GEOTECHNICAL SPECIAL PUBLICATION NO.

106

Geotechnical
Measurements
Lab and Field
PR O C EED IN G S OF S E SS IO N S OF G E O -D E N V E R 2 0 0 0

SPONSORED BY
The G eo-Institute o f the A m erican S o ciety o f C ivil E ngineers

A ugu st 5 -8 , 2 0 0 0
D enver, C olorado

EDITED BY
W. A llen Marr

I*

American Society
INSTITUTE of Civil Engineers
1801 ALEXANDER BELL DRIVE
RESTON, VIRGINIA 20191-4400
Abstract: The papers included in this Geotechnical Special Publication deal with Geotechnical Special Publications
geotechnical measurements. The first group focuses on laboratory measurements
and their use to further our understanding and characterize geotechnical behavior. 1 Terzaglii Lectures 32 Embankment o f Dams-James L. Sherard
2 Geotechnical Aspects o f Stiff ami Hard Contributions
The second group focuses on the application of methods to measure actual Clays 33 Excavation and Support fo r the Urban
performance of the geotechnical elements of constructed facilities. 3 Landslide Dams: Processes, Risk, and Infrastructure
Mitigation 34 Piles Under Dynamic Loads
4 Tiebacks fo r Bulkheads 35 Geotechnical Practice in Dam Rehabilitation
Library of Congress Cataloging-in-Publication Data 5 Settlement o f Shallow Foundation on 36 Fly Ash fo r Soil Improvement
Cohesionless Soils: Design and 37 Advances in Site Characterization: Data
Geo-Denver 2000 (2000 : Denver, Colo.) Performance Acquisition, Data Management and Data
Geotechnical measurements : lab and field : proceedings of sessions of Geo- 6 Use o f In Situ Tests in Geotechnical Interpretation
Engineering 38 Design and Performance o f Deep
Denver 2000 : August 5-8, 2000 / sponsored by the Geo-Institute of the American
7 Timber Bulkheads Foundations: Piles and Piers in Soil and
Society of Civil Engineers ; edited by W. Allen Marr. 8 Foundations fo r Transmission Line Soft Rock
p. cm. (Geotechnical special publication ; no. 106) Towers 39 Unsaturated Soils
Includes bibliographical references and index. 9 Foundations & Excavations in Decomposed 40 Vertical and Horizontal Deformations o f
ISBN 0-7844-0518-2 Rock o f the Piedmont Province Foundations and Embankments
10 Engineering Aspects o f Soil Erosion, 41 Predicted and Measured Behavior o f Five
1. Engineering geology— Measurement— Congresses. 2. Soils—Testing— Dispersive Clays and Loess Spread Footings on Sand
Congresses. 3. Rocks—Testing—Congresses. I. Marr, W. Allen. II. Title. III. Series. 11 Dynamic Response o f Pile Foundations- 42 Serviceability o f Earth Retaining Structures
TA703.5. G417 2000c Experiment, Analysis and Observation 43 Fracture Mechanics Applied to Geotechnical
12 Soil Improvement: A Ten Year Update Engineering
624.1 '51 —dc21 00-042132 13 Geotechnical Practice fo r Solid Waste 44 Ground Failures Under Seismic Conditions
Disposal V7 45 In Situ Deep Soil Improvement
14 Geotechnical Aspects o f Karst Terrains 46 Geoenvironment 2000
Any statements expressed in these materials are those of the individual authors and do not 15 Measured Performance Shallow 47 Geo-Environmental Issues Facing the
necessarily represent the views of ASCE, which takes no responsibility for any statement Foundations Americas
made herein. No reference made in this publication to any specific method, product, process 16 Special Topics in Foundations 48 Soil Suction Applications in Geotechnical
17 Soil Properties Evaluation from Centrifugal Engineering
or service constitutes or implies an endorsement, recommendation, or warranty thereof by
Models 49 Soil Improvement fo r Earthquake Hazard
ASCE. The materials are for general information only and do not represent a standard of 18 Geosynthetics fo r Soil Improvement Mitigation
ASCE, nor are they intended as a reference in purchase specifications, contracts, 19 Mine Induced Subsidence: Effects on 50 Foundation Upgrading and Repair for
regulations, statutes, or any other legal document. ASCE makes no representation or Engineered Structures Infrastructure Improvement
warranty of any kind, whether express or implied, concerning the accuracy, completeness, 20 Earthquake Engineering & Soil Dynamics II 51 Performance o f Deep Foundations Under
suitability, or utility of any information, apparatus, product, or process discussed in this 21 Hydraulic Fill Structures Seismic Loading
publication, and assumes no liability therefore. This information should not be used without 22 Foundation Engineering 52 Landslides Under Static and Dynamic
first securing competent advice with respect to its suitability for any general or specific 23 Predicted and Observed Axial Behavior o f Conditions-Analysis, Monitoring, and
application. Anyone utilizing this information assumes all liability arising from such use, Piles Mitigation
including but not limited to infringement of any patent or patents. 24 Resilient Moduli o f Soils: Laboratory 53 Landfill Closures-Environmental
Conditions Protection and Land Recovery
Photocopies: Authorization to photocopy material for internal or personal use under 25 Design and Performance o f Earth Retaining 54 Earthquake Design and Performance o f
circumstances not falling within the fair use provisions of the Copyright Act is granted by Structures Solid Waste Landfills
ASCE to libraries and other users registered with the Copyright Clearance Center (CCC) 26 Waste Containment Systems: Construction, 55 Earthquake-Induced Movements and
Transactional Reporting Service, provided that the base fee of $8.00 per article plus $.50 per Regulation, and Performance Seismic Remediation o f Existing
page is paid directly to CCC, 222 Rosewood Drive, Danvers, MA 01923. The identification 27 Geotechnical Engineering Congress Foundations and Abutments
for ASCE Books is 0-7844-0518-2/00/ $8.00 + $.50 per page. Requests for special 28 Detection o f and Construction at the 56 Static and Dynamic Properties o f Gravelly
permission or bulk copying should be addressed to Permissions & Copyright Dept., ASCE. Soil/Rock Interface Soils
29 Recent Advances in Instrumentation, Data 57 Verification o f Geotechnical Grouting
Acquisition and Testing in Soil Dynamics 58 Uncertainty in the Geologic Environment
Copyright © 2000 by the American Society of Civil Engineers 30 Grouting, Soil Improvement and 59 Engineered Contaminated Soils and
All Rights Reserved. Geosynthetics Interaction o f Soil Geomembranes
Library of Congress Catalog Card No: 00-042132 31 Stability and Performance o f Slopes and 60 Analysis and Design o f Retaining
ISBN 0-7844-0518-2 Embankments II Structures Against Earthquakes
Manufactured in the United States of America.
61 Measuring and Modeling Time Dependent
Soil Behavior
89 Recent Advances in the Characterization of
Transportation Geo-Materials
Preface
62 Case Histories o f Geophysics Applied to 90 Geo-Engineering fo r Underground
Civil Engineering and Public Policy Facilities
63 Design with Residual Materials: 91 Special Geotechnical Testing: Central Measurements are essential to geotechnical practice. Measurements give us the
Geotechnical and Construction Artery/Tunnel Project in Boston, means to obtain realistic and meaningful numbers to use in design. The giants in
Considerations Massachusetts geotechnical engineering excelled at taking the right measurements and knowing
64 Observation and Modeling in Numerical 92 Behavioral Characteristics o f Residual
Analysis and Mode! Tests in Dynamic Soils what to do with them.
Soil-Structure Interaction Problems 93 National Geotechnical Experimentation
65 Dredging and Management o f Dredged Sites Geotechnical measurements include observations of the physical state of
Material 94 Performance Confirmation o f Constructed subsurface materials and how that state responds to changes in the physical
66 Grouting: Compaction, Remediation and Geotechnical Facilities environment. These measurements may be carried out in the laboratory on
Testing 95 Soil-Cement and Other Construction
67 Spatial Analysis in Soil Dynamics and Practices in Geotechnical Engineering
representative samples of the materials. They may be made in situ as either a
Earthquake Engineering 96 Numerical Methods in Geotechnical measurement of a property or an observation of the change of physical state, such as a
68 Unsaturated Soil Engineering Practice Engineering: Recent Developments movement or a change in pore water pressure.
69 Ground Improvement, Ground 97 Innovations and Applications in
Reinforcement, Ground Treatment: Geotechnical Site Characterization Geotechnical measurements have provided the life blood for advances in
Developments 1987-1997 98 Pavement Subgrade, Unbound Materials,
70 Seismic Analysis and Design fo r Soil-Pile- and Nondestructh’e Testing
modern geotechnical engineering. Karl Terzaghi writing in 1936 said, “I came to the
Structure Interactions 99 Advances in Unsaturated Geotechnics United States and hoped to discover the philosopher’s stone by accumulating and
71 In Situ Remediation o f the 100 New Technologies and Design coordinating geological information....It took me two years o f strenuous work to
Geoenvironment Developments in Deep Foundations discover that geological information must be supplemented by numerical data which
72 Degradation o f Natural Building Stone 101 Slope Stability 2000 can only be obtained by physical tests carried out in a laboratory.” Later in his
73 Innovative Design and Construction for 102 Trends in Rock Mechanics
Foundations and Substructures Subject to 103 Advances in Transportation and
career, Terzaghi began to rely on measurements of field performance during
Freezing and Frost Geoenvironmenfal Systems Using construction to help bridge the gaps in data from the field and lab work performed
74 Guidelines o f Engineering Practicefo r Geosynthetics during the planning and design stages of a project.
Braced and Tied-Buck Excavations 104 Advances in Grouting and Ground
75 Geotechnical Earthquake Engineering and Modification The papers included in this Geotechnical Specialty Publication deal with
Soil Dynamics HI 105 Environmental Geotechnics
76 Geosynthetics in Foundation geotechnical measurements. The first group focuses on laboratory measurements and
106 Geotechnical Measurements: Lab <£ Field
Reinforcement and Erosion Control their use to further our understanding and characterize geotechnical behavior. The
Systems second group focuses on the application of methods to measure actual performance of
77 Stability o f Natural Slopes in the Coastal the geotechnical elements of constructed facilities. These papers provide insight to
Plain
78 Filtration and Drainage in
the state of our practice at this point in time.
GeotechnicaUGeoenvironmental
Engineering These papers were prepared and presented in conjunction with the GeoDenver
79 Recycled Materials in Geotechnical 2000 Geo-Institute Speciality Conference held on August 5-8, 2000 at the Denver
Applications Marriot Technological Center. For the preparation of this GSP, all papers were peer-
80 Grouts and Grouting: A Potpourri o f
reviewed by at least one anonymous reviewer in accordance with the standards of
Projects
81 Soil Improvement fo r Big Digs ASCE and the Geo-Institute. Required revisions were made by the authors prior to
82 Risk-Based Corrective Action and final acceptance and publication. All papers are eligible for discussion in the ASCE
Brownfields Restorations Journal of Geotechnical and Geo-Environmental Engineering and for ASCE awards.
83 Design and Construction o f Earth The editor wishes to express sincere thanks to the anonymous reviewers for their
Retaining Systems
otherwise thankless efforts to provide accurate and timely reviews.
84 Effects o f Construction on Structures
85 Application o f Geotechnical Principles in
Pavement Engineering
86 Big Digs Around the World
87 Jacked Tunnel Design and Construction
88 Analysis, Design, Construction, and
Testing o f Deep Foundations
 Contents

PID Controlled Combined Axial Torsional Testing System for Cohesive Soil..................1
Dayakar Penumadu and Douglas Mandeville
Effect of Sample Preparation on Steady State.................................................................. 16
Seyed Abolhassan Naeini and Mohammad Hassan Baziar
Quantifying Void Ratio Variation in Sand Using Computed Tomography.....................30
Khalid A. Alshibli, Susan N. Batiste, Roy A. Swanson, Stein Sture,
Nicholas C. Costes, and Mark R. Lankton
Load-Deformation Performance of Peat Soil Under Large Concrete Plates..................44
Mark Vakher
Modeling Grain Crushing and Debonding of Soils........................................................... 56
Jaroslav Feda
Practical Applications of Time Domain Reflectometry (TDR) to Monitor
and Analyze Soil and Rock Slopes.................................................................................65
Neil Anderson and David Welch
Comparison of TDR and Inclinometers for Slope Monitoring........................................80
Charles H. Dowding and Kevin M. O’Connor
Performance Monitoring for a Critical Structure Built Within a Landslide..................91
Barry K. Myers, L. Randley Squier, Mark P. Biever and Rex K.H. Wong
Wharf Movements Due to Tidal Fluctuations................................................................. 109
Jonathan E. Thomas and Wilbur J. Morin
Inclinometer Data Analysis for Remediated Landslides................................................. 126
William R. Stevens and Bill E. Zehrbach
Stabilization of Fill Slopes Solved Using Deep-High Capacity Tiebacks.......................138
Chester A. Carville and Ronald J. Tanenbaum
Strain Measurements of Concrete Piers in Expansive Soils....................... 151
Thomas A. Chapel and John D. Nelson
Pore Pressures in Partially Saturated Soil: A Case History of the Diamond Valley
Lake Project........................................................................................................................164
Michael G. Smith, Russell L. Jemigan, and Arleen A. Arita
Performance of an Anchored Sheet-Pile Wall.................................................................. 179
Shailendra N. Endley, Wayne A. Dunlap, David M. Knuckey, and
Karun Sreerama
Subject Index...................................................................................................................... 199
Author Index...................................................................................................................... 200

vii
 PID Controlled Combined Axial Torsional Testing System for Cohesive Soil

Dayakar Penumadu, Member ASCE1, and Douglas Mandeville, Member ASCE12

Abstract

The present paper summarizes the details o f a testing system developed for
performing combined axial torsional loading on hollow cylinder cohesive
specimens. The axial-torsional loading system used in this research employs a MTS
loading frame with Testware-SX control. The frame is operated hydraulically
through the use o f a servovalve that feeds actuators which can independently control
the axial load/displacement and/or torque/angular rotation. Stress paths ranging
from triaxial compression to pure shear to triaxial extension can be obtained using
this testing apparatus by rotating the fixed inclination of the major principal stress
(p) with the axis of rotational symmetry. The ratio between the torsional shear
stress increment and the vertical stress increment is maintained a constant value to
achieve a desired p value. A new approach using PID control was implemented to
achieve precise control. To implement this into the control procedure, calculated
control channel was used throughout the test using updated specimen dimensions.
Undisturbed hollow cylinder specimens were obtained using high water content
kaolinite slurry, and custom built consolidometer. Typical test results for stress path
corresponding to p = 30 degrees were presented for isotropically consolidated kaolin
clay under undrained conditions.

Previous Hollow Cylinder Axial-Torsional Devices

Triaxial testing of hollow cylinder soil specimens subjected to torque was

first performed by Cooling and Smith [1936]. In their study, torque was applied to
unconfined hollow cylinder specimens to evaluate the resistance o f soil under pure
shear conditions. The use o f specimens having a hollow cylinder shape in soil
testing became more widely used recently. The effects of the intermediate principal

1Associate Professor, 128 Rowley Laboratories, Civil and Environmental

Engineering, Clarkson University, Potsdam, NY 13699-5710.
2 Staff Engineer, 10015 Old Columbia Road, Suite A-200, GeoSyntec Consultants,
Columbia, MD 21046.

l
2 GEOTECHNICAL. MEASUREMENTS GEOTECHNICAL MEASUREMENTS 3

stress, CT2, and the effects of principal stress rotation was studied by Broms and authors were able to follow complex stress paths using manual control of the
Casbarian [1965]. They concluded that the principle of superposition could be used apparatus.
to predict the shear strength and pore pressure parameters for any rotation o f the
principal stress axes. Problems associated with past test data have been discussed recently
[Tatsuoka et al. 1986, Sayao and Vaid 1991, Frost and Dmevich 1994], The major
In the mid 1960’s, Saada introduced an automated testing apparatus which issues that primarily deal with testing procedures are: 1) measurement of axial and
could be used to perform hollow cylinder axial - torsional testing [Saada 1967 and torsional forces without any coupling effects or cross-talk, 2) neglecting the
1968], The inclination of the principal stresses created by subjecting the soil inclusion of half the specimen weight in the calculation of vertical stress at the
specimens to a combination o f axial and torsional stresses was used to study center o f the specimen, 3) correction of forces carried by the membrane and filter
anisotropy in soils. This device was used in many subsequent investigations to paper strips for a given cell pressure, 4) provision of full friction surfaces to the
examine the behavior o f anisotropic clays [Saada 1967, Saada and Zamani 1969, specimen to avoid slippage at the interfaces with end-platens, and 5) obtaining
Saada and Ou 1973, Saada and Bianchini 1975]. Saada discusses his findings as homogeneous and repeatable hollow cylinder soil specimens.
well as the advantages and limitations of hollow cylinder testing in a state of the art
paper [Saada 1988], Ishibashi and Sherif [1974] developed a torsional simple shear Axial - Torsional Shear Frame and Control Software
apparatus to study soil liquefaction.
The axial-torsional loading system used in this research employs a MTS
loading frame with Testware-SX control. The frame is operated hydraulically
Hollow cylinder specimens in Saada’s studies were obtained by coring full
through the use of a servovalve that feeds actuators which can control the axial
cylinder specimens. Preparing samples in this manner can lead to handling and
load/displacement and/or torque/angular rotation. Along the axial direction the
some disturbance to the specimens. In this research, specimens were made using the
system is capable of applying ±9.8 kN load or 12.7 cm displacement (full stroke).
slurry consolidation method and trimming of the specimens was avoided. Saada’s
Along the torsional axis, the system is capable o f applying ±113 N-m torque or ±
device was limited to load control. This prohibits the possibility of investigating the
45° rotation. The stress or strain controlled actuators can be programmed in
post failure conditions of the soil. Pore pressure equilibration and strain rate
Testware-SX to have accurate, automated, and repeatable control o f the axial and
become major issues at failure under load control. Failure is sudden in load control
torsional loads or displacements. With this testing system, any desired stress path
mode; the strain rate becomes very fast. In clay specimens, the pore pressure may
involving the rotation of the major principal stress can be accurately achieved for
not have a chance to equilibrate when this happens.
soil specimens. The MTS system (Fig. 1) uses a closed loop servo-control for
achieving the desired stress or strain path. The control loop has a very fast update
Lade developed a torsion shear apparatus with an outside diameter of 22 cm,
rate of 5 kHz and is capable o f supporting up to ten input signals and eight
an inside diameter of 18 cm and a height of 5 cm [Lade 1975]. Further examination
calculated control channels. The controlling elements are the personal computer and
revealed that these specimen dimensions created nonuniform stresses and strains in
the digital controller; they determine what the actuator does. The control signal
the hollow cylinder [Lade 1976] and a specimen height of 25 cm was found to
being generated represents the action the actuator should take to apply load or
create a uniform stress state in the hollow cylinder [Lade 1981]. Lade’s work with
deformations to the specimen. The control signal is transmitted to the controlled
kaolin clay using the torsion shear apparatus revealed that the failure surface
element consisting of the servovalve, the hydraulic actuator, and the specimen itself.
obtained from experiments on K« consolidated hollow cylinder torsion shear tests
The feedback is a response from the appropriate sensor(s) that indicates how the
could be modeled using an isotropic failure criteria [Hong and Lade 1989], Hong
controlled element has performed for the new conditions o f variables (e.g. stress or
and Lade also examined the strain increment and stress directions in torsion shear
strain). The digital controller then reacts to the relative difference between the
tests on Ko consolidated kaolin clay. They observed that the clay behaved as an
desired signal and the feedback signal and adjusts the control signal to correct the
elasto-plastic material [Hong and Lade 1989].
difference. The control software specifically developed to perform combined axial
torsionsal tests on kaolin clay will be discussed in a later section.
A hollow cylinder torsion shear testing device which was capable of
applying different internal and external cell pressures was introduced in 1983 [Hight
Axial - Torsional Shear Cell
et al. 1983]. Their hollow cylinder device tested sand specimens with an inside
diameter of 20.3 cm, an outside diameter of 25.4 cm and a height of 25.4 cm. The An axial torsional triaxial cell was custom designed and fabricated in a
large sample size was necessary to be able to monitor displacements over a central previous study by Abrantes [1998] under the direction of the primary author. The
gauge length and measure pore pressure of the specimen at its mid height. They main characteristics of this cell are: the ability to apply normal and shear stresses on
used finite element analysis to determine the optimum specimen dimensions. The a hollow cylinder specimen without coupling or slippage, individual control of pore,
 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 5

cell, axial and shear stresses to the specimen, and ease o f assembly. The cell can
accomodate specimens with an inside diameter o f 7.1 cm, an outside diameter of
10.2 cm and a height o f 20 cm. In this research, the pressure on the inside and the
outside of the hollow cylinder specimen were always kept equal. Radial drainage is
provided through holes drilled on the sides of the end caps and through the use o f
filter paper strips. Porous plastic plugs were placed in the drainage holes to prevent
them from clogging with clay during saturation and consolidation. The filter paper
pattern, following the recommendations of Berre [1982 was used for compression,
extension, and torsion shear tests. For (3 equal to 0°, 30°, and 45°, the filter paper
pattern suggested for triaxial compression was used because the specimen was being
compressed axially. For [3 equal to 60° and 90°, the filter paper pattern
recommended for triaxial extension was used because the specimen was being
extended axially. The lack o f prior recommendations in the literature regarding
filter paper for three-dimensional testing lead to this. The filter paper was orientated
on the sample such that it would unwind with an increase in torque. To provide
frictional resistance at the end caps, short, metal tabs were glued to the end caps.
These tabs are very thin and have a height of 5mm and caused minimal disturbance
to the specimen during the process of attaching the end platens. This cell was also
used to perform tests on solid cylinder specimens by employing a different set o f
end caps. The solid cylinder end caps have radial drainage and are designed to be
used as lubricated ends with thin latex membranes and vacuum grease. The
specimens had a diameter of 10 cm and a height o f 10 cm (H/D = 1.0). The
lubricated end caps have a diameter o f 10.8 cm and allow the specimen to deform as
Slot Conditioner Type Range/Capacity a right cylinder during shearing, resulting in negligible frictional end boundaries
J1 AC ^ LVDT (axial displacement) 12.7 cm during the application of deviator stress. A complete drainage path from the sample
J2 DC ► Load Cell 9.8 kN to the drainage lines was provided using filter paper strips. Lubrication is provided
J3 DC ► Extensometer 12% Full Scale through the use of 0.6 mm thick membranes and a thin layer of vacuum grease. The
J4 DC —► Torque 113 N-m inclination o f filter paper strips used in this research were based on the
J3 Uv/ —► Angle Rotation +/- 45 degrees recommendations o f Berre [1982],
Jo np
uu ^ DifTeiential Tidiisducci 5.G kPs
J7 DC
jg _^ rVII
“ LUdu UCU A design modification was made to the cell to allow for long-term drained
jo DC testing. Over the course o f several days required for drained testing, air was found
Tin pr to diffuse into the confining fluid, across the latex membrane, and into the pore
111 water lines. An impermeable air bladder was added between the cell and the
J12 _Servovalve Connector (E/P Transducer) pressure supply to prevent the diffusion o f air across the latex membrane and into
J13 __Servovalve Connector (Torsion) the sample. A thin layer o f silicone oil was added to the top of the burette, part of
J14 ^ Servovalve Connector (Axial) the volume change device. This was done to prevent the loss o f water through
evaporation during drained testing and minimize air diffusion under back-pressure.
Digital Controller

Hollow Cylinder Specimen Preparation

Fig. 1: Combined Axial Torsional Test System
Undisturbed hollow cylinder kaolin clay specimens were obtained using the
slurry consolidation method. The slurry had an initial water content of 125%. The
 GEOTECHNICAL MEASUREMENTS 7
6 GEOTECHNICAL MEASUREMENTS

slurry was consolidated one dimensionally. A custom built consolidator was used to
consolidate the specimens. The consolidator was lined with Teflon to minimize
surface friction effects. Using double drainage, specimens were consolidated under
a vertical stress o f 207 kPa (30 psi) for approximately 48 hours under Ko conditions,
well beyond the end o f primary consolidation. The specimens had a final inside
diameter of 7.1 cm, an outside diameter o f 10.2 cm, and a height of 20.3 cm. These

Height (cm)
dimensions provide a zone o f uniform stress and strain distribution by minimizing
the effects of the frictional end platens in axial - torsional testing [Lade 1981].

In a previous research program performed using this consolidator [Abrantes

1998], a difference in the water contents of 4.5 % was noticed for hollow cylinder
specimens after slurry consolidation. The slurry consolidation procedure involved
pouring kaolinite slurry into the teflon lined consoUdometer and subjecting it to 210
kPa by a loading piston at the top. Drainage was allowed through the top and
bottom of the consolidometer. For these specimens, the water content varied from
43% (top of sample) to 47.5% (bottom of sample) at the end of K<> consolidation
(Fig. 2). The average water content was 45.7% and the void ratio varied from 1.13
at the top of the specimen to 1.24 at the bottom of the specimen. This was attributed
to friction due the large height to wall thickness ratio. After isotropic consolidation Water Content (%)_____________

fTI
in the cell, the water content, and correspondingly, the void ratio, was constant Consolidated from top only • • » - - Consolidated from both ends
across the height. However, this results in a change in the cross sectional area
because the change in the void ratio is different at each point along the height o f the W ater C ontent V ariation A cross H o llo w Cylinder S pecim en H eight
sample. At the top o f the sample, the void ratio changes from 1.13 to .947 and from
1.24 to .947 at the bottom of the specimen. This corresponds to the cross sectional
area of the specimen ranging from 39.4 cm2 at the top to 36.1 cm2 at the bottom.

An alternative method was devised in an attempt to minimize the water

Volumetric Strain (%)

content variation across the height of the specimen. The main idea was to
consolidate the sample by applying vertical stress from both ends o f the slurry
consolidometer. The modified procedure involved consolidating from one direction
for several hours, then remove the base and use two pistons and consolidate from
both directions simultaneously. The reason for the specimen to be consolidated in
one direction initially was to transform the slurry to a semi-solid state for facilitating
consolidation from both directions. The resulting variation o f the water content
using this method is also shown in Fig. 2. The average water content was 45.2%.
The water content is still higher in the middle, but is more uniform in the
representative zone free from the end platens. Full cylinder specimens were also
prepared using the slurry consolidation method. Powdered kaolinite was mixed with
de-aired water to make a slurry with an initial water content of 155%. The slurry

Square Root of Time

Fig. 3: Isotropic C onsolidation Curve (2 7 6 kPa)
8 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 9

was consolidated one dimensionally at 207 kPa (30 psi). The final dimensions of A typical example of volume change behavior for a hollow cylinder
the cylinder were 10 cm (4 in) in diameter and 10 cm (4 in) in height. The final specimen during isotropic consolidation is shown in Fig. 3. The consolidation
water content was 44%. The maximum water content variation across the specimen process was allowed for 14 hours which is well beyond the time of completion for
height was less than 2%, resulting in very uniform specimens. the primary consolidation.

Specimen Assembly In drained triaxial testing, the consolidation curve is also used to find the
appropriate deformation rate during shearing. Bishop and Henkel’s [1960]
The triaxial cell used for this research permits easy assembly of both hollow procedure was used to find the appropriate strain rate to ensure that no excess pore
and full cylinder specimens with minimal disturbance. The following steps were pressure is built up during the application of shear stress. Using their method, the
used in the assembly process for hollow cylinder specimens: time to failure was found to be 35 hours. Assuming that failure occurs at about 10%
1) The bottom cap is attached to the base of the cell. axial strain, a strain rate o f .0045 %/min (deformation rate of .0009 cm/min) was
2) A latex membrane with a diameter of 7.1 cm and a thickness of .6 mm is sealed used. This rate was found to be slow enough to prevent the generation of excess
to the inside of the bottom cap. pore pressures during shearing. A strain rate o f .05%/min was used for the
3) The specimen is lowered around the membrane and is seated on the bottom cap. undrained tests and permitted equalization of the pore pressures during undrained
4) The top cap is placed on the specimen and the inner membrane is sealed around testing.
the inside of the top cap.
5) Filter paper, used to provide radial drainage, is placed on the specimen. The Calculation of Stresses and Strains
filter paper is designed so that no corrections are needed for the axial load. The
angle of the filter paper differs for compression and extension tests [Berre 1982]. The kaolin clay hollow cylinder specimens were subjected to a combination
Great care was taken to ensure that the filter paper is always in contact with the of axial load and/or torque at the end of isotropic consolidation. During testing, the
porous plugs in the top and bottom caps. inner and outer pressures (Pj and P0) were kept equal throughout testing. Fig. 4
6) At this point, the outside membrane is placed around the specimen and is sealed shows the states of stress produced on a hollow cylinder specimen subjected to a
with O-rings. combination o f axial load and torque. After correcting for the inner and outer
7) The pore lines are attached to the top cap. membrane stiffness, piston friction, piston uplift, and system compliance, Eqs. 2
8) The Lucite chamber is set into place and the top cap is attached to the piston. through 9 were used to calculate various stresses and strains on the hollow cylinder
specimen [Saada 1988]. A single load cell measured the axial load and torque. No
After the specimen is fully assembled in the triaxial cell, the saturation measurable “cross-talk” was noticed between the axial load and torque readings.
process was started. The system is flushed with CO2 and then purged with de-aired
water. The active back pressure method is used to dissolve any CO2 trapped in
solution. A back pressure of only 35 kPa was sufficient to reach full saturation in ( 2)
o, = 2 D2\ + OVceU
.
few hours. Full saturation was assumed when the B-value (Au/Acrceii) reached a z x (Ro
2 -R?)
value of 0.99. The next step in the process was to perform isotropic consolidation.
Valves connecting pore lines to the volume change device were closed and the cell (3)
pressure was increased steadily. When the cell pressure reached the desired level, O. = Or
the drain lines were opened and the consolidation process started. During
consolidation, the volume change of the specimen was measured with a burette and
the change in height of the specimen was measured with a dial gauge. This was (4)
required to calculate the correct dimensions of the specimen at the end of
consolidation. At the end of consolidation, Eq. 1 was used to find the new outside
radius and the new inside radius of the hollow cylinder [Tatsuoka, 1986]: 3M,
2x ( R l - R } ) <5>
Ro •w initial ( 1)
AH
(6)
GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 11

(7)

Up + »,
( 8)
K-R,
<T22-tfe
20(Rl-R?)
(9)
CTu-ar 3H ( R 20 - R ? )

Eqs. 10 to 12 were used to find the incremental principal stresses in terms of

the vertical stress, Actz, and a ratio term called the k value. The k value is equal to
the ratio between the torsional shear and the vertical stress increments (k= Axez/Aaz).

(10)
A ct, = - ^ ( \ + yJ\ + 4 k 2 )

ii
>
q

o
OD

(12)
A ct3 - A<Tz(1 Vl + 4 * 2 )

Control Software and Typical Test Data

In this research, the angle of inclination o f the major principal stress in

relation to the axis of rotational symmetry, p, is controlled to vary from 0 to 90
degrees. When p is equal to 0 degrees, the specimen is in triaxial compression state
of stress. For P equal to 45 degrees, the specimen is in a state o f pure torsional
shear. At P equal to 90 degrees, the specimen is in triaxial extension mode. These
three cases are relatively easy to control in Testware-SX. The desired loading rate is
entered into the control program and either the shear stress is held at zero for tests
corresponding to p equal to 0 and 90 degrees or the load is held at zero for tests
corresponding to P equal to 45 degrees. However, when p equals 30 or 60 degrees,
a combination of axial and torsional loads is applied to the specimen. The specimen
undergoes proportional loading, with a constant ratio between the axial stress and
the torsional shear stress increments to maintain a constant p throughout the test.
Fig. 4: State of Stress in Hollow Cylinder Specimen
12 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 13

The algorithm required to perform a combined axial - torsional test is fine adjustments to the torsional shear stress on a continual basis as opposed to
discussed below. Eq. 13, provides a unique relationship between P (the inclination switching on and off when a certain limit is exceeded as typically done in the past
o f the major principal stress) and k (ratio of stress increments). research. The PID values have a large influence on the degree of control maintained
during testing. In this testing, using the Auto-Tune feature o f TestStar and other tuning
techniques, P was found to be 11,000, I to be 1700 and D to be zero. The relatively
t-tan(2 fl) = ^ =k (13) high P and I terms were necessary to have the system respond in a sufficient manner.
2 A ct, These coefficients were found to maintain the best control during both undrained and
drained testing.
If p is equal to 30°, the corresponding value for k is 0.866. The ratio between the
Control Method Description 4>' Su(kPa) e,r(%) Ut(kPa) P max p min
torsional shear stress increment and the vertical stress increment is 0.866. If p is
equal to 60°, the corresponding value for k is -0.866. To implement this into the PID Strain Control in Axial Direction 31.69 79.7 8.2 163.9 31.80 29.41
Stress Control in the Torsional
control procedure, the following expression (Eq. 14) is maintained by the calculated Direction
control channel throughout the test:
Table 1: Failure Parameters for p=30°

0.866 • Aar. - A r& = 0 (14) Use of this new control method prevented the peaks and valleys in the vertical and
shear stress that usually occur with the limit detector. This can be seen in the
variation of torsional shear in Fig. 5. The axial strain rate for this test was
Eq. 14 is equal to zero at the start of the application of axial/torsional stresses; it ,05%/min; the torsional stress was not applied at a constant rate because of the use
should remain close to zero throughout the test to maintain a constant inclination of of PID control. The new method with PID control reached the desired value of p
the principal stress at 30°. The axial stress was increased under strain control mode from the beginning o f the test. For an entire test, the value o f P varied only about a
for all the testing at the appropriate strain rate (for undrained tests the criteria being degree in either direction from the target value (Fig. 6). Table 1 gives a summary of
equilibration of excess pore pressure). A new closed-loop method was devised in the friction angle, undrained shear strength, major principal strain at failure and the
this research using PID control for the torsional axis to maintain the desired excess pore pressure generated at failure. The ability to maintain the desired value of
relationship between the vertical and shear stress increments. This ensures constant P throughout the test is of particular concern when testing over consolidated clays
inclination of the major principal stress at all times. This was accomplished by which can reach failure at low strains.
setting up a new control mode in Testware-SX. The algorithm is based on adjusting
the output channel to match a target command using three terms: a proportional term Summary and Conclusions
P, an integral term I, and a differential term, D, using Eq. 15:
The description o f an automated torsional shear testing system was presented. This
system was used to perform experiments on kaolin clay. The hollow cylinder
0( t ) = Pe(t) + l \e(t')dt' + D — (15) specimens were obtained using a two-stage slurry consolidation technique. Issues
J dt associated with specimen preparation, and assembly in a custom fabricated axial-
torsional test cell were discussed. The stress paths involving fixed inclination of
The e(t) term represents the error term and is the difference between the target major principal stress (P) were performed. This approach uses updated specimen
value and the input value at any instant of time. The PID algorithm minimizes this dimensions while undergoing deformations for calculating the inclination o f major
difference by adjusting the output value. Recall that the relationship between the principal stress. Test results for p = 30° were presented demonstrating a precise
torsional shear stress increment and the vertical stress increment should satisfy Eq. 14. control of P (variation of less than 1° ) using PID algorithm.
The target for control is to keep this equation at zero throughout the test. The vertical
stress increment, Aaz, is being increased in strain control and the torsional shear stress Acknowledgements
increment, Atez, has to be adjusted accordingly. As this equation deviates from zero, the
PID control either increases or decreases the torque, which changes the torsional shear This material is based upon work supported by the National Science Foundation
stress and makes the value of Eq. 14 closer to zero. With a 5 kHz update rate, true under Grant No. 9872618.
stress control can be achieved for the torsional shear stress. The PID control makes very
14 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 15

4) Broms, B. B., and Casbarian, A. O., 1965, “Effects of the Principal Stress Axes and of
the Intermediate Stress on Shear Strength,” Proc. 6th International Conference on Soil
Mechanics and Foundation Engineering. Montreal, Vol. 1, pp. 179-183.
5) Cooling, L. F., and Smith, D. B., 1935, “The Shearing Resistance of Soils,”
Proceedings. Institute of Civil Engineers. Vol. 3, pp. 333-343
6) Frost, J. D., and Dmevich, V. P., 1994, “Towards Standardization of Torsional Shear
Testing,” Dynamic Geotechnical Testing II, ASTM STP 1213. Ronald J. Ebelhar, Vincent
P. Dmevich, and Brace L. Kutter, Eds., ASTM, Philadelphia, pp. 276-287.
7) Hight, D. W., Gens, A., and Symes, M. J., 1983, “The Development of a New Hollow
Cylinder Apparatus for Investigating the Effects of Principal Stress Rotation in Soil,”
r.potechniaue. Vol. 33, pp. 355-383.
8) Hong, W. P., and Lade, P. V., 1989, “Elasto-Plastic Behavior of K« Consolidated Clay
in Torsion Shear Tests,” Soils and Foundations. Vol. 29, No. 2, pp. 127-140.
9) Hong, W. P., and Lade, P. V., 1989, “Strain Increment and Stress Directions in Torsion
Shear Tests.” Journal of Geotechnical Engineering. Vol. 115, No. 10, pp. 1388-1401.
10) Ishibashi, I. and Sherif, M.A., "Soil Liquefaction by Torsional Simple Shear Device,"
Journal of the Geotechnical Engineering Division. ASCE, Vol. 100, No. GT8, Aug., 1974,
pp. 871-888.
11) Lade, P. V., 1975, “Torsion Shear Tests on Cohesionless Soil,” Proceedings, 5th Pan-
American Conference on Soil Mechanics and Foundations Engineering. Buenos Aires, Vol.
1. pp- 117-127.
12) Lade, P. V., 1976, “Interpretation of Torsion Shear Tests on Sand,” Proceedings. 2nd
36 International Conference on Numerical Methods in Geomechanics. Blacksburg, Va., pp.
381-389; errata in Journal of the Geotechnical Engineering Division, Proceedings of the
34 ASCE, 1978, Vol. 104, No. GTl.pp. 173-174.
32 13) Lade, P. V. 1981, “Torsion Shear Apparatus for Soil Testing,” Laboratory Shear
...... ,,,j, ■
Testing of Soil. ASTM STP 740, R. N. Young and F. C. Townsend, Eds., ASTM, pp. 145-
~ 30 163.
w
<D 14) Saada, A. S., 1967, “A Stress Controlled Apparatus for Triaxial Testing,” Proceedings,
g 28 Control with PID algorithm
cn Strain Control - axial (.05%/min) Journal of Soil Mechanics and Foundation Division. ASCE, Vol. 93, SM6.
t 26 Stress Control - torsional (PID control) 15) Saada, A. S., 1968, “A Pneumatic Computer for Testing Cross-Anisotropic Materials,”
¿3 24 Materials Research and Standards. Vol. 8, No. 1, pp. 17-23.
a> 16) Saada, A., S., and Baah, A. K., 1967, in Proceedings, Third Pan-American Conference
m 22
on Soil Mechanics and Foundations Engineering. Caracas, Vol. 1, pp. 67-88.
20 17) Saada, A. S., and Zamani, K. K., 1969, in Proceedings, Seventh International
Conference on Soil Mechanics and Foundation Engineering. Mexico, pp. 351-359.
18*1 18) Saada, A. S., and Ou, C. D., 1973, “Strain - Stress Relations and Failure of
16 Anisotropic Clays,” ASCE Journal of Soil Mechanics and Foundations. Vol. 99, No. SMI 2,
pp. 1091-1111.
0 1 2 3 4 5
19) Saada, A. S., and Bianchini, G. B., 1975, “Need to find Title of this one,” Journal of
Axial Strain (%) Soil Mechanics and Foundations. Vol. 101, No. GT11, pp. 1151-1164.
20) Saada, A. S., 1988, “Hollow Cylinder Devices: Their Advantages and Limitations,”
Fig. 6: Variation o f (3 value Throughout the Test
Advanced Triaxial Testing of Soil and Rock. ASTM STP 977, Robert T. Donaghe, Ronald
References C. Chaney, and Marshall L. Silver, Eds., American Society of Testing and Materials,
Philadelphia, 1988, pp. 766-795.
1) Abrantes, A., 1998, “The Effect of Principal Stress Rotations in the Undrained Shear 21) Sayao, A., and Vaid, Y. P., 1991, “A Critical Assessment of Stress Nonuniformities in
Behavior o f Kaolin Clay Using Axial Torsional Testing,” M.S. Thesis. Clarkson University,
Hollow Cylinder Test Specimens,” Soils and Foundations. Vol. 31, No. 1, pp. 60-72.
Potsdam, NY.
2) Berre, T., 1982, “Triaxial Testing at the Norwegian Geotechnical Institute,”
22) Tatsuoka, F., Shoji Hara, K., and Shinji Prdhan, T., 1986, “Failure and Deformation of
Geotechnical Testing Journal, A STM GTJ. Vol. 5, March/June, pp. 3-17.
Sand in Torsional Shear,” Soils and Foundations, Vol. 26, No. 4, pp. 79-97.
3) Bishop and Henkel, 1960, The Measurement of Soil Properties in the Triaxial Test.
 GEOTECHNICAL MEASUREMENTS 17

void ratio for characterizing the flow - slide behavior of sands. Further
investigation of undrained response and liquefaction of saturated soils were
based on monotonic triaxial tests. (Castro et al 1982.1987,1992, Poulos et al.
1985. Ishihara 1993, Zlatovic and Ishihara 1994. Baziar and Dobry 1995.
Verdugo and Ishihara 1994, Chemeau et al 1997, Mitchell and Seed 1996)
The results from these work have generally confirmed that the steady
state line of saturated sand is influenced primarily by the factors such as:
density, drainage condition, level of confining stress and effective stress path
However , other researchers have pointed out that the undrained steady state
EFFECT OF SAMPLE PREPARATION ON STEADY STATE strength is not only dependent on the void ratio of the soil mass, but is also
affected by the effective consolidation stress (Alarcon-Guzman et al 1998).
Studies by De Gregorio (1990) indicated that the dry pluviation technique
provides a lower location of the steady state line in the e - cr . plane than the
Naeini, Seyed Abolhassan 1 moist tamping and moist vibration sample preparation method Experimental
Baziar. Mohammad Hassan,Associate member ASCE2 results pressented by Marcuson el al ,(1990) indicates that the steady state line
is somehow affected by the sample prepartion method They found that the
steady state line obtained from the test specimens, prepared by means of the
ABSTRACT remolded continuously wet pluviated technique was slightly below the steady
state line obtained from that of the specimens prepared by moist tamping. In
Undrained Triaxial compression tests were performed on reconstituted
addition , Dobry and his Co- workers have tested specimens of silty sand
samples of Ardebil sand to evaluate the effect of the specimen preparation
prepared by sedimentation under water This procedure results in a layered
procedure on the position of the steady state line under monotonic loading.
sample that is not uniform Their results have indicated that the steady state
Tests were performed on saturated samples by three distinctive methods of
line is strongly affected by the sample preparation procedure (Vasquez an
sample preparation, namely, “wet tamping”, "water sedimentation”, and
Dobry 1988, Baziar 1991 , Baziar and Dobry 1995 )
undercompaction In all cases, the loosest state of structure are tried to be
obtained The weakest samples were those formed by water sedimentation The work presented here, are the results of undrained - strain
method because of non-uniformity of the sample. The steady state line of controlled monotonic triaxial tests on saturated samples compacted to a density
specimens prepared by water sedimentation method lies under those of wet by different sample - preparation techniques such as:” wet tamping”, “water
tamping and undercompaction methods The test results indicate that the sedimentation”, and “undercompaction” to study the effect of sample
normalized steady state strength of a given material tends to vary to some preparation on the position of steady state line .
extent depending upon the fabric formed by different modes of deposition.
Based on the data presented, it is found that the liquefaction behavior of
TESTING MATERIAL
specimens tested in triaxial equipment may be significantly affected by the
method of sample preparation. Ardebil sand collected from a liquefied zone located in the state of
Ardebil is used throughout this study (Baziar et al 1998). Individual
INTRODUCTION particles are subrounded and the predominant minerals are feldspar and
quartz Maximum and minimum void ratios are obtained 1 16 and 0 625,
Steady state in cohesionless soils represents a condition of deformation
repectively The gradation is uniform with a coefficient of uniformity of
at constant volume, pore pressure, velocity, shear and mean normal effective
187 The particle size distribution curve along with selected index
stress Steady state is an outgrowth of Casagrande's (1936) concept of critical
properties of the samples are shown in Fig (1) and Table 1.

SAMPLE PREPARATION
(l)ColIege of Engineering, Imam Khomeini International University, Qazvin, Iran.
The specimen preparation procedure for triaxial strength testing most
(2)College of Civil Engineering, Iran Uni\. Of Science anil Technology, Tehran, Iran. commonly described in the literature requires the sample to be saturated and
Associate Professor

16
18 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 19

then densified to the required density by some means (l ee and Litton 1969, streched to the inside face of a split mould which is attached to the base
Lee and Seed 1967 , Finn Pickering and Hransby 1971 ) pedestal of the cell
Each portion of the slightly moist sand is strewned to a predetermined
height in six lifts. At each stage of the lifts, tamping is applied lightly with a
U.S. Standard Sieve Number
small flatbottom tamper Because of capillary effects between particles, the
moist sample can be placed at a very loose state well in excess of the maximum
void ratio of the dry sample, which is determined by the procedures stipulated
m the ASTM standard test method

Percent (Jc re -a r by '.'Jtagl-t

In water sedimentation method , a predetermined amount of sample is
dumped in to a mould filled with water to a certain height by using a funnel
with an opening of 0.3 cm in diameter A time of 30 minutes is allowed, until
the material has completely settled The soil was allowed to fall from the
funnel which is kept 1-3 mm above the water level This stepwise deposition
is repeated six times to construct a complete test sample This porcess creates
a layered specimen
For undercompaction , the specimens are constmcted by compacting wet
materials in layers to a selected percentage of the required dry unit weight of
the whole specimen Undercompaction is used since it is generally recognized
that when a typical soil is compacted in layers, the compaction of each
succeeding layer can further densifv the soil below it The undercompaction
method uses this fact to achieve a uniform specimen by compacting the lower
layers to a lower dry unit weight than the desired final unit weight by a
predetermined amount which is defined as "precent of undercompaction
Fig 1- Grain Size Distribution Curve lor Ardebil Sand (Uni)”

Table I- index Data Desired dry unit weight - Dry unit weight of layer used during compaction
Uni_
Soil type Unified soil D50 D10 Cu Cc F.C Gs
classification ydry Desired dry unit weight
(mm) (%) (t/m 3) This value of (Uni) is linearly varied from a maximum at the bottom layer
to zero at the top layer of the specimen If the value of undercompaction has
Ardebil sand SP -S M 0.18 0.075 1.87 1.37 10 2.73 1.25 been selected appropriately, a uniform dry unit weight (void ratio) throughout
the specimen will be achieved (Ladd 1978) Therefore, the key point in this
method of specimen preparation is the selection of the of undercompaction
All of the specimens in this study are prepared by three distinctive value for the bottom layer
methods of sample preparation , namely , “wet tamping”, “water All lest specimens prepared by undercompaction method had an initial
sedimentation”,and “undercompaction”. In all cases, the loosest state of water content of 9% and were compacted with 4% undercompaction (Uni =
structure are tried to be obtained The basic requirements for the two 4%) value in six layers in a split compaction mould attached to the triaxial cell
methods are, firstly to obtain homogeneous sample with uniform distribution (see Ladd 1978) All samples were 5 08 mm (2in) in diameter and 101.6 mm
of void ratio, and secondly to be able to prepare samples of the lowest possible (4in) height
dry density (highest porosity)
In wet tamping method , six equal preweighed portions of sample are
mixed with de- aired water at a water content of 9% . The membrane is
20 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS

TESTING TECHNIQUE
Once a sample was formed by any method of sample preparation, a
vacuum of 35 Kpa was applied to produce confining pressure Then the forming
mold was dismantled The specimen height and diameter were measured at
several heights on the specimen to compute its void ratio The triaxial cell was
then assembled and filled with deaired water Carbon dioxide was run into the
specimen for 45 minutes and then the sample was saturated with deaired water.
The cell and back pressures were then continuously increased, while
maintaining a difference of 35 Kpa, until the desired back pressure was
achieved .
A B-value greater than 0 97 was typically obtained at a back pressure of
200Kpa The volume change was recorded during isotropic consolidation
from 100 to 485 Kpa so that the void ratio and relative density at the end of

T ABLE 2. Summary of The Tests Performed on Ardebil Sand

consolidation could be calculated When the specimen was hilly consolidated
(which required approximately 60 minutes), the drainage valves were closed,
and the sample was subjected to monotonic loading. The axial deformation
and pore pressure in the specimen were recorded during loading. Tests were
conducted with strain control at a rale of 1% /minutes.
TEST PROGRAM
To investigate the effects of sample preparation on steady state line of a
silty sand, a total of nine isotropically consolidated, undrained monotonic
compression triaxial tests (CIUC) were performed using Ardebil sand , as
listed in Table 2.
Specimens were prepared by using each of the three methods, as
described earlier, and the steady state line (SSL) was obtained In a CIUC test,

CIUC: Consolidated Isotropically Undraincd Compression

a continuous increase of pore pressure when the axial stress is increased,
indicates that the specimen is contractive, while a decrease in pore pressure
build up indicates that the specimen is dilative.In all CIUC'tests performed
here, the specimens were contractive

e<\ is the void ratio before the consolidation,

i, is the void ratio after the consolidation
In all tests, the consolidation ° pressure varied from 250
to 660 Kpa. It is noted that during consolidation, the nine specimens
experienced significant reductions in their overall void ratio As listed for Test
9 in Table 2, the largest decrease was from 111 to 0.905.
TEST RESULTS
The results of the liquefaction tests performed on the .Ardebil Sand, using
three different sample preparation techniques are presented in Table 2
Figs. 2 to 4 illustrate the deviatoric stress and induced pore pressure in
terms of the axial strain, and also the effective stress paths for different
methods of sample preparations
Fig 2 shows the stress-strain behavior for the Ardebil sand for specimens
prepared by watre sedimentation In these tests.the specimens were isotropically
22 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 23

Devatoric Stress (Kpa)

Deviatorlc Straaa (Kpa)

700

Excess Pore Water

600

Pressure(Kpa)
500
Excass Pore Water

400
300
200
100
0
0 2 4 6 8 10 12
Axial Strain (%)

P ' (Kpa)

F.C.=10 % ,Effective Street:»)- 400 , b)-2P0, c) 150 (Kpa) F.C.=10% .Effective Stress :a)-485, b)-275, c ) - 100 (Kpa)

Fig 2 - Monotonic Triaxial Test on Ardebil Sand Fig 3 - Monotonie Triaxial Test on Ardchil Sand

( Water Sedimentation Method ) (Undrecompaclion Method )

24 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 25

consolidated to 300-600 Kpa As seen in Figure 2 Test (b), the shear stress q.

Devatoric Stress (Kpa)

as well as the pore pressure, increased up to a strain of about l%. At this axial
strain, the specimen failed with peak strength of about 196 Kpa At an axial
strain of about 3%, the deviatoric stress and pore pressure remained constant
up to an axial strain of 15%. Both curves of stress and pore pressure in Figure 2
show that the soil were contractive during the test
Fig 3 shows the results of C1UC test for undercompaction method In
these tests the specimen were isotropically consolidated to 250-660 Kpa As
seen in Figure 3 test (b), the shear stress and the pore pressure increased up to a
strain of about l%. At this axial strain, the specimen failed with peak strength
of about !20Kpa, and at 6 to 11% axial strain exhibited almost the same
strength and pore pressure.
In Fig 4 test (b), at strain of 1% , the specimen failed with maximum
strength of about 125Kpa, and at 7% strain exhibited almost the same strength
Excess Pore Water

and pore pressure up to an axial strain of 11% In these tests, the specimens
Pressure(Kpa)

were isotropically consolidated to 295-550 Kpa

The results of nine C1UC tests on Ardebil sand showed that they were
contractive That is, the peak strength and steady state strength are distinctively
different The development of large strains at steady state for water sedimented
specimens was about 4% to 7%, w hile for the same kind of tests .this value is
6% to 7% for undercompaction and wet tamping methods respectively.The
observed differences between large strains in water sedimentation and two
other methods is due to the non-uniform packing of the particles in each layer
when water sedimentation is used
The values of a us and <j>us for three methods of sample preparation are
listed in Table 2. These values for undercompaction and wet tamping methods
are almost the same ( a us = 25.64° and 25.17" respectively), while for water
sedimentation is 27.35°. Undercompaction and wet tamping methods produces
almost an isotropic and homegeneous fabric. Whereas, the method of water
sedimentation roughly duplicate the natural or artificial sedimentation
processes through water, and the fabric produced is anisotropic and layered
This method likely represents to some extent hydraulic fill process in the field
To compare the SSL’s of this three sets of tests values of steady state
strength versus the initial confining pressure for different method of sample
preparation are plotted in Fig 5. The residual strength increased from
undercompaction to wet tamping and from that to water sedimentation The
test data arranged in Fig.5 indicates that the normalized steady state strength of
a given material tends to vary to some extent depending upon the fabric formed
by different modes of deposition (see Baziar and Dobry (1991) , lshihara
(1993) and zlatovic (1994)). This is considered as a manifestation of the effect
F.C.=10% .Effective Stress :a)-385, b)-280, c)- 245 (Kpa) of the fabric created by the different methods of deposition in preparing the
test specimens.
Fig 4 - Monotonic Triaxial Test on Ardebil Sand
(Wet Tamping Method )
26 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 27

Fig 6 Shows the relationship between the void ratio and the log of the
effective mean stress, as the samples deformed at large strains. The steady state
line of samples prepared by water sedimentation method lies under that of wet
tamping and undercompaction methods. In another words, for a given effective
mean stress, the sample prepared by water sedimentation is more liquefiable
than the sample prepared by two other methods.

CONCLUSIONS

From the study described in the preceeding pages, the effects of the method of
sample preparation on the liquefaction characteristics of saturated Ardebil sand
under strain-controlled monotonic triaxial test conditions my be summarized as
follows:

1) The liquefaction behavior of reonstituted saturated specimens tested in

triaxial equipment will be significantly affected by the method of specimen
preparation

2) Because of the same grain size distribution of the samples in this study ,
the slopes of the steady state lines formed by different techniques were all
about the same , which is consistent with previous data (castro et al 1982 ,
poulos 1981 , 1985) , and is an important point that counts in practical
Fig 5 - Variation Of Undrained Steady State Slrengtlul ( Sus )Wilh
selection of steady state strength
Effective Vertical Consolidation! Pressure For Ardebil Sand

3) The undercompaction mothod is able to make looser sample than the other
two method of sample preparation

4) The steady state line of samples prepared by water sedimentation method

lies under that of wet tamping and undercompaction methods .

5) The steady state strength for wet tamping and undercompaction were
roughly the same , but tends to vary for water sedimentation

Fig 6 - Comparision of Steady State Line For Three Methods :

a- Undercompaction
b- Wet Tamping
c- Water Sedimentation
28 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 29

REFERENCES 13) De Gregorio , V.B (1990): “ Loading Systems, Sample Preparation, and
Liquefaction, “Journal of Geotechnical Engineering, Vol 1 16, No.5, PP
1) Alarcon - Guzman. A..Leonards, G. A and Chameau J.I (1988)
805 - 821
“Underained Monotonic and Cyclic Strength of Sands," Journal of
Geotechnical Engineering, ASCE, Vol 114, No. 10, PP 1089 - 1109 14) Dobry, R (1991) “ Soil Properties and Earthqueke Ground Response, ",
Guest Lecture, Proceedings Tenth European Conference on Soil Mechanics
2) Baziar, M l! (1987) “Influence of Testing Technique on The Steady and Foundation Engineering, Florence. Italy
State Line of a Sand,” Master Thesis, Department of Civil Engineering,
Rensselaer Polytechnic Institute, Troy N Y 15) Finn ,W D L .Pickering, D J and Bransby, P L ,(1971 ) : “ Sand
Liquefaction in Triaxial and Simple Shear Tests , “ Journal of The Soil
3) Baziar, M.H. (1991): “Engineering Evaluation of Permanent round Mechanics and Foundation ASCE, Vol 97, No. SM4, Proc .Paper, 8039,
Deformations Due to Seismically - Induced Liquefaction,” Doctoral Thesis, PP 639 - 659
Rensselaer Polytechnic Institute. I S A
16) lshihara, K. (1993) “Liquefaction and Flow Failure During
4) Baziar, M H and Dobry, R (1991) "Liquefaction Ground Deformation Earthquakes”, The 33 rd Rankine Lecture, Geotechnique, Vol 43, No.3,
Predicted From Laboratory Tests,” Proc. Second International Conference PP 351 -415
on Recent Advances in Geotechnical Earthquake Engineering and Soil 17) Ladd, RS (1978): “Preparing Test Specimens Using
Dynamics Undercompaction”, Geotechnical Testing Journal, GTJODJ, V o l, No 1,
5) Baziar, M.H., and Dobry R. (1995) “Residual Strength and Large - PP 16-23
Deformation Potential of Loose Silty Sands , “J G.T. ASCE, Vol 121 , 18) Lee, K L , and Seed, FI B., (1967): “ Dynamic Strength of
No. 12, PP 896- 906 Anisotropically Con Solidated Sand”, Journal of The Soil Mechanics and
6) Baziar, M H , Razeghi, H R., and Neaini, A H. (1997): “Determination Foundations Division, ASCE, Vol 93, No SM5. PP 169- 190
of Liquefaction Potential by Monotonic Triaxial Test” Research Report, 19) L.ee. K.L , and Fitton, J.A.. (1969) ” Factors Affecting The Dynamic
Iran University oF Science and Technology Strength of Soil", in Vibration Effects on Soils and Foundation, ASTM, Stp
450, Philadelphia, PP 71-95
7) Baziar, M H , Naeini, A H, and Ziaie Moayed, R (1998) “Geotechnical
Damaje During the 1997 Ardebil Earthquake in lran“Proceedings of The 20) Marcuson, III, W.F., Hynes, M E, and Franklin, A G (1990) “
First IRAN - JAPAN Workshop on Recent Earthquakes in Iran and Japan , Evaluation and Use of Residual Strenght in Seismic Safety of
PP . 137-146 Embankments”, Earthquake Spectra, 6. PP 529-572

8) Casagrande, A. (1963) "Characteristics of Cohesionless Solis Affecting 21) Poulos. S J , Castro, G., France, J W., (1985): “Liquefaction Evaluation
The Stability of The Slopes and Erath Fills “Journal of The Boston Society Procedure”, Journal of Geotechnical Engineering, Vol 111, No. 6, PP 772
of Civil Engineering. -792

9) Castor ,G.(1969) : ” Liguefaction and Cyclic Deformation of Sands , A 22) Verdugo, R and lshihara, K (1996): “The Steady State of Sandy Soils”,
Critical Review “ Harvard Soil Mechanics Series ,No 88 Japanese Geotechnical Society, Vol 36, No.2 , PP 8 1 -9 1
23) Vasques - Herrera, A and Dobry, R (1988) : ” The Behavior of
10) Castro, G. Poulos, S.J (1977): “Factors Affecting Liquefaction and
Undrained Contractive Sand and Its Effect on Seismic Liquefaction Flow
Cyclic Mobility”, Journal Of Geotechnical Engineering Division Proc
Failure of Earth Structures, “ Department of Civil Engineering, Rensselaer
ASCE, Vol. 13, No GT6, PP.501 - 516
Polytechnic Institute, USA.
11) Castro, G , Poulos, S.J. France, J.W , and Enos, J L. (1982): 24) Zlatovic, S and lshihara, K, (1994) “The Loosest State of Silty Soils”,
“Liquefaction Induced by Cyclic Loading", Geotechnical Engineering Inc Normalization, Proc. 49 Th Japanese Conference On Civil Engineering,
Report Submited To National Science Foundation Sapporo. 3, PP 332 - 333
12) Castro, G . Keller, T O and Seed, H B (1992): “Steady State Strength
Analysis Of Lower San Fernando Dam Slide’, Journal Of Geotechnical
Engineering, ASCE, Vol 18 , NO 3, pp 406 - 427
 GEOTECHNICAL MEASUREMENTS 31

Introduction
The mechanical behavior o f granular materials is highly dependent on the
arrangement o f particles, particle groups and associated pore space. These
geometric properties comprise the so-called material fabric. In the literature, fabric
analysis techniques are mainly classified as destructive (e.g., specimen stabilization
and thin-sectioning), and nondestructive techniques (e.g., magnetic resonance
imaging, ultrasonic testing, x-ray radiography, and x-ray computed tomography) X-
ray radiography has been used to trace density change of soil samples (e.g., Roscoe
1970, Vardoulakis and G raf 1985, Vardoulakis et al. 1985). However, it suffers
Q uantifying Void Ratio Variation in Sand Using Com puted Tomography from the limitation o f not providing a three-dimensional (3-D) radiograph.
Consequently, most o f x-ray radiograph techