor bas
b em mentt co onsstrucctioon. Thhe inittiall reeadiing
coomppletion n of
o SCC
S C in nstaallaation
n. FForrce in strrutss (zzero
o reeadding g) area seet tto thet daay whhen SC CC
arre rreadd affter thee strrutss arre in
nstaalled. in
nstaallattionn iss co omp pletedd. For
F I-4 4 siide att fin nal
T le 1 O
Tabl Observa
ation
n sch
hedu
ule. exxcav vattionn staagee, th
he wal
w ll mov
m vem mentt shhow w veery
Dayy A vity
Acti y smmalll cu urvvatu
ure whw ich meean ns th wall is tilt
hat thee w
lik
ke a bloc
b ck. Thee mov m vem op is 58m
mentt att thee to 5 mm
1000 Ex
xcavate to leevell -3..2m and
d coonstrruct baseemeent
annd the
t mooveemeent at thee bo otto 1 mm. This
om is 15m
1200 Ex
xcavate to leevell -4..8m and
d coonstrruct baseemeent m ns that
mean t t thhe walw ll mov
m vem mentt iss th
he ccom mbin na-
1277 Ex
xcavate to
t leevel -6.3
3m at
a miiddle area of
o thee sitte tio
on ofo slig
s ghtly y ovverrturn
n annd slid
dingg.
F incclin
For nom meteer I-22, the am mou unt off wall
w
1555 Ex
xcavate to
t leevel -6.3
3m unde
u er baasem
ment B1--A
m emeent is less
move l s thhan I-44 as thee deepthh off ex
xcav va-
1788 Ex
xcavate to
t leevel -7.9
9m unde
u er baasem
ment B1--B on is shaallowerr. The
tio T e m maxiimu um gro ounnd mov
m ve-
m t iss 400mm
ment m. Most off th he mov
m vem mennts occcur
w n th
when he exccav vatio on levvel reaachh -33.2m m befo
b ore
5.. O
Obsserv
vation
nal ressults coonsttrucct B B1--A. Affterr B1-A A iss coonsstruucted, the
t
in
ncreemeentaal mov
m vem mentt duue to
t furt
f therr exxcav vatiion
As ttherre aree in
A ncliinom metterss and
a strrain
n gaug
g gess
is veery small. Th his beccausse the
t waall is i supp portted
in
nstaalled
d in
n th
he site,, thee ob
bseervaationn are pres
p senntedd
att rellativelly shallow w deepthh, so
s the can ntileever ty
ype
in
n tw wo paarts first,, th he gro ound d mmov vemmennt and
a d
m emeent is rest
off move r trainnt. In addditio
on, thee exxcav va-
seecoond forrce in stru
s ut.
P. Tanseng and V. Namwiset / Performance of SCC Retaining Wall Used With Top-Down Construction Method 363
tion lev vel beetweeenn twwo sid des aree un neqquall. The
T e peecteed to be in ncreeasee ass thhe loaad on S3 3 will
w
eaarthh pressurre at
a II-4 sid de is
i grea
g aterr th han eaarthh reedisstrib
buteed to
t adja
a acen nt sstru
ut. How
H wev ver, thee sttrut
prresssuree at
a I--2 side
s e; ther
t refo ore, the grea
g aterr fo orcee lo
oad on S22 an nd S4S are noot in ncreeasee siignnificcanttly.
onn w walll at
a I-2
I traansffer thrroug gh baasem men nt and
a d Th he rea
r sonn off noo load inccreaasinng may
m y coomee fro om
pu w I-4
ushh the wall 4 inn thee op
ppoositee diirecction. he resiistaancee of
th o 150 0mm m plaain co oncrretee sllab
w ch cast
whic c t ag gainnst SCCC waall befo
b fore strrut S3 is
reemo ovedd.
-1 I-2 side
e
-5
-7
-9
0 50
0 100 150 20
00
60
40
30
I-2
2
20
I-4
4
10
0
0 50
0 100 150 20
00
8 Ob
F ure 8.
Figu bserv
ved grouund mov
m vemeent at
a exccavaation
n of the
t soil
s 50
S1
S
S
S5
S
S6
30 S
S7
m
movvem
mentt
S
10
exxcaavatted groug undd elev e vatiion aand max m ximu umm Figurre 9.. Meeasurred strut
s forcce an nd maxim
m mumm horrizon
ntal
grrouund mo oveemeent from incl i lino omeeterrs. It can
c bee m emen
move nt veersus tim
me.
seeenn thaat whe
w en thet exccavaatio on is
i mmadde at
a I--2 side
s e
too level -3..2mm thhe mmax ximu um gro ounnd mov
m vem men nt iss 6.. Bac
B k aanallysiis for
fo stif
s ffneess parram
metterss
25m mm. Th hen folllow w byb th he exccavatioon tot leve
l el -
4.8m m at a I-4
I sid de. After
A r exca e avaation n, thee wall
w l Afterr th
A he eexcavaatio on is ccommpleeted d, thhe oobsservved
m vem
mov mentt inccreaasee to 48mm m. Whe
W en thet ex xactionn diispllaceemeentss arre com mpaared d with
w h displlaceemeent
- 3m at I--2 sid
reeachh -6.3 de, thee m max xim
mum m wall
w l frrom
m FE EMM att th he dessignn sttagee. The
T e caalcu ulatted
m vem
mov mentt is nott inncreeasee. HowH wevver, thee ex xcavva-- grroun nd mooveemeent is 1.664 tim t me larg
l ger thaan the t
tion at I-4 4 side
s e too leeveel -7.9- 9m cau usee slligh
htlyy ob bserrveed oone. TheT im mporrtan f tor is tthe stiiff-
nt fact
inncreeasee in
n grrounnd m mov vemmen nt. neess paaram metters ussed inn FE EMM att desi
d gn staage
T
The fo orcee in
n teemp poraary strrut meeasu ured
d with
w h w ch is
whic i low wer thaan it i hhas to be.. Th herrefoore, in
sttrainn gaug
g ge sho owss th hat thee sttrut forrce increasee thhis ressearrch,, th b ck anaalyzzes by
he bac y uusing ob- o
w h th
with he exca
e avaationn deptd th. Th he mmaxximmum m sttrutt seerveed groound d mov
m vem mentt arre perf
p form medd. InI the
t
foorcee off all seevenn sttrutts iss 45
50 kN
k perr strrut wh hichh baackk anal
a lysiis, 3D D finnite eleme men nt pro ograam
iss looweer than
t n th he preedicctio on witth FEMF M (FE EMM PL LA
AXIS S 3D Tun T nneel iss ussed.. Thhe mod
m del paaram me-
prreddictiion is 678 kkN)). The T ob bserrvedd max
m ximu umm teer arre sim
s ilarr to thee on ne uusedd in
n thhe desi
d ign stag ge.
