PREDICTION OF GROUND DEFORMATIONS CAUSED BY UNDERGROUND
CONSTRUCTION OF THE TREN URBANO IN RIO PIEDRAS
by
Angela L. Alba-Carb6
B.S., Civil and Environmental Engineering
University of Puerto Rico, 1995
Submitted to the Department of
Civil and Environmental Engineering in Partial Fulfillment of
the requirements for the Degree of
MASTER OF SCIENCE
in Civil and Environmental Engineering
at the Massachusetts Institute of Technology
February, 1998
copyright 1998, Massachusetts Institute of Technology
All Rights Reserved
Signature of Author
Dpartment of Civil and Environmental Engineering
December day, 1997
Certified by
Professor Andrew J. Whittle
Thesis Supervisor
Accepted by
Professor Joseph M. Sussman
Chairman, Departmental Committee on Graduate Studies
FeB:";X
M-1
PREDICTION OF GROUND DEFORMATIONS CAUSED BY UNDERGROUND
CONSTRUCTION OF THE TREN URBANO IN RIO PIEDRAS
by
ANGELA L. ALBA-CARBO
Submitted to the Department of Civil and Environmental Engineering on December 1 st, 1997 in
partial fulfillment of the requirements for the Degree of Master of Science in Civil and
Environmental Engineering
ABSTRACT
One of the principal factors affecting the design of deep excavations in cohesive soils is
the control of ground deformations, in order to minimize damage to adjacent facilities and
mitigate the costs of underpinning. The goal of this thesis is to estimate the magnitudes of
ground movements for Section 7 of the Tren Urbano project in San Juan, Puerto Rico, and to
relate these movements to the stratigraphy of the surrounding alluvial soils and proposed
methods of construction.
The area around the proposed Rio Piedras Station is underlain by thick alluvial deposits
comprising the Hato Rey formation. A review of the site investigation data has led to a
simplified interpretation of complex stratigraphy at the site. Predictions of excavation-induced
ground movements are made by non-linear finite element analyses, incorporating a relatively
simple effective stress, elasto-plastic soil model (HS). Input stiffness and strength parameters for
this model are estimated from available laboratory test data, while (Darcian) groundwater flow is
controlled by field permeability measurements. Finite element calculations of real-time coupled
flow and deformation are performed using the PLAXIS, a commercially available PC based
code. The calculations focus on a simplified geometry, based on preliminary designs for cutand-cover approaches to the Rio Piedras Station, supported by a diaphragm wall and cross-lot
bracing. The analyses evaluate the effects of selected input parameters (wall embedment length,
soil stiffness and permeability properties) on predicted wall deflections and ground deformations.
The predictions of maximum wall deflections and maximum surface settlements are in very good
agreement with published empirical data, for excavations in similar types of soil.
However, predictions of settlement distributions are unrealistic and do not match
empirical case history data. In principle, the calculations of ground movements can be improved
by using soil models which replicate more closely the non-linear stiffness properties of Hato Rey
soils. The application of these more complex analyses can only be justified if there is sufficient
test data for selecting model parameters. The thesis recommends specific types of laboratory and
field test that should be carried out for this purpose.
Thesis Supervisor:
Title:
Prof. Andrew J. Whittle
Associate Professor of Civil and Environmental Engineering
ACKNOWLEDGMENTS
I wish to express appreciation to those individuals and organizations that made this thesis
possible:
First of all, I would like to thank Dr. Antonio Gonzalez for involving me with the Tren
Urbano project. Ever since we met at the University of Puerto Rico, he has been a source of
support and inspiration.
My advisor, Professor Andrew J. Whittle, for his careful review of this thesis. He was
always very encouraging and supportive of my work. It has been a great experience working
with him.
I should also like to thank Professor Charles C. Ladd and Professor Herbert Einstein for
been extremely generous of their time and giving me their input during several stages of my
thesis.
I would like to thank the Tren Urbano GMAEC for providing the financial support of this
research. The Tren Urbano research experience has been very rewarding in many different
aspects.
Special thanks to my office mate, Lucy Jen, for all her help and good advises.
Mary and Carolyn for their kind words and interesting conversations.
My family in Puerto Rico for being so loving and supportive. Papi y Mami, los quiero
mucho.
Ram6n, my husband, friend and companion.
experiences. Mi Vida,
eres mi inspiraci6ndivina.
Thanks for sharing with me all my
Angela L. Alba-Carb6
TABLE OF CONTENTS
Abstract
Acknowledgments
Table of Contents
List of Tables
List of Figures
2
3
4
6
7
1. Introduction
1.1 Overview of Tren Urbano
1.2 Overview of Rio Piedras Section
1.3 Methods for Predicting Underground Movements
1.4 Selection of Finite Element Program
1.5 Objectives and Scope
10
11
12
13
14
16
2. Geology of Section 7 Alignment
2.1 Stratigraphy of the San Juan Metropolitan Area
2.2 Bedrock
2.2.1 Cretaceous Rocks
2.2.2 Tertiary Formations
2.2.2.1 Fajardo Formation
2.2.2.2 Aguada Formation
2.3 Quaternary Deposits
2.4 Near Surface Deposits
2.5 Ground Water Conditions
2.6 Seismic Conditions
2.7 Overview of Engineering Properties of Hato Rey Deposits
27
27
28
29
30
30
31
32
34
36
36
37
3. Site Investigation for Rio Piedras Station
3.1 Introduction
3.2 Scope of Site Investigation
3.3 Stratigraphy
3.3.1 Section 7
3.3.2 Rio Piedras Station
3.3.3 Index Properties
3.4 Groundwater Conditions
3.4.1 Watertable and Equilibrium Pore Water Pressures
3.4.2 Hydraulic Conductivity
3.4.2.1 Field Test Data
3.5 Strength and Deformation Properties
3.5.1 Test Procedures
3.5.1.1 Field Tests
3.5.1.2 Laboratory Tests
47
47
47
49
49
50
50
51
56
56
56
57
57
57
59
3.5.2
3.5.3
3.5.4
3.5.5
4.
Undrained Strength, Su
Drained Strength Parameters
Compressibility and Shear Stiffness
Consolidation Properties
Finite Element Model for Rio Piedras Excavation
4.1 Introduction
4.2 The Hard Soil (HS) Model
4.2.1 Summary of Formulation
4.2.2 Input Parameters for Hato Rey Soils
4.3 Idealized Braced Excavation
4.4 Results
4.4.1 Base Case Analysis
4.4.2 Effects of Individual Parameters
4.4.2.1 Effects of Unloading Modulus Eurref
4.4.2.2 Effects of Poisson's Ratio Vur
4.4.2.3 Effects of Permeability Anisotropy
4.4.2.4 Effects of Wall Length
4.4.2.5 Summary of Results
4.5 Practical Interpretation of Results
4.5.1 Measured Soil Movements Published in Literature
4.5.2 Comparison between Predicted and Empirical Data
60
62
63
65
109
109
109
111
112
115
121
121
123
123
124
124
125
125
126
126
127
5. Summary, Conclusions and Recommendations
5.1 Summary and Conclusions
5.2 Main Sources of Uncertainty
5.3 Recommendations
180
180
183
183
References
186
Appendices
188
A. Finite Element Analyses Case Studies
B. Vertical Coefficient of Consolidation Plots
C. Computations of E' from Oedometer Tests
188
197
211
LIST OF TABLES
2.1
2.2
Bedrock Formations in San Juan Metropolitan Area
Comparison of Index Properties for Hato Rey Alluvial Sediments
3.1 Description of Borings
3.2 Piezometer Locations Phases I and II, and III
3.3 Piezometer Locations Phase IIII Pumping tests
3.4 Consolidation Tests
3.5 Description of Triaxial Tests
3.6 Permeability Results
3.7 Typical Range of Values of Vertical Coefficient of Consolidation
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
Hard Soil Model Parameters
Diaphragm Wall Parameters
Anchors Parameters
Parametric Study
Ground Movement Predictions for Rio Piedras at H = 22m
Comparisons of Wall Deflections 8wmax/H (%) at each Excavation Step
Comparisons of Ground Settlements 8vmax/H (%) at each Excavation Step
Comparisons of Uplift/H (%) at each Excavation Step
130
130
130
131
131
132
132
133
LIST OF FIGURES
1.1
1.2
1.3
1.4
1.5
1.6
Phase I: Tren Urbano
Final System Configuration
Section 7 Alignment
Rio Piedras Station Section- Microtunnel
Rio Piedras Station Transversal section North- Cut and Cover
Examples of Semi-Empirical Design Charts
2.1
2.2
2.3
2.4
18
19
20
24
25
26
Map of the Greater Antilles
Map of Puerto Rico
Hato Rey Formation
Location of Puerto Rico with Respect to Major Geographic and Tectonic
Features
Seismic Hazard Map of 1987 for Puerto Rico
Maps of Acceleration for 50, 100 and 250 Year Exposure
41
42
43
71
3.3
3.4
3.5
Boring Locations of Geotechnical Investigations
Geoprofile of Section 7: from boring 219 (stn. 217+25)
to boring P11 (stn.229+17
Legend of Geoprofile
Stratigraphy of Rio Piedras Station
Natural Water Content
73
74
75
76
3.6
Liquid Limit, WL
77
3.7
3.8
3.9
3.10
3.11
3.12
3.13
3.14
3.15
3.16
3.17
3.18
3.19
3.20
3.21
3.22a
3.22b
3.23
3.24
Plastic Limit, wp
Plasticity Index, Ip
Plasticity Chart for Phases I and II
Plasticity Chart for Phase III Investigation
Liquidity Index
Selection of Unit Weight
Site Layout of Pump Test (Phases I and II)
Results of Pump Test (Phases I and II)
Site Layout of Pump Test No.1 (Phase III)
Site Layout of Pump Test No.2 (Phase III)
Results of Pump Test No. 1 (Phase III)
Results of Pump Test No.1 (Phase III)
Groundwater Elevations: Section 7 and Rio Piedras Station
Initial Stress Profile
Permeability Results from Slug Tests
Measured Standard Penetration Test N-Value
Undrained Strength from Measured Standard Penetration Test
Pocket Penetrometer Results
Pressuremeter Strength Results from Correlations
78
79
80
81
82
83
84
85
86
87
88
89
90
92
93
94
95
96
97
2.5
2.6
3.1
3.2
44
45
46
Undrained Strength Results from Laboratory and Field Tests
Undrained Strength Results from UUC and CIUC Tests
Backcalculation of Preconsolidation Pressure for Hato Rey Soils
Backcalculation of Overconsolidation Ratio for Hato Rey Clay
Relationship between Cohesion Intercept and Preconsolidation Pressure
Estimation of Friction Angle for Mohr-Coulomb failure Criterion
Consolidation Curves
Compression as function of Water Content (w) and Void Ratio, eo
Young's Modulus from CIUC, M6nard Pressuremeter and Oedometer
Tests
3.34 Vertical Coefficient of Consolidation Profile
3.25
3.26
3.27
3.28
3.29
3.30
3.31
3.32
3.33
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
4.10
4.11
4.12
4.13
4.14
4.15
4.16
4.17
4.18
4.19
4.20
4.21
4.22
4.23
4.24
4.25
4.26
4.27
4.28
4.29
4.30
4.31
98
99
100
101
102
103
105
106
107
108
Hyperbolic Stress-Strain relation in Primary Loading for a Standard Drained
and Triaxial Test
134
Relationship between Cohesion Intercept and Preconsolidation Pressure
135
Estimation of Stress History for Hato Rey Clays
136
Estimation of Friction Angle for Mohr- Coulomb Failure Criterion
138
Determination of Dilatancy Angle,
139
Determination of the Primary Loading Stiffness, E5 0ref
140
Determination of Poisson's Ratio, Vur
141
Determination of the Unloading/Reloading Stiffness, Eurref
142
Determination of m Parameter from CIUC Tests
143
Determination of m Parameter
144
Comparison between CIUC Triaxial Tests and Single Element HS
Predictions
145
Boundary and Initial Conditions for Ideal Case
146
Finite Element Mesh
147
Geometry, Support Conditions, and Excavation Sequence
148
Flow Boundary Conditions at each excavation stage
149
Wall deflections for Base Case Analysis
150
Predicted Surface Settlements for Base Case Analysis
151
Predicted Horizontal Surface Displacements for Base Case Analysis
152
Curve Fits for Bending Moment Calculations
153
Predicted Bending Moments for Base Case Analysis
155
Effective Lateral Earth Pressures for Base Case Analysis
156
Total Lateral Earth Pressures for Base Case Analysis
157
Pore Water Pressures on Wall for Base Case Analysis
158
Effect of Unloading Modulus Eur on Wall Displacements
159
Effect of Unloading Modulus Eur on Ground Surface Settlements
160
Effect of Unloading Modulus Eur on Horizontal Surface Displacements
161
Effect of Poisson's Ratio on Wall Displacements
162
Effect of Poisson's Ratio on Ground Surface Settlements
163
Effect of Poisson's Ratio on Horizontal Surface Displacements
164
Effect of Anisotropic Permeability on Wall Displacements
165
Effect of Anisotropic Permeability on Ground Surface Settlements
166
Effect of Anisotropic Permeability on Horizontal Surface Displacements
Effect of Wall Length on Wall Displacements
Effect of Wall Length on Ground Surface Settlements
Effect of Wall Length on Horizontal Surface Displacements
Maximum Lateral Movements Ratio, 8wmax/H (%)
Maximum soil Settlements Ratios, 8vmax/H (%)
Comparison of Wall Uplift Ratios
Observed Maximum lateral Movements and Soil Settlements reported in
Literature (Clough and O'Rourke, 1990)
4.40 Summary of Measured Settlements and Horizontal Displacements Adjacent
to Excavations in Stiff to Very Hard Clay (Clough and O'Rourke, 1990)
4.41 Dimensionless Settlement Profiles Recommended for Estimating the
Distribution of Settlement Adjacent to Excavations in Different Soil Types
(Clough and O'Rourke, 1990)
4.42 Maximum Lateral Movements for Rio Piedras
4.43 Maximum soil Settlements for Rio Piedras
4.44 Heave Inside Excavation for Rio Piedras
4.32
4.33
4.34
4.35
4.36
4.37
4.38
4.39
167
168
169
170
171
172
173
174
175
176
177
178
179
1.
Introduction
Tren Urbano is a high-capacity, rapid transit regional rail system to be constructed
in the metropolitan area of Puerto Rico. Phase I of Tren Urbano consists of an L-shaped
route, extending from Bayam6n east to Rio Piedras and north from Rio Piedras to
Santurce and includes approximately 14 kilometers of double track guideway and 15
stations (Figure 1.1). The Rio Piedras Alignment, Section 7, is entirely underground,
consisting of 1500 meters and will be constructed by cut-and-cover excavations and
tunneling techniques.
One of the principal factors affecting the design of these sections is the control of ground
deformations, in order to minimize damage to adjacent facilities and mitigate the costs of
underpinning.
This is especially true for construction in congested areas where the
potential for damage to adjacent buildings, utilities, etc., can lead to very expensive
remedial measures because of uncertainties in predicted deformations.
The design
specifications address these problems by requiring stiff pre-stressed bracing of cut-andcover excavations, and installation of a permanent tunnel canopy by pipe jacking
('umbrella' construction method).
However, there remain large uncertainties in the
prediction of ground movements and the effectiveness of these construction procedures.
The aim of this project is to estimate the magnitudes of ground movements for
Section 7 of the Tren Urbano project, and to relate these movements to the stratigraphy of
the surrounding alluvial soils and proposed methods of construction.
1.1
Overview of Tren Urbano
The Tren Urbano project is divided into several phases: Phase I, Phase IA, Phase
II, Phase Ill, and Phase IV. Figures 1.1 and Figure 1.2 shows the Tren Urbano Phase I
and the final system configuration, respectively.
Phase I of Tren Urbano is an
independent viable project that does not depend on any future extensions of the system.
However, the planning and design of Phase I anticipates these future actions and nothing
in the design of this phase will preclude these extensions in the future. The Phase I
Alignment of Tren Urbano can be described as follows (Figure 1.2):
1. From the Bayam6n River to PR-21 and the De Diego Ave. Interchange (Sta. 182+60),
the alignment runs slightly above existing grade through the
6 5 th
Infantry Expressway
corridor.
2. From PR-21 (De Diego Station) to PR-3 at Rio Piedras, the alignment is elevated and
proceeds through highly developed areas such as Villa Nevirez, Veteran's Hospital
Area, Centro Medico, etc.
3. From PR-3 at Rio Piedras up to Central Ave. (PR-17), the route descends
underground as it proceeds through the downtown area of Rio Piedras.
4. From Central Ave. to the end of the trackway at Sagrado Coraz6n, an elevated
structure is proposed crossing the Hato Rey center and "Milla de Oro" area.
Approximately 40% of the alignment is at or near grade. The remainder of the alignment,
aside from short below-grade sections in the Jardines de Caparra, Centro Medico, and Las
Lomas areas and the underground section through Rio Piedras, is generally elevated
above roadway rights-of-way.
1.2
Overview of Rio Piedras Section
Section 7 alignment of the Tren Urbano involves approximately 1,500 meters of
underground tunnels and two stations: Rio Piedras Station and University of Puerto Rico
Station. It connects to the Villa Nevairez (Section 6) on the south, and the Hato Rey
(Section 8) to the north. The alignment passes Barriada Venezuela, crosses PR-3 near
station 219, proceeds to go underground at Calle Julian E. Blanco and continues under
Rio Piedras until Station 231 near Calle Mariana Bracette (Figures 1.3a-d). Much of the
Rio Piedras alignment runs beneath (or slightly to the east of) and parallel to the Ponce de
Le6n Avenue.
This section will be all constructed within deep alluvial deposits of
interbedded stiff clays and silts referred to as the Hato Rey formation. This alignment
was developed to optimize service to the traditional town center of Rio Piedras and the
University of Puerto Rico (UPR). The section also includes a junction to accommodate a
future extension of the system to Carolina Centro (Phase II of Tren Urbano), which runs
due east from Rio Piedras Station along an elevated alignment parallel to PR-3.
The Rio Piedras Station will be sited beneath existing buildings and utilities
adjacent to the Commercial Center Historic District, along Ponce de Le6n Avenue. The
existing buildings consist of old masonry structures, which could be very sensitive to
deformations.
The approaches to the station will be constructed by cut-and-cover
excavations, with a maximum depth of approximately 23 m, while the station itself will
be excavated by tunneling techniques. To minimize subsidence impacts, the preliminary
design proposes a very strong and stiff "roof' of jacked pipes to be installed between the
station and the buildings.
The microtunnel arch will provide a continuous shield,
supporting and protecting the crown of the excavation. The underground mined station
work will occur between Sta. 219+90 and Sta. 221+15. The overburden depth over the
crown of the station varies between 7m and 5m. The station opening, approximately 14m
high by 19m wide will be arch shaped. Figures 1.4 and 1.5 illustrate typical crosssections of the Rio Piedras Station and North Approach Cut and Cover Sections, based on
preliminary designs.
1.3
Methods for Predicting Ground Movements
Existing techniques for predicting ground movements fall into two categories:
semi-empirical methods and finite element methods. Semi-empirical methods, which are
based on field data collected from case histories, provide a useful guide for estimating a
likely wide range of movements, but cannot be used reliably for site specific predictions.
Figure 1.6 shows empirical charts of the distribution of ground surface settlements for
braced excavations in: a) stiff to very hard clays; and b) design chart for wall movements
after Clough et. al. 1989. As a first approximation, ground conditions along the Rio
Piedras alignment can be classified as stiff clays. Hence, the results in figure 1.6a show
that maximum expected ground settlements for a 20 m deep excavation are 8v-6 cm
(2.5").
More reliable predictions can be achieved by using advanced finite element
analyses that incorporate realistic modeling of soil behavior, proper definition of the site
stratigraphy and relevant soil properties, and realistic simulation of all pertinent
construction activities, which occur during the excavation process.
Finite element
analyses were first applied to braced excavations by Clough et. al. (1972), Christian and
Wong (1973) and have now gained widespread acceptance through their capability to
model complex construction sequences, and to incorporate detailed site-specific
properties of the structural system and surrounding soils. In principle, the finite element
method offers a comprehensive tool for analyzing multiple facets of excavation
performance ranging from the design of the wall and supporting system, to predictions of
ground movements and the effects of construction activities such as dewatering, ground
improvement, etc. The MIT geotechnical group has extensive experience in applying
these advanced numerical simulations for excavations and has reported results from two
detailed case studies at well instrumented sites: 1) Post Office Square garage in Boston
(Whittle et. al., 1993a), and the World Trade Convention Center in Taipei (Whittle et. al.,
1993b).
In both cases, the finite element predictions were performed using site
characterization data provided prior to construction, while the simulated excavation
sequence was based on the actual record of construction activities. The analyses were
compared with all available field monitoring data (i.e., wall deflections, vertical and
lateral soil deformations, strut loads, pore water and lateral earth pressures) in order to
assess all aspects of the prediction. No attempt was made to correlate or adjust input
parameters to improve agreement with the measured data. Appendix A summarizes these
case studies.
The results show very good agreement with the measured excavation
performance and provide strong support for further applications of these methods for
Tren Urbano project.
1.4
Selection of Finite Element Program
The recent studies of excavation performance at MIT have used the ABAQUSTM
program, a commercial workstation based finite element program.
The stress-strain-
strength properties of soft clays were described using the MIT-E3 effective stress soil
model (Whittle and Kavvadas, 1994; Whittle et. al., 1994). In contrast, the current study
uses PLAXISTM
,
a commercially available PC-based finite element program, together
with much simpler soil models. The reasons for this choice are as follows:
1. Soil conditions at Rio Piedras comprise alluvial deposits of clays, sands and clayey
sands, which are highly overconsolidated. There is limited laboratory data on these
materials and no existing documentation to show the advantage of using sophisticated
soil models. The lack of laboratory data also makes parameter selection difficult even
for relatively simple soil models.
2. ABAQUS is a very general finite element program that is widely used in the
aerospace, defense and oil industries.
However, its application within the
geotechnical community is mainly for research purposes, and it is not widely used in
the construction industry. In contrast, PLAXIS is designed as user-friendly software
for geotechnical problem solving and has many features similar to ABAQUS for the
purpose of modeling excavation behavior: i) It models concurrently the soil
deformation and flow of groundwater, and hence can model the effects of partial
drainage during construction.
Excavation and dewatering activities alter the
groundwater regime in the soil, which sets up a transient flow condition (partial
drainage) and induces time dependent deformations of the soil mass. ii) PLAXIS
includes a range of simple elasto-plastic models (EP) of soil behavior. This research
uses the Hard Soil model (Schanz and Vermeer, 1996), which was originally
developed to describe the behavior of sand, gravel, and heavily overconsolidated
(stiff) cohesive soils. This model follows the logic of the Duncan-Chang model (HS:
Schanz and Vermeer, 1996), but also incorporates plasticity theory.
The key
characteristics of the Hard Soil model are:
1. Stress dependent stiffness, according to a power law.
2. Hyperbolic relationship between strain and deviatoric stress.
3. Distinction between primary deviatoric loading and unloading/reloading.
4. Failure behavior according to the Mohr-Coulomb model.
These features represent a set of reasonable assumption for modeling soil
conditions at the Rio Piedras site.
1.5
Objectives and Scope
The aim of this project is to estimate the magnitudes of ground movements for
Section 7 of the Tren Urbano project, and to relate these movements to the stratigraphy of
the surrounding alluvial soils and proposed methods of construction.
This project
involves three main tasks:
1. Characterization of the site stratigraphy, ground water conditions and engineering
properties of the principal soil layers along the Section 7 alignment in Rio Piedras
2. Development of numerical analysis models to represent a typical cross-section of the
project, simulate the proposed construction sequence and represent the deformation and
flow properties of the soils
3. Interpretation of the predicted ground movements
No previous geotechnical engineering studies were conducted specifically for the Rio
Piedras Section of the Tren Urbano prior to those conducted through Tren
Urbano/GMAEC.
However, there have been previous studies of local soil conditions
(Deere, 1959; Kaye, 1959; and Monroe and Pease, 1977). The site characterization is
based on geotechnical exploration programs already carried out for Section 7 alignment
(GMAEC, GDR 1996). The geotechnical field investigations consisted of two phases:
1. Phase One, which included 6 borings along the present Rio Piedras alignment between
Sta. 217+00 and Sta. 225+00 (current alignment). Six additional borings were drilled
east of the Rio Piedras alignment near Highway PR-3.
2. Phase Two, which included 10 borings along the revised proposed tunneled section
between Sta. 217+00 and Sta. 230+00.
Additional field and laboratory tests were performed during Phase Two investigation,
index tests', unconfined compression tests, borehole permeability (9) and pressuremeter
tests (3). Also, supplemental geotechnical investigations for Rio Piedras were conducted
from August through the first week of October 1996.
The analyses focus on two typical cross-sections of the cut-and-cover braced
excavations at the North and South approaches of the Rio Piedras Station.
Special
consideration is given to the representation of soil properties in the numerical analyses.
The predicted ground deformations are compared with published data from projects using
similar construction techniques and available empirical correlations.
The tests include Atterberg limits, particle size gradation, etc. However, there is no information given on
typical unit weights.
Atlantic Ocean
San Juan Bay
Santurce
Cataifo
San Juan
Bayamdn
Carolina
UPR Station
Guaynabo
PiedrasStation
Legend:
-
Alignment
....... Section 7
SStation
Figure 1.1 Phase I: Tren Urbano
LEGEND
Isene
+It
W
Umw
PHASE I
PHASEIA
AAAA PHASES
M M I PHASE4
I
OLD
SAN JUAN
Figure 1.2 Final System Configuration
I11 1 ,)/
((I , , III
,x
II
,/'
IC
I'(O1(( I)1 III lN
IIN
AvlI 11
R
ON PIEDRAS
IDRAS STATION
STATION
RIO
221+00
I
DE DIEGO
O
TREN URBANO
Figure 1.3a Section 7 Alignment
ALIGNMENT PLAN AND PROFILE
STA 217 75 TO STA.221-50
LOCALLY PREFERRED ALTERNATIVE
ROBLES SIREET
Figure
("- "H1
' 'l ., I
I
/ /
1I
\'
T)
I
O
i; rI
i!i
'
,t;
SALDOAA STREET
A- 1 .. ..
. .
OR. JOSE
O
TREN URBANO
Figure 1.3b Section 7 Alignment
ALIGNM[NT PLAN ANI PROFILE
SIA.221- 50 10 STA.225V-25
ILOCALLY PREFERRED ALfRNAIIVE
I I*,I
II
rr ~
*~lru
'1,1
oo
-,il
nt
C-4
0n
i Looom V
0
TREN URBANO
Figure 1.3c Section 7 Alignment
AlIGNMrNT P.LAN AND) PROFILE
SIA.225125 10 SIA.229100
IOCALLY PREFERRED AI.ERNATIVE
---
djs~~
k)
Figure 1.3d Section 7 Alignment
UILC:NC
ZXS;G
XXISflNG gUILCING
'0 REMAIN
MT'- AIR -,-ACt*
WA---' ZAR!E
SE7*,E:-N1 S.S ANC %iICRONNE'L7OESIGNEC - ' :A7Z--. ANO ]WvR7
,H4E NA7, -,A-. viGRArE3 -RCUCHTHE~VICRCUNNE..14'JG :NST.RUClCN
ACK 'RAINACE
ASS EI'72 SYSTTY. -41S :,NC7 N SHALL 2E
BE-1NO HE ARCMI t---JRAL
FINISH
SECTION
CZ
0::L~.
£
\RZC~
P102
Figure 1.4 Ri'o Piedras Station Section- Microtunnel
-
EXISTING
SUILDING
rO REMAIN
EXISTING BUILDING
TO REMAIN
CLERESTORY
REFER TO
ORAWING
A-UO201
PONCE DE
LECN
----
EMERG
STAIRS
= =,=---U.P E
CONCOURSE
LEVEL
MEZZANINE LEVEL
ESCALATIO
PS-34
.I
SECTION
SCALE- :200
G
A-RP2
A-RP105
A-RP107
A-RF 109
A-RP206
Figure 1.5 Rio Piedras Station Transversal Section North- Cut and Cover
PLTFORM LEVEL
TOP:F RAIL
0.5
0
c0.
t
I
*
Distance from Excavation
d
' H
Max. Excavation Oeoth
2.5
2.0
1.5
1.0
0.
c
c
So 0.
0.
0.
Sa12
e ,<
E
,
0.
i
°ol
/
O.24
Zone for High Horizontal
Stiffness of Sugoort
.3
-Zone
0.5
/
0.6/
/
0.
0.8
for Low 'orizontat
Stiffness of Suogort in London
/
0.4-
/
Legend:
* London YMC (48)
London NPY (48)
A Neasden (45)
SBell Common (52)
0 South Cove (24)
a
A
O
O
Clay
State St. (22)
Churchill Square (I5)
Columbia Center '19)I
Houston Bldqs -53.54)
I'st Nat'I Bank (4e)
Summary of Measured Settlements and Horizontal Displacements Adjacent to Excavations in Stiff to Very Hard Clay.
(a)
10
30
50 70 100C
(EI)/(r,h)-
300 500 700 1000
3000
Increasing System Stiffness
(b)
(from Clough et a, 1989)
Figure 1.6 Examples of Semi-Empirical design
Charts
2.
Geology of Section 7 Alignment
The island of Puerto Rico is the easternmost and smallest of the Greater Antilles (Figure
2.1), a chain of large islands that comprises, from west to east, Cuba, Jamaica, Haiti, Dominican
Republic, and Puerto Rico. The islands are bounded on the north by the Atlantic Ocean and on
the south by the Caribbean Sea.
Puerto Rico is roughly rectangular in shape, measuring
approximately 110 miles from east to west and 37 miles from north to south. Puerto Rico is
approximately 1,000 miles southeast of Miami, Florida, and 425 miles north of Venezuela. The
climate of Puerto Rico is subtropical marine; and it lies in the zone of the trade winds which
blow steadily from the northeast (Kaye, 1959). The city of San Juan is located on the eastern part
of the north coast of Puerto Rico (See Figure 2.2).
2.1 Stratigraphy of the San Juan Metropolitan Area
The following is a summary of the geologic units of the San Juan Metropolitan area ,
based on three main sources of information Deere (1955), Kaye (1959), and Monroe (1976). The
stratigraphic sequence can be summarized as follows (from oldest to youngest):
(1) Bedrock
The bedrock is comprised of Cretaceous' rocks and Tertiary 2 formations. No rocks older than
Upper Cretaceous have been recognized in Puerto Rico (Deere, 1955).
The rocks of Upper
Cretaceous age consist of a great range of pyroclastic, sedimentary, extrusive, and intrusive
igneous rocks. The Tertiary formations can be subdivided into the basal Rio Guatemala group
SThe Cretaceous period took place from 144 to 66.4 millions of years ago (total duration of 78 millions of years).
2 The tertiary formations are 66.4 to 1.6 millions of years of age.
(middle to upper Oligocene), the Aguada limestone (lower Miocene), and the Aymam6n
limestone (lower Miocene).
(2) Quaternary Formations 3
These sediments were deposited during the Pleistocene and Recent epochs; and they include,
from oldest to youngest:
(a) Old Alluvium (Hato Rey Formation); which consists of thick deposits of clay, sand,
clayey sand, sandy clay, and occasional beds of gravel.
(b) Santurce Sand; which is comprised of red and tan clayey silty sand, pure white quartz
sand, partially cemented calcareous sand and quartz sand, and stiff red and gray
mottled sandy clay.
(c) San Juan Formation; consisting of calcareous sandstone
(d) Floodplain Sediments; comprising mostly silty clay
(e) Lagoonal Sediments
(f) Sand Covered Lagoonal Sediments
2.2 Bedrock
The older rocks, that is to say, those of Cretaceous and early Tertiary age, are highly
deformed and faulted.
They comprise a sequence of volcanic flows, pyroclastics, and
sedimentary rocks, but many of the latter consist largely of reworked volcanic material. Into
these rocks have been intruded plugs, dikes, sills, and larger subjacent bodies that range in
composition from granodiorite porphyry to diabase (Kaye, 1959).
3 Quaternary is the name for recent deposits (not older than 1.6 millions of years ago).
Unconformably overlying the Upper Cretaceous and lower Tertiary complex in the San
Juan area is a sequence of sands, clays, marls, and limestones of early Miocene age, which have
been tilted to the north and faulted on a small scale but which are otherwise undeformed. These
rocks probably underlie most of the coastal plain alluvium (Kaye, 1959).
Section 7 alignment: In the northern part of this section of the alignment, the top of decomposed
and weathered calcareous limestone was encountered. The Geotechnical Data Report (GMAEC 4,
GDR, April 1996) also states that at the southern limits of the project, the alluvial soils may also
be underlain by the Eocene and Paleocene Rio Piedras siltstone (sedimentary & volcanic rocks).
This formation may be what Kaye (1959) refers to as the Fajardo Formation. Near the old town's
plaza, the Rio Piedras siltstone was found at about 35m of depth. The contact between the
Cretaceous siltstone and the Tertiary calcareous limestone has not been established even though
the evidence is that it should lie at depths below 40 m within the town of Rio Piedras.
2.2.1
Cretaceous Rocks
Cretaceous rocks comprise the bedrock only along the southern boundary of the San Juan
Metropolitan area. In general, no particular problem has been associated with the Cretaceous
rocks. The upper part of the rock has been weathered to soft decomposed rock and residual soil.
There exists a gradational and often irregular contact between residual soil, soft decomposed
rock, and sound rock.
4 General
Management Architectural and Engineering Consultant
2.2.2
Tertiary Formations
The major portion of the San Juan area is underlain by the Tertiary rock formations ranging from
zero to more than 100 ft. The great irregularity in depth to rock is due to erosion and solution.
The character of the Tertiary formations is also extremely variable.
2.2.2.1
Fajardo Formation
At the southern limits of Section 7 Alignment, the alluvial soils may be underlain by the
Eocene and Paleocene Rio Piedras siltstone (GMAEC, GDR, April 1996). This formational unit
could be what Kaye (1959) refers to as the Fajardo Formation (early Eocene or late Paleocene).
Most characteristic of the Fajardo formation is a relatively soft, very well bedded, light colored,
non-fissile aphanitic 5 rock of rather low density. The color is light yellow, light tan, white, pink,
and various shades of red. The beds are generally about 2 inches thick. Thin intercalations,
usually less than 1 inch thick, of light-gray to white kaolinitic clay are common between siltstone
beds.
The rock is everywhere well jointed, and it breaks readily into rhomboidal-shaped
flagstones and blocky fragments bounded by smooth joints and bedding planes. Despite the fact
that the rock seems to be deeply weathered, it does not decompose readily to soft clay but retains
its cohesive strength and form even when wet. This peculiarity probably accounts for the fact
that in the San Juan area its outcrops generally form prominent topographic highs. Most of the
Montes de Hatillo, south and east of Rio Piedras are cuestas of these rocks. Besides the rather
soft ashy siltstone there is also cream-colored to white thin bedded chert and a siliceous siltstone.
5A
rock in which individual crystals are too small to be identified without the aid of a microscope.
No fossils have been found in the Fajardo formation and it is judged to be late Paleocene or
Eocene in age (Kaye, 1959). No formational unit of the older complex 6 higher than the Fajardo
formation has been recognized in the San Juan area. Between the time it was deposited, probably
in the Paleocene or Eocene, and the onset of the deposition of the next youngest rocks in the
early Miocene, a profound revolution took place during which occurred most of the folding,
faulting, and intrusion of igneous material that characterizes the older complex.
2.2.2.2
Aguada Formation
The basal Tertiary formation in the San Juan area is the Aguada formation that lies
directly upon the eroded surface of the Cretaceous rocks. Because of its depth it normally is not
of concern in foundation work.
The formation consists predominantly of non-carbonate rocks (Kaye, 1959).
It is
composed of sands, gravels, shale, marl, and interbedded limestone . It outcrops in the middle
and southern portion of the San Juan Metropolitan area (Deere, 1955). The beds assigned to the
Aguada formation crop out as soil-covered round slopes and tepee-shaped hills east and west of
Rio Piedras. This topographic form is in contrast with that developed on the overlying dense
Aymam6n limestone, which generally forms karst hills with steep rocky slopes. The Aguada
formation possesses the following characteristics: 1) It underlies the dense limestone of the
Aymam6n formation; 2) it possesses a fauna of Miocene aspect like the overlying Aymam6n
formation; and 3) it consists essentially of interbedded pure limestone and softer chalky to marly
6 Older complex: consist of an unknown thickness of sedimentary and volcanic rocks of late Cretaceous to Paleocene
or early Eocene age.
7 Limestone, and particularly dense limestone, occurs only sparsely and generally as thin lenses (Kaye, 1959).
limestone, although up to 10m of basal sand, gravel, and shale occur where the formation rests
directly on the Cretaceous rocks.
The Aymam6n limestone, which overlies the Aguada formation, is the formation that
forms the bedrock throughout the majority of the San Juan Bay area.
The Aymam6n limestone was originally a hard, finely crystalline, dense limestone.
However, during the period that it was subjected to sub-areal erosion, extensive solution reduced
it to karst topography characterized by caverns, fissures, and sinkholes. Collapse of many of the
openings took place, causing brecciation of the limestone. Clay filled many of the channels and
caverns and became mixed with the limestone fragments. Because of the modified character of
the rock, the term bedrock in the San Juan Bay area does not necessarily designate a firm sound
rock into which piles cannot be driven.
Normally erratic driving resistance is encountered
(Deere, 1955).
The entire San Juan area was at one time covered by these formations (Aguada &
Aymam6n) but they have been almost completely removed by solution and erosion.
The
remnants show collapse structures, clay-filled fissures and cavities, and in general the effects of
solution.
2.3 Quaternary Deposits
The last one million years of geologic time (from the end of Tertiary to present), have been
characterized by: a) severe climatic changes; and b) formation of glacial ice-sheets covering large
areas. Four main advances of the glaciers separated by periods of warmer climate took place.
During the glacial stages, the precipitation was largely retained by the glaciers.
Consequently, the amount of water in the ocean decreased and as a result, there was a general
lowering of the sea level throughout the world {e.g., last glacial age: lowering of 70m to 100m}.
Stream gradient and energy increased resulting in downcutting through and reworking up
previously deposited coastal plain sediments.
Each of the advances of glaciers was followed by a period of warmer climate, probably of
much longer duration than the periods of glacial condition. The sea level rose to a position even
higher than it is now (from 20m to 50m above the present sea level) and finer grained deposition
generally occurred.
In Puerto Rico, nearly every deposit along the coast has a direct relationship with the
change in sealevel corresponding to the various glacial and interglacial ages.
Volcanic Activity and Crustal Stability: At present there are no active volcanoes in Puerto
Rico or the Virgin Islands (since 60 million years ago). However, the frequent earthquakes,
which occur throughout the island, are evidence of continued crustal instability. Puerto Rico is
located within an active seismic area with major fault zones located in the Puerto Rico trench
between the Caribbean and North American plates. There are four major tectonic features near
Puerto Rico capable of generating a major earthquake, these are the Puerto Rico Trench, Mona
Passage, the Anegada Trough, and Los Muertos Trough.
Recent uplift: Meyerhoff (1933) concludes that the uplift has been differential with perhaps
1.5m to 3. lm uplift in the northeastern part of the island and 4.6m to 9. 1m along the western and
southern coasts. The differential uplift in Puerto Rico is probably a combination of the general
lowering of sea level and slight uplift and tilting. This has been of immense importance from the
foundation engineering viewpoint because the lowered water table has allowed soil above the
ground water level to dry out, forming a stiff desiccated crust several feet in thickness.
2.4 Near Surface Deposits
In San Juan and its suburbs (Santurce, Hato Rey and Rio Piedras) the hills and the adjacent
lowlands, which are not covered by more recent lagoonal or flood plain sediments, show
exposures of thick deposits of clay, sand, clayey sand, sandy clay, and occasional beds of gravel.
These soil deposits are commonly referred to as the old alluvium, the blanket sands of Puerto
Rico, and/or the Hato Rey formation (Figure 2.3). The formations are striking in appearance
with colors ranging from white to reddish brown, often in a mottled pattern. The light brown,
brown, and reddish brown color comes from oxidation. The sand is mostly quartz sand, a product
of older Cretaceous bedrock in the interior of the island. Secondary structure is commonly
present in the form of joints, concretions, and occasional slight cementation by iron oxides.
These deposits are noted to be of variable thickness, but probably less than 100m. The presence
of occasional stratification, sand pockets, and lenses; the lack of marine fossils; and the areal
relationship with the definitively fluvial type of deposit where the Hato Rey formation overlaps
the Cretaceous rocks to the south, suggest very strongly the fluvial origin of the deposits,
probably been deposited as a piedmont alluvial plain.
The Hato Rey formation is quite extensive and can be traced westward along the north
coastal plain for more than 50 miles. In that area it blankets the eroded Tertiary rocks, although
solution remnants of limestone (haystack hills, "mogotes") potrude above the general level of the
coastal plain as hills of "circumalluviation" (Deere, 1955).
The sediments rest unconformably upon the eroded and weathered surface of the Tertiary
formations in most of the area but south of the Tertiary-Cretaceous boundary they overlap the
Cretaceous rocks a short distance. The gravels are present only near the Cretaceous boundary.
Away from the Cretaceous boundary, in the area of the Tertiary formations, the sediments are
predominantly clay and quartz sand which have been derived from the Tertiary rocks. The
Aguada formation (lower Miocene) in particular contains many beds of quartz sands and was no
doubt an important source of sediments.
Age: The Hato Rey formation is younger than the erosion surface (eroded Caguana
peneplane) of the Tertiary and Cretaceous rocks upon which it rests, and older than the cemented
dune sands, floodplain sediments, and lagoonal deposits which in many areas overlie it. It predates the period of erosion in which valleys were carved through the Hato Rey deposits and were
infilled with floodplain sediments.
These valleys were eroded during the time of lowered
sealevel of middle Wisconsin 8 and the floodplain and lagoonal sediments were deposited with
the rising sealevel, of late Wisconsin time. Therefore, the Hato Rey formation must be at least
pre-middle Wisconsin (Deere, 1955). Meyerhoff (1927) dates the Caguana peneplane as late
Pliocene, and the uplift and dissection of the peneplane as early Pleistocene.
The Hato Rey
formation could have been deposited contemporaneously with and following the dissection, then
it is reasonable to date the Hato Rey formation as early Pleistocene.
Section 7 alignment: The entire Rio Piedras alignment consists of Older Alluvial (i.e.
Hato Rey) deposits that are Pleistocene and Pliocene silty and sandy clays with interbedded
sands. The predominant soil types encountered during the site investigation were clays and silts
with variable amounts of sand typical of cut-and-fill structures in floodplain areas. The silty
clays and clayey silts both contain some sand and are highly pre-consolidated by desiccation. The
sand grains are fine to medium in size and consist almost entirely of clear quartz, although the
grains are often stained tan or red by iron oxide. The sequence of the different types of soil is
present in an erratic fashion. The sediments are normally in a dense or compact stage. The soil
8 Late
Pleistocene
deposits along this section of the alignment rest unconformably over the bedrock at depths
estimated from 30 to 100m.
2.5 Ground Water Conditions
Local experience indicates that the limestone underlying the clayey coastal plain sediments
comprises a confined bedrock aquifer that, locally, contains ground water under artesian pressure.
In these areas, considerable ground water inflow can occur into excavations. However, there are
no artesian conditions expected in the vicinity of Rio Piedras according to Capacete (personal
communication, 1997).
An unconfined watertable aquifer resides within the soil deposits overlying the limestone
bedrock.
Perched ground water conditions are also possible within the highly variable,
discontinuous, lenticular sand layers that are found within the alluvium (which is generally more
clayey in nature). Some ground water monitoring data (GMAEC, GDR, April 1996) show large
variations over time and could be caused by leaking sewer and water utilities in the vicinity.
2.6 Seismic Considerations
Puerto Rico is located within an active seismic area with major fault zones located in the
Puerto Rico trench between the Caribbean and North American plates. There are four major
tectonic features near Puerto Rico capable of generating a major earthquake, these are the Puerto
Rico Trench, Mona Passage, the Anegada Trough, and Los Muertos Trough. Figure 2.4 shows
the location of Puerto Rico with respect to major geographic and tectonic features.
Approximately 300 earthquakes are registered in Puerto Rico each year, but only a few are felt by
people (GMAEC, GDR, April 1996). The earthquake activity in the Caribbean strongly suggests
that the main sources of possible damaging earthquakes are active and that although no
predictions can be made, this area is susceptible to future earthquake shocks of the same or
higher magnitude than those experience in the past, such as the ones in the years 1844, 1867,
1906 and 1918. The strong earthquakes of 1943 (magnitude 7.5) and 1946 (magnitude 8.1),
whose epicenters were located in the Puerto Rico Trench Fault Zone, did not cause any damage
on the Island due to the high attenuation that took place due to their location (GMAEC, GDR,
April 1996). Figure 2.5 shows the seismic hazard map (1987) for Puerto Rico (Earth Scientific
Consultants and W. McCann and Associates, Inc., 1994). Figure 2.6 shows maps of acceleration
for 50, 100 and 250 year exposure (Earth Scientific Consultants and W. McCann and Associates,
Inc., 1994). Most of the island has values in excess of 0.25g (Earth Scientific Consultants and
W. McCann and Associates, Inc., 1994).
2.7 Overview of Engineering Properties of Hato Rey Deposits
There are two main sources of information on the physical, index and engineering
properties of the Hato Rey deposits: i) PhD thesis at University of Illinois by Deere (1955); and
ii) Geotechnical Data Reports (GDR) of GMAEC for Tren Urbano Section 7 (GDR, 1996a,b).
The following paragraphs summarize the main observations made by Deere (1995) and highlight
some similarities and differences reported from the Section 7 site investigation.
Water contents measured on 70 samples from typical borings in the Hato Rey formation
ranged from 15 to 48 %. The great majority of the values are in the range from 30 to 35%.
Deere (1955) explains that the wide spread in values is due primarily to variations in texture of
the samples, which range from clayey sands to clays. Beneath the water table, the materials are
generally saturated. The GDR data (Table 2.2) show less scatter with water content equal to
approximately 29 + 7%.
Deere (1955) performed a limited number of Atterberg limits tests (15 samples) showing
more uniform results. The liquid limit ranges from wL= 7 8 to 105%, (with an average of 92%)
while the plastic limit, wp= 31 to 47% (ave. 38%). Hence, the plasticity index Ip= 41 to 62%.
The majority of the samples plot close to or slightly above the A-line in the Casagrande
classification chart (Casagrande, 1955) and are classified as high plasticity clays (CH) while
some samples below the A-line are silts of high compressibility. The GDR data (Table 2.2) show
less scatter and lower values with wL= 53 + 17 % and wp= 24 ± 7%, where most of the data plot
above the A-line in the Cassagrande classification chart.
The activity ratio was determined on 4 samples and ranges from A=0.95 to 1.60, and hence
are classified as normal to active sediments. Two samples were studied by X-ray diffraction and
differential thermal methods to determine their clay mineralogy. One of the samples showed
poorly associated kaolinite with some illite 9. The other, which was studied by X-ray diffraction
method' o , gave indications of being a member of the kaolin family but not kaolinite. This could
possibly be poorly associated kaolinite, or perhaps halloysite.
The cohesive sediments of the Hato Rey formation are normally of stiff to very stiff
consistency. Unconfined compressive strengths of over 70 samples from typical borings range
from 1.0 to 5.5 TSF and average 2.9 TSF. The lower values were associated with sand pockets
or joints in the sample that caused local failure during testing. The N-values obtained from the
standard penetration test range from 15 to over 100. However, the samples normally fall in the
9 Work by Dr. E. Grim, (Deere, 1955)
10 Work done by the Laboratorio Industrial in Puerto Rico, (Deere, 1955)
range N= 20 and 40. This would indicate a consistency of very stiff to hard, with unconfined
compressive strengths on the order of 2.0 to 4.0 TSF or greater.
The wide scattering of the values is said to be caused by joints, the wide range in textural
composition, and the random sand pockets all typical properties of the Hato Rey sediments.
Since settlements are not a serious problem, the compressibility characteristics of the sediments
have not been extensively investigated at the time of Deere's thesis and the results of only a few
consolidation tests were available. A review of these data shows that the applied consolidation
stresses were significantly smaller than the pre-consolidation pressure (there is no well defined
yield in the e-logav' data). By using the empirical relationship, CC=0.009(wL-10) (Terzaghi and
Peck, 1948), to estimate the compressibility of normally consolidated clays, Deere (1955)
estimates C,=0.72. By doing curve fitting to the measured data he estimates a preconsolidation
pressure, up'= 20 TSF. Other test results suggest much lower values than this, op'= 3-5 TSF.
Deere also suggests that the sediments of the Hato Rey formation possess many
characteristics that are normally associated with swelling clays. Terzaghi (1955) states that clays
which have been investigated with relation to heave phenomena have a plasticity index greater
than 30 and a liquidity index close to or below zero, are intensively jointed, and have a water
table located at a depth of 15' or more below the surface. The author argues that the Hato Rey
formations meet these conditions in most respects with the exception in some areas of the deep
water table and intense jointing.
Deere (1955) states that no heave phenomena have been
observed to date, but the probability that swelling could be of importance under certain
conditions should be realized. Swelling properties may be of particular importance for proposed
deep excavations in the Hato Rey deposits associated with Section 7 of the Tren Urbano.
Table 2.1 Bedrock Formations in San Juan Metropolitan Area
Thickness,
Description
Stratigraphic
feet
unit
Fajardo
Formation
Light colored ashy siltstone, siliceous siltstone and
chert, interfingering graywayke, conglomerate, and
3,000+
early Eocene?
or late
Paleocene?
impure limestone.
Aguada
Age
Friable sandstone, clay, and concretionary limestone
325
early Miocene
Thick-bedded, light colored, dense limestone
950+
early Miocene
Formation
Aymam6n
Formation
Table 2.2 Comparison of Index Properties for Hato Rey Alluvial Sediments
Property
Deere (1955)
GDR (1996a)
GDR (1996b)
Water Content, w (%)
15 - 48
29 ± 7
29 ± 8
Plastic Limit, Wp (%)
31-47
24 ± 7
23 ± 5
Liquid Limit, WL (%)
78 - 105
53 ± 17
52 ± 14
Compressibility
0.72
0.38
0.38
Index,
C,* (of N.C. clay)
* Estimated from empirical relation, Cc=0.009(wL-10%) after Terzaghi and Peck (1948)
°
800
85.
75
i.i•_
i i"._L
ii
.
PS
igr
I
,ii~
i:
:
: :
:'
.. .. .
.
: :
..
.
.."
I
.: ." "
+
:
:::
:
\.
ii::
i
.1
"
"
:
iJ
- -.
ri-i.~- -inMAC
I:-:
!
S:
"
-"
: i
I
?A~
i+
LlA
r~ri:il-:-.;l_:_
-::
I-
+i~ ::I
:i
ii~i::''i
::!]i'
j:l::::
':1.:
:.
::
1:.
;::.:i
ii:!:i:i
1:
N1
CI
(i
A...N:::::::::':...
..... V
I
-LAI UN L A):iii
!!
i::.
i::;:
i:iii::i
ii:::
iii:i!
;i.:.i:
oii!;!i
:ii
i i ::
:i::ii::i:::iiiiii
:.
:i:ij
:i
: -:
B;i:LIC I : i:i;.
.i
~t
'i~ii:i:iii:,
i
S
Figure 2.1 Map of the Greater Antilles
::
:
,
REIUBL C
OFj
: :
i-:----::-.ir::__ii':::
.:_i::i~~~~~i
7'
A
.;::.
::;.i
: ::
::.
:::i
AA
: :ii:
:
!:
"]:::::.:
.::
....
::
::::[
"!
": :" "6'::.
' " (o(F S/PA
: . ...
.:
ii
oftAIGreterAntPUETOll
-
:.
.
. ..
r..
ii
Cayman .I
I
:
.
l'iii: j ;i.i
I
cc<,
,I.
NRO:
7"::_ '.;- A :':.
A
:
I
ANDRO
.
.•
.....
"ineS
..
iii : ':iiii
i :iiii
:
--
: + ' .....::::
_ii :
-:.;:T:;:.lii"i.:ii;.i
: i. •
i_;i ,,I.:::
-.
J
65'
70"
E AA
iiiii'
iiii
:ii iilj :
. . .":
ii .. iiiiii:iiiii
",::iiiREPU
RICO
i'~i "i
+ Mons
"ii~~iii
i:ii iliiii
i
riiil.:il R IC
~O y Ii-d:i i
~!=.
i".i""
"[!!:" ! ! i ..:[i!;!::"
L. . . . . . .
. . . . .. . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . .. . . .
:00
I 00
. . . . . . . . . .
.00
•.
I
Atlantic Ocean
Hato Rey Formation
Alignment
Station
San Juan Bay
Catafio
/
San Juan
/
'4
Q
" *"
®
\
4
4
4*
4*
1'.
tUP Station
4*
.
4*
Rio PiedrasStation
'i
.4 I.
iI.
.4
%4
*
-f
N.Z
L°
Figure 2.3 Hato Rey Formation
r
, ,
,
.1
//
* 4/I
I.-.4-
4*4
* ,
A,'I
*4,
4*
4*
" ,
...
|
/
4*
•
-'
/
"
iN
I
LESSER
",ANTILLES
TROUGH
701
65'
s60
Location of Puerto Rico with respect to major geographic and tectonic features. The
Caribbean Plate extends to the east of the Lesser Antilles and to the north to Puerto Rico and
Hispaniola. The oceanic deeps flanking the island such as the Puerto Rico Trench. Los Muertos
Trough and Anegada Passage are all geologically active features and are the principal causes of
the island's earthquake hazard (after McCann and Sykes. 1984).
Figure 2.4 Location of Puerto Rico with Respect to Major Geographic and Tectonic
Features
s55*W
'6'
S0
Year
Exposure
33 9
45%6
,,5e
C~
33%.+
I,
ft
16
M.
Calculated probabilities that any given area will experience gound accelerations e.%ceeding 0.1g in a 50 year period.
100
70
%
N
Year
Expourm
50 %
I
G~aw
Calculated probabilities that any given area will experience round accelerations exceeding 0.1g in a 100-year period.
Figure 2.5 Seismic Hazard Map of 1987 for Puerto Rico (Earth Scientific Consultants
and W. McCann and Associates, Inc., 1994).
-
1994 54) Year E..inure Variation = 1L
'^ ^^
-600
"'^
-67.0
"^^
-67.00
--6.
-66.00
1994 100 Year Exposure Variation
1
-6.50)
-65.00
0.6
01.79%g
(1.75%g
0. 7 1%g
07
0.
og
0.6 3 %g
47
0.
0.
1%g
0.55%g
%g
O.JS%g
33
O.%g
,3 1%g
17.3%
-6&00
6i '
-67.00
-663.0
-66.00
0 .1 9 %g9
R0.17%gyo
66&0O
1994 250 Year Exposure Variation0.6
-o.isfg
0..1 % g
S
-6&800
-67.50
-67.00
-66).,0
-66.00
-6 .SO
-6&(0
Map of acceleration not to be exceeded at a probability of 90 % as zalculated by the program scisrisk
III and the parameters described above. In this case the recommended gound variabilirt oif 0.6 is used. Note that
most of the island has values in excess of .25. and that extreme values c!canr' !\ccd ,(30 on the island of Puero
Rico and 0.35 on Vieques
Figure 2.6 Maps of Acceleration for 50, 100 and 250 Year Exposure (Earth Scientific
Consultants and W. McCann and Associates, Inc., 1994).
3.
Site Investigation for Rio Piedras Station
3.1
Introduction
The site characterization is based on geotechnical exploration programs already carried
out for Section 7 alignment, that were intended for final design and construction of the section as
a design-build contract (GDR 1996a, 1996b). Additional geotechnical evaluations and design,
possibly including additional geotechnical field investigations and laboratory testing, will
necessarily be required to complete the project (GDR, 1996a).
3.2
Scope of Site Investigation
Geotechnical investigations for the Rio Piedras Section were started in March 1995 as
part of the preliminary geotechnical field investigation for the Tren Urbano.
No previous
geotechnical engineering studies were conducted specifically for the Rio Piedras Section of the
Tren Urbano prior to those conducted through Tren Urbano/GMAEC. However, there have been
previous studies conducted for other reasons near the Rio Piedras alignment (Deere, 1959; Kaye,
1959; and Monroe and Pease, 1977). Geotechnical field investigations were completed in two
phases (I and II) ending during February 1996 (GDR, 1996a).
Also, a final phase (III) of
supplemental geotechnical investigations for Rio Piedras were conducted during August,
September and the first week of October 1996 to obtain additional information for prospective
bidders (GDR, 1996b).
The purpose of the Phase I investigation was to study the general subsoil conditions in
the Rio Piedras area in order to recommend possible underground construction techniques for the
Rio Piedras. The Phase II investigations included field and laboratory tests, such as, pocket
penetrometer tests on all samples, limit and grading tests for detailed soil classification,
unconsolidated-undrained (UU) triaxial compression tests, and slug tests to estimate the
hydraulic conductivity (test performed in standpipe piezometers installed in nine borings). The
Phase III investigations include further laboratory testing (UUC, CIUC triaxial shear and
consolidation tests), Menard type pressuremeter tests, piezometer installations and two additional
pumping tests.
The Phase I investigation program included 6 borings along the present Rio Piedras
alignment drilled to depths varying from 15 to 40m (Table 3.1), between Sta. 217+00 and Sta.
225+00 (see Figures 3.1a-b).
Six additional borings were drilled east of the Rio Piedras
alignment at and near Highway PR-3 in order to investigate the possibility of a future extension
of the alignment to the town of Carolina (Phase I borings are identified by the number of the
station closest to each boring).
Phase II consisted of drilling ten borings along the revised proposed tunnel section
between Sta. 217+00 and Sta. 230+00. These borings were drilled to depths well below the
proposed top of rail along this section of the alignment, typically in the order of 30m. These
borings are identified as P-borings (P-l to P-11).
Continuous sampling was done in the P-
borings from above the proposed tunnel crown to below the rail elevation.
The Phase III investigation program included six additional geotechnical test borings to
augment the previous studies. The borings (designated B 101- B 107) were drilled at midpoints
between previous borings (Phase I and II) at areas of interest based on subsurface conditions
encountered in the previous borings, and at the locations of important underground construction
elements such as tunnel station and shafts. The borings were drilled to depths of 20.3 to 30.6m
below existing grade and included continuous sampling from approximately 3m above crown
elevations to 2m or greater below invert grades. Table 3.1 gives a brief description of borings
done for all investigations.
The test borings were made using conventional truck-mounted CME geotechnical drill
rigs 83mm hollow stem continuous flight auger casing. In this method the borings are advanced
by turning the auger into the ground a desired amount distance. Sampling below the bottom of
the auger is attained by inserting the sampling apparatus within the auger, eliminating the need
for casings. Two types of samples were taken: 1) standard (Terzaghi) 35mm I. D. split spoon
(disturbed) samples, and 2) 75mm inside diameter, thin-walled, stainless-steel, Shelby tube
samples ("undisturbed"). The undisturbed samples were obtained by forcing the Shelby tube
samples into the ground using a static force or downward pressure and was pulled out also using
a static pull. The Shelby tube samples were sealed in the field with wax, end caps and tape and
shipped to soil laboratories for testing.
3.3
Stratigraphy
3.3.1 Section 7
Figures 3.2 and 3.3 show the stratigraphy for the entire Section 7 Alignment. This figure
confirms the erratic nature of the deposits, which comprise a series of erratic layers of clays, silts
sands and sandy clays. The composition of the predominant elements of the alluvial deposits can
be described as follows: 32% clayey sand, 24% clay, 13% sandy clay and 11% sand. The clayey
layers, which are heavily overconsolidated, contain considerable amount of silts and sands. In
general, there is great scatter in the data due to the wide range in textural composition of the
Hato Rey deposits.
Since a lot of spatial variability exists it is not possible to identify
characteristic continuous horizontal layers. In the following sections the groundwater conditions
and index and engineering properties are discussed.
3.3.2 Rio Piedras Station
Figure 3.4 illustrates the stratigraphy near the Rio Piedras Station.
About 26% of the
stratigraphy are comprised of clays, 13% sandy clays, 2% silty clays and 2% gravelly clays
(total: 43% of clayey material).
Approximately 3% are sands and 44% clayey sands, totaling
about 47% sands. Also, included are 5% silts, 2% fill and 1% organic.
The subsequent numerical calculations approximate the complex stratigraphy by a single
(homogenized) layer using average properties of a sandy clay (or clayey sand).
3.3.3 Index Properties
In order to classify samples according to the USCS and refine visual-manual soil
classifications on the boring logs, selected samples were used for Atterberg Limits (ASTM
D4318), natural moisture content (ASTM D2226), gradation or sieve (ASTM D422 and D 1140),
and hydrometer (ASTM D422) analyses.
The Atterberg Limits are very useful for soil
identification and classification (Lambe and Whitman, 1969). They describe the consistency of
fine-grained soils with varying degrees of water content.
The limits can be used for quick
estimates of engineering properties through empirical correlations.
There is some scatter in natural water content but it decreases with depth (see figure 3.5),
with w= 29+ 7%. More scatter is shown the measured liquid limits (Figure 3.6), which range
from wL= 5 3 +17%. In general wL decreases with depth.
The plastic limits of the alluvial
deposits (Figure 3.7) exhibit similar degree of scatter as the natural water content, with wp=24
±7%. In all Atterberg limits plots there is wide scattering of values that might be due to the
broad range in textural composition of the deposits. There is wide scatter in the plasticity index
(Figure 3.8), with Ip=29+ 15%.
Figures 3.9 and 3.10 show the Casagrande classification chart for Phases I, II and III.
The majority of the data plot close to or above the A-line with most of the samples plotting as
clays of low plasticity (about 50%) and clays of high plasticity (about 50%).
Figure 3.11 illustrates how the Liquidity Index varies with depth. A great amount of
scatter is evident, with a Liquidity Index equal to 0.17± 0.26.
Figure 3.12 shows dry and saturated unit weight values from several tests (GDR 1996b).
Average saturated and dry unit weight of 17.85±0.94 kN/m 3 and 13.35+1.1 kN/m 3 , respectively,
were selected.
3.4 Groundwater Conditions
During the field investigations, piezometers were installed in two of the Phase I borings
and in nine of the Phase II borings along the alignment. The ground water levels in general were
monitored during the period from March 1995 to March 1996. During Phase III investigation
piezometers were installed in all six borings. The piezometers were screened at various depths to
monitor piezometric ground water levels in sandy strata encountered at various depths in the
borings.
The instal!ati~ns used open standpipe piezometers (50mm diameter) with slotted
screens and filter packs, were set up as monitoring points along the length of the alignment.
Additional piezometers were installed as part of the groundwater pumping tests carried
out during the three phases of geotechnical investigation.
Locations of the piezometer
installations are listed on Tables 3.2 and 3.3.
A pumping test was performed during October 18 through 21, 1995 next to boring 226.3,
near Sta. 225+00. It was performed to obtain preliminary indication of the difficulty of creating
a drawdown given the erratic characteristic of the soil deposits that will be encountered in
tunneling. The results from that pump test were included in GDR (1996a). The test consisted of
drilling a pumping well to a depth of 30.4 meters. The well casing was perforated and wrapped
with filter fabric from 5m below the ground surface to a depth lm above the base of the
borehole. The annular space was packed with sand instead of gravel. Three piezometer clusters
were installed at 2.3m (226.3-P1), 4.6m (226.3-P3) and 9.2m (226.3-P2) radial distance from the
well (Figure 3.13). These piezometers were installed at depths of 15.8, 21.0 and 26.8 meters at
the 226.3-P1 monitoring well; 21.3 and 25.9 meters depths at the 226.3-P3 monitoring well and
at 15.8, 19.8 and 26.8 meter depths at the 226.3-P2 monitoring well. A submersible pump with
intake at 28 meters deep was used. A maximum drawdown was created in the well to a depth of
26.8 meters and pumping was continued for 80 hours.
The results of the tests (see Figure 3.14) are only applicable in the immediate vicinity of
the test. Test well pumping produced only a small drawdown in the adjacent, approximately
0.67m and the following conclusions were reported in GDR (1996a):
1. The computed permeability of 3x10-3 cm/sec is equal to the highest values obtained in the
small-scale falling head tests in the piezometers of the exploration program. Under the
condition of a very small drawdown, the permeability computation is highly unreliable' and a
value exceeding 10- 3 cm/sec probably represents only a minor portion of the overburden
soils.
2. The very small drawdown and the substantial flow to the test well suggest that there may
have been local recharge directly into the well, presumably rainfall on the surface and from
leaking shallow utilities seeping toward the water table that affected the test.
Two supplemental pumping tests were performed as part of the Phase III geotechnical
investigation during 3 September through 4 October 1996. The two new pumping test locations
were selected to investigate aquifer conditions in the vicinity of relatively sandy soils
encountered near the intersection of Ponce de Le6n and Gaindara Avenue (supplemental Pump
Test No. 1), and in the vicinity of the south shaft area for the proposed mined Rio Piedras Station
(supplemental Pump Test No. 2).
The tests were performed to evaluate the impact of
groundwater on the construction, so that appropriate construction methods can be chosen, and
groundwater control schemes designed and estimated.
Pump Test No. 1 was performed within the section of twin mined tunnels at about Station
223+24, just north of where the tunnel alignment passes under Dr. Jose Gindara Ave (Figure
3.15). Test Well PW-1 was installed at Station 223+23.95. Pump Test No. 2 was performed at
the south end of the proposed Rio Piedras Station at about Station 219+61 (Figure 3.16). Test
Well PW-2 was installed in Calle Arzuaga at Station 219+81.49.
Both tests were designed with 30m deep, fully screened, 150mm diameter pumping
wells, installed to maximize potential yield from the formation strata. Both tests included
1It was also stated that the borings logs indicate
cm/sec would not be unreasonable.
the presence of SP lenses, and that in that case, an average of
-3
10
multiple piezometers screened at "shallow", "middle", and "deep" horizons in monitoring wells
at 3 m and 9 m radii from the pumping wells in relatively sandy strata. The pumping wells were
installed using a Mobile B90 drill rig equipped for mud rotary drilling. The holes were advanced
using a 35.6 cm (14") diameter tricone drill bit attached to an approximately one foot diameter
by ten foot long drill rod stabilizer.
Development was accomplished by a combination of
airlifting, swabbing, surging, and pumping.
The piezometers were installed using Acker AD2 and BK 50 drill rigs using 8.26cm (31/4") hollow stem auger casing.
Shallow, intermediate and deep screened piezometers are
referred to by the designations A, B, and C, respectively, at given radii from the test wells.
Both tests were intended to create maximum stress on the alluvial aquifer system by
pumping at the highest possible rates, thereby creating maximum possible drawdown. Pump
Test No. 1 took place during 27-30 September 1996 and was initially pumped at a rate of
approximately 100 gpm. During the first hour of the test the rate had to be throttled back to a
constant 70 gpm (to maintain pump operations), for a total test duration of 72 hours.
The
sustained drawdown in the pumping well at 70 gpm was equivalent to a drawdown of
approximately 24m in the pumping well (i.e. approximately 5m above the pump intake). Test
No. 2 took place during 1-3 October 1996 and was initiated at the same maximum pumping rate
of about 100 gpm. But the pumping rate was quickly throttled back after casing storage was
depleted. Maximum drawdown was maintained to the pump intake at approximately 29m depth
without pump operation problems during this test. After several hours the test maintained a more
or less constant rate of approximately 20 to 25 gpm in contrast to the higher rate of the first test.
The results of these two pumping tests, shown on Figures 3.17 and 3.18, were included in
GDR (1996b). The conclusions reached in the report were the following:
1. That a properly constructed and developed well in the area of the test at Gindara Ave.(pump
test No. 1) will give good yield, and produce significant drawdown.
2. That a properly constructed and developed well in the area of the test at the Rio Piedras
Station (pump test No.2) will provide low to moderate yield, and produce drawdown.
Tighter well spacing and more extended pumping time will probably be necessary for
groundwater control than in the area of the pump test at Gindara Ave.
3. The two supplemental pump tests show well yields moderately greater than would be
anticipated from the soil descriptions and laboratory tests. The drawdowns observed during
the tests indicate greater communication horizontally than would be expected from the soil
data. In both tests, the screened horizons exhibited different responses to pumping. The
maximum drawdowns measured at r= 3m Test No.1 at (78.5 hrs.) were 2.1m, 7.8m and
11.7m, for the shallow, intermediate and deep piezometers, respectively.
Equivalent
drawdowns due to Test No.2 (at r= 3m, 53 hrs.) were about 0.5m, 2.9m and 1.0m,
respectively.
4. From a dewatering viewpoint, the soils above and below the tunnel and station construction
present a complex series of heterogeneous, anisotropic strata of erratic continuity.
5. Because of the anisotropic nature of the soils, distinct vertical gradients have been observed.
Typically, the higher the piezometer screen, the higher the water level observed. It appears
that the source of recharge is shallow, from infiltration, leaking utilities and possibly from
flow off the hills in surficial aquifers. This vertical component to the groundwater flow may
result in perched water conditions that could impact construction operations.
The report states that it is not feasible to fully design an effective dewatering system in
advance because of the variability of the soil and groundwater conditions along the alignment of
Section 7 of the Rio Piedras Contract of Tren Urbano, and in view of vertical flow and other
departures from conventional dewatering assumptions. The observational approach is strongly
recommended along with installation of additional wells for further testing.
3.4.1 Water Table and Equilibrium Pore Water pressures
Figure 3.19a shows the ground surface and groundwater elevations for all three of site
investigations (data from the pumping tests are also included). It can be seen that for most cases
there is variability within five meters. Within the Rio Piedras Station area (see Figure 3.19b),
which has an average ground elevation of El.+26m 2 , the water table elevation changes from
13.41m to 18.59m, with an average elevation equal to El.+16m.
Figure 3.20 shows the initial stress profile using the dry unit weight above the water table
and the saturated unit weight below the water table. In-situ effective vertical stresses (c'vo) were
calculated using the following equation: o'vo = avo - uo, were (vo is the vertical total stress, and uo
is the pore water pressure.
(Note: at the Rio Piedras Station, the ground elevation is
approximately 26m and the groundwater averaged an elevation of 16m).
3.4.2 Hydraulic Conductivity
3.4.2.1 Field Test Data
During Phase II, permeability tests were performed in each of the nine P-series borings where
piezometers were installed on clayey and sandy soils. Both rising head and falling head slug
tests were performed. The tests were carried out using the following procedures: i) the rising
head test consisted of bailing out water within the piezometer down to an elevation close to or
slightly below the screened section (to the base of the piezometer, if possible), and then
measuring the time taken for the water pressures to rise to the static level. ii) The falling head
test consisted of adding water to the piezometer up to the ground elevation or top of the
piezometer and measuring the time taken for this level to return to the static water elevation. The
results from these tests (Figure 3.21) shows an average hydraulic conductivity, k=3 x 10-4 cm/sec,
with maximum and minimum values equal to 1.3 x10-3 cm/sec and 2.7 x10-6 cm/sec,
respectively. In general higher values of k are measured in falling head tests. Most of the high k
results correspond to silty and clayey sand. The permeability computed in the first pumping test
(Phase I and II), k=3x10 -3 cm/sec, is very similar to the maximum value for slug tests.
Additional slug tests were performed during the Phase III investigations as part of a future
addendum, but were not available at the time of this study.
3.5
Strength and Deformation Properties
3.5.1 Test Procedures
3.5.1.1 Field Tests
Field tests performed on Section 7 include Menard Pressuremeter (PMT) and Standard
Penetration Test (SPT). Standard Penetration Tests (SPT) were performed in conjunction with
split spoon samples in accordance with ASTM D1586. Each test consists of recording the
number of blows required to drive the sampling spoon a distance of 30 cm into the ground using
a 64kg hammer falling 75cm. The Standard penetration N-value were recorded as the number of
2Mean Sea Level elevations
blows required to advance the sampler 30cm beginning 15cm below the top of the sample
interval.
In order to obtain in-situ stress-strain characteristics of soil with depth, three pressuremeters
were successfully completed during the Phase III investigations. Strength measurements were
obtained in terms of pressuremeter modulus (E) and the pressure limit values. The designations
for these borings are PM-1, PM-3 and PM-P6. PM-1 was drilled adjacent to boring B106 near
the middle of the proposed UPR Station (see Figure 1.3). PM-3 was drilled adjacent to GDR
Phase Two boring P-2 at the south shaft area for the Rio Piedras Station. PM-P6 was drilled next
to GDR Phase Two boring P-6 near the twin mined tunnels at the intersection of Ponce de Le6n
and Gindara Avenue.
The pressuremeter borings were advanced using similar hollow stem auger drilling
equipment as described above for the sample borings. After drilling to the target test interval, a
clean, smooth-walled borehole test interval was established by pushing a Shelby tube. If the soil
stratum was too stiff to advance the Shelby tube then the test interval was accomplished by slow
rotary drilling fluid ahead of the auger casing. After creating a smooth borehole wall below the
casing, the pressuremeter was lowered into position and the expansion test conducted.
Pressuremeter tests were performed at 3m intervals in each of the three pressuremeter borings
(GDR, 1996b). The pressuremeter used was a Menard G-AM Pressuremeter. In this type of test,
the loading component is a dilatable cylindrical probe set in place within the soil at the desired
level of testing in a previously drilled borehole. Once installed, the prohb is submitted to an
increasing pressure applied in equal incremtPs. A. each pressure stage the volume changes of
the probe are recorded at specific time intervals. The pressure-volume relationship is then drawn
up for subsequent determination of material properties.
3.5.1.2 Laboratory Tests
A combination of split spoon and Shelby tube samples were used for laboratory testing in
order to determine strength characteristics of soil strata encountered in the borings. In general,
the same types strength tests were performed in the three Phases of investigations. However,
based on requests from prospective bidders, additional testing were performed in Phase III with
emphasis on CIUC and UUC triaxial testing. The tests were performed in accordance with
referenced ASTM specifications.
During Phases I and II, the laboratory strength tests performed consisted of Pocket
Penetrometer (PP), Unconfined Compression (UC) tests, and Unconsolidated-Undrained (UU)
triaxial tests. The laboratory strength tests performed in Phase III include unconfined
compressive strength (UC) tests (ASTM D2166), unconsolidated-undrained (UU) triaxial tests
(ASTM D2850), consolidated undrained (CU) triaxial tests (ASTM D4767), and laboratory vane
shear tests (ASTM D4648). Shelby tube samples were tested both in Puerto Rico by Jaca &
Sierra (UU, consolidation and laboratory vane shear) and in Miami by Ardaman & Associates
(CIU).
Penetrometer tests were performed on cohesive split spoon samples that exhibit certain
amount of cohesion. The unconfined compression test (ASTM D 2166-72) was performed on
the best quality sample recovered in the split spoon sample 3. The unconfined strength values
measured from the split-spoon samples are used to compare relative strengths and as an
additional index property for classification and identification purposes. More accurate values of
strength are obtained from thin-walled Shelby tubes. Shelby tube samples were sent to Ardaman
& Associates, Inc., in Florida for triaxial testing. The UU tests were performed in order to obtain
information helpful for the design and construction of the tunnels.
The CIU triaxial tests were only performed on several samples from the first two borings
drilled, B 106 at UPR Station and B 102 near the middle of Rio Piedras Station, because of the
long turnaround time for this type of test (GDR, 1996b). The tests were performed in order to
provide supplemental triaxial compression information for the design and construction of the
stations and tunnels in accordance with requests from prospective bidders. The CIU and UU
tests were run on tube samples taken from depths of interest related to the vertical tunnel
alignment (i.e., near the crown, springline or invert elevations).
3.5.2 Undrained Strength, S,
Results from measured Standard Penetration Tests are included in Figure 3.22a, and show
the following main features:
1 Above El.+10m, N remains relatively constant, varying between N=10 to 30.
2 Much more scatter exits below El. 10m. SPT N increases with depth, with high values, N 2
190, and low values N = 10.
Undrained shear strength, so, can be estimated from SPT N values using two empirical
correlations, i) so = 0.13N (ksf) (Terzaghi and Peck, 1948); and ii) s" = 4.4N (KPa) (Simpson, B.
et al., 1979). Based on these correlations, s, = 100kPa above El.+10m, Figure 3.22b, s" =
200kPa at El.+10m increasing to s -= 600kPa at E1.-8m 4 .
3 At best these samples are highly disturbed due to the hammering process by which they were obtained.
4 Pocket Penetrometer results show great scatter (see Figure 3.23), with s. = 25 to 240 kPa (qu=50 to 480 kPa; and
can be classified as medium to hard clays)
Figure 3.24 shows the undrained strength, su, from the pressuremeter tests estimated from the
following empirical correlations:
1 Baguelin et al 1978:
s,, (TSF) =
2
P -P
12
Amar and Jezequel, 1972:
su(KPa) =(
3
P - O
+
+25
Centre d'Etudes Menard, 1967:
s,, (TSF) =
P- P
5.5
The first two of these correlations tend to give approximately constant su with depth, with an
average su= 140 KPa.
Figure 3.25 summarizes the undrained strength results from all the laboratory tests and
correlations for field tests.
Although the results show a large scatter, su 5 200 kPa above
El.+10m. Below this elevation, the principal source of scatter are SPT-N data points, and these
may reflect stratigraphic variations not seen in laboratory tests on small samples of soil.
The selection of the undrained strength is highly influenced by the quality of the samples and
test procedures. The conventional practice of conducting shear tests tends to be highly empirical
and often unreliable because it neglects to account for the following three principal factors that
affect the measured su: i) anisotropy, ii) strain rate (or time to failure), and iii) sample
disturbance.
Pocket penetrometer tests (PP) and Unconfined compression (UC) tests were
performed on highly disturbed samples (split-spoon samples). The pocket penetrometer tests
give a rough approximation of the strength of a cohesive soil and should be used as an additional
index property for classification and identification purposes.
Standard Penetration tests give
very poor measure of su in low overconsolidated soils and can only be used in stiff,
overconsolidated soils.
The M6nard pressuremeter test results are similarly affected by the
influence of stress release that is associated by the pre-bored hole of the test.
More accurate values of strength are obtained UUC and CIUC triaxial tests performed on
thin-walled Shelby tubes.
The use of UUC testing to estimate su depends on the following
compensating errors: i) increased su due to neglecting anisotropy and rate of failure; and ii)
decreased su due to sample disturbance. Better estimates of undrained strength may be attained
using CIUC test results. In order to select s,, the values of the laboratory vane were not included
because they were too low compared to the rest of the undrained tests. Also excluded are the
results from the SPT N correlations because below El. +10m the data show large scatter and a
trend different from that of the undrained tests. Selection of undrained strength, shown on
Figure 3.26, involves UUC and CIUC tests, since these tests were performed on the best samples
available. There is some degree of scatter, with an average su= 150 kPa (147±48 kPa)
3.5.3 Drained Strength Parameters
Based on the estimates of undrained shear strength, su, it is possible to the preconsolidation
pressures, a'p, by invoking the SHANSEP 5 equation for normalized soil shear strengths:
si
= S UP
where su is the undrained strength and a'vo,,
is the in situ vertical effective stress. Typical
values for S and m ranges from 0.22±0.03 and 0.8±0.1, respectively. Values of S=0.25 and
5 Stress History And Normalized Soil Engineering Properties, (Ladd and Foott, 1974).
62
m=0.8 were chosen. Results (Figure 3.27) show maximum ;'p= 1000 kPa and minimum o',=
300 kPa (average 600 kPa). Backcalculated overconsolidation ratio (OCR), shown on Figure
3.28, vary mostly from OCR= 8 to 1.5 (few samples at El.+25m plot with OCR>>8).
The undrained cohesion, c', can be estimated using an empirical correlation developed by
Mesri&Adel-Ghaffar (1993), shown on Figure 3.29. By previously selecting an average a'p=
600kPa, two values of cohesion, c', are determined from Figure 3.29: 1) an upper bound value of
c' equal to 60 kPa (2 < OCR < 5); and 2) a lower bound value of 16 kPa (10 5 OCR 5 20). The
average from the two estimates c'= 40 kPa was selected and used in the analyses described in
chapter 4.
Effective stress measurements in CIUC triaxial tests provide the only source of information
on drained strength parameters for Hato Rey soils. Figure3.30a, report drained friction angles
0'= 30.30, 32.90 and 35.80. The cohesion intercept (c') is very difficult to estimate from these
limited CIU data.
Once the drained cohesion intercept (c'), was selected, an estimate of the drained friction
angle, 0' cann be made from CIU effective stress data. Figure 3.30b shows results at maximum
o)'-o'3 using an intercept of c'=40 Kpa 6 . The interpreted drained friction angle varies from
0'=24.60 to 24.90.
3.5.4 Compressibility and Shear Stiffness
Figure 3.31 shows compression data from conventional incremental load oedometer tests
on Shelby tube samples from Phase Ill site investigations. The data are plotted in conventional
e-logloc'vo space.
Results show how compressibility of samples varies as a function of the
vertical confining effective stress. The data show a characteristic non-linear response, but no
well-defined yield point corresponding to a pre-consolidation pressure (i.e., O'p larger than
maximum c'vo).
Figure 3.32 shows the compressibility, mv, of the soil specimens at the in-situ vertical
effective stress a'vo, as functions of the initial void ratio and water content (Appendix C shows
the calculations of my from the measured data). In general mv increases as w (or eo) increase
with typical values ranging from mv = 0.0002m 2/kN at eO = 0.6 to my = 0.005m 2/kN at eo = 1.0.
These values correspond to the range expected for very stiff overconsolidated clays (see table
3.7)
Young's Modulus, E, can be calculated from results of: i) pressuremeter expansion tests, ii)
CIUC tests (conventionally reported at 50% strength), and/or iii) oedometer tests.
=
and E= (1+ v')(1- 2v') da'V
doa'
v
(1- v')
dE,
In this third case, the drained Poisson's ratio must be assumed, while cases 1 and 2 are
assumed undrained (vu = 0.5). Results from all of these tests are shown in Figure 3.33. Note that
the minor principal stress a' 3 in figure is equal to the cell pressure at the end of consolidation,
o'cell in the CIUC tests. For the M6nard pressuremeter and oedometer tests
3'=
',vo.
The
highest values correspond to the pressuremeter tests followed by the CIUC tests. The oedometer
data give very low moduli compared to the CIU and M6nard pressuremeter tests.
6 a' = c'cos4'=33.5KPa; tana' = sino'
3.5.5 Consolidation Properties
The vertical coefficient of consolidation can be calculated from laboratory oedometer
tests using two methods: log-time and square-root time method. Appendix B shows log-time
(B1-B7) and square-root time/ settlement (B8-B 14) curves for all tests.
It was difficult to
determine cv since in most cases (even for clays) the curves were of the form corresponding to
clayey silts or silts (Head, 1980).
For clayey silts, the initial convex-upwards portion of the primary curve was passed
before any readings could be taken. This indicates that settlements would occur quite rapidly
and would not be expected to cause long-term problems. For silts, a typical log-time/ settlement
curve is concave-upwards from the start (t50 <0.1 min). This indicates very rapid consolidation.
The doo0 (100% primary) point cannot be determined by the conventional method, and the
square-root-time curve is of little use for determining the do (0% primary) point because there is
no linear portion evident. For more details on how to estimate cv for these types of soil refer to
Head (1980).
There is large scatter in results (Figure 3.347), and significant differences between values
from the two methods. Much lower results are obtained from the square-log-time method, with
cv varying from 0.35 to 6.3 m2/y. Approximately 90% of data can be classified as clays of
medium plasticity (Table 3.7). However, more scatter and much higher results are determined
from the log-time method with cv varying from 0.28 to 130 m2/y. Approximately 55% of data
can be classified as clays of low plasticity, 27% as medium plasticity clays, 9% as high plasticity
clays, and 9% as silts (Table 3.7).
7
c, values are reported for the last two load increments in each test.
Table 3.1 Description of Borings
Boring
219
P-1
220.6
B101
P-2
B102
P-3
P-4
P-5
224.9
B104
P-6
225.8
226.3
B105
P-8
B106
P-9
B107
P-10
P-11
Station
217+35
218+34
218+97
219+00
219+61
220+22
220+51
221+45
222+02
222+84
223+28
223+42
224+06
224+93
225+08
226+23
226+56
227+36
227+77
228+30
229+17
Ground
Elevation
Offset from Maximum
m, MSL
Center Line Depth, m
22.93
25.16
28.43
27.55
27.02
25.47
25.2
24.19
23.14
22.32
21.84
21.05
20.34
21.11
21.49
25.23
26.12
27.5
29.73
28.34
27.48
On bold: Phase III Investigation
Not on bold: Phases I and II
2R
6L
12R
6.35L
11.5L
12.03L
15.5L
5L
4L
6R
5.95L
24L
2R
19.5L
0.2R
3L
11.94L
17.5L
1.01L
16.5L
13L
15.4
24.85
38.41
30.5
33.49
27.3
26.11
35.01
27.57
30.64
23.5
24.54
35.98
30.43
20.3
27.59
30.6
24.49
22.9
24.54
26.07
Table 3.2 Piezometer Locations: Phases I, II and
Boring
Station
Ground Location
m
elevation Elevation
m, MSL
m
219
217+35
22.93
7.73
P-1
218+34
25.16
6.87
BO11
219+00
27.55
-1.406
P-2
219+61
27.02
2.63
B 102
220+22
25.47
7.53252
P-4
221+45
24.19
4.37
P-5
222+02
23.14
9.42
224.9
222+84
22.32
5.02
B104
223+28
21.84
2.3328
P-6
223+42
21.05
8.25
226.3
224+93
21.11
-4.79
226.3*
224+93
21.11
-0.19
2 2 6 .3 (pl)*
224+95
21.11
5.26
2 2 6 .3 (p1)*
224+95
21.11
0.08
2 2 6 .3 (pl)*
224+95
21.11
-5.72
226.3(p2)*
225+02
21.11
5.26
226.3(p2)*
225+02
21.11
1.29
226.3(p2)*
225+02
21.11
-5.72
B105
225+08
21.49
11.2792
P-8
226+23
25.23
4.8
B106
226+56
26.12
-2.836
P-9
227+36
27.5
10.84
B107
227+77
29.73
18.6048
P-10
228+30
28.34
12.18
P-11
229+17
27.48
12.24
*: Piezometers of Pumping Test for Phases I and II
Time period of averaged groundwater elevations
III Investigations
USCS
Time
Classification Period'
Months
SC
9
CL
1
SP to SP-SC
1
CH
1
SM-SC
1
SC
1
SC
1
SM
7
SC
(5 days)
SC
1
SP-SM/SP-SC
9
SP-SM/SP-SC
9
5
5
5
5
5
5
SM
(5 days)
SC
1
ML
1
SM-SC
1
SC
1
SC
1
SC
1
GWT,
Elevation
Average, m
13.16
17.09
13.41
14.66
14.33
18.59
17.23
17.91
18.34
18.17
15.64
14.83
16.41
16.04
13.86
16.27
15.024
14.19
16.77
15.47
12.32
13.88
23.08
19.55
18.79
Table 3.3 Piezometer Locations Phase III Pumping Tests
Elevation
Station
Elevation Screen
well/
Interval,m
to
Average,m
m
from
piezometer
4.185
223+24
16.91
-8.54
PW-1
223+24
9.46
8.24
8.85
PW1-3A
2.375
223+24
3.14
1.61
PW1-3B
-8.57
-7.2
PW1-3C
223+24
-5.83
8.12
8.73
9.34
223+29
PW1-9A
2.95
1.73
2.34
PW1-9B
223+28
-8.63
-7.26
223+30
-5.89
PW1-9C
P-6
223+42
11.05
8.05
PW-2
219+81
21.97
-3.48
9.245
11.91
12.52
219+62
13.13
PW2-3A
219+61
PW2-3B
PW2-3C
219+61
-1.95
-3.17
-2.56
PW2-9A
219+56
13.28
12.07
12.675
PW2-9B
219+57
10.1
6.88
8.49
221.2
-0.12
-1.525
BO11
219+00
-2.93
USCS
Static
Water
Level,m
17.97
19.06
18.73
14.52
18.89
18.57
14.45
18.59
13.45
17.27
16.1
13.63
17.44
14.46
SC-ML/SM-SC
SC-CL
SM
SM-SC
SM
13.61
SP/SP-SC
SM-SC
SM
SP-SM
NO-SAMPLE
SC
SM
Table 3.4 Consolidation Tests
Reference Boring
Station
Elevation of Sample (m) oY'vo, UCSC
Name
From
To
(kPa)
Cl
B101
218+99.76
19.63
19.02
147
CL
C2
B102
220+21.65
19.22
18.61
117
SM
C3
B102
220+21.65
9.01
8.55
221
CL
C4
B104
223+28.12
17.27
16.66
87
SC-SM
C5
B104
223+28.12
13.92
13.31
117
SM
C6
B106
226+56.41
19.41
18.8
125
SC
C7
B 107
227+77.31
17.54
17.49
165
CL
Table 3.5 Description of Triaxial Tests
Reference Boring
Station
Elevation of Sample,
m
To
Name
From
T1A
B102
220+21.65
17.39
16.78
T1B
B102
220+21.65
16.17
15.75
TIC
B102
220+21.65
17.39
16.78
T2A
B102
220+21.65
10.23
9.62
T2B
B102
220+21.65
10.23
9.62
T3A IB106
226+56.41
22.16
21.55
T3B
B106
226+56.41
22.16
21.55
T3C
B106
226+56.41
22.16
21.55
Table 3.6 Permeability Results
Boring Station
Falling Head
Rising Head
m
From, To,
From,
To,
(m)
(m)
(m)
(m)
15
18
218+34
15
18
P-1
P-2
219+61
21
24
21
24
P-4
221+45
17
20
17
20
P-5
222+02
11
14
11
14
P-6
223+42
10
13
10
14
P-8
226+23
19
20
19
20
16
17
15
15
P-9
227+36
16
16
13
13
P-10 228+30
P-11
229+17
12
15
12
15
-3
Pumping Test (Phase I and II) k=3x 10 cm/sec
UCSC
SC
CH
SC
CH
CH
SC-CH
SC-CH
SC-CH
USCS
CH
CL
SM
SC
SC
SC
SM
ML
CL
vo
(kPa)
149.6
163.0
149.6
211.6
211.6
76.1
76.1
76.1
'cell
(kPa)
68.7
137.3
207.0
103.0
206.0
34.3
68.7
138.3
Permeability, k,
Rising Head Falling Head
(cm/sec)
(cm/sec)
2.86 x 10-4
2.56 x 10- 4
2.64 x 10-4
1.6 x 10
5
6.7 x 10
3.7 x 104
3.9 x 10.
8.1 x 10
4.2 x 10- 4
1.3 x 10.3
4
5.6 x 10
7.4 x10
5.4 x 10
3.2 x 10
7.5 x 10
2.7 x 106
.
1.7 x 10
2.1 x 10
Table 14.5.
SOME TYPICAL VALUES OF COEFFICIENT OF VOLUME COMPRESSIBILITY
Description of
compressibility
Coefficient of colume
compressibility, m,
(m-'MN)
Very high
Above 1.5
Very organic alluvial
clays and peats
High
0.3-1.5
Normally consolidated
alluvial clays (e.g.
estuarine clays)
Medium
0.1-0.3
Fluvio-glacial clays
Lake clays
Upper 'blue' and weathered
'brown' London Clay
Low
0.05-0.1
Boulder clays
Very stiff or hard
'blue' London Clay
Very low
Below 0.05
Table 14.6.
Clay types
Heavily overconsolidated
'boulder clays'
Stiff weathered rocks
TYPICAL RANGE OF VALUES OF COEFFICIENT OF CONSOLIDATION AND
COMPRESSION INDEX FOR INORGANIC SOILS
Soil type
Plasticity
index
range
Coefficient of consolidation
c, (m:/year)
undisturbed
Clays -
montmorillonite
high plasticity
medium
plasticity
Compression
index
C,
remoulded
Up to 2.6
Greater than 25
0.1-1
25-5
1-10
15 or less
10-100
above 100
low plasticity
Silts
About 25-50%
of undisturbed
values
0.8-0.2
From Lambe and Whitman (1979)
Table 14.7.
TYPICAL VALUES OF C,,
Soil type
C3,,
Normally consolidated clays
0.005-0.02
Very plastic clays
0.03 or higher
Organic clays
0.03 or higher
Overconsolidated clays
(Overconsolidation ratio
greater than 2)
Less than 0.001
From Lambe and Whitman (1979)
Table 3.7 Typical Range of Values of Vertical Coefficient of Consolidation
W
g2r1-
-
A.NOTE:..
ADDONAL
FOR
Gl20/121Tl
T DRAWN
REFER
SURVEY INFORMATKON.
PW2BORINGS
BLOWUP
RIO PIEDRAS CONTRACT
GEOTECHNICAL INVESTIGATOI
BORING LOCATIONS
STA.217+50 TO STA.225+00
Figure 3. 1a Boring Locations of Geotechnical Investigations
. , ,
-
NOTE,
RIO PIEDRAS CONTRACT
GEOTECHNICAL INVESTIGATION
BORING LOCATIONS
STA.225+00 TO STA.232+50
Figure 3. lb Boring Locations of Geotechnical Investigations
C/1
0
0
0
0
CD
01
01
CD
0
0-
0
0~~~I
.1,
III
ALLI
~
I~
______
full "1~y1 j
Jil
T- r"N'
A
1
2a
c
iii
111
_Pj
14
3
LThLLThLLLLI AWLJWLL
IIIIIIIIE :IWINhI~lH~~
LLIJ L~LALA~LLLW Li LLIJLIA
: ;ilhFlhJIrJIEAK
n.
Fk
RUNPII
1111j
ffiLLLtL~LLLLWLA
liiiIi~ I~~UIIIIIIllhIIJI~
2Ti
L
Ifi'i'i
LA
Elevation,m
SANDY CLAY
FILL
SAND &GRAVEL
clayey sand
ORGANIC
sandy silt
GRAVELLY CLAY
FAT CLAY
CLAY
PAVEMENT
SC WITH LIMESTONE
POROUS Limestone
CLAYEY SAND & GRAVEL
GRAVEL
sand,..
SILT
CLAYEY SILT
Figure 3.3 Legend of Geoprofile
SILTY CLAY
-f-
Z-
---F]t_
65 5-
Elevationo
IN
-u
-u
po
-u
Natural Water Content (w), %
10
0
-10
-20
Figure 3.5 Natural Water Content, w
Liquid Limit (wL), %
100
80
60
40
20
0
120
30
0
O
O
CO
8
20
o
0
- ..---------------------------
°0
oo ' 0-El -
CC"
0
O
8
0
0
00
O
0
0
---------- --................................
a
--------0
0: 0
:0
--- - -- - - --- -- - - - - -- - -----0 -- - - - - --- - - - -- - ----
0
-10
----
-20
Figure 3.6 Liquid Limit, wL
o
Phases I & II
0
Phase II
Plastic Limit w , %
I
I
I
I
0:
I
0
0
00
coo
I
0O
0
o
0
n
°
0
D
O0
O
-----------------0---E -----------------------------------------~~----------.....
o
0.
0
0
---------
[]
[]
[]~
-10
- -- - --
Figure 3.7 Plastic Limit, w
---------------
[]
0
~
[][
-
-
o
Phases I & II
0
Phase ll
-
Plasticity Index (I ), %
P
10
20
30
20
o 10
0
0=
-10
Figure 3.8 PLasticity Index, I
40
50
60
70
80
Plasticity Chart
120
100
00
60
×
x
oA-Line
oo
or
40
-or
20
-----
x>
or OL
0
20
A
0
40
-ML
60
Liquid Limit (wL), %
Figure 3.9 Plasticity Chart for Phases I and II
80
100
120
Plasticity Chart
oo
/?
~ja
X
acQ,
ct
Y
U
51
V1
CrJ
PI
0
10
20
30
40
50
60
Liquid Limit (wL), %
Figure 3.10 Plasticity Chart for Phase III Investigation
70
80
90
Liquidity Index, %
-0.5
30
I
I
1.5
0.5
0
I
I
I
0
00
Q
0
O3
0O
0
20
O
.......0* 00
O
e O[ 0
lid
C30
0
O
00-
.
.
.--...
Li
0:
[ []
.
E
. ]. .
-
-
10
..- ........
O
[]:
OO
[O
0 O]
- -
0
0
-- - - - - -
-...
----------------------------------------------
Figure 3.11 Liquidity Index
Unit Weight, kN/m3
14
Figure 3.12 Selection of Unit Weight
16
18
<I
ii
ii
!
WELL
.57
I
BORING
I
226.3,
<ELEV
.7
Figure
Site3.13Layout of Pump Test (Phases I and
S84
r',
2111)
9.2m
+20
+20
+15
+15
+10
+10
+5
0
-10
0
'I'll'
LEGEND
STATIC (PRE-PUMPING)
1 2 3 4 5m
-
-10
WATER LEVEL
WATER LEVEL LOWERED BY PUMPING
Figure 3.14 Results of Pump Test (Phases I and II)
SNORTH
P-6
R=35.51 M (116.5 FT.)
AVE. PONCE DE LEON
P9-MEDIUM
R=9.14 M (30 FT.)
P9-SHALLOW
R=9.14 M (30 FT.)
P9-DEEP
P3-SHALLOW
R=3.05 M (10 FT.)
/1
A
14 iVI 3
P3-MEDIU IM
S..
m (10 FT_
R=3.05
. .05. m(1
F. .
!
+
Rf=.
~
P3-DEEP
R=3.05 M (10 FT.)
WELL-UPR
Figure 3.15 Site Layout of Pump Test No. 1 (Phase m)
r
I.)
R=62.48 M (205 FT.)
+
PLAZA DE RECREO
P9-SHALLOW
R=9.3 M (30.5 FT.)
I
P9-MEDIUM
R=9.3 M (30.5 FT.)
PE
R=20.27 M (66.5 FT.)
P3-MEDIUM
R=3.45 M (11.33 FT.)
WELL PLAZA
.
+
CALLE ARZUAGA
-
P3-DEEP
R=3.52 M (11 FT.)
P3-SHALLOW
R=3.52 M (11 FT.)
AVE. PONCE
DE LEON
NORTH
Figure 3.16 Site Layout of Pump Test No.2 (Phase IIm)
35.5m= RADIAL DISTANCE FROM TEST WELL (TYP.)
3mn
0-72
F
CPM
.. . . A
I'L
3I,0
TOWV.
LEVEL
APPROXIMATE
OF TUNNELCROWN
o0 SA. 223 , 24
000
00
a
APPROXMATE
LEVEL
OF TUNNELINVERT
O 4TA. 223 + 24
-
BENIONITESEAL(YP.)
SCREENINTERVAL
(TYP.)
LEGEND
T
F
igr
STATIC
(PRE-PUMPING)WATERLEVEL
WATERLEVELLOWERLDBY PUMPING
3SAND
FILTER
PACK
(TYP.)
GRAPHIC SCALE
SETION LOOKING SOUTH
Figure 3.17 Results of Pump Test No.1 (Phase III)
ov
tom
VERTICAL
0
HORIZONTAL
I
2
3
4
___
5m
20.3m- RADIALDISTANCE
IROM TEST WELL(TYP.)
-
-
_-
5.B1s
PPU
0=22 GPM
--
.
'TFS!
[E.-I L
_
)
~9A
C)
2
(CW
*20 -.
00
f t0 ...
II I-
-----
--
--
----
APPROXIMATE
BASE OF
STATION
EXCAVATION
SA. 219+61
0
-----
ENTONIIESEAL (TYP.)
SCREEN INTERVAL(IP.)
SANDFILIER PACK (TYP.)
SSATIC (PRE-PUMPING)WATERLEVEL
-
WATERLEVELLOWEREDBY PUMPING
GRAPHIC
Figure 3.18 Results of Pump Test No.2 (Phase III)
o
HORIZONTAL(
1
SCALE
5m
_4
4
s
30
25
20
t'
----------------
A
** .
. . .
.
. .
.
. .
. .. .
A
AA
...........
. ... . .. .
.
.
A.
*AX
A
x
AA
A
IL
A
10
;
. . . . .. . .. . . L
A
------ ------
[- -
- -
- -
- -
- -
-
- ---- Ground Elevation
A
from Piezometers
X
from Pump Tests
(Phase Il)
I
216
'
I
'
I
'
218
''
i
' '
220
''I
''
222
Figure 3.19a Groundwater Elevations: Section 7 Alignment
224
226
228
230
(Station, m)
I
I
I
A
I
I
I
I
I
I
i
II
I
I
I
Xx
I
J
I
I
I
I
I
I
A
x
El.+16m
Rio
LO
15
0
5
........................
-+--
..........................
A
Piedras Station
x
218
219
220
221
Figure 3.19b Groundwater Elevations: Rio Piedras Station
222
Ground Elevation
from Piezometers
from Pump Tests
(Phase HI)
223
(Station, m)
224
In-Situ Stresses and Pore Pressures (kPa)
0
100
25 - ----------
20. . .
200
.
.
400
500
.
.
15
...
------------
10
I
600
---......... - - Total Stress
S -- - Pore Pressure
Effective Stress
-----------
-----------..
.. ......
300
\
:-:
5.
\
-5
Figure 3.20 Initial Stress Profile
92
Hydraulic Conductivity, k (m/day)
0.001
14 111
-
0.1
0.01
1
A
I1 111 I I
0
-Y
I
I I
I
I
I
I
A
I I II
O
1
"
II
A
Field Pumping
6I
Min.=2.7x10-6 cm/s=
0.0023 m/day
Tests =3x10 "3 cm/s=
2.6 m/day
10-4 cm/s=
0.26 m/day
1'-Average=3x
0
A
0
Max.=1.3x10 3 cm/s=
1.1 m/day
AO
A
o
10
-6
Rising Head
Falling Head
10-5
SIIII~II
.....
.I O
I
•1A
OAI
I
,.,,1
0.0001
0.001
Hydraulic Conductivity, k (cm/sec)
Figure 3.21 Permeability Results from Slug Tests
0.01
SPT N
30
0
1
200
100
50
I
200
150
o
Phases I & II
W
Phase ii
o
o
O
O
o
b o0
0
0 0
o
10
---
-------------- ----------------- C
---- ---------
-- 0------
O0
O
O 0O
0
------------ -- -
00
O
o0
0
-10
0
Figure 3.22a Measured Standard Penetration Test N-Value
s (KPa)
U
0
200
30
- *
....
400
600
:
:
=100
800
0
C
1000
1200
o
s =0.13N (ksf)
0
s =4.4N (KPa)
CI] 0
20
0
0
c0
m o
10
0
ED 13
10
-- -
0
O
EQ
-
CID
S -- 0:
EO
-10
0
0: E
0
-
::
:
Ew0
-o
o
0
O
o
Figure 3.22b Undrained Strength from Measured Standard Penetration Test
Su, KPa
150
100
o0
20 -
S
o:
o
8
0
0
0
O
oo00 O:
0
0:
0
0
9----
0
8
0
(9 o
0
0
0
0
I000
---o 0
250
200
0
oo
0
0
0
00
0
o8
0
------- -------------0
10 ...... O--8
a
[-:
0
0
00
o
i
0
o
0
0o
E]
o
.....
.. .... ... .... ... .... ... ...
E1-0 ....
0
o
-10
o
.~-0
----- -- -- -- -- -- -- -0 -
-20
Figure 3.23 Pocket Penetrometer Results
i
o
Phases I & II
O
Phase M
Strength, s (kPa)
100
200
400
300
500
700
600
I
30
I Line 1 Correlation
-
----
I
I
i
- - - - -
-
25 L~.~~...
03
Baguelfn et. al. 1978
0
Amar and Jezequel, 1972
V
Centre d'Etudes Menard, 1967
20
E9
O
-----
-----
-~U
......
.
.
"
e
E
-------
-o
--------
10
m ,1
v-----------
--
-
'.'7
f
v
.-----------.-------.m-
-
----------- --- -- - - -
.-------.---
5
EB
BO
[]
,
:V
1
0
.V
[]m~--~o:i
.. . ..
. .
.
v
------ -- - --------------------------
i-------------------.
.
. .
.
.
.
.G •
.
.
.
Vv..
.
.
.
. .
v
.
.
.
.
.
.
.
.
-5
- --
. .
.
, . .I
.
. ..
I-I
-
I
I
I
III I
I
I
I
i
-10
Figure 3.24 Pressuremeter Strength Results using Correlations
1
I
I
Su, KPa
Su, KPa
400
200
800
600
1000
30
-- I
+
x
+x
o
A
S+
x
A
+O x0x o
>1
250
200
150
100
50
.+
++
o
20
o+
x
O+xA+
o
:
00
:o
........
X
+
A
----------X
Ao1 +
+0
x
A
x
+
x
AO
+ AXX
+
+
:X-
:x
x
10
8+
X
0
x
+
x
x
------- 8--------.--- oo:
+ A
X
0
o
X
x
0
x
X
x
x
A
X
+ +
o
Ax 0
+4,
---0-- +X
x
x
A
*
A
+
Ao
0
A
+
:x
X
x
--
-10
-- - - - - -
-- - -
-- - - - - - - -
...
A
UUC
0
o
CIUC
1
0
A
x
+
I__
i.-
,
__A____._._I_
-20
Figure 3.25 Undrained Strength Results from Laboratory and Field Tests
UC
lab. Vane
P.P.
PMT
0.13N
4.4N
30
25
2 5
Su, KPa
100
50
0
- - - - - - - - - - - .-
150
- - - - - - - - -- - - - -.
-.
- - - - --..
- - - - - - - - - - - - - - - - - - - --..
. . - - - - - - - - - - - - - - - - - -
20
o
0
0
• =15 -.
-----------------------------------
10
5
A
--------....
---..
--.--..
.
-----------------------A
Su (KPa)=118.3-1.1 (Elev.
A
*
UUC
CIUC
Linear Regression
Figure 3.26 Undrained Strength Results from UUC and CIUC Tests
Preconsolidation Pressure, a' (KPa)
p
0
200
-- ------ - -
400
-
-
-
-
-
--
600
-
-
-
-
-
- - -
800
-
-
-
-I
--
-
1000
-
-
-
A
i -
--
1200
-
--
-
25
15A
9
10
A
-
10
5
10
-.----------....
-
..--------------
.--.------------ ----------- . - .--...
-A
--
--
A--
- S
Suu - S
-'vo
<'vo )
s=0.25, m=0.8
Undrained Tests:
UUC and CIUC
-5
Figure 3.27 Backcalculation of Preconsolidation Pressure for Hato Rey Soils
100
OCR
0
20
40
60
OCR
80
100
120
x
x
x
A
x
x
. .
...........
x:x
x
A
x
X
xxX
A
. --. ...
__J..............
.
...
x
x
x
x
Based on assumed o' =600 KPa
-- - - - - - - - .-.-..-. .....
.
.
.
.
.
.
.
.
x
A
UUC
-0--
CIUC
---x-Mnard
Figure 3.28 Backcalculation of Overconsolidation Ratio, for Hato Rey Clay:
10,000
I
' ;'""
.I '
I
-. -"""
i i . -.
1I
~ ''--~--
I
I I I (rllL
r
o-n in ap/a'
Range of
1
0 2 to 5
* 10 to 20
1000
-- 0.10
o
10 (Table 1)
0
10
100
0
0
CL
S\C.
0.I
:)
IC
I
i
i
111
100
I
I
I
I
I I I I III
I
I
uuuul
1000
0.024
! I 1 1
t
i
,
! 111
tutu
10,000
I
I
I
I
I I III
I I liii
IO,000
a', kPa
Figure 3.29 Relationship between Cohesion Intercept and Preconsolidation Pressure
(Mesri and Adel-Ghaffar, 1993)
102
200
150
100
50
00
50
100
150
200
250
350
300
p', KPa
400
200
150
100
50
0
0
50
100
150
200
250
300
350
p', KPa
400
0
50
100
150
200
250
300
350
p', KPa
400
200
150
100
50
0
Figure 3.30a Estimation of Friction Angle for Mohr-Coulomb Failure Criterion
103
150
T1B
B102
163.0
T1C
B102
149.6
..............
sifn=-tan a'
-3
II
50
....
.........
.... .....
...
0
R =
0.90
a=
22.60
=
24.60
200
Reference
Boring
T2A
B102
T2B
B102
S150
S100-
(2)
' KPa
1
211.6
............ ............
.........
..
......
-.........
211.6
--
-
..
50 ------ -- ;j~;----...
R=
0.99
ac=
22.60
0'=
24.60
200
Reference
Boring
o' , KPa
T3A
T3B
B106
B 106
76.1
76.1
T3C
B106
76.1
150
100
- -- -
50 -...
-----
--
--
i~'~~
(3)
i
------.................................
--------
---------
R, =
0.98
a=
22.80
=
24.90
S'
0
50
100
150
250
200
(KPa)
2
/
p'= (a'+ o')
300
350
400
Figure 3.30b Estimation of Friction Angle for Mohr-Coulomb Failure Criterion
104
Vertival Effective Stress, o' (KPa)
Vertival Effective Stress, o' (KPa)
VC
100
0
vc
1000
1.2
0
I
.
.
I .I .I
I
0.05
-
0.8
0.1
--
x-
100
. . .
I I I
II
II
II
I
II
I I
---
--
-o
('I
0.6 ------
0.15
0-
~
0.2
0.2 k
0.25
±_1
0.3
(b)
Figure 3.31 Consolidation Curves
1I
1
I IL
1000
0.0008
0.0007
0.0006
0.0005
0
2
0.0004
0.0003
0.0002
0.0001
0.2
C
0.4
0.6
Void Ratio, e
0.8
1
0
0.0008
0.0007
-
----
---
- - - ----------------- ------------------------- - - - - -- - - - - --. . . . .. . . . . .
-
.
0.0006
2
0.0005 -
0
2,
0.0004
0.0003
- ----------------------- - ----------------
----
--
--
---------------
i
_.......---.....
.- ----------- ---------------------------------
--------------------.....................
............... ...................
....
.
...............
0.0002
.
--------------------------------------------------------------------- -i
. .
.
... .
.
.
..
. ,
I
,:.. . ..
.
0.0001
Water Content, w (%)
Figure 3.32 Compression as a function of Water Content (w) and Void Ratio (e)O
106
' ~ ' 'I '1 'J
106
Test
'3
E
cell
E5 0
vo
voo
CIUC
Phase Il
E PMT
E'= (1+ v'X1- 2v')
( -
v') m,
A Oedometer
10 5
O
[]
E-
Ii
0
O
-
O
O
[]
[]
O
O
O
D] D]
A]
-~-- --O --
10 4
O
I I
100( \
0
50
a
MnrPicplSrsI(
150
200
100
250
Minor Principal Stress, a' (KPa)
3
300
350
Figure 3.33 Young's Modulus from CIUC, Menard Pressuremeter and Oedometer
107
Vertical Coefficient of Consolidation, c
L
A
A
A
A
A
Vertical Coefficient of Consolidation, cV (m2/y)
120
100
80
60
40
20
0
(m2 /y)
0
2..
Square-Root
4
I .. l .l I
6.
I .l
.
.
A
Log-Time
A
.
8.
.
.
10
12
A
Square-Root
A
Log-Time
4 LA
A
--------A
4
A
..
J..
. ..
. ......
. .. . .....
...
. ..
...
. . ...
- - - - - - - - --- - --- -----...
I
.. ..
. . ..
.......................
. ...........
.
- - -- - - - - - - - - , .
A
-A
..
..
. . ..............
L...........
..............
- -- - - - - - - - - - . .
...........
A
.............
..............
.....
i
-------------
. . . . . .
. . . . -.--
A -------------------------......
--- - - - L-- - --- ---------------------------------
A
(a)
Figure 3.34 Vertical Coefficient of Consolidation (c) Profile
4.
Finite Element model for Rio Piedras Excavation
4.1
Introduction
This chapter describes the development of a finite element model for simulating the
performance of braced excavations within typical Hato Rey soil conditions expected in the
vicinity of the Rio Piedras Station.
Development of the model focuses mainly on the
representation of soil behavior in the finite element analysis through selection of input
parameters for simplified elasto-plastic models. Complete predictions of wall deflections and
ground movements are presented for an idealized braced excavation whose dimensions
approximate the cut-and-cover sections at the north and south approaches to the Rio Piedras
Station. The predictions are evaluated through comparisons with empirical data for excavation
in similar stiff soil conditions.
4.2
The Hard Soil (HS) Model
Calculations of ground movements were accomplished using the PLAXIS 1 program,
which models concurrently the deformation of the soil and flow of groundwater, and hence, can
simulate the effects of partial drainage during construction.
Due to the lack of laboratory test data from which to estimate material properties, a
simple elasto-plastic model, referred to as the 'Hard Soil' (HS) Model, is used to describe soil
behavior (a sophisticated soil model is only appropriate when high quality input data are
available).
The HS model provides a simple framework for characterizing the behavior of
almost incompressible soils like sand, gravel, and heavily overconsolidated (stiff) cohesive soils.
A commercially available PC based non-linear finite element program supplied by A.A.
Balkema, Rotterdam, Netherlands.
109
Shear stress-strain properties are based on the well known Duncan-Chang model (Duncan
and Chang, 1970), but HS also includes formal definitions of loading based on plasticity theory.
The basic characteristics of the HS model and their relevance to the Rio Piedras conditions are as
follows:
1. Stiffness is a non-linear (power law) function of confining pressure. This enables the
model to describe stiffness variations with depth in a single soil unit.
2. Hyperbolic relationship between shear strain and deviatoric stress. This attribute more
reliable predictions of ground movement distribution around excavations, compared to linear
elastic models.
3. Distinction between primary deviatoric loading and unloading/reloading.
This is
essential for excavation problems where parts of the soil mass unloads while other regions
undergo loading.
4. Failure is described by the Mohr-Coulomb criterion. This is the simplest
representation of shear strength for a cohesive-frictional material.
As discussed in chapter 3, the Rio Piedras Station area consists of erratic layers of
overconsolidated clays, silts, sands, and sandy clays, which are not horizontally connected
continuously. In order to conduct the numerical calculations, an averaged soil profile is needed.
As a consequence, a profile consisting of one layer of overconsolidated material is used as a
simplified solution.
110
4.2.1 Summary of Formulation
One of the most important characteristics of the HS Model is the hyperbolic relationship
between the vertical strain, e1, and the deviatoric stress, q, in primary triaxial loading (Duncan &
Chang, 1970):
E
q
for q<qf
(4.1)
This relationship is presented in Figure 4.1, where the parameter Ei is the initial Young's
modulus for primary loading, and can be obtained from the following expression:
ref c'cot q'-'
E = Eref
4.
r(4.2)
P
Where:
1. Eire f is a reference Young's modulus corresponding to a reference pressure, pref
2.
G'3
is the minor principal stress, which is the confining pressure in a triaxial test.
3. m is a constant exponent that determines the variation of stiffness with confining pressure.
For stiff clays PLAXIS (1997) recommends a default value, m=0.5.
For unloading and reloading stress paths, the stiffness is linear with secant modulus, Eur, given
by:
u
=Eref
ref
c'cot-
p
3
refr
(4.3)''
(4.3)
where Eurref is the reloading modulus at a reference pressure pref. The Plaxis manual recommends
a default value, Eur re f = 4Eso0 ref ,where E5 0 ref is the secant modulus in first loading at q/qf=0.5.
4.2.2 Input Parameters for Hato Rey Soils
A list is given below of the material parameters used by the Hard Soil model (values in brackets
correspond to recommended values):
1.
c', drained cohesion
2.
(p', drained angle of internal friction
3.
'P, dilatancy angle (for stress states at failure)
4.
Eso0r f, Secant Young's Modulus for primary loading at q/qf= 0.5.
5.
var, Poisson's ratio for unloading/reloading (Vur = 0.1)
6.
Eur ref, Unloading/reloading Young's Modulus (Eurrf = 4 E 50 ref)
7.
m, exponent controlling pressure dependency of stiffness (m = 0.5)
8.
Rf, is the failure ratio which relates the asymptotic limit of the hyperbolic stress-strain
law qa and the maximum shear stress, qf defined by Mohr-Coulomb (Rf=qf/qa = 0.9)
Table 4.1 summarizes the 7 input parameters selected for the HS model to represent the Hato
Rey soils at Rio Piedras:
1. Cohesion (drained), c': Due to the lack of tests, the cohesion intercept, c', was estimated
based on correlations relating c' to a'p, the pre-consolidation pressure (Mesri&AdelGhaffar,1993), shown in Figure 4.2. As there is no reliable pre-consolidation pressure data
for the Hato Rey soils, an approximate value is estimated indirectly from lab. measurements
of undrained shear strength, using the SHANSEP 2 equation (after Ladd and Foott, 1974):
si = S
2 Stress
"
(4.4)
History And Normalized Soil Engineering Properties
112
where s, is the undrained strength and a',vo is the in situ, vertical effective stress. Typical values
of the constants, S = 0.22+0.03, and m = 0.8±0.1. Figure 4.3 shows the values of a'p computed
from CIUC, UU and M6nard pressuremeter strength measurements using average values (S=0.25
and m=0.8). There is large scatter in the data and no well defined trend with depth. For the
selected a'p, the correlations of Mesri and Adel-Ghaffar (Figure 4.2) bound the drained cohesion,
16 5 c' < 60 KPa. The average from the two estimates, c'= 40 Kpa was selected for the Rio
Piedras analysis.
2. The drained angle, (p' is estimated from the CIUC triaxial data, together with the empirical
estimate of cohesion, c'= 40 kPa. Figure 4.4 shows the linear correlations used to estimate
4'
from three sets of CIU triaxial tests. Linear regression analyses give 0'= 24.60 - 24.90, with
regression coefficients R2 = 0.90- 0.99.
An average value 0'=24.60 is used in the finite
element model.
3. Dilatancy angle, '
Figure 4.5 compares CIU measurements from one test with, a' 3 = 100 kPa (Test T2A) with
single element simulations of undrained shearing using the HS model, with T = 0 ° - 30.
Figure 4.5b shows that the measured stress-strain shear response can be well described by the
HS model with Y = 3o. However, there are large discrepancies between the computed and
measured effective stress paths (for all 4 values). This result reflects limitations of the HS
model in characterizing shear induced pore pressures at small strains.
4. The primary loading stiffness (secant) parameter, Eso0r ef , is estimated by comparing results
from CIUC tests, M6nard Pressuremeter (MPM) tests, and Oedometer data (Section 3.5.5).
Figure 4.6 shows linear correlations through each of these data sets as function of the
113
confining pressure. Assuming a reference pressure prf = 100 kPa, the parameter Eso0ref = 20
MPa 3 is well defined from both CIU and MPM data sets.
5. Unloading/reloading Poisson's ratio , v
Figure 4.7 compares the computed and measured CIU (T2A) shear behavior at a cell
pressure, a'3 = 100 KPa using Vur = 0.1, 0.2, 0.25.
The best approximation of the laboratory results was accomplished with a Poisson's ratio of
0.1.
6. Unloading/reloading stiffness , Eur e f
Figure 4.8 compares the triaxial results at
Y'3=
100 kPa (Test T2A) with the single element
calculations using the HS model and Euref/Eso fe = 2, 3, 4 and 6.
A ratio Euref/E0 ref= 2
provides the best match to the measured stress-strain behavior and also significantly
improves the prediction of the effective stress path.
7. Power stiffness law, m
The Plaxis manual recommends a default value, m= 0.5 for overconsolidated, stiff clays.
This parameter can also be estimated from laboratory triaxial tests as follows:
The non-linear stiffness function for the Hard Soil model is given by
E 50 = E 50
a,+ccot
P ref +c'cot0'
(4.5)
when a'3 = P'ref, then E50 = Es0 ref . Hence, by taking logs on both sides of this equation we
can also get:
(4.6)
.
log E50ref = mlog Pref +c'cot
0' )
3 Results from the Oedometer tests were not considered due to the reasons already explained in section 3.5.4.
114
Hence, m can be computed directly from the triaxial test data as shown in Figure 4.9. The
linear correlation in this case gives m= 0.86, but with a relatively low regression coefficient
R2= 0.735. Figure 4.10 compares the computed and measured values of E 50 as functions of
confining pressure
Y'3
for m= 0.5 and 0.86. Given the large scatter in the measured data
there is little to support the selection of either value of m. Hence, m= 0.5 is used in all
subsequent calculations.
8. The Failure ratio, Rf = 0.9 is used in accordance with Plaxis recommendations.
Using the input parameters listed on Table 4.1, single element calculations were performed
using the following cell pressures: 50, 75, 100, 150, 200 and 300 kPa. A comparison of the
single element calculations and all the triaxial tests (Tla through T3C) is shown in Figure 4.11.
In general, there are more similarities between the shear stress-strain paths than the effective
stress paths. However, this similarity only occurs at low axial strain levels (a < 5%) and the
shear stress-strain paths become very different as the axial strain increases (Ea > 5%)4 .
4.3
Idealized Braced Excavation
Figure 4.12 summarizes the initial conditions considered in the finite element model.
The Rio Piedras site has a level ground surface. The groundwater table is located at a depth of
10m, and initial pore pressures are assumed hydrostatic. The analysis uses saturated unit weight
of 17.84 kN/m 3, both, above and below the water table, as Plaxis does not consider the capillary
stresses above the water table. The analyses recognize that hydrostatic conductivity represents
one of the least certain properties at the site and hence, the base case analysis assumes k = 3x10 -3
cm/sec (2.6m/day), corresponding to the maximum expected permeability at the site.
4 For
finite element excavation calculations, characteristic shear strain levels are expected to be less than 5%.
115
The idealized geometry comprises a plane strain excavation with half width, B/2 = 11 m,
(similar to that expected at Rio Piedras Station approaches), and is supported by a 0.9 m thick
diaphragm wall. The analysis assumes that the reinforced concrete diaphragm wall has elastic
properties (Table 4.2), and is wished-in-place (i.e., the installation of the wall has no effect on
stresses or pore pressures in the surrounding soil). The wall has a total length, L = 33m, such
that at the maximum excavation depth, H = 22m, the embedment ratio L/H = 1.5. In order to
minimize the effects of the boundaries on predicted ground movements, the finite element mesh
extends far from the excavation to a total depth of 100 m and laterally to a distance of 200 m
from the centerline. The Plaxis code limits the total number of 15-noded and 6-noded triangular
elements that can be solved in a given analysis 5 . Hence the selection of the mesh must be
carefully tailored to the proposed excavation sequence.
Figure 4.13 shows the mesh (1,739
nodes and 782 6-noded elements) which models an excavation sequence with 2.75m vertical
steps. The excavation is braced internally by rigid supports equally spaced at intervals of 2.75m
vertically (nominal properties of the supports are given in Table 4.3).
The analyses assume a simplified construction sequence, (Figure 4.14) comprising the
following steps: 1) The soil is initially excavated unsupported to a depth h,= 2.75m; 2) the wall
is propped at the surface and excavation proceeds to a depth h equal to 2.75 m; 3) a second level
of support is installed at a spacing of 2.75m; and 4) step 3 is repeated until the excavation
reaches a total depth H equal to 22 m. At each excavation stage partial drainage can occur over a
specified time period related to the total expected corstruction duration, t= 240 days.
The
following paragraphs discuss these assumptions in more detail:
1. Plane Strain Model
5 For 6-noded
elements the number of elements in the mesh was restricted by a limitation of the maximum element
number available in Plaxis equal to 800 elements.
116
Although most real excavations have geometries that are three dimensional, the current
calculations consider a simplified planar geometry. Plane strain assumptions apply for situations
where cross section, soil properties, and loading scheme are approximately uniform over a
significant length such that displacements in this direction are assumed to be zero. Plane strain
analyses can be a good approximation for linear excavations (such as those for transit lines), but
will overestimate movements at the corners of an excavation. This type of analysis gives similar
settlements to 3-D analyses but greater horizontal deformations (Hashash, 1992).
2. Material Models
The Hato Rey soils are simulated using HS model, with input parameters described in section
4.1. The diaphragm wall has linear elastic properties and is modeled using two rows of solid (6noded triangular) elements were used because wall thickness is non-negligible compared to other
dimensions in the excavation problem. The wall is non-porous 6 and has a unit weight, - 23.6
kN/m 3. A rough interface 7 is assumed in the analyses. This assumption is only significant for
excavations approaching collapse (Desai 1988, Bakker and Vermeer, 1986).
3. Initial Conditions
The initial, in-situ state of stress is computed in the program input sub-menu. After entering a
coefficient of lateral earth pressure, Ko, for all soil layers Plaxis computes and shows the initial
stress state. This stress state is characterized by an initial vertical stress o'vo and an initial o'ho
which are related by Ko in the following way:
0VO
= (Oave -z)-
a ho = Ko .
P
(4.7)
vo
6 Non-porous elements have only displacements degrees of freedom
in Plaxis.
7 i.e. strength at the interface is identical to the strength of the adjacent soil.
117
where y'ave is the average weight above the stress point, z is its depth below the surface and Pi is
the initial pore pressure in the stress point.
The analyses assume Ko= 1, this represents a
reasonable average value given the lack of experimental data, erratic layering and geological
history of the Hato Rey soils.
4. Groundwater Conditions
Plaxis distinguishes three types of pore pressures states: initial Pi, ultimate Pu and excess Pe. The
initial pore pressure state (or initial steady-state pore pressures) represents the pore pressures in
the undeformed (initial) situation. The ultimate pore pressure state (or ultimate steady-state pore
pressures) may be used to study the influence of lowering or rising water tables on the
deformation and stability of the soil body. In general, it can be used to study soil response due to
a change from one pore pressure state to another pore pressure state. In this way, the steady state
pore pressure, Ps, can be defined by:
Ps= Pi + C(Pu-Pi)
(4.8)
where C=O implies that pore pressures are set to the initial level, while C=I, corresponds to
ultimate conditions.
The distribution of steady-state pore pressures is determined by the
boundary conditions, the geometries, and permeabilities of the different soil layers. Both pore
pressure distributions are independent of deformations. In contrast, excess pore pressures are
caused by deformation itself and represent pore pressures due to undrained soil response and
consolidation. The active pore pressures, Pa, are defined as follows:
Pa= Ps + Pe
(4.9)
The pore pressures can be imposed in two ways: 1) phreatic line input, or 2) groundwater
calculations. When pore pressures are entered by means of a phreatic line, the pore pressures at
a stress point are simply calculated by multiplying its depth below the phreatic line by the unit
118
weight of water.
For many geotechnical problems, steady-state pore pressures are nearly
hydrostatic and it is not necessary to performed detailed pore pressures calculations. Instead, a
phreatic surface is specified and the pore pressure is taken to be hydrostatic along vertical lines
beneath the phreatic surface. The alternative approach using a groundwater flow module enables
any steady state flow field to be imposed. The Plaxis groundwater flow module is needed in
order to simulate the lowering of the water table within the excavation. Boundary conditions in
these calculations are:
1. Prescribed piezometric head8 , H, referred to as 'open node'.
2. Zero flux ('closed node')
Figure 4.15 indicates the flow boundary conditions used at each stage of the excavation. The
Plaxis program assumes P= 0 at the excavated grade in each step of the analysis.
5. Bracing System
The bracing system is modeled as an elastic-plastic spring. One end of the spring is connected to
a node in the mesh and the other end is fixed. The input parameters in Table 4.3 correspond to
an approximation of an incompressible, non-yielding bracing system.
6. Calculation Sequence
The following calculation steps are performed to simulate the construction sequence of the
excavation:
1. Unsupported excavation:
A undrained analysis 9 is carried out in which the soil
elements at the top of the excavation are removed (switched off). At this stage, the
total excavation depth H= 2.75 m.
8 H=
P/yw + y, where y is elevation.
9 'plastic calculation'
119
2. After equilibrating the stresses at H= 2.75m a consolidation 10 analysis is performed,
in order to allow for partial drainage. Assuming a total construction time t= 240 days,
each stage of excavation is allowed to consolidate for 30 days.
3. Excavation and installation of strut 1 & 2:
A plastic calculation is performed to
model the excavation of another 2.75 m, so the total excavation reaches a depth of 5.5
m. The first anchor (Om depth) is activated as well as the second anchor (2.75m
depth).
4. Consolidation: Same as step 2
5. Excavation and installation of strut 3: A third plastic calculation is performed to
model the excavation of another 2.75 m, so the total excavation reaches a depth of
8.25 m. The third anchor (8.25m depth) is activated.
6. Consolidation: Same as step 2
7. Excavation and installation of strut 4: A plastic calculation is performed to model the
excavation of another 2.75 m, so the total excavation reaches a depth of 11.0 m. The
fourth anchor (1 im depth) is activated.
8. Groundwater calculation: The water inside the excavation is lowered 1 m. A flow
calculation is carried out to determine the pore pressures for the second groundwater
situation. A plastic calculation is done to activate this groundwater situation.
9. Consolidation: Same as step 2
10. Steps 7, 8 and 9 are repeated until the excavation reaches 22m of depth.
10 The governing equations as used in Plaxis follow Biot's theory
120
4.4 Results
4.4.1 Base Case Analysis
This section presents results from numerical analyses of the idealized excavation in Rio
Piedras. Using the base case input parameters listed in Table 4.1, with k = 2.6m/day and Eurref =
80 MPa (4E 50 ref) as recommended in the Plaxis manual.
Figure 4.16 presents the wall
displacements for the base case analysis at each stage of excavation as a function of the
excavated depth.
During the first, unsupported excavation phase, the wall deforms in a
cantilever mode with maximum deflections at the top (6 wmax = 4mm). Thereafter, movements
are constrained by the rigid bracing system and maximum wall deflections develop below the
current grade level (i.e., the wall deforms by bulging below the excavation level). Table 4.6
(Case 1) presents the ratio of maximum wall movements and total excavation depth, 8 wmax/H.
Maximum lateral wall deflections increase from 8wmax/H = 0.1 - 0.16% at early stages of the
excavation to 0.24% (5.2 cm) at H = 22m. Surface settlements around the excavation are of
great practical importance in estimating potential damage to surrounding facilities. Figure 4.17
illustrates the settlements at the various excavation stages. The analysis also predicts that the
wall itself moves upward by almost 3 cm at the final excavation grade.
Maximum ground
settlement increase from 6vmax/H = 0.08 - 0.04% at early stages of the excavation to 0.15% at H =
22m (Table 4.7), which is comparable with field data reported in literature. The analyses predict
settlements of approximately 45% of the maximum value (Figure 4.17) occurring at locations
very far from the excavation (x - 200m). The inward horizontal surface displacements towards
the excavation are presented in Figure 4.18. Maximum lateral displacements equal to 80% of
maximum surface settlements occur approximately 60m from the wall.
Bending moments in the wall can be estimated by curve fitting the wall deflection profile
using power law series (see figures 4.19a - h). After a mathematical expression" is obtained by
curve-fitting, the bending moments of the wall are calculated using the following equation:
M = -y'EI
(4.10)
Y"- c2y
-2
(4.11)
ox
where y is the wall displacements at an elevation x and EI - 1,400 MN m 2/m.
Figure 4.20 shows the bending moment distributions for each excavation stage. At H =
22m, maximum moments occur at a depth of approximately 8m (El.+18m) and below the
excavated grade at El.+2m, with Mmax = 700 kNm/m.
These predicted bending moments are approximately one-third of the plastic moment
expected for a heavily reinforced 0.9m thick concrete diaphragm wall (Mp = 2.0MNm/m;
Hashash and Whittle, 1992). Therefore it is unlikely that the wall will fail during excavation at
Rio Piedras.
Figures 4.21, 22, and 23 compare earth and pore water pressures on both the inside and
outside faces of the diaphragm wall at the beginning and end of excavation (H = 22m). The
effective lateral earth pressures (Figure 4.21), o'v, decrease on the excavated side, while o'v tend
to increase from El.+22m to El.+7.5m. Below El. 7.5m the final vertical effective stresses are
lower than the initial vertical effective stresses. The total lateral earth pressures (Figure 4.22)
show similar trends than the effective lateral earth pressures. At the end of excavation (H =
22m) the maximum flow velocity at the excavated grade, v = 1.6 m/day. The predicted inflow of
water into the excavation 1 2 is 17.5 m 3/d/m.
8
11y= mo + m*x + ... + ms*x +
12
mg*x 9 ; y = wall displacement and x = elevation.
Q = Velocity*Area; Area = B/2*lm.
4.4.2 Effects of individual parameters
A series of parametric analyses were conducted in order to evaluate the effects of
individual soil properties on the predicted excavation performance. The following paragraphs
investigate of the unloading modulus Eurref, Poisson's Ratio Vur, anisotropic
hydraulic
conductivity and wall embedment length. The undrained analyses are listed numerically in Table
4.4.
4.4.2.1 Effects of Unloading Modulus E,,ref
Figures 4.24 - 4.26 compare predictions for analyses with Eurrf = 40, 120 MPa
(Eurref/Esoref = 2, 6), respectively, with results for the base case parameters. Figure 4.24 shows
that the deflected wall mode shapes are similar, however, the unloading stiffness has a major
influence on the magnitudes of the initial cantilever deflections (at H = 2.75m). At the final
excavation depth, H = 22m, maximum wall deflections occur below the excavated grade (at
El.+Om) and range from 8wmax = 4.9 - 5.6cm (15% variation for all three cases). There is large
effect of reducing the unload modulus on the predicted vertical movements of the wall for H 5
8.25m (Figure 4.25). The Case 2 analysis shows ground surface heave extending throughout the
retained soil at H = 8.25m. While results for Case 3 (Eurref = 120 MPa) predict much smaller
wall uplift with heave extending less than 40m. At later stages of the excavation, the effects of
Eure f are relatively small, such that at H = 22m maximum surface settlements range from 5vmax =
3.2 - 3.7cm (the largest settlements occurring for Case2). The unload modulus has a major
impact on the far field settlements (at x = 200m) where, 8, decreases from 1.7cm (Case 3) to
0.7cm in the Case 2 analyses (i.e., more than 59% reduction). On the other hand, there is very
limited impact on vertical wall movements when unload stiffness is increased above the base
123
case value. In all three cases, maximum settlements tend to occur at a lateral distance of 40 to 60
m from the wall. The effects of varying the soil stiffness on horizontal surface displacements are
shown on Figures 4.26. By decreasing the unload modulus (Case 2), maximum horizontal
displacements are reduced but they also occur further from the excavation (at about 80 m from
the wall), while increasing Eure f (Case 3) has the opposite effect. At H = 22m, the maximum
6
hmax
range from 2.5cm to 2.7cm and are thus little affected by uncertainties in the unload
modulus.
4.4.2.2 Effects of Poisson's Ratio Vur,,
Analysis Case 4 assumes a Poisson's ratio, Vur = 0.25, a value which is more typical of
measured effective stress paths in Ko-swelling experiments (Pestana, 1994), than the value (Vur =
0.1) recommended and used in the base case calculations.
Figure 4.27 shows how the wall
deflections change with Vur (Cases 1 and 4). There is no impact on deflections at the early stages
of the excavation, however, during the last 4 excavation stages of excavation the maximum
movements decrease by about 5%. There is also no effect on the settlement predictions for H <
1 im, Figure 4.28. However, increases in Vur are linked to a 13% to 20% reduction in 8vmax at H =
16.5 - 22m. There is also a decrease in wall uplift and in horizontal surface displacements
(Figure 4.29) by about 10%.
4.4.2.3 Effects of Permeability Anisotropy
As discussed in Chapter 3, there is a discrepancy in hydraulic conductivity measured in
small-scale, slug tests and the large-scale field pumping tests. The base case analysis considers
an upper bound estimate of k, which corresponds to the worst case situation in terms of wall
124
stability and inflow rates.
Analysis Case 5 considers the situation where the hydraulic
conductivity is anisotropic with kh = 10k, = 2.6m/d (i.e., kv = 0.26m/day similar to slug test
data). Figure 4.30 compares the predicted wall deflections for the base case and Case 5 analyses.
There is no effect on maximum wall deflections during the first 5 stages of excavation (H <
13.5m). However, for H > 16.5m, anisotropic flow causes larger subgrade wall movements than
the base calculation. At H = 22m, Case 5 predicts
wmax
= 6.5cm (wmax/H = 0.3%). In general,
anisotropic flow reduces the predicted ground surface settlements (about 50% during the last 5
stages of excavation) however, there is a corresponding increase in predicted uplift of the wall
20% to 70% higher than the base case calculations (Figure 4.31). Figure 4.32 shows a 20%
decrease in predicted horizontal surface displacements with anisotropic permeability.
4.4.2.4 Effects of Wall Length
This section considers the effects of increasing the length of the wall from L= 33 to L= 44m
(Case 6). Results were compared up to an excavation depth of 19.25m (7 stages of excavation).
Figure 4.33 shows that the wall length has very limited impact on the predicted wall defections.
Wall length does reduce the maximum surface settlements and all far field settlements by
approximately 15%, while the uplift of the wall decreases by 30% (Figure 4.34). Wall length has
minimal effects on the horizontal surface displacements (Figure 4.35).
4.4.2.5
Summary of Results
Tables 4.5 - 4.8 summarize movement predictions for the Rio Piedras excavation for all cases
that were evaluated. Figure 4.36 summarizes the predicted maximum lateral deflection ratio,
Swmax/H as function of the excavation depth for all six analysis cases (Tables 4.1, 4.4). In all
125
cases, the deflection to depth ratio increases from a minimum
8 wmax/H
= 0.05 - 0.1% at H = 5m
to 0.22% - 0.30% at H = 22m. The minimum wall deformation occurs for Case 2 using a low
unload modulus, while the highest value is predicted in Case 5, with anisotropic permeability.
Overall, the individual factors considered in these parametric studies have little influence on the
predicted wall deflections. Figure 4.37 shows the maximum soil settlements ratios, 8vmax/H as
function of the excavation depth for H > 8m, reaching a maximum range 8vmax/H = 0.07 - 0.17%
at H = 22m.
Contrary to the effect on wall deflections, decreasing soil stiffness increases
maximum soil settlements, while increasing the stiffness has no effect on settlements.
Anisotropic permeability has a major effect, showing much lower settlement ratios than the other
five analyses.
Figure 4.38 compare the uplift ratios of the wall with excavation depth, H. The largest
wall uplift is predicted for Case 5, with anisotropic permeability, while the smallest uplift occurs
when the embedment of the wall increases (Case 6). It is very important to emphasize the uplift
predictions are not very realistic and are a result of the limitations of the Hard Soil model.
Further discussion on how these predictions compare with data published in literature is
discussed in the next section
4.5
Practical Interpretationof Results
4.5.1 Measured Soil Movements Published in Literature
Semi-empirical design charts provide a useful guide for estimating a likely range of
movements, based on ground deformation data collected from specific excavation histories.
Clough and O'Rourke (1990) provide a through review of the available techniques for estimating
126
soil movements including data from excavations supported by diaphragm walls. Figure 4.3913
shows their summary of maximum wall movements and soil settlements for excavations in stiff
clays, residual soils, and sands. The data can be summarized as follows:
1. The horizontal movements tend to average about 0.2% of the excavation depth.
2. The vertical movements tend to average about 0.15% of the excavation depth.
3. No significant correlation was found between maximum movements and different types of
wall.
There is a lot of scatter in the data, especially for the more extensive database of wall
deflections.
Figure 4.40 shows the distributions of soil surface settlements and horizontal
displacements for excavation sites in stiff to very hard clays, involving horizontally supported
concrete diaphragm and soldier pier walls, and other support systems. The settlements are only a
small percentage of excavation depth, (6vmax 5 0.3%), but are distributed over three times the
excavation depth from the wall (d/H 5 3). Records of horizontal ground movements are more
variable, and show two distinct zones of movement.
The majority of the horizontal
displacements fall within a triangular boundary with the same dimensions as those pertaining to
the settlements. The second zone includes excavations affected by their support systems. In
general, the horizontal movements will tend to equal or exceed their vertical counterparts, with
an upper bound of 2.5 times the vertical movements.
4.5.2 Comparison between Predicted and Empirical Data
Figure 4.42 compares the predicted maximum lateral wall deflections for the Rio Piedras
13
The cases reported in the charts are due to the basic excavation and support process.
Displacements caused by
ancillary construction activities were removed.
127
excavation with the empirical range proposed by O'Rourke and Clough (1990). All six analyses
predict Swmax/H = 0.25 ± 0.3%, which is in good agreement with measured data in the literature.
Maximum soil settlements (Figure 4.43) also match published data, with an average ratio 8vmax/H
= 0.13%. From Figure 4.40 it can be estimated that the expected maximum settlement for stiff
clays suggested by Clough and O'Rourke (1990) is 0.3% H (i.e. for H = 22m, 5vmax = 6.6 cm).
For Rio Piedras, all finite element analyses settlement results are below 3.7 cm, with an average
of 2.8 ± 0.8 cm. These results are also below the allowable maximum settlement specified in the
Tren Urbano contract, which is 3.8 cm (Capacete, 1997; Personal Communication).
According to Clough and O'Rourke (1990), maximum horizontal movements are below
0.3% of the excavation depth (i.e. 6.6 cm at H = 22m). The numerical experiments show that
horizontal surface movements are 2.5 ± 0.2 cm.
Figure 4.41 compares the distribution of surface settlements and horizontal movements as
proposed by Clough and O'Rourke for design purposes.
Their recommendations show a
triangular distribution with zero settlement at a lateral distance of d = 3H (d = 66m at Rio
Piedras).
The finite element analyses predict maximum ground settlements and horizontal
surface movements at d = 40 to 60m (1.8H - 2.7H) from the wall. At boundaries far from the
excavation, 200m (9 H), significant amount of settlement develops.
predicted heave inside the excavation for Rio Piedras.
Figure 4.44 show the
Although, significant heave is to be
expected at the base of an excavation in this type of stiff soil, the amounts of heave estimated
represent an upper bound.
In general, it seems that the finite element results give reasonable estimates of wall
deflections, ground settlements and horizontal surface movements. However, the distribution of
the settlements and horizontal surface movements tend to be unrealistic.
128
A more realistic
distribution of deformations can be obtained by improving the modeling of soil behavior
(Whittle and Hashash, 1994). Though the role of the constitutive model is very important, the
use of a more sophisticated soil model is only justified when high quality input data are
available. This is why more refined and sophisticated laboratory tests are highly recommended
if more reliable predictions are needed.
129
Table 4.1 Hard Soil model parameters
Input Parameter
Hato Rey Formation
Base Case Analysis
40 kPa
40 kPa
24.60
24.60
30
30
20,000 kPa
20 MPa
0.1
0.1
40,000 kPa I
80 MPa'
Power Stiffness Law, m
0.5
0.5
Failure Ratio, Rf
0.9
0.9
k = 0.26 - 2.6 m/d,
kh = kv = 2.6m/d
Cohesion, c'
Angle of Internal Friction, (p'
Dilatancy, y
Primary Loading Stiffness, E 5 0ret
Unloading/Reloading Poisson's Ratio, Vur
Unloading/Reloading Eurref
Hydraulic Conductivity, k
Table 4.2 Diaphragm Wall Parameters 3
Diaphragm Wall
Input Parameter
Young's Modulus, E
2.3 x 107 KPa
Poisson's Ratio, v
0.1
Total Unit Weight, Yt
23.6 KN/m 3
Table 4.3 Anchors Parameters
Fixed-end Anchors
Input Parameter
Stiffness, K4
1.15 x 108 KN
Ultimate Force
1,000 KN/m
1Eurref recommended in Plaxis version 6.31 is equal to 4 Esoref80,000 KPa
Eurref recommended in Plaxis version 6.31 is equal to 4 Ecrc f-8C,000 KPa
Model: elastic and Type of materi,. tun-porous
3 Material
EA
4 K =
L-D
where: L= effective Length=B/2; and D= distance between struts
130
Table 4.4 Parametric Study
Parameter Change
Case
1
Base case, table 4.1
2
Eurref = 40 MPa
3
Eurref = 120 MPa
4
Vur
5
kh= 10kv; kv = 0.26m/day
6
L = 44m
=
0.25
Table 4.5 Ground Movement Predictions for Rio Piedras Excavation at H = 22m 5
6
6
6
vmax/H,
Uplift,
Heave,
H,
8hmax, 6
vmax,
wmax,
Case
wmax/H,
Uplift/H,
Heave/H,
mm
mm
mm
m
m
m
%
%
%
%
Case 1
Case 2
Case 3
33.1
37.2
32.1
52.2
48.8
56.1
26
25
27
22
22
22
27.9
18.7
31.5
166
199
162
0.15
0.17
0.15
0.24
0.22
0.26
0.13
0.09
0.14
0.75
0.90
0.74
Case 4
27.5
49.6
23
22
27.3
141
0.13
0.23
0.12
0.64
Case 5
16.4
65
22
22
50.9
458
0.07
0.30
0.23
2.08
Case 6
21.3
42
19
19.25
21
0.11
0.22
0.11
5 Except
6
6
Case 6, which is at H = 19.25m
Maximum horizontal surface displacements
Table 4.6 Comparisons of Wall Deflections 68,,,/H, (%) at Each Excavation Step
Case 1
Case 2
Case 3
Case 4
Case 5
Case 6
(1) 2.75
0.160
0.080
0.160
0.160
0.160
0.240
(2) 5.50
0.120
0.040
0.120
0.120
0.120
0.120
(3) 8.25
0.107
0.080
0.133
0.107
0.107
0.107
(4) 11.0
0.140
0.100
0.140
0.140
0.140
0.140
(5) 13.75
0.176
0.128
0.176
0.160
0.176
0.176
(6) 16.50
0.200
0.160
0.200
0.187
0.200
0.200
(7) 19.25
0.217
0.194
0.229
0.206
0.240
0.229
(8) 22.0
0.240
0.220
0.260
0.230
0.300
Excavation
Depth, m
Table 4.7 Comparisons of Ground Settlements vma,/H, (%) at Each Excavation Step
Excavation
Case 1
Case 2
Case 3
Case 4
Case 5
Case 6
(1) 2.75
0.080
0.000
0.080
0.080
0.080
0.080
(2) 5.50
0.040
0.000
0.040
0.040
0.040
0.040
(3) 8.25
0.027
0.000
0.027
0.027
0.027
0.027
(4) 11.0
0.040
0.020
0.060
0.040
0.020
0.040
(5) 13.75
0.080
0.064
0.080
0.064
0.032
0.064
(6) 16.50
0.107
0.107
0.107
0.093
0.040
0.093
(7) 19.25
0.126
0.137
0.126
0.103
0.057
0.114
(8 22.0
0.150
0.170
0.150
0.130
0.070
Depth, m
132
Table 4.8 Comparisons of Uplift
Excavation
(%) at Each Excavation Step
uplift ma/H,
Case 1
Case 2
Case 3
Case 4
Case 5
Case 6
(1) 2.75
0.080
0.080
0.080
0.080
0.080
0.000
(2) 5.50
0.160
0.240
0.120
0.120
0.160
0.080
(3) 8.25
0.187
0.267
0.160
0.160
0.187
0.107
(4) 11.0
0.180
0.240
0.180
0.160
0.220
0.120
(5) 13.75
0.176
0.192
0.176
0.160
0.224
0.112
(6) 16.50
0.160
0.160
0.173
0.147
0.227
0.107
(7) 19.25
0.149
0.126
0.160
0.137
0.240
0.114
(8) 22.0
0.130
0.090
0.140
0.120
0.230
Depth, m
133
asymptote
Axial Strain, e1
q qf
Rf
Where:
qf is the ultimate deviatoric stress, which is derived from the Mohr-coulomb failure criterion.
As soon as q=qf, the failure criterion is satisfied and perfectly plastic yielding occurs
according to the Mohr-Coulomb model.
* Rf is the failure ratio (Rf <1).
* c' is the cohesion
* (p' is the drained angle of internal friction
*
Figure 4.1 Hyperbolic stress-strain relation in primary loading for a standard drained triaxial test
134
10000
Ocn in a'p/- n Range of
0 2 to 5
0 10 to 20
1000
C
-
0.10
I0 (Table I)
100
0
a.
-
_U
0
0.1' 1(
)
'
"
rtlt
"
'
100
"
llI
1000
0
I 1t rlill
10,000
i
i
i t
a
i
O100000
'a;, kPa
a'P = 600 kPa
16 kPa < c' < 60 kPa
c' (average) = 40 kPa
Figure 4.2 Relationship between Cohesion Intercept and Preconsolidation Pressure
(Mesri & Adel-Ghaffar, 1993)
135
a' (KPa)
p
0
200
400
x
x
25
600
800
I
:
---- ---- ----------........
'A
X
X
X:
20
1200
1000
e
------------------------
----------- -------------- ------ ------15 15
i
x
x
x
-------S10
10 - ----------------------
--- -
-------
--
----
*-
'Assumed
A
iY'
=600 KPa
pi
>
-5--------------..........
- - --- --------- -------------- ------5
-
x
x
x................
X
S
vo
I
-5
m
S
a'vo
-10
i
a vo)
*
A
UUC
CIUC
x
M6nard
s=0.25, m-0.8
Figure 4.3a Estimation of Stress History for Hato Rey Clays
136
OCR
0
20
40
60
OCR
80
100
120
(a)
Figure 4.3b Estimation of Stress History for Hato Rey Clays
C1
II
Cu
I
200
0
50
100
150
200
250
p= (a' + 'h) / 2 (KPa)
300
350
400
Figure 4.4 Estimation of Friction Angle for Mohr-Coulomb Failure
Criterion
138
120
100
t3>
!
II
0'
20
0
250
p'= (a'V+ a' R) / 2 (KPa)
I
120
b) Shear Sttess-Strain
i------
100
bV
Ic
vI
----------------------
------------ ---------- ------
- -- -- -
----------------------------
-----------------
------------
-------
o
Measured Data
HS Model Predictions:
------------
20
nI
l
- -------------
l
--
4 = lo
S = 20
V=30
l
4
Axial Strain (%)
Figure 4.5 Determination of Dilatancy Angle, xy
139
5 104
I
--
+-
CIUC
--E-
PMT
/
- - A - - Oedometer
I
I
I
I
I
I/
I
-
4 104
At
' =100 KPa:
3
E50ref=20,000 KPa
/
a
[]l
]
/
I
ii
3 104
---------
......... .. . ... ..........
2 104
/
I /.. ......
..
/
*
--- -- - .
1 104
I
)
50
100
150
200
250
Minor Principal Stress, a' (KPa)
I
300
Figure 4.6 Determination of the Primary Loading Stiffness, Esref
140
.-.- . . . ..- . .-.
350
120
.
.
a) Effective Stress Path
100
-------
80 ----------------
_
-------------------------------
------------
-
-
--- ---------
cI
0
80
0
0
50
120 0
50
100
150
p'0 -(o'--+-o')2 (KPa)
200
250
120
100
100
oM
200
150
0v
0
0
1
2
3
AxialRatio,
StrainV(%)
Figure 4.7 Determination of Poisson'---s
Ul"
141
0
e
= 0.2
=0.25
4
250
cOO
60 .
.
.
100
S0 -
1000
120
2
50
0- - -
- -
- -
- - - - - -
12--0250-
100
- -
- -
- - -
-
150
- - -
- -
---
200
-- -
) /2 (KPa)
p- ((T' + ;'h
- -- - -
250
-
-
- - --- -
- -
b)Shear Stress- Straini
050
60 ...................-- ..............................---......
0
0
rf
r
Measured Data
HS Model Predictions:
o--..
40
ref 2E ref
.E
..
ref
-E
SE
50
0
3
Axial Strain (%)
4
Figure 4.8 Determination of the Unloading/ Reloading Stiffness, E
142
ref
ref=
ur
2
ref
50
ur
SE
01
E
4E
ref
ur
0.2
I
I
1
I
I
I
I I
I I
I
I
I
I
I
I
il
I
I
I I
i
I
i
R2 = 0.73
0.1-
0-
I
------ ---------- ----------------------
----------------
-0.1-
-
..........
------- -----------------~ ~ ---------- ~------------- ---
-0.3 i
-0.2
..
.
..
.
-0.15
.l
.
.l
-0.1
l.
l
-0.05
.l
l.
;
0
0.05
Where:
ref(CIUC)
y=log 50=
E5 oref
x
ref
log50
03 (CIUC) + c'cot
= log log
p'ref +c'cot0'
20,000
l' o '3 (CIUC) + c'cot '
= log
100 + c'cot 0'
y= mx+b; slope, m=0.86
Figure 4.9 Determination of m Parameter from CIUC Tests
143
------ -- ---------
0.1
0.15
0.2
51
- - I-
*
CIUC
*
Menard
A
Oedometer
-
- - -
4 104
-
= 0.86
----- ----------m
-----.--.--------------- ..--
3 104
0
I,
x/
- - - - -0 - -
2 104
+
-------------- ---- ------- -: r
*
-
1 104
:
*
----------
0
-------
50
100
. .
.
A
A
0)
.-... .
----------
g*
0(
--- -
A
150
200
a 3 (KPa)
Figure 4.10 Determination of m Parameter
144
--
250
300
350
--T--T--Measured
Dat
--
7-T------v T-T
--T I-
Measuredata
----------- !
- -----------
I
I
I--- -
HS Model Predictions
.
. ...------ --. -
HS Model Predictions
.---.--
--------- ---it-- -- -- ----- --- --- --- ---
A'
------ --
100
--- ------
150
200
250
p- (o' + o')/2 (KPa)
x-0
~
e
~c--0
2
Axial Strain
-0
--
3
(%)
Figure 4.11 Comparison between CIUC Triaxial Tests and Single Element HS Model Predictions
-i--B--
4
B/2= 11 m
ao',
(KPa)
100 m
200 m
Figure 4.12 Boundary and Initial Conditions for Idealized Excavation
-
Mesh Scale
02
-I
Figure 4.13 Finite Element Mesh
4
[m]
6
[10
I
S
h = 2.75
h = 2.75m
3)
Figure 4.14 Geometry, Support Conditions, and Excavation Sequence
Centerline
Ground Surface
H=y
---* Water
------- Table
(y = 90m)
H = 90m
Figure 4.15 Flow Boundary Conditions at each excavation stage
I
'
'
'
i
Ii
'
i
I
'
'
'
I
'
'
i
5
0.01
'
'
'
I
I
'
'
'
i
i
i
i
Wall Displacements,
0.02
0.03
I
'
'
'
'
i
i
I
i
_
25
Excavation Depth, H (m)
0.04
Figure 4.16 Wall Deflections for Base Case Analysis
150
I
20
15
10
'
0.05
0.06
I
I
5
II
10
15
20
2
Excavation Depth, H (m)
Distance from Back of Wall, m
120
0.06
160
Base Case: Rio Piedras
HS Model: Table 4.1
K = K = 2.6m/d
x
y
E = 4E 0
0.04
SStrut Spacing: 2.75 m
Strut Spacing: 2.75 m
Toe of Wall: El. -7 m
0.02
---------------
-----------.-----.------------------
= H (m)
0
------------.. ------
---
- - - ----
13.75
16.5
19.25
-0.02
-0.04
I
i
I
I
[
I
Figure 4.17 Predicted Surface Settlements for Base Case Analysis
22
3
2.5
2
1.5
1
0.5
0
10
15
Excavation Depth, H (m)
5
Distance from Back of Wall, m
80
120
0.03 -
20
160
0.025
0.02
0.015
0.01
0.005
0
-0.005
K
-0.0 1
,
,
,
I
,
,
,
I
Figure 4.18 Predicted Horizontal Surface Displacements for Base Case Analysis
152
0.006
0.004
a) Excavation Step: 2.75m
b) Excavation Step: 5.5 m
00055 --------------
-
0.0035
'''''''
0.005
E
0.0045
80.003
0.004
.... .............-
0.0025
-------------------
0.0035
0.002
~
0.003
|1llII-
0.01
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
-----------------
- -
-------
----------------
0.008
.......I. II
.................
-I
-.
0.006
. . .
----------------------------
0.004
0.003 .
-10
-5
-5
. 0.I.
0
n fM'3"
5
10
15
Elevation. m
25
30
iI
.10
-10
. .
-5
0
Figure 4.19 Curve Fits for Bending Moment Calculations
(General Expression for Curve Fits: y = m0 + m *x + ... m *x8 + m *x9)
153
5
. . ll.. . !ll:I!,ll:
.. . . . . I
10
15
Elevation, m
20
25
30
0.035
0.025
r
' '
'' I I -
16.5
fm
Excavation Step
f) Excavation StepF 16.5:m
e): Excavatioo Step:
0.03
S---------------
0.02
.......
-------...............
0.025
...................
.........
.015
. .........
0.02
..........
0.015
Q0.01
.. . . .
0.01
.....
...........
..................
----- - - -
0.005
-------------------5
01
1--------
1i
-- -------
--------------------------
0.005
0
------
0
10
-5
0
5
10
15
20
25
-1 0
3C
0.05
-5
0
5
10
15
20
25
30
0.06
h) Excavatation Step: 22: m
g)Excavation Step: 19.25 m
0.05
0.04
0.04
.0.03
0.03
Q0.02
----------- --
0.02
-1,--------------
I
0.01
0.01
I
. ...
...
.I
... I ...
. .....
.....
-,-----....
....................................
.............I--------------------..-...
I.-..T.....-.-.....
.....-..
...
....
.....
....
....
55
-10 -5
-5 0
0
10
15
20
25
30
^ "
0-10
.i
0
-10
-5
0
Elevation, m
Figure 4.19 Curve Fits for Bending Moment Calculations
8
+ m *x9 )
(General Expression for Curve Fits: y = m + m *x + ... ms*x
154
5
10
15
Elevation, m
20
25
30
Bending Moment, KN-m/m
800
-800
24
20
16
Figure 4.20 Predicted Bending Moments for Base Case Analysis
155
Effective Lateral Stress, o' KPa
h
-400 -350 -300 -250 -200 -150 -100 -50
30 ,,
,,,I ,,
,
,
l, Ii
0
Effective Lateral Stress, o' KPa
h
-50 -100 -150 -200 -250 -300 -350 -400
0
11111,,,,
" I
" "'
"
I""
I" " I"t
--------.
-,
20
2.......
,
..
....... ...............
..
-
20
20
-,
-
5-----------------
-
20 -------
lo
Fu
4
4
Fu
........
.21..
t-Lra
u
... -----s--sfo
/At Full Excivation Dept
B seC--s------
x
-
-
4 --
0......-------------
30
on
e
..
--
. 0
10
A ------
[Retained Soil|
"1
Excavated Soi
,
-0
Figure 4.21 Effective Lateral Earth Pressures for Base Case Analysis
5
Total Lateral Stress,
-600
I I
-500
-400
I I I I I I I I I I I
I
-300
I I
h -
-200
I I
Total Lateral Stress,
KPa
II
-100
0 0
-100
I I
-200
-300
h
KPa
-400
-500
-600
- .
25
20
15
tor
10 o
5
0
0
-5
-5
-10
-10
Figure 4.22 Total Lateral Earth Pressures for Base Case Analysis
Pore Pressure, u
-200
-250
I
1 1
-150
I1
-100
' I
Pore Pressure, u
KPa
1
0
-50
I
I
I
II
0
-50
I
'
I
I
I
I
-100
I
I
I
I I
-150
I
............ .......... ...............
-
-----.--------
----
/
-250
-200
I
I
I
......................
I
Soi-
5
-A
......... 0
..
.
.....
.
~....................
...
---
-I-----------
...
I
--....-------
- -j.......
..........
I
-------
E
o 10 ------
I I I
SExcavated
Retained Soil
20
KPa
j
--
..........
~'~~
- - - - - - - - - ---
5
I
i
.
U
0
0
U
------------ ..............
. . .. . . . . .
-5
-5
At Full Excavation Depth
At Full Excavation Depth
I
-10
I
I
Ii
1
II
I
I
III
1111
I
I
11
,
1
II
I
II II I I II i
Figure 4.23 Pore Water Pressures on Wall for Base Case Analysis
1
i
11111111
I
I
I
I
I
I
I
1
11
1 1
i
II11
I I
I
I
-10
Wall Displacements, m
0
0.01
0.02
0.03
0.04
Wall Displacements, m
0.06
0
0.01
0.02
0.03
0.04
0.05
Rio Piedras Excavation
HS Model: Hato Rey Soils
i.
Figure 4.24 Effect of Unloading Modulus E ur"ref on wall Displacements
*
Base Case
0.06
Distance from Back of wall, m
Distance from Back of wall, m
40
0
0.04
, ,
,
80
I
'
160
120
I_
I '
I
Rio Piedras Excavation
HS Model: Hato Rey Soils
Base Case
80
0.04
0.03
0.03
Low E U-r, Case 2
0.02
0.02
0.01
.I \
0
\
2.75
-
,-
-- -
--
0
-0.01
-0.01
-0.02"
-0.02
-0.03 -
-0.03
-0.04
-0.04
'
Figure 4.25 Effect of Unloading Modulus E ur ron Ground Surface Settlements
120
Distance from Back of Wall, m
Distance from Back of Wall, m
80
0.03
0.03
0.025
0.025
0.02
0.02
0.015
0.015
0.01
0.01
0.005
0.005
0
0
-0.005
-0.005
-0.01
-0.01
Figure 4.26 Effect of Unloading Modulus on Horizontal Surface Displacements
120
Wall Displacements, m
0.01
0.03
0.02
0.04
0.05
12
8
Figure 4.27 Effect of Poisson's Ratio on Wall Displacements
162
0.06
Distance from Back of Wall, m
0
40
120
160
0.04
0.03
0.02
0.01
0
-0.01
-0.02
-0.03
-0.04
Figure 4.28 Effect of Poisson's Ratio on Ground Surface Settlements
163
Distance from Back of Wall, m
160
120
0.03
0.025
22
0.02
0.015
/
0.01
0.005
0
2.751
Rio Piedras Excavation
HS Model: Hato Rey Soils
Base Case
-0.005
v =0.25, Case 4
or
,
-0.01
Figure 4.29 Effect of Poisson's Ratio on Horizontal Surface Displacements
164
Wall Displacements, m
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
Figure 4.30 Effect of Anisotropic Permeability on Wall Displacements
165
Distance from Back of Wall, m
0
0.06
80
40
I
160
120
1'
I
'
Rio Piedras Excavation
HS Model: Hato Rey Soils
Base Case
k =10k =2.6m/d, Case 5
0.04
S0.02
_
0
-0.02
22
-0.04
I
I
Figure 4.31 Effect of Anisotropic Permeability on Ground Surface Settlements
166
Distance from Back of Wall, m
120
160
0.03
0.025
0.02
0.015
0.01
0.005
0
-0.005
-0.01
Figure 4.32 Effect of Anisotropic Permeability on Horizontal Surface Displacements
Wall Displacements, m
0.01
0.02
0.03
0.04
0.05
Figure 4.33 Effect of Wall Length on Wall Displacements
168
0.06
Distance from Back of Wall, m
0
120
40
160
0.04
0.03
0.02
0.01
0
-0.01
-0.02
-0.03
-0.04
Figure 4.34 Effect of Wall Length on Ground Surface Settlements
169
Distance from Back of Wall, m
120
160
0.03
0.025
0.02
0.015
0.01
0.005
0
-0.005
-0.01
Figure 4.35 Effect of Wall Length on Horizontal Surface Displacements
170
0.35
-e-- Case 1
0.3
--
Case 2
-E
Case 3
-
Case 4
A-- Case 5
0
Case 6
0.25
0.2
0.15
0 .1
------ ------
0.05
0
5
10
15
20
25
Depth of Excavation, H (m)
Figure 4.36 Maximum Lateral Wall Deflection Ratios, 8
wmax
171
/H
0.2
-
0.15
ase o
---
-
1
0.1
0.05
i
0
5
10
15
Depth of Excavation, H (m)
20
Figure 4.37 Maximum Soil Surface Settlement Ratios, 8 vmax/H
172
25
0.3
0.25 -
0.2 --- --
0
0.15 -.........
0.1 -...
-- 8-0 .0 5 ------ -- ----------------------- --
Case 1
-- Case 2
05 .Case 3
-- I
Case 4
A
Case 5
-
Case 6
0
0
5
10
15
Depth of Excavation, H (m)
Figure 4.38 Comparison of Wall Uplift Ratios
173
20
25
8 Soldier Pile a Lo,
or Sheetpiles
• 13
E 20 0
E
O
00
phrom Wails
r
SOSoil Nail Walls
g160
z
31
49
13
A Orilled Pier Walls
0 Soil Cement Walls
-
> 120-
,'"
0
S.0
j 80S4
o
-
00
03
40
0o
x
40
30
20
10
0
DEPTH OF EXCAVATION H,(m)
Soldier Pile & Log
or Sheetpiles
O Diophrom Walls
&Drilled Pier Walls
U
200
E
56,240 mm
-
3
N 56
E
60 160 LLI
80-
3
40
m
0
0
0
0
10
2
30
20
DEPTH OF EXCAVATION H, (m)
40
Figure 4.39 Observed Maximum lateral Movements and Soil Settlements reported
in Literature (Clough and O'Rourke, 1990).
174
d
Distance from Excavation
Max. Excavation Depth
1.0
0.5
O~ 0o,o
0
.1IS o O
0.1C c 0.2 0.3E
' H
2.5
2.0
1.5
vo
1 Ha
e O--OP_ *- _
O
3.0
---
3.5
0
O0
0
0
0.4
0.1
0.2 -
/
0o
0./
0
0.4
/
io
U.;) .
V .
*OLa
0.6
0.7[
0.8 /
V
AJ
/0
Zone for High Horizontal
3Stiffness
of Support
Zone for Low Horizontal
Stiffness of Support in London Cloy
Legend:
* London YMC (48)
* London NPY (48)
A Neosden (45)
* Bell Common (52)
O South Cove (24)
0
A
O
0
0
State St. (22)
Churchill Square (15)
Columbia Center (19)
Houston Bldgs (53,54)
Ist Nat'l Bank (44)
Figure 4.40 Summary of Measured Settlements and Horizontal Displacements
Adjacent to Excavations in Stiff to Very Hard Clay (Clough and O'Rourke, 1990).
175
d/H
1.0
2.0
v
8vM
0.5 -
1.0
a) Sands
d/H
1.0
1.5
2.0
1.5
2.0
2.5
3.0
vm
0.5
b) Stiff to Very Hard
Clays
d/H
0
Or
0.5
1.0
I
0.5
-Settlement
envelope
c) Soft to Medium Clays
Figure 4.41 Dimensionless Settlement Profiles Recommended for Estimating
the Distribution of Settlement Adjacent to Excavations in Different
Soil Types (Clough and O'Rourke, 1990).
176
200
S*
E
Case 3
*
Case 4
A
Case 5
Case 6
150 -
8wmax/H=0.5%
0
S100
-
6
/H=0.2%
wmax
50 -
0
10
20
30
Depth of Excavation, H (m)
40
Figure 4.42 Maximum Lateral Movements for Rio Piedras Excavation
177
200
D
Case 3
*
Case 4
A
Case 5
*
Case 6
150
6
/H=
vmax
100
50
0
5
10
15
20
25
Depth of Excavation, H (m)
30
Figure 4.43 Maximum Soil Settlements for Rio Piedras
178
35
40
500
400
o
Case 1
o
Case 2
O
Case 3
*
Case 4
A
Case 5
A
300
200
100 H
,
0
5
10
15
20
Depth of Excavation, H (m)
.
.
.
I'
25
Figure 4.44 Heave Inside Excavation for Rio Piedras at H = 22m
179
5.
Summary, Conclusions and Recommendations
5.1
Summary and Conclusions
One of the principal factors affecting the design of deep excavations in cohesive soils is
the control of ground deformations, in order to minimize damage to adjacent facilities and
mitigate the costs of underpinning. This is especially true for construction in congested areas
where the potential for damage to adjacent buildings, utilities, etc., can lead to very expensive
remedial measures because of uncertainties in predicted deformations. In a constrained urban
environment the influence of movements may be the most significant design issue and may
severely impact the support system and construction methods. It is therefore essential to make
reasonable predictions of ground deformations prior to construction, and to design lateral earth
support systems to constrain the movements within acceptable limits. The goal of this project is
to estimate the magnitudes of ground movements for proposed cut-and cover excavations in
Section 7 of the Tren Urbano project, and to relate these movements to the stratigraphy of the
surrounding alluvial soils.
Chapter 1 presents an overview of the Tren Urbano project, focusing on Section 7 in Rio
Piedras Section. The fundamentals of finite element analyses and semi-empirical methods for
predicting ground movements were described.
The main advantage of (non-linear) finite
element analyses is their capability to model complex construction sequences and include
detailed site-specific properties of the structural system and surrounding soils. This thesis uses
Plaxis, a commercially available, PC based, non-linear, finite element program, which is fully
capable of simulating the coupled soil deformation and groundwater flow that occurs due to the
excavation.
180
Chapter 2 describes the geology of the San Juan Metropolitan area. The information was
obtained from a small number of studies. The entire Rio Piedras alignment consists of Older
Alluvial deposits (Pleistocene and Pliocene age) comprising silty and sandy clays, with
interbedded sands. These Hato Rey sediments are highly pre-consolidated by desiccation, and
highly heterogeneous such that continuous units cannot be identified in adjacent boreholes. The
soil deposits along this section of the alignment rest unconformably over bedrock at depths
varying from about 30 to 100m. Perched ground water conditions are possible within the highly
variable, discontinuous, lenticular sand deposits within the mainly clayey alluvium.
Chapter 3 presents information on field and laboratory test results and selected
engineering properties. The initial geotechnical site investigations were completed in two phases
(ending in February 1996), with a supplemental program (Phase III) in August- September 1996.
The strength, flow and compressibility properties of Hato Rey formation were carefully
evaluated. For depths up to El.+Om, an average undrained strength, su = 150 kPa, was obtained
from the more reliable UUC and CIUC triaxial tests. Results from empirical correlations and
backanalyses show that drained cohesion, c' = 40 kPa, drained friction angle approximately
equal 24.60. In general the compressibility my increases with the insitu water content w (or void
ratio eo) with typical values ranging from my= 0.0002m 2/kN at eo= 0.6 to my = 0.005m 2/kN at eo=
1.0. Groundwater conditions are hydrostatic with the water table located at El.+16m (near Rio
Piedras station).
Upper and lower bounds of permeability, k = 2.6m/d and k = 0.26m/d,
respectively, were estimated form slug and field pumping tests.
Chapter 4 describes the Hard Soil (HS) model selected for the numerical experiments and
the modeling procedure used in Plaxis. The input material properties selected for Rio Piedras are
summarized in Table 4.1. The analyses focus on simplified excavation geometry, soil profile
and construction sequence (see Figures 4.12 - 4.14). The idealized geometry comprises a plane
strain excavation with half width, B/2 = 1im supported by a 0.6m thick diaphragm wall. Due to
the lack of laboratory test data from which to estimate material properties, a simple elasto-Plastic
model, referred to as 'Hard Soil Model' (HS), is used to model soil behavior.
In order to
understand fundamental mechanisms controlling soil and wall movements, a series of numerical
analyses were conducted, to investigate effects of individual parameters including the elastic soil
stiffness properties, anisotropic permeability (reducing kv relative to kh) and wall embedment.
The examination of the effect of these parameters leads to the following main observations:
1. The analyses show maximum wall deflections in the range
wmax
= 4.8 to 6.6cm. The largest
movements occur when anisotropic permeability is included in the analyses, while smallest
movements occur when the elastic soil stiffness is reduced. All cases follow the same trend
with decreasing rate of deflection with increase in excavation depth. The normalized wall
deflections, 8wm~/H, vary from 0.22% to 0.30% at the final excavation grade (H = 22m).
2. The analyses predict maximum settlements in the range 8vmax = 1.5 to 3.7cm. Anisotropic
permeability causes a reduction in settlements, while the largest settlements occur for the
lowest elastic soil stiffness. The normalized ground settlements, 8vm/H, varies from 0.07%
to 0.17% at H= 22m.
3. All of the analyses predict significant upward movement. These predictions are considered
unrealistic and arise due to limitations of the HS soil model.
In general, the finite element analyses results on magnitudes of maximum wall deflections,
ground settlements and horizontal surface movements have very good agreement with empirical
ground movements data from excavation histories (Clough and O'Rourke, 1990). However, the
distribution of the settlements and horizontal surface movements is not realistic.
182
5.2
Main Sources of Uncertainty
The main sources of uncertainty in achieving reliable analytical predictions of soil
deformations can be attributed to several factors:
1. Limitations in the site investigation and geometric approximations. The initial conditions in
the ground (stratigraphy, initial stress state, and ground water flow regime) can play a
significant role in the modes of deformation that occur.
2. Uncertainties in the selection of engineering properties (strength, flow and compressibility
properties) as a result of inadequate laboratory and field characterization.
3. Constitutive representation of soil behavior using the HS model in the PLAXIS program.
Soil modeling has an important role in the predictive accuracy of finite element analyses.
More realistic distribution of surface settlements and horizontal surface deformations can be
achieved by using more realistic soil models to describe soil behavior. However, the use of a
sophisticated soil model is only appropriate when high quality input data are available. The
complex stratigraphy for Rio Piedras brings other problems to the modeling process. The wide
variability in texture, composition and appearance of Hato Rey soils makes the selection of
model input parameters for an averaged soil profile very difficult.
5.3
Recommendations
Control of ground movements is likely to be a significant issue in Rio Piedras since
predictions from the simplified analyses are similar in magnitude to the allowable movements.
This represents a problem, especially if damage does occur (i.e. cracking of structure, etc.).
More refined analyses can be achieved if more reliable data is available. The principal problem
is how to determine representative soil behavior given the complex layering of Hato Rey soils.
183
The borehole data (Figures 3.2 - 3.4) show main layers are clays, sandy clays and clayey sand,
all of which can be sampled.
Characteristic layers are 1 to 2m thick.
By taking almost
continuous samples from 1 borehole (and x-raying), 3 or more main horizons can be identified to
perform the necessary laboratory tests. The main problem to estimate permeability values, k, is
the lack of layer continuity, which already affected previous pumping tests on Rio Piedras. This
behavior can be modeled by calibrating k more carefully using kv, kh, assuming horizontal
layering and estimating the bulk k (similar to varved clay).
Additional site investigation is needed primarily to define the deformation (stiffness)
properties of the Hato Rey deposits and to select input parameters for advanced constitutive
models. The additional tests suggested include:
1. Cross-hole seismic shear wave testing between to boreholes to estimate shear stiffness profile
in the field. This tests could be done using conventional equipment (e.g., Stokoe and Woods,
1972) or using seismic cone devices (Campanella et al., 1986).
2. Additional consolidation and triaxial shear tests:
A. The triaxial shear tests should be performed on samples that are re-consolidated under Ko
conditions to the estimated in situ stress state.
B. A combination of undrained and drained compression and extension shear tests.
C. Consolidation tests using high pressure triaxial equipment to measure compression and
lateral stress (Ko) properties at stress levels exceeding the pre-consolidation pressures, ('p.
Permeability and consolidation properties should be estimated from Constant Rate of Strain
(CRS) consolidation tests.
3. Ko estimated from filter paper suction measurements (Chandler and Gutierrez, 1986) on
undisturbed soil samples.
4. A series of Ko-consolidated Direct Simple Shear tests, using a Geonor direct simple shear
device, to provide essential independent validation of model predictions.
The main focus, in general, is to perform laboratory testing on high quality soil samples.
Undisturbed samples should be obtained using a Denison core barrel (3.5" O.D.) The selection
of test specimens should be guided by x-ray inspection of sample quality and uniformity. The xrays also provide data on desiccation and other features of the macro-fabric.
In addition to improving the available soil data, advanced constitutive modeling is essential
to obtain more realistic predictions of ground movements.
The MIT-S1 effective stress soil
model (Pestana, 1994) is well suited for characterizing the behavior of the stiff Hato Rey alluvial
deposits. If the proposed program of laboratory tests are carried out it will be possible to use this
type of advanced soil model to refine the predictions describe in this thesis.
185
References
Bakker, K.J. and Vermeer, P.A. (1986). "Finite Element Analysis of Sheet Pile Walls," Int.
Symposium on Numerical Models in Geomechanics, NUMOG III, Niagara Falls, Canada, pp.
455-462.
Campanella, R.G., Robertson, P.K. and Gillespie, D. (1986). "Seismic cone penetration test,"
Proc.In-Situ '86, Blacksburg, VA, 116-130.
Capacete, J.L. (1997) Personal Communication
Chandler, R.J. and Gutierrez, C.I. (1986). "The filter paper method of soil suction
measurement," Geothecnique, 36(2), 265-268.
Christian, J.T. & Wong, I.H. (1973). "Errors in simulating Excavation in Elastic Media by
Finite Elements." Soils and Foundations, 13(1), 1-10.
Clough, G.W., Smith, E.M. and Sweeney, B.P. (1989). "Movement Control of Excavation
Support Systems by Iterative Design." Foundation Engrg Proceedings Congress, ASCE,
Evanston, Ill., pp. 8 6 9 -8 8 4 .
Deasi, C.S. (1988). "Case Studies through Material Modelling and Computation," Case
Historiesin GeotechnicalEngrg, 2nd Int. Conference, St. Louis, Vol. II, pp. 1551-1565.
Deere, D.E. (1955). "Engineering Properties of the Pleistocene and recent sediments of the San
Juan Bay Area, Puerto Rico," PhD Thesis, University of Illinois.
Duncan, J.M., and Chang, C.Y. (1970). "Nonlinear Analysis of Stress-Strain in Soil," ASCE J.
of the Soil Mech. And Found. Div. Vol. 96, pp. 1629- 1653.
Earth Scientific Consultants and McCann, W. (1994). "Seismic Hazard Map for Puerto Rico
1994," The Seismic Safety Commission of Puerto Rico.
FTA and Puerto Rico Highway and Transportation Authority, FinalEnvironmental Impact
Statement, Tren Urbano, San Juan MetropolitanArea
GDR (1996a). "Geotechnical Data Report, Tren Urbano, Section 7; April 1996",.
GDR (1996b). "Geotechnical Data Report, Tren Urbano, Section 7; October 1996",.
Hashash, Y.M.A (1992) "Analysis of Deep Excavations in Clay," PhD Thesis, Massachusetts
Institute of Technology, Cambridge, MA.
Head (1980). Manual of "Soil Laboratory Testing", Chapter 14, Wiley
Kaye, C.A. (1959). "Geology of the San Juan Metropolitan Area, Puerto Rico" USGS Prof.
Paper317-A.
Lambe, T. W. and Whitman, R. V. (1969)."Soil Mechanics" John Wiley and Sons.
McCann, W. and Sykes, L.R. (1984). "Subduction of Aseismic Ridges Beneath the Caribbean
Plate: Implications for Tectonics and Seismic Potential of the Northeastern Caribbean," JGR 89
(B6), 4493-4519.
Mesri & Adel-Ghaffar (1993). JGE, ASCE, 119(8).
Meyerhoff, H.A. (1927). "Tertiary Physiographic Development of Porto Rico and the Virgin
Islands." Geol. Soc. America, Bull, Vol. 38, pp. 5 5 7 -5 7 6 .
Meyerhoff, H.A. (1933). "Geology of Puerto Rico." University of Puerto Rico Mon., Ser. B.,
No. 1, 306 p.
Monroe, W.H. (1976). "The Karst Landforms of Puerto Rico." USGS, Professional Paper 899.
Monroe, W.H. and Pease, M.H. (1977). Geologic Map of the San Juan Quadrangle, Puerto
Rico; US Geological Survey.
186
O'Rourke, T.D and Clough, G.W. (1990). "Construction Induced Movements of Insitu Walls."
Proc.ASCE: Design and Performance of Earth Retaining Structures,439- 470.
Pestana, J.M. (1994) "A unified constitutive model for clays and sands," PhD Thesis,
Massachusetts Institute of Technology, Cambridge, MA.
Schanz, T. and Vermeer, P.A. (1996). "An Adequate Model for Sand Soil Behavior".
Simpson, B., Calabresi, G., Sommer, H., and Wallays, M. (1979)."Design Parameters for Stiff
clays" Proceedings of the Seventh European Conference on Soil mechanics and Foundation
Engineering, Brighton, UK, Vol. 5, 91- 124.
Stokoe, K.H. and Woods, R.D. (1972). "In-situ shear wave velocity by cross-hole method,"
Journalof the Soil Mechanics and FoundationsDivision, 98(5), 443-460.
Whittle, A.J. (1996). "Prediction of Excavation Performance in Clays." To appear BSCE J. of
Civil EngineeringPractice.
Whittle, A.J., Hashash, Y.M.A. (1994). "Soil Modeling and Prediction of Deep Excavation
Behavior," Proc. of the InternationalSymposium on Pre-failure Deformation Characteristicsof
Geomaterials,Sappore, Japan, Vol. 1,589-594.
Whittle, A.J., Hashash, Y.M.A. & Ladd, C.C. (1993b). "Analysis of excavations in Taipei
clay," Research Report R93-12, Department of Civil & Environmental Engineering, MIT,
Cambridge, MA, 78p.
Whittle, A.J., Hashash, Y.M.A. & Whitman, R.V. (1993a). "Analysis of a Deep Excavation in
Boston," J. Geotech. Engrg., ASCE, 119(1),69-90.
187
El.
f4.9m-j
Steel Column
-i
Roof Slab0.8m
El.
SLevel
-1
0.28m
El
1 .5 i
evel -2
El.
-0.28m
Fill
(0.6 to 4.0m)
(HI&A
El.+2.1 m
El.+ Om
Boston City
Basin I)aum
S
Data
u 0 Measured
S
.L E.+2.7m m
---
!
1987)
..
.. .....
.. .......
v4
Assumed in
F.E. Model
0
4.
hmi
Level
Clay
i10 to 15m1n
r
0.28m
,
El.
7.6n
-3
Level
5
* e-
..........
-4
lydrostatic
Pore Pressure
..
(~I)
A
....
0.28m
rCd
0a
El.
-10.7
El. l.
-31.rin
Lve
Lev
El.- 10.7m
Sand
-5
0.30m
Level
- 1 6.81
Load\
Deari i i e
Level
......... ......
.
...............
.....
-6 _
a'
-7
0.20m
0 .30in
Gravel
Element
1-.
(LBE-E)
-20.7m
Sound ,rgi
Till
to 11.6m)
(1.5
5
Wealhered
Are ill ite
I
100 150
1 50
Pore Pressure, u
200
Pa
( k
0
)
Measured
Data.
Lab. Oedomeier
Tests (II&A, 1987)
Assumed in
F.E. Models
ix
EI.-I 8.3mi
W. e.. .
,t
,.1
10.3 to 5.8n
0.30m
tr
800
600
400
200
Effective Stress (kPa)
illite
Figure Al Garage structure and initial soil conditions at Post Office Square, Boston. Principal
uncertainties in the ground conditions included the in-situ ground water pressures and engineering
properties of the underlying till (glaciomarine) and weathered argillite. The excavation follows a
top-down sequence. The perimeter diaphragm wall is braced internally by cast in-situ concrete floor
slabs with internal columns suplported by load-bearing elements.
'-4
0
a)
C
on
¢3
la
C
irn
(I-
.. 9
*Reference
Configuration
*Undrained
Excavation
.2
iia~
7
-20.7
-Undrained
I-I-
-4. 6
..
1-Undrained
Excavation
-45 Days Dewatering
-- 9. 1
60 Days Dewatering
*Roof Slab Installation
*ist Floor Installation
Stage 12
=tage 10.
Undrained
.
t1i-12.2
Excavation
3.475 Days Dewatering
4
-Undrained Excavation
_.Z-a.6
-8.5
-6.7*30 Days Dewatering
-2nd Floor Installation
- t.3rd
Floor Installation
State 16
IStage
Undrained
-
-Undrained Excavation
6-7.6
-9.6
2.2
-16.*4th
*25 Days Dewatering
Floor Installation
-18 .3
-21.3
6th
Undrained
Excavation
Floor Installation
Stage 2
Excavation
"
30 Days Dewatering
-- 13.7
-1
*16.8
19(
I 35 Days Dewatering
.
-12.2] 5th Floor Installation
I "t.
" .6
cage 22
tUndrained
Excavation
Excavation
30 Days Dewatering
I.--1 -=7t
18.3
a
1Sge
3/
ath
Floor Installation
r
*Permanent Dewatering
Figure A2 Construction sequence used in finite element model. The simulation includes detailed
modelling of construction berms and de-watering schemes based on the actual record of site
activities.
189
0
10
20
10
30
Lateral Wail Deflection (mm)
50 605
_30 _ 0o "-s----0 _I 10 20
-0 50
___~
'7-
0 X~
irl-l
X
I -4*
X.
IoO a~
a
a
11
CX:"
-4-~
.......
Q
.. " -"....
I
....-
-15
-15
-15
Bottom of Wall
-20
0
U
X
m;
-IU
I
0
xi
Ox;
Stae
I
i
(
J*
Stage O
19
0
- O "
-25
Finite Element Results:
Base Case Prediction
-
-
-----
- Modified
Analysis
Mod. Analysis Stage 321
o! . i
.I
oi
4.
-5
-10
.o
-15
-20
Inclinometer Location
Stage
Plan
C*
-10 0
I
10
i
20
I
28
6
-25
30 40 50 60
Lateral Wall Deflection (mm)
Figure A3 Comparison of predicted lateral wall deflections with inclinometer measurements. The
base case analysis showed good agreement with cantilever type movements during the inital
excavation, but underpredicted measured wall deflections after stage 10/11. The original (base case)
analysis was modified to incoporate two important factors: a) floor slab shrinkage (which controls
the support stiffness); and b) improved modelling of flow conditions in the weathered argillite.
Results from this modified analysis are generally in very good agreement with the measured wall
deflections.
190
10
0
20
Distance Behind Wall (m)
20
10
40 0
30
.
Finite Element Results:
SBase Case Prediction
-30 -
-20
-40-
.....
--
- Modified Analysis
Modified Analysis Stage 32
Stage
--
-30
10
Stage
19
...... ............
-...
!
Surface Settlement Point Plan
•i40
0
0
_In
-30S
o:
O.....
0
-20
--------..-.......
,
-
-40
....
_
.-
0
-10
.
..
-
-20-
-50
. ..-
..........................
-10 .
40
30
.....
..
I Stage 28
S40
30
20
10
Distance Behind Wall (m)
50
Figure A4 Comparison of predicted and measured surface settlements. Although there is a large
scatter in the near field data (within 15m of the wall) the modified analysis gives very reasonable
description of the overall settlement trough.
:
Effective Stresses (kN/m )
Pore Pressure, u (kN/m 2 )
11
M(01
4f00
Xl-i-i-
6(00)
3l00
u
--
200
-- i--i
r
40
IFI1
5(
r
rTTI-
Rcinforcd Cunmclc
Diapluagsll WAll
+.
t
2 1
25 2
STIFF SIL.TY CLAY
wilh diin fine sand lenses
hsicbWIcddec4
40
VliIY I)I'NSI SIILTY SAND
& S'TIFF (CLAY I.AYI+I(S
I
Best Ei
j
Date:
Juiine1987
1975
li rosaic
Ev !
Stan of Constfructtoi
Maxiniii Drawdowi
Shc II.v.
cigai
9Novc,11
985
(GWI @
Bi-lk
a'
.e: PI
Pi . l i
.
U
;/(L
) for
,,
j
I asedn
i
3
P e&.
(1990)
(1975)
Pi i e'
, & Mol
. ;Wu
i,
i
Figure A5 Typical section and site conditions, World Trade Convention Center, Taipei. Soils
cbnditions are typical of K I Sungshan deposits of soft clay. Principal unknowns are the in-situ pore
pressures (which are non-hydrostatic due to past history of groundwater pumping) and engineering
properties of the lower clay. Laxboratory lest data show similar normalized soil properties (strength
and deformation) for both clay layers,which are inconsistent with cone resistance measurements.
-3.5
Excvaion (45 days)
Stage 2
-6.0
I
E xcavauon 8 days)
SStage
4
kN/m
226
-
176 kN/
-5.0
-8.5
2nd Suut Lveli
Prestressing (30 days)
3rd Strut Lavel
Prcsressmng f5 days;
Stage 11
SStage 5
Berm Rcmoval (10 days)
Stage 12
Figure A6 Idealized excavation sequence used in finite element model. The perimeter diaphragm
wall is braced by three levels of pre-stressed steel struts.
193
a
us
i
.
S............-...
Case 2
S
615
I
2()
SIag
25
i
-
I
j
J.
•
s la
ata, SI-3, SI-5
.
Analysis: Case I
Analysis: Case 2Stages
0
10
20
30
40
50
60
70 0
11 & 12
10
20
30
40
50
60
700
20
40
60
80
t100
Lateral Wall Defleclion, 8W (mim)
Figure A7 Comparison of predicted lateral wall deflections with inclinometer measurements in the
south and west walls. The analysis d(oes not describe accurately the initial cantilever deflection and
ultimately overestimates the measured wall deflections by up to 30%. Much of this discrepancy is
related to the uncertain properties of the lower 'stiff clays' which was partially demonstrated in the
results of modified analysis (case 2).
20
Excavation Stage
6
4
2
1.40
S
.
1.2
10
12
Data: (MaxMin)
-- *-- West (Keelune
.-. "South
(Hsin Yi)
Z 1.0
0.8
Anaivses:
. -
-_
-----
-----
Case I
---- Case 2
-..-, L-----
S0.6
0.4
A-,
ILevel 1 1--~1_~_~~_
*
----
I
~___~_~~~~~
~~_~~~~~_____~~_~~~~_~_ - -~-- ---i
...
.,I
--, -,
-----.-----.
.
U
0.0
3
.......
i
. .-
- ---------
*
1.
..................
......
...
....
........
.....................
..
.........
.........
........
..
.....
1.8I . ..................
r
....
--,.-..
--..
--'-..
- --..-.---:
I
..
- .......
......
4
1.6
1.4
Vt
1
___._--_- t- _
F_.-/-
1.0
' ::
.
_
_
......
_
evel 2
0.4
0.2
:.
H-
0.6
I
,---,
:
... tLJ !
0.8
I~iii
--- "
A
I
-
-
Z 1.2
-0.8lr
0.4
8
_-'--------[
10
12
Excavation Stage
Figure A8 Comparison between predicted and measured strut loads. The analysis gives
fiasonable predictions of the strut forces measured by pairs of vibrating wire strain tgauges at all
three levels of bracing.
195
0
5
1110
---- - ----------
15
15 ------------- ---.--------
:•
i:
/
i
Ii
20
\
-----------
L
25 .....
-3025
30
500 400
300 200 100
0
0 100 200 300 400 500
Total Lateral Pressure,a (kPa) Total Lateral Pressure, oh (kPa)
Figure A9 Comparison of predicted and measured total lateral earth pressures. The figure
compares total lateral pressures measured on both the inside and outside faces of the diaphragm wall
at the beginning and end of excavation. The predictions are in excellent agreement with the
measured data. The contrast between these predictions and the wall deflections in figure A2.7
suggests that stiffness properties of the upper soft clay deposit require further investigation.
196
B. Vertical Coefficient of Consolidation Plots
__
r_
Cv, Log-Time Method
0.25 tsf
0.016
0.0165
0.017
0.0175
0.018
0.0185
0.019
0.0195
0.02
0.0205
0.021
0.1
Cv, Log-Time Method
0.5 tsf
IUKNUIE
0.028
0.029
0.048
0.031
0.05
0.032
0.052
0 033
0.054
0.034
0.056
0.035
0.036
1
10
100
1000
0.058
0.1
Cv, Log-Time Method
2 tsf
.
0.066 t
.
| .
0.071
0.076
1 s J l
1
100
1000
10000
0.091
..........
Cv, Log-Time Method
4 tsf
Cv, Log-Time Method
0.096
0.136
0.106
0.146
0.156
0.166
..
.-
0. 186
0.146
0.196
0.156
0.206
0.166
Time, min
100
1000 10000
0.176
-
10
100
Time, min
0.136
1
10
Time, min
0.126
-
0.086
10
0.116
t
0.081
0.1
0.046
0.03
Time, min
0.096
Cv, Log-Time Method
1 tsf
1000 10000
0.216
0.1
1
10
Time, min
Figure B I Cv Calculation Using Log-Time Method for Sample C I
100
1000
10000
0.1
1
10
Time, min
100
1000 10000
Cv, Log-Time Method
0.25 tsf
0.029
Cv, Log-Time Method
0.5 tsf
0.05
0.03
0.086
0.052
0.091
0.054
0.096
0.031
0.032
0.033
0.056
0.034
0.101
0.058
0.106 I 111111111
1-11111111
FIII
1111111
0.035
0.1
1
10
100
Cv, Log-Time Method
0.1
1000
1
Time, min
Cv, Log-Time Method
2 tsf
10
100
Time, min
1000 10000
0.1
1
10
100
11rIEm
1000 10000
Time, min
Cv, Log-Time Method
4 tsf
Cv, Log-Time Method
8 tsf
0.146
0.207
0.151
0.212
0.156
0.217
0.275
0.222
0.28
0.26
0.265
0.27
M_
TT_
TMF-T-T
0.161
""' """"'
~t~ttt~ttm
0.166
0.1
1
10
Time, min
100
0.285
III"l
I ititlnhr
1000
10000
0.1
1
10
100
Time, min
Figure B2 Cv Calculation Using Log-Time Method for Sample C2
1000 10000
0.1
1
10
100
Time, min
1000 10000
Cv, Log-Time Method
0.25 tsf
0.011
0.0115
0.012
0.0125
0.013
0.0135
0.014
0.0145
Cv, Log-Time Method
1 tsf
Cv, Log-Time Method
0.5 tsf
0.018
0.0294
0.0304
0.019
_-
0.02
I
IrI
0.0314
__.
0.0324
0.021
0.0334
0.022
0.0344
0.023
0.1
1
10
100
1000
10000
0.1
10
1
100
0.1
1000 10000
Cv, Log-Time Method
2 tsf
0.0433
0.0443
0.0453
0.0463
0.0473
0.0483
0.0493
0.0503
0.0513
0.0697
c
3
0.0797
Time, min
100
1000 10000
0.0988
0.1038
0.1088
0.1138
0.1188
0.1238
0.1288
0.1338
0.0847
10
100
Cv, Log-Time Method
8 tsf
Cv, Log-Time Method
4 tsf
0.0747
1
10
Time, min
Time, min
Time, min
0.1
1
10
1000 10000
100
Time, min
Figure B3 Cv Calculation Using Log-Time Method for Sample C3
1000 10000
0.1
1
10
100
Time, min
1000 10000
Cv, Log-Time Method
0.5 tsf
Cv, Log-Time Method
1 tsf
0.058
0.059
0.06
0.061
0.062
0.063
0.064
0.065
0.066
0.1
1
10
100
0.1
1000 10000
10
100
1000 10000
Time, min
Cv, Log-Time Method
4 tsf
Cv, Log-Time Method
8 tsf
0.165
0.17
0.175
0.18
0.185
0.19
0.195
0.2
0.1
1
Time, min
1
10
Time, min
Figure B4 Cv Calculation Using Log-Time Method for Sample C4
100
1000 10000
i HN
0.1
1
10
100
Time, min
1000 10000
Cv, Log-Time Method
0.031
Cv, Log-Time Method
0.5 tsf
0.052
0.053
0.054
0.055
0.056
0.057
0.058
0.059
0.06
0.032
0.033
0.034
0.035
0.036
0.037
0.1
1
10
100
1000
10000
0.078
0.08
0.082
0.084
0.086
0.088
0.09
0.092
0.1
1
10
Time, min
0.115
!
0~
100
1000 10000
Cv, Log-Time Method
2 tsf
Cv, Log-Time Method
4 tsf
0.165
0.18
0.13
0.185
0.135
0.19
10
Time, min
100
1000
10000
0.1
1
10
100
Time, min
Figure B5 Cv Calculation Using Log-Time Method for Sample C5
10
100
1000
10000
Cv, Log-Time Method
8 tsf
0.215
0.22
0.225
0.23
0.235
0.24
0.245
0.25
0.175
0.125
1
1
Time, min
0.17
.1
0.1
Time, min
0.12
t'r
Cv, Log-Time Method
1 tsf
1000
10000
0.1
1
10
Time, min
100
1000
10000
Cv, Log-Time Method
0.25 tsf
I
0.0165
Cv, Log-Time Method
1 tsf
Cv, Log-Time Method
0.5 tsf
0.032
0.052
0.053
0.054
0.055
0.056
0.057
0.058
0.059
0.06
0.033
0.0175
0.034
0.0185
II
-
10
100
0.035
0.0195
0.036
0.0205
0.037
0.038
0.0215
0.1
1
10
100
0.1
1000 10000
1
0.078
0.079
0.08
0.081
0.082
0.083
0.084
0.085
0.086
0.125
0.13
0.108
0.135
0.14
0.145
0.116
0.15
0.118
0.155
0.16
0.12
Time, min
100
1000 10000
1000 10000
Cv, Log-Time Method
8 tsf
0.114
10
100
Cv, Log-Time Method
4 tsf
0.11
1
10
Time, min
0.112
0.1
1
Time, min
Time, min
Cv, Log-Time Method
2 tsf
0.1
1000 10000
0.1
1
10
Time, min
Figure B6 Cv Calculation Using Log-Time Method for Sample C6
100
1000
10000
10
100
Time, min
1000 10000
I
Cv, Log-Time Method
0.25 tsf
0.0025
0.009
0.0035
0.01
0.0045
0.011
Cv, Log-Time Method
0.5 tsf
Cv, Log-Time Method
1 tsf
0.015 i
0.016 "
0.017
-If
0.0055
0.0065
0.018
i
i
0.012
i9i i i "Hi"
H
0.019
0.0075
0.013
0.02
0.0085
0.014
0.021
0.1
1
10
100
1000 10000
0.1
1
Time, min
10
100
1000 10000
0.1
1
Time, min
10
100
1000 10000
Time, min
I
0.023
0.027
0.029
0.031
0.033
10
100
1000
0.05
0.033
0.035
0.037
0.039
0.041
0.043
0.045
0.047
0.049
0.051
0.025
Time, min
Cv, Log-Time Method
8 tsf
Cv, Log-Time Method
4 tsf
Cv, Log-Time Method
2 tsf
10000
0.055
0.06
0.065
0.07
0.075
0.08
0.085
0.1
1
10
Time, min
Figure B7 Cv Calculation Using Log-Time Method for Sample C7
100
1000
10000
0.1
1
10
Time, min
100
1000 10000
0:
0.0020.0040.0060.0080.01 0.012
0.0140.0160.018 -
Cv, Square-root Time Method
1 tsf
Cv, Square-root Time Method
0.5 tsf
Cv, Square-root-Time method
0.016
0.031
0.021
0.036
0.041
0.026
0.046
X
0.031
X
X
0.051
x
x
0.036
x-~
0.056
0.041
0.061
0
0.2
t^(1/2)
0.4
0.6
0
0.5
t^(1/2)
t^(1/2)
4
Cv, Square-root Time Method
2 tsf
4
Cv, Square-root Time Method
8 tsf
0.15
I I I
Cv, Square-root Time Method
4 tsf
0.046
0.051
0.056 X
0.061
0.066
:
i
0.076
0.086
0.16
0.096
0.106
0.17
0.116
0.126
0.071
0.076
0.081
0.086
0.091
0.096
0.18
1 I
I I
0.136
X
0.146
0.19
0.156
0
0.5
0.2
0.166
i1
0
1
2
t^(1/2)
t^(1/2)
I
Figure B8 Cv Calculation Using Square Root-Time Method for Sample Cl
3
4
4
2
tV(1/2)
6
Cv, Square-root-Time method
.25tsf
0.029
0.03
Cv, Square-root Time Method
0.5 tsf
X
XXXXX
0.031
0.032
i
0.03
0.051
0.035
0.061
0.04
0.071
0.045
0.033
0.05
0.034
0.055
0.035
XXX
__
0.081
xx
Sxx
X
XX
1
1.5
2
0.5
1
t^(1/2)
1
t^(1/2)
1.5
1.5
2
0
2
Cv Square-root Time Method
4 tsf
0.15
0.16
0.17
0.18
0.19
0.2
0.21
0.22
0.23
0.5
x
-X-XX
0.101
0111
011
0
t^(1/2)
Cv, Square-root Time Method
2 tsf
x-
0.091
0.06
0.5
0.088
0.098
0.108
0.118
0.128
0.138
0.148 0.158
0.168
0
Cv, Square-root Time Method
1 tsf
0.5
1
t^(1/2)
1.5
2
Cv, Square-root Time Method
8 tsf
0.2
----
----
0.22
-
0.24
0.26
-~-xxcx
0
0.5
x
1
t^(1/2)
Figure B9 Cv Calculation Using Square Root-Time Method for Sample C2
1.5
1.5
X xXX
0.28
XX
X
0.3
2
0
0.5
1
t^(1/2)
1.5
2
Cv, Square-root Time Method
0.25 tsf
0.0105
0.011
0.0115
0.012
0.0125
0.0130.0135
0.014
0.0145
Cv, Square-root Time Method
0.5 tsf
0.013
x
0.017
X
0.019
0.021
0.023
0
0.5
1
t^(1/2)
1.5
2
0.15
0.16
0.17
0.18
0.19
0.2
0.21
0.22
0.23
0.033
0.038
j
X
X
X
X
0.048
0
0.5
1
t^(1/2)
1.5
0
0.5
1
t^(1/2)
1.5
2
Cv, Square-root Time Method
4 tsf
Cv, Square-root Time Method
2 tsf
0.043
X
0.015
X
2
i
X0
0.5
1
1.5
t^(1/2)
Figure B 10 Cv Calculation Using Square Root-Time Method for Sample C3
2
Cv, Square-root Time Method
Cv, Square-root-Time method
Cv, Square-root Time Method
0.5 tsf
0.033
0.035
0.037
0.039
0.041
0.043
0.045
0.047
0.0296
0.0306
X
0.0316
0.0326
0.0336
0.0346
0
0.5
1
t^(1/2)
1.5
0.046
0.051
X
-
0.056
X-X-X
0.066
0.071
0.5
1
t^(1/2)
1.5
2
Cv, Square-root Time Method
2 tsf
4 tsf
0.096
0.126
X
X XXXX X
0.136
X
0.146
0
Figure B I
0.5
1
t^(1/2)
1.5
2
0
0.5
1
1.5
2
8 tsf
0.136
0.146
0.156
0.166
0.176
0.186
0.196
0.206
0.116
xx
0.5
Cv, Square-root Time Method
0.106
X XXX
0
t^(1/2)
Cv, Square-root Time Method
0.046
0.056
0.066
0.076
0.086
0.096
0.106
0.116
XLXXXX
0.061
0
2
1 tsf
0.041
1
1.5
t^(1/2)
Cv Calculation Using Square Root-Time Method for Sample C4
2
X
X-X
0
0.5
1
t^(1/2)
1.5
2
Cv, Square-root Time Method
1 tsf
Cv, Square-root Time Method
0.5 tsf
Cv, Square-root-Time method
0.0315
0.034
0.032
0.039
0.044
0.0325
0.049
0.033
0.0335
X
X
X-XX
X
0.054
x
0.059
x
0.064
0.034
0
0.5
1.5
1
t^(1/2)
2
0.095
0.105
X
0.168
x
0.135
0.5
1
t^(1/2)
1.5
X-
0.178
0.188
0.198
0.125
0
1.5
1
t^(1/2)
2
2
SqareRoo-Tme
Figre 12 vsinClcuaton
0
T
x
0.5
1
t^(1/2)
X
ethd fr Smpl C
Figure B12 Cv Calculation Using Square Root-Time Method for Sample C5
0.5
XX
x
1
t^(1/2)
1.5
2
0.176
0.186 X
0.196
0.206
0.216
X
0.226
X-x0.236 0.2460 .2 5 6 -----------
X
1.5
X
Cv, Square-root Time Method
8 tsf
0.118
0.128
0.138
0.148
0.158
0.085
XXX
0.5
Cv, Square-root Time Method
4 tsf
Cv, Square-root Time Method
2 tsf
0.115
0
0.054
0.059 X
0.064
0.069
0.074
0.079 0.084
0.089 0.094
0
2
0
0.5
1
t^(1/2)
1.5
2
Cv, Square-root Time Method
0.5 tsf
Cv, Square-root-Time method
0.0165
0.017
X
0.0175
Cv, Square-root Time Method
1 tsf
0.016
0.031
0.021 X
0.036
0.041
0.026
0.018
x
X
0.0185
0.019
Xx
0.036
X
0.046
x
0.031
XXX
0
0.5
1
tA(1/2)
1.5
0.5
1
t^(1/2)
1.5
---
-------------------
0
2
0.076
0.117
0.061
0.066
0.086
0.137
c
0
0.096
X
S0.106
X
X
X
X
1
t^(1/2)
1.5
XX
XX
XXxX
1.5
2
XX
XX
X
0.177
0.116
2
1
t^(1/2)
X
0.157
XX
0.197
0.126
0.5
0.5
Cv, Square-root Time Method
8 tsf
Cv, Square-root Time Method
4 tsf
0.056
0.081 0.086 0.091
X
----
Cv, Square-root Time Method
2 tsf
0.071
0.076
X--X
0.061
0
2
X
0.056
0.041
0.0195
X
0.051
X
0
0.5
1
t^(1/2)
Figure B 13 Cv Calculation Using Square Root-Time Method for Sample C6
1.5
2
0
0.5
1
t^(1/2)
1.5
2
Cv, Square-root Time Method
0.25 tsf
0
0.001
0.002
0.003
0.004
0.005
0.006
0.007
Cv, Square-root Time Method
0.5 tsf
0.0077
0.0082
0.0087
0.0092
0.0097
0.0102
0.0107
0.0112
0.0117
x
X Xx-,
0
0.5
X
1
X
1.5
2
0.0133
0.0143
0.0153
0.0163
XJ
0.031
0.033
0.035
0.037
0.039
0.041
0.0223
x
0.0243
0
0.5
1
1
1.5
2
1.5
2
X
0
0.5
1
1.5
2
t^(1/2)
Cv, Square-root Time Method
8 tsf
0.0481
x
0.0531
x
0.0581
.
.
rrr!4
-v.....
0.0430.045
x
i-x
t^(1/2)
X
0.0193
0.5
Cv, Square-root Time Method
4 tsf
0.0213
X_
XXX
0.0173
X
0.0183
0
0.0203
0.0263
X
t^(1/2)
Cv, Square-root Time Method
2 tsf
0.0253
X
x
XI
t^(112)
0.0233
Cv, Square-root Time Method
1 tsf
0.0631
0
1
2
t^(1/2)
Figure B 14 Cv Calculation Using Square Root-Time Method for Sample C7
3
4
0
0.5
1
tV(1/2)
1.5
2
C. Computations of E' from Oedometer Tests
Cl
2
0
0
100
1L
. Kpa
300
200
600
500
400
. . ..
.... . . .
100
200
300
400
500
600
700
= 147 KPa
a'
a.
S= 117 KPa
m =0.00036 m/KN
m= 0.00071 m2/KN
E= 2.500 KPa
S
0
800
700
dYv.Kpa
800
::::::::::E= 1,266 KPa
0.05
0.
.si.
)5
15
0.1
T -:': ...............
15.
h
i
.*1]
.
....
:I
. .
.
.
1
I7
..
I
i.
.. ...
.....
....
.
: .. .:
--:-
.. ......
....-......
. ....
. ...
..
0 .2
:-
'' "
i.:.;..
i.....i.:...
.
.. ....
... ....
I
......
..
.. .
0.
0.3
a',
C3
0
100
200
300
Kpa
C4
500
400
700
600
800
0
0
. .......
O' ,Kpa
100
200
300
400
500
600
700
800
= 221 KPa
2
m = 0.00015 m /KN
0.02
.-.-. -..
....
-.
-.
,-
0.2
E= 5,888KPa
T
0.04.
7. .... .. ..
. ..
0.3C4'
0.05-*-
i
...
.
.....
i.
...
i.:
...
....
..
...
.:
.
....: ......
........
.....
c, Kpa
0.06
i :.
..
.: .-i.. -
...
.i..i -i
..
.. .
0:
0.12
.. . ...............
0.14
Figure C
E' Computations from Oedometer Tests
211
E= .2.11
KP
..........
Ub
0
C6
o' v,Kpa
100
200
300
500
400
600
800
700
o'v
,Kpa
0
100
200
300
400
500
600
700
800
0
0.
' - 125 KPa
ov117 KPa
2
m = 0.00047 m /KN
2
0.0003 m /KN
m
0.02K
E= 1,886 KPa
E= 3,045 KPa
0.05
0.04
0.06
.'
. . .-
0.15
..
. ...
i
-
0.08.....
0.1
---
.......
0.2
0.14i
i
i
-
0.16
C7
0
o'v, Kpa
100
200
300
400
500
600
700
-
800
0'
......-
0.01 ,5
o'v= 165 KPa
2
m=
S 0.00011 m /KN
E= 8,486 KPa
dea
dcr '
0.02:
0.03 L
0.o
0.06
-
(1+ v'X1 - 2v')
O-') mv
F
(v'=0.2)
.
0.07
-
0.08
Figure C2 E' Computation from Oedometer Tests
212