sttrutt fo
orcee iss lo
oweer th hann thhe pre
p dicted d on ne be-
b - Th he wor
w rk sstarrt with
w h sellecttionn off suuitabble Yo oung g’s
caausse of
o th he con ntacct betw
b weeen bra b acin
ng and
a baase-- m ulu
modu us. T Thee ap pprroxiimaatio
on of s ear straain in
o she
m t sllab is not
ment n cloose.. Th heree arre 5 to
o 100mm m gap
g p sooil due
d e too ex xcav vationn is doone by assum ming th hat
leeft beetween n base
b emeent slab aand teemp poraryy thhe disttorttion a le of the gro
n angl g und d ffrom m thet
brracing g; th
hereeforre, wh w mo
hen thee wall ovee to
o cloosee veertical axxis is t app
i the proximmatee sh heaar sttraiin. By
thhe ggap earrth preessuure red ducee annd aalso
o strut forrce.. asssummin ng thiss, the t sh p filee caan be
hearr sttraiin pro
T gaap is
The i notn moodeelled d in F M, so thee sttrutt
n FEM pllotted as sho own n in
n Fiigurre 10
1 anda d Fiigurre 11 for
foorcee frrom
m FE EM is hhig t observ
gherr thaan the ved
d onne. SC CCC in nstaallaationn stag
s ge an nd exccavatioon staage
The strrut S3
T 3 iss reemo oveed first
f t fo w deningg
or wid reespeectiivelly. Th he av veraage ap ppro oximmat
ate sheear
w king
work g area
a a. The
T lo oad on n S2 2 aand S4 4 arre ex-- sttrain
n ata SCC installlatioon staage iss 1 peerceent
364 P. Tanseng and V. Namwiset / Performance of SCC Retaining Wall Used With Top-Down Construction Method
In
n bac
b ck anaalyssis prroceedu ure, Eu su =25 50,
5000, annd 7750 0 was trieed aas inp
i ut parp rammeteers for
so
oft clay y too FEM
F M. The
T e grrouund mo oveemeentss fro om
eaach triial aree co om
mparred wiith observ vedd grou
g und
m emeent an
move nd choo
c osee thhe best
b t fitt onne ffor suiita-
blle stiff
s ffness parram meteers. Figurre 131 is tthe baack
annalyysiss rresu f m FE
ults from EM co omp parred with
w
obbserrveed ddataa. From
F m thet fig guree, You
Y ung’’s mod m du-
lu
us para
p ameeterr Eu su of o 55000 givvess th
he bbestt fitt to
obbserrvaationn mov
m vemmentt. T
The value is higgherr th han
esstim
matiion on ne froom sellf-bborinng pressureemeeter
Figure 10. Shear strain profile: SCC insstallation stage.
daata thhis maay be th he rresuult off so oil--stru
uctuure
nteractiionn in FE
in EM.
A the
As t SC CC waall tipp is no ot emb
e beddded d innto
sttiff claay layyer; theerefforee, thet stiiffn
nesss off sttifff
Figure 11. Shear strain profile: excavation for basement stage. cllay maay nnott play any y roole.. How
H weveer, thee baack
annalyysess arre also
a o do one to verrifyy th
hat the
t stiffneess
off thhe clay
c y is inssen nsitiive to thee moveemeentss. The
T
Eu su off stiff clay c y at 7550, 100 00, andd 155000 weere
triied as inpput stiffness. Fiigu ure 14 sho owss thhat the
t
caalcuulatted grooun nd mov
m vem mentt is inssensitiive to the
t
sttiffn
nesss off sttiff claay. Thiis prov
p ved d that tthe excca-
P. Tanseng and V. Namwiset / Performance of SCC Retaining Wall Used With Top-Down Construction Method 365
vaatioon doe
d es not
n havve any
y in
nflu
uencce on
o stifff clay
c y giive thee bbestt fitt off grrouund mooveemeent betweeen
laayerr. FEEMM an nd ob bserrvattionn foor A-ddireectiion.. For B-
diirecction, thee grrou und mooveemeent forrm FE EM is
lo
oweer th hann th
he obso serv
ved on ne. Thhis maym y be
b be-
b
caausee of i epeendeencce stifffneess off
o thee sttresss ind
M hr-C
Moh Coullom m del. Thhe stifffneess for
mb mod f sheeariing
paath wh o cur alo
hichh occ ongg B-di
B recction n iis high
h her
th
han it shoould d be.
b Ho oweeverr, th
he gro ounnd mov
m ve-
m t in
ment n A--dirrecttion
n is moore imp nt thhan in B-
porrtan
fore, th
diirecction; ttherefo he Eu su off 50 0 iss su
uitab
ble
fo
or prac
p cticaal purp
p pose.
1 Com
F ure 14.
Figu C mpariison of FEM
F M and d obsserveed reesultts by
y usin
ng
Eu su =5500 for soft
s clay d varry Eu su of stifff clay
y and y at
000, and
7500, 10 a 1500 0.
In case
c e of o groound d mov
m vem mennt duri
d ing SC CCC
coonsstructio on, th he app proxximmatee ssheaar straain iss
abbouut 1 peerceent.. Inn thhis case
c e th he sstifffneess of thee
sooil is low
l wer thaan tthe stiiffness ussed forr ex xcav
va--
tion anaalyssis beccause larg l ger sheear strrain n occcurr inn
th T baack anaalyssis is donne by
he ssoill. The b app plyingg
vo oluumeetricc strrain
n too so
oil elem
e men nt w
wheere thee SC CCC
iss loocatted. The
T am
mou unt off ap ppliied vo olummettricc Figure 16. Back analysis result: SCC installation stage.
sttrainns are calcc culaated d from
f m SC CC installlationn
reecorrd as sh how wn iin Fig guree 15. Th he soft
s ft clay
c y
p ameeter Eu su of 50,, 10
sttiffnness para 00, annd 150
1 0
arre uused for
f inp put parram meteer for
f FEM M. Th he soil
s l T s reeseaarch
This h also
a o sttudyy influ
i uen nce off so oil-
coondditioon is a umeed to be undrrainned duee to
i assu o sh
hortt sttiffn
nesss oon forrce in sttrut. In n the
t stuudy
y, the
t
peeriood of
o con
c stru uctiion.. Th
he shea
s ar sstren ngtth of
o sooil-- Eu su off 250, 500 5 0, an nd 7500 arre used
u d. F ure 17
Figu
ceemeent slu urryy is asssumme to o be 5 kP
Pa. is thhe plot
p t beetw ween n strutt fo
orcee from
fr m F FEMM and
a
obbserrveed sstruut forc
f ce. It caan be b seeen thaat the t
Soil Injectted Voluumetrric caalcuulatted strrut forrcess reedu
ucess ass thhe Yo oungg’s
vvolum
me grouut sttrain m ulu
modu us inncrreasse. Thiis sstud
dy tryy too obbtaiin the
t
( )
(m^3) (m^33) (
(%)
0
0.77 0.13 16.90 su ffness paaram
uitaablee sstiff metter forfo m moveemeent
prrediiction;; therefore, forf thee prropo osed Eu su =
50 s ft cllay, it is on
00 of soft o tthe connseervaativve side as
th s ut fforcce from
he stru m FE EM aree laarger thaan the t
3
3.08 1.07 344.90 obbserrveed oone..
9. References
Abstract. At present, due to the need for engineering works of large scale, we have the challenge of solving problems related to
deep excavations, where instability problems often occur. In this paper is presented the analysis and validation of the excavation
of the service well PS – 39, of the executive project of collectors for treatment of sanitary sewage, located in the city of Rio de
Janeiro. This well was constructed using a contention system of a wall of secant column, ring structure type of “shotcrete”
covering the columns, and a slab of "jet grouting" anchored with passive tiebacks. The soil parameters of the different layers that
compose the geotechnical profile, necessary for numerical modeling, were determined based on the geotechnical information
available of the field instrumentation, laboratory data and specific literature review, considering the Mohr-Coulomb constitutive
model. The results and interpretation, of the stresses and displacements during the excavation process, were determined using the
PLAXIS 3D (2012) program. These results were validated with the measurements determined by instrumentation; then the
numerical model will predict rationally the behavior of the set soil-structure during the excavation process. This allows to solve
similar problems in advance, allowing proper planning.
Keywords. deep excavation, numerical modeling, Mohr-Coulomb constitutive model, field instrumentation data
The structure of primary containment consists For analysis of the results is available the reg-
of the secant columns, which have a diameter of isters of measurements for the last stage of
0.80 m, and are superimposed 0.20 m, in alter- excavation, of the horizontal displacements
nate columns were positioned 12 reinforcing bars registered with the inclinometer I-01. Also is
(ø= 25 mm), these columns were performed with available the registers of piezometric heads,
the technique of "wet -mixing" around the determined during the period of excavation
perimeter of the well to a depth of 19 m, which (consisting of 18 stages of excavation), regis-
was modeled with a solid element with linear tered with the piezometer P-03, at an average
elastic behavior. The second containment struc- depth of 13.50 m, corresponding to the layer of
ture is formed by a ring of shotcrete with 0.20 m silt.
of thickness, and built as the excavation pro- The parameters used for modeling, according
gressed, this structure was modeled by plate to the model of Mohr Coulmb were determined
elements with elastic behavior. The third struc- based on the results of field and laboratory test,
tural element was a base plate with thickness of 3 data from field instrumentation, research results
m, made with 67 columns of "jet -grouting" with with similar soils in different areas of the city of
nominal diameter of 1.20 m and 32 columns with Rio de Janeiro, and specific literature review.
nominal diameter of 1.50 m, this plate was These parameters allow to validate the model
anchored with 25 passive tiebacks in the founda- with measurements from field instrumentation.
tion ground. The jet-grouting plate was modeled The parameters used in the modeling are pre-
with a solid element, and passives tie-backs with sented in the Table 1(a) and Table 1(b).
a combination of embedded piles and spring
connections.
J.C.S. Tito and C. Romanel / Numerical Analysis 3D of a Deep Circular Excavation in the City of Rio de Janeiro 369
E9_Active -5
-10
-15
-20
-300 -200 -100 0 100 200 300
Steady seepage pore water pressure (kPa)
Figure 5. Evolution of the pore pressure on the active side
and the passive side.
literature review, were obtained the values used analysis of two and three dimensions, constitut-
in the 3D numerical analysis. ing a tool to make a lot of analysis, preventing
Once established the values of the soil pa- the occurrence of failures. This tool allows to
rameter, of the four layers corresponding to the determine an approximate representation of
stratigraphic profile of the well PS-39, required reality by modeling, using numerical analysis.
by the model Mohr Coulomb and using the Also allowing the drafting and designing of
PLAXIS 3D program, the maximum displace- quality projects, obtaining results with adequate
ments were determined in the control line. In the speed to the requirements and possibly offering
last stage of excavation, the maximum horizontal better cost, therefore constitutes a competitive
displacement has a value of 0.0143 m at eleva- tool.
tion -5.5 m, representing a ratio of the horizontal
displacement/ wall height of 0.000918. And 5. References
according to the inclinometer readings, the
maximum horizontal displacement at the eleva- Almeida, M. S. S., Futai, M. M. and Lacerda, W. A. 2008.
tion of -7.2 m, has a value of 0.0147 m. The Laboratory Behaviour of Rio de Janeiro Soft Clays - Part
1: Index and Compression Properties. Soils & rocks, Vol.
difference between the displacements determined 31, pp. 69-75.
with the program and field measurements is Barrios, J. E. M. et al. 1997. NOTAS. Excavaciones en
0.000403 m, representing a percentage of devia- condiciones complejas. Bogotá: Editorial Escuela Co-
tion of 2.74 % relating to the horizontal dis- lombiana de Ingeniería.
placement registered in the inclinometer. Brinkgreve R. B. J., Engin E. and Swolfs W. M. 2012.
PLAXIS 3D. Material Models Manual 2012. Holanda:
Concerning to the earth pressures acting on Plaxis bv.
the secant columns, of the active horizontal Christian, J. T. and Wong, I. H. 1973. Errors in simulating
pressure side. For the elevation of - 13.2m excavation in elastic media by finite elements, Soils and
(corresponding to the bottom of the excavation) Foundations. The Japanese Geotechnical Society, Vol.
the horizontal effective stress is 246.87 kN/m2, 13, No. 1, pp. 1-10.
Cintra, J. C. A., Aoki, N. 2010. Fundações por estacas,
for the elevation of -16.2 m (corresponding to the projeto geotécnico. São Paulo: Oficina de Textos.
base of the secant column) the stress is 189.21 Craig, R. F. 2004. Mecânica dos Solos, Sétima Edição. Rio
kN/m2. For the passive horizontal pressure side, de Janeiro: LTC – Livros Técnicos e Científicos Editora
the horizontal effective stress for the elevation of Ltda.
2.80 m (corresponding to the top of the secant Duncan, J. M., and Clough, G. W. 1971. Finite Element
Analyses of Port Allen Lock, Journal of the Soil Me-
column) is 1.439 kN/m2, and for the elevation of chanics and Foundation Division, ASCE, SM8, pp.1053-
-16.2 m (corresponding to the base of the secant 1068.
column) is 64.296 kN/m2. Formigheri, L. E. 2003. Comportamento de um Aterro Sobre
According to the 3D numerical analysis of Argila Mole da Baixada Fluminense
the excavation, considering steady state flow, the Dissertation - Departamento de Engenharia Civil,
Universidade Católica do Rio de Janeiro, Rio de Janeiro.
effect of excess pore pressures in the wall of Massad, F. 2005. Escavações a céu aberto em solos tropicais
secant columns is greater in the soil content side, Região Centro-Sul do Brasil. São Paulo: Oficina de
in the clay layer. This is due to the undrained textos.
behavior of these soils, the discharge process and More, J. Z. P. 2003. Análise Numérica do Comportamento de
the drawdown of the water table, generating Cortinas Atirantadas em Solos
-
Departamento de Engenharia Civil, Universidade
positive pore pressures (excess pore pressure of Católica do Rio de Janeiro, Rio de Janeiro.
suction) on the boundary of the excavation, Morgenstern, N. R. and Eisenstein, Z. 1970. Methods of
increasing the effective stress. In the establish- Estimating Lateral Loads and Deformations, Conference
ment of the boundary conditions for the analysis, Proceeding Paper ASCE Specialty Conference on Lat-
the inside of the well acts as a drainage system, eral Stresses in the Ground and Design of Earth-
Retaining Structures, pp.51-102.
generating a progressive drawdown of the water Palmer, J. H. L.V. and Kenney, T. C. 1972. Analytical Study
level, as the excavation progresses; contributing of a Braced Excavation in Weak Clay, Canadian Ge-
to the horizontal displacement, at the last stage of otechnical Journal, Vol. 9 (2), pp. 145-164.
the excavation, in the interface of the layers of Pereira, C. D. D. 2008. Aplicação de Jet-grouting em
soft clay and sandy silt, in an amount of escavações profundas em solos moles
Dissertation - Departamento de Engenharia Civil,
0.00088 m.
Universidade do Porto, Portugal.
The finite element method has many ad- Potts, D. M. 2003. Numerical analysis: a virtual dream or
vantages due to its generality and ease of deter- practical reality?. Géotechnique 53, No. 6, pp. 535–573.
mination of clusters of complex calculation in
J.C.S. Tito and C. Romanel / Numerical Analysis 3D of a Deep Circular Excavation in the City of Rio de Janeiro 373
Abstract. The soil nailing is one of the most effective provisions for the control and reduction of the tunnel pre-convergence and
face extrusion and its use has become integral part of modern design methods. In the years, improvements have been introduced
in the technological aspects and efforts have been made to better understand the mechanical behaviour of new materials and
equipments. Some innovative technical solutions are discussed in this paper, devised to improve the reinforcing action of Glass
Fiber Reinforced Polymer (GFRP) pipes and of soil nails. The effectiveness of these solutions, compared with traditional ones, is
assessed on the basis of laboratory and on site pull out testing.
Keywords. glass fibre reinforce pipes, soil nailing systems, pull-out test, tunnel face
2. Technical aspects about GFRP pipes and the etching could be equal to 2 mm in a 60/40
soil nails pipe, having 60 and 40 mm respectively as
external and internal diameters. A drawback of
Specific details are given on innovative soil this procedure is the cutting of external glass
nailing systems: firstly on a new kind of fibre- fibres and the loss in content of continuous fibres
glass pipe characterized by an external corrugat- effectively bearing tensile actions. Therefore, the
ed profile and secondly on a soil nail consisting pipe tensile strength is reduced with respect to
of an internal fibreglass pipe and an external the case of pipe with a not improved adherence.
geotextile sheath devised to contain the injected Alternatively a new product was introduced,
cement grout. Both products are patented by Elas named “Corrugated” GFRP pipe (VTR-CRG by
Geotecnica S.r.L. (Segrate-MI, Italy). Elas Geotecnica S.r.L.). It is obtained introduc-
ing, before the polymerisation of the resin in the
2.1. Glass fibre reinforced pipes production process, a phase of preforming on the
aligned fibres, that creates a corrugated external
The GFRP pipes used for soil nailing at the profile. The depth of the transversal shrinkage is
tunnel face are obtained through an industrial equal to 1.7 mm in the 60/40 pipe (Figure 2). In
manufacturing process called “pultrusion” this way, the aim to improve the bonding adher-
(Goldsworthy, 1971), that ensures to the final ence is reached, while at the same time the
product a uniform distribution of perfectly longitudinal continuity of the glass fibres, hence
aligned fibres. In general, the pultrusion of the pipe overall tensile strength, are maintained.
reinforced polymers is similar to the extrusion The fibre content by weight can be measured
process used for metals, with the addition of
after heating treatment, at 800°C for 8 hours. In
tensile forces applied to the glass fibres to
the case of standard pipes, with spiral groove, it
guarantee their alignment before the polymerisa-
tion of the matrix polyester resin. is also possible to distinguish between the total
The fibre content, usually expressed as per- content of glass and the content of glass from
centage by weight or volume of reinforcing longitudinal continuous fibres only. Figure 3
fibres with respect to the total weight or volume, shows the assembly of fibres after the heating
influences the tensile strength of the pipe. It is treatment on the two different pipe samples: the
therefore considered as reference parameter for presence of cut fibres is evident in the standard
the prediction of the pipe mechanical properties. pipe while in the corrugated pipe the fibres are
In particular, a higher tensile strength is associat- continuous.
ed with a higher fibre content, according to a
relation that can be assumed as roughly linear.
The pultrusion process enables the production of
pipes having 60-65% range in fibre content,
which is almost twice the content that can be
reached by other manufacturing techniques.
In addition to the tensile strength, the re-
sistance to pull out of the GRFP pipe is another
parameter used to identify the soil nail perfor-
mance. In fact, in the soil nailing the pipes are
driven or grouted inside drilled boreholes and
remain unstressed until the soil movements (for
instance those induced by the tunnel face extru-
sion) mobilize tensile forces which are trans- Figure 1 Fibreglass reinforced pipes: 60/40 AM pipe with
ferred to the ground through friction along the spiral groove
interface. For a better performance the lateral
surface of the pipe needs to be treated so to
improve the adherence with the surrounding
mass.
The adherence is traditionally improved by
etching a spiral groove along the external surface
of the pipe (Figure 1). For instance, the depth of
376 C.L. Zenti and D. Sterpi / Laboratory and In-Situ Testing for the Identification
(b)
Figure 3 Glass fibre mass after heating treatment: from (a)
60/40 AM pipe with spiral groove and (b) 60/40 CRG
corrugated pipe
nature and conditions (Zenti et al, 2008, Renda et The tensile and the bond strengths are used at
al., 2012). In all the tested cases, the pull out load the design stage, in both numerical and analytical
reached values higher than those measured with approaches. In the first case for instance, when in
standard soil nails, tested in the same construc- the mesh a distinction is made between ground
tion sites to get values significant for a compari- mass and nails, these properties are used as
son. In some cases the pull out load was meas- mechanical parameters of the nail elements
ured as almost 10 times higher, as in the case (tensile strength) and of the interface elements
reported in Section 3.2. (bond strength). Also when the concept of
To date, 24 m is the maximum length of these “equivalent material” is assumed to simulate the
soil nails used in tunnelling applications. As an presence of soil nailing in a homogeneous
example of application, a documented case equivalent ground mass, the bond properties are
history concerns a tunnel being excavated in used first to convert the mechanical action of the
Southern Italy (max. overburden 65 m, tunnel soil nailing in a confinement pressure exerted at
section 150-170 m2) in highly weathered soft the excavation face, and then to convert this
rock mass and in presence of high pore water confinement in an improvement of the mechani-
pressures (Renda et al., 2012; Sterpi et al., 2013). cal properties (i.e. the cohesion) of the ground
In the most severe conditions, the tunnel face mass ahead of the excavation (Grasso et al.,
was effectively supported by an average of 50 1989; Osgoui and Oreste, 2007; Bobet and
nails for reinforcement and 4 additional nails Einstein, 2011).
improved with a coaxial drain for coupled
reinforcement and drainage actions. The nails 3.1. Laboratory pull out tests
had sub-horizontal direction, 20 m length and
superposition of 10 m along the tunnel axis. A first laboratory investigation was carried out in
the past by applying the pull out load with a
hydraulic jack. The evidence of positive results
3. Pull-out test
suggested to design a test set-up for standardiza-
Together with the tensile strength, the bonding tion of this type of tests. The pull out tests were
properties of GFRP pipes and soil nails are carried out with the following objectives:
fundamental characteristics for the design of to assess the response of the reinforcing
reinforcing systems and should be assessed by elements to a tensile load;
both laboratory and in situ tests. to verify the achievement of a sliding condi-
In particular, laboratory tests are necessary to tion, represented by a cumulative residual
identify the interaction between the GFRP pipe displacement at the attainment of the maxi-
and the injection mixture and to highlight the mum pull-out load;
possible better performance of one nail type with to identify the parameters for the calculation
respect to others, under prescribed testing condi- of the adherence properties, according to
tions. These can be suitably devised to investi- standard ACI 440.3R-12.
gate the influence of various factors on the nail The tested GFRP nails are constituted by
performance. In addition, the possibility to carry CRG corrugated pipes characterized by circular
out series of tests under controlled and repeatable hollow section of outer diameter Øout = 60 mm,
conditions represents the basis for a commercial internal diameter Øint = 40 mm and total length L
certification of the product. The pull out is = 750 mm. The hollow section of GFRP pipes
probably the most convenient and widely used was previously filled with resin to prevent
testing technique, though not the only one (Chu breakage during the test execution.
and Yin, 2005). The samples were made by placing each
Field pull out tests are basically aimed at as- GFRP pipe inside a metal formwork, of 300 mm
sessing the bonding properties in actual on site as internal diameter, 410 mm in height, and 0.8
conditions. Since it is proved that the effective- mm in thickness, and by pouring a cement
ness of the reinforcement system depends also on mortar. The large ratio (equal to 5) between the
the nature and condition of the soil being rein- diameter of the grout block (300mm) and the
forced (Mair, 2008; Wong et al., 1999), the outer diameter of the pipe (60mm), together with
optimal design has to be based on these field the reduced wall thickness of the formwork
tests results. (0.8mm), allow to limit the mechanical contribu-
378 C.L. Zenti and D. Sterpi / Laboratory and In-Situ Testing for the Identification
tion of the formwork itself to the pull out re- The loading frame is constituted by four steel
sistance, i.e. its possible confinement action. columns, acting as supports and vertically
The embedment length of the GFRP pipe in- driving the transversal plate: this optimizes the
side the cement mortar block is = 300 mm stiffness of the frame in the loading direction.
(Figure 5). The ratio between the embedment A single type of injection mixture was used
length and the outer diameter of the GFRP for all the GFRP pipes. The grout mixture is
anchor (/Øout=5) allows to approximate the prepared with a water/cement ratio equal to 0.45,
adherence between the pipe and the cement in particular with a dosage of 25 kg of cement,
mortar as a uniformly distributed effect. 11.25 l of water and 0.2 kg of high plasticizer.
The tests were carried out on a total of 6 The density of mixture was found to be equal to
samples: 3 tests 24 hours after casting and 3 tests 1883 kg/m3 and its strength characteristics were
48 hours after casting. They were performed assessed by simple compression tests on cubic
using the electromechanical testing machine specimens of 100 mm side, carried out 24 and 48
Schenck (maximum load equal to 1000 kN), able hours after casting. The results showed mean
to apply on the samples tensile and compression values of compression strength equal to 6.6
loads. For the correct positioning of the sample it N/mm² and 22.4 N/mm², respectively 24 and 48
was necessary to design a suitable testing hours after casting.
framework (Figure 5). The pull out tests were carried out as stated in
section B.3 of the standard ACI 440.3R–12.
Since this standard refers to fibre reinforced
polymer composites bars for concrete and ma-
sonry structures, it was necessary to change the
testing set up to fit: (a) the case of a fibreglass
pipe and (b) the application to the underground
reinforcement. The necessary changes are related
to the size of the mortar block in which the pipe
is embedded and the clamping method. The free
head of the GFRP pipe has been inserted in a
protective steel cylinder case for gripping and
application of the tensile load. This was neces-
sary in order to extend the gripping surface, thus
reducing the risk to damage the GFRP element.
The pull-out tests were performed under dis-
placement control (0.02mm/s), with continuous
load and displacement monitoring, till the com-
plete pull out. The results (Figures 6 and 7) show
that the CRG corrugated pipes offer a high pull
out resistance that slightly increases with time as
a consequence of the mortar hardening (24-48 h).
250
CRG-A
225
CRG-B
200
CRG-C
175
LOAD [kN]
186 kN
150 129 kN
159 kN
125
100
75
50
25
0
0 10 20 30 40 50 60 70 80
DISPLACEMENT [mm]
Figure 5 Sketch and pictures of the test equipment and the Figure 6 Laboratory pull out tests on CRG corrugated pipes
sample after 24 hours
C.L. Zenti and D. Sterpi / Laboratory and In-Situ Testing for the Identification 379
250
CRG-F 218 kN
for the activation of a compression stress in the
225
200
CRG-G 199 kN 203 kN grout, thereby providing a pull out performance
CRG-H
175
better than the one with standard pipes.
LOAD [kN]
150
125 3.2. In situ pull out tests
100
75 The performance of the innovative nails is shown
50 here with reference to field tests carried out at a
25
tunnel face in highly weathered soft ground
0
0 10 20 30 40 50 60 70 80 (Figure 9). The reinforcing element is inserted
DISPLACEMENT [mm] and grouted into the borehole so that its head
Figure 7 Laboratory pull out tests on CRG corrugated pipes extends from the excavation face (Figure 10).
after 48 hours The borehole axis should be driven in a direction
normal to the surface, so to apply a condition of
The maximum tangential stress at failure Wmax, pure traction during the pull out.
defined as the bond strength, can be calculated
according to Equation (1)
F pull out
W max (1)
Cb "
where: Fpull-out is the pull-out load at failure; Cb
is the embedment
length of the sample within the cement grout
block. Considering the diameter of the tested
pipes (60mm) and the embedment length (a) (b)
(300mm), the bond strength is then evaluated Figure 8 Fracture patterns after pull out tests: (a) standard
assuming the maximum value of load. AM and (b) corrugated CRG pipe samples
Table 1. Pull-out load and bond strength The grout is a mixture of ordinary cement,
water and bentonite, in the ratio 100:60:4 by
Time Load – bond strength weight. The head is prepared using a threaded
Sample
[h] [kN] [kN/m2] [N/mm2] steel pipe, which acts as a protective case and
CRG-A 24 129 2290 2,290 widens the lateral surface of the nail where the
CRG-B 24 159 2810 2,810 load is applied, to reduce the risk of damaging.
CRG-C 24 186 3295 3,295
CRG-F 48 199 3527 3,527
CRG-G 48 203 3590 3,590
CRG-H 48 218 3857 3857
671,6 kN
reaction force is transferred to a stiff steel plate 600 CRG3
AM1
placed against the excavation face. 500
AM2
GFRP
The values of pressure are controlled by an 400 Corrugated
GFRP AM3
Standard
analog manometer and the displacement of the 300 Profile Profile
head is measured by optic differential levelling. 200 78,2 kN 28,4 kN
MAX PULL-OUT RESIDUAL
For each load increment two measurements are 100 LOAD LOAD
presence of large voids), characteristics of the Proc. Int. Congr. Progress and innovation in tunnelling,
Toronto, Vol. 2, pp.671-678.
grout and quality of injection. Juran, I. & Elias, V. 1991. Ground anchors and soil nails in
Nevertheless, it should be reminded that in retaining structures. In: Foundation Engineering Hand-
general the nail performance and in particular the book, 2nd Edition (ed. H.Y. Fang), 868-905.
bond properties depend on the site nature and Kim, J.-S., Lee, I.-M., Jang, J.-H. & Choi, H., 2009. Grouta-
bility of cement-based grout with consideration of vis-
state, and that on-site testing results cannot be cosity and filtration phenomenon. Int. J. Numer. Anal.
straightforwardly extended to other different Meth. Geomech., Vol. 33, pp. 1771–1797.
sites. For this reason specific field testing pro- Lunardi, P., 2008. Design and construction of tunnels,
grammes are always advisable. analysis of controlled deformations in rock and soils
(ADECO-RS), Springer.
Mair, R.J. 2008. Tunnelling and geotechnics: new horizons.
4. Conclusion Géotechnique, Vol. 58, pp. 695-736.
Mair, R.J. & Taylor, R.N. 1997. Bored tunnelling in the
An innovative Glass Fibre Reinforced Polymer urban environment. Proc. 14th Int. Conf. Soil Mech.
pipe is obtained by a particular forming process Found. Engng., Vol. 4, pp. 2353–2385.
that creates a corrugated external profile while Ng, C.W.W. & Lee, G.T.K. 2002. A three-dimensional
parametric study of the use of soil nails for stabilising
avoiding, at the same time, the cutting of the tunnel faces. Computers and Geotechnics, Vol.29,
glass fibres. The resulting pipe offers high pp.673-697.
bonding properties with the surrounding cement Oreste, P.P. & Dias, D. 2012. Stabilisation of the excavation
grout. Moreover, innovative systems for soft face in shallow tunnels using fibreglass dowels. Rock
ground nailing have been described and com- Mech. Rock Engng., Vol. 45, No. 4, pp. 499–517.
Ortigao, J.A.R. 1996. FRP applications in geotechnical
pared with traditional ones. The new soil nailing engineering. Proc. ASCE 4th Materials Engineering
technique provides high pull out strength, due to Conference, Washington DC, Vol.1, pp.535-544.
the possibility to apply high injection pressures Osgoui, R.R. & Oreste, P. 2007. Convergence-control
without soil fracturing and loss of grout, and to approach for rock tunnels reinforced by grouted bolts,
create homogeneous nails with uniform adher- using the homogenization concept. Geotech. Geol.
Engng., Vol. 25, No. 4, 431–440.
ence. The performance of these solutions is Renda, D., Grasso, P., Rizzo, F., Pinchiaroglio, L., Eusebio,
assessed on the basis of laboratory and on-site A., Pescara, M., Osgoui, R., Bondanelli, L., Ballabeni,
testing. F., Aguglia, F., Zenti, C.L. & Sterpi, D. 2012.
Improvement in soft ground tunnelling using an
5. References innovative technique. Proc. Int. Symp. Advances in
Ground Technology and Geo-Information – IS-AGTG,
Singapore: Research Publishing, pp.197-205.
ACI 440.3R-12, 2012. Guide Test Methods for Fiber-
Schlosser, F. 1982. Behaviour and design of soil nailing.
Reinforced Polymer (FRP) Composites for Reinforcing
Proc. Int. Symp. Recent Development in Ground Im-
or Strengthening Concrete and Masonry Structures.
provement Techn., pp. 399–413.
American Concrete Institute.
Sterpi D., Rizzo F., Renda D., Aguglia F. & Zenti C.L. 2013.
Anagnostou, G. & Serafeimidis, K. 2007. The dimensioning
Soil nailing at the tunnel face in difficult conditions: A
of tunnel face reinforcement. Proc. 33rd ITA AITES
case study. Tunnelling Underground Space Techn., Vol.
World Tunnel Congress WTC 2007 (Prague, Czech Re-
38, pp.129–139.
public), pp. 291–296.
Wong, H., Trompille, V., Subrin, D. & Guilloux, A. 1999.
ASTM D 4435-04, 2004. Standard Test Method for Rock
Tunnel face reinforced by longitudinal bolts: analytical
Bolt Anchor Pull Test. American Standard Testing Mate-
model and in situ data. Geotechnical Aspects of Under-
rials.
ground Construction in Soft Ground. Balkema: Rotter-
Bobet, A. & Einstein, H.H. 2011. Tunnel reinforcement with
dam, pp. 457–462.
rockbolts. Tunnelling Underground Space Techn., Vol.
Wong, H., Subrin, D. & Dias, D. 2000. Extrusion movements
26, No. 1, pp. 100–123.
of a tunnel head reinforced by finite length bolts – a
Cheng, Y.M., Choi, Y., Yeung, A.T., Tham, L.G., Au,
closed-form solution using homogenization approach. Int.
A.S.K., Wei, W.B. & Chen, J. 2009. New soil nail mate-
J. Numer. Anal. Meth. Geomech., Vol. 24, pp. 533-565.
rial – Pilot study of grouted GFRP pipe nails in Korea
Zenti, C.L. 2006. Laboratory and Field Investigation on the
and Hong Kong. J. Materials in Civil Engng., Vol. 21,
Behaviour of Swelling Grouts for Underground Excava-
No. 3, pp. 93–102.
tion, Ph.D. Thesis, Politecnico di Milano, Italy.
Chu, L.M. & Yin, J.H. 2005. Comparison of interface shear
Zenti, C.L., Sterpi, D. & Bonomi, C. 2008. Laboratory and
strength of soil nails measured by both direct shear box
field investigation on improved soil nails. Proc. World
tests and pullout tests. J. Geotech. Geoenv. Engng., Vol.
Tunnel Congress WTC 2008, Agra, India, Vol.1, pp.263-
131, No.9, pp.1097–1107.
271.
Goldsworthy, W.B. 1971. US Patent 3556888 Pultrusion
Zenti, C.L., Cassani, G. & Sterpi, D. 2012. Technical
machine and method.
solutions for soil nails in tunnel face reinforcement and
Grasso, P., Mahtab, A., Pelizza, S. 1989. Reinforcing a rock
drainage. Proc. World Tunnel Congress WTC 2012,
zone for stabilizing a tunnel in complex formations.
Bangkok, Thailand, Vol. 2 CD-ROM.
This page intentionally left blank
Soil-Structure Interaction, Underground Structures and Retaining Walls 383
V.M. Ulitsky et al. (Eds.)
IOS Press, 2014
© 2014 The authors and IOS Press. All rights reserved.
Subject Index
3D roughness parameter 119 fractured rock 329
adjacent foundations 35 Frankfurt limestone 28
anchor forces 273 freezing and thawing of soils 184
back analyses 359 frost heave 345
barrettes 189 frost heave problems 184
base-isolated structures 235 gate shaft structure 329
bearing capacity of driven piles 242 geosynthetic reinforcement 227
bearing capacity of piles 68 geotechnical limit state design 103
bridge strengthening 208 geotextile-reinforced soil walls 345
building basement 353 glass fibre reinforce pipes 374
building settlements 176 ground improvement 18, 220
buildings’ deformations 147 ground motions 132
calculation of retaining structures 302 groundwater cut-off systems 251
case studies 259 hardening soil model 28
centrifuge testing 35, 318 high-rise building(s) 3, 189, 195
clayey slope 353 historic buildings 290
constitutive soil models 273 Hoek–Brown failure criterion 329
contact stress mapping 119 influence of thermal conditions 140
control of deformations 345 instrumentation 359
creep and shrinkage of concrete 176 interaction of foundations
crushable sands 54 and freezing soil 147
cryogenic processes 147 interaction of structure with freezing
damaged dikes 132 and thawing soils 203
deep excavation(s) 259, 279, 337, 367 interaction old pier footing 208
deep foundations 195 lateral impact 11
deep tunnels 111 limiting soil pressure 35
deformations of buildings 80, 279 load distribution of piles 189
DEM analyses 155 long-term settlements 80
depth of frost penetration 345 Ménard pressuremeter tests 220
dewatering field tests 337 marine soil 155
diaphragm wall 353 model walls 345
differential settlement 195 modified Cam-Clay model 318
dike protection 132 Mohr-Coulomb constitutive model 367
drilled displacement columns 18 monitoring 359
dynamic cone penetration test 242 monitoring data 290
earth retaining structures 284, 310 monitoring of buildings 279
earthquake conditions 296 mortar-column-improved ground 195
earthquake impact 235 Napoli underground 337
embankment 18 non-linear spring model 265
excavation(s) 35, 353 nuclear power plant foundations 214
failure mechanism 68 numerical analyses 259
FE analyses 318 numerical modeling 132, 290, 367
FEM analysis 189 numerical soil-structure-interaction
field instrumentation data 367 analysis 353
field tests 68 oblique uplift capacity of a drilled pile 155
flexible retaining walls 265 oil tanks 54
384
Author Index
Adekunte, A. 251 Khalimov, O.Z. 147
Akçakal, Ö. 259 Koçak, B. 259
Al-Asali, B.E. 111 Krajewski, W. 28
Alibekova, N.T. 242 Kuribayashi, K. 132
Amornfa, K. 189 Kuroda, S. 132
Ariaratnam, K. 318 Lengkeek, H.J. 296
Ardianto, A. 329 Leppla, S. 28
Asadollahi, M. 11 Leung, C.F. 35
Autuori, S. 337 Lisyuk, M.B. 3
Avellan, K.C. 103 Liu, F. 155
Awwad, T. 111 Lombardi, D. 169
Baykal, G. 119 Madhav, M.R. 161
Becci, B. 265 Maffei, C.E.M. 302
Besseling, F. 296 Melentijevic, S. 310
Bilgin, Ö. 273 Mirdamadi, A. 11
Bochkarev, N.P. 279 Morev, I.O. 242
Bogov, S.G. 279 Nabokov, A.V. 184
Briaud, J.-L. 11 Namwiset, V. 359
Carni, M. 265 Nappa, V. 169
Carvalho, E.M.L. 176 Nicotera, M.V. 337
Chao, Z. 235 Nishi, T. 132
Cheang, W. 189 Ohsumi, T. 132
Danziger, B.R. 176 Ong, C.W. 318
da Rosa, L.d.M.P. 176 Ong, D.E.L. 35
Durgunoğlu, T. 259 Paramonov, M.V. 203
El-Mossallamy, Y. 68 Paramonov, V.N. 184
Elahi, H. 290 Pavan Kumar, P.V.S.N. 161
Fernández, J.L. 310 Phienwej, N. 189
Flora, A. 169 Pramusandi, S. 329
Furuichi, H. 132 Putera Agung M, A. 329
Gandhi, S.R. 353 Romanel, C. 367
Gniel, J. 18 Russo, G. 337
Gonçalves, H.H.S. 302 Sabermahani, M. 290
Guazzelli, M.C. 302 Sadek, S. 195
Haberfield, C. 18 Sakharov, I.I. 203
Haiyang, Z. 235 Sato, A. 345
Hamidi, B. 220 Schlee, J. 227
Hara, T. 132 Shashkin, A.G. 3
Hasan, S. 195 Shashkin, K.G. 3
Hashimoto, H. 140 Sivapriya, S.V. 353
Hayashi, H. 140 Smolak, V.Y. 279
Hazarika, H. 132 Sokolić, I. 208
Jacquard, C. 127 Sterpi, D. 374
Jiang, M. 155 Su, T. 318
Jurado Cabañes, C. 284 Sunaryo, B. 329
Katzenbach, R. 28 Suzuki, T. 345
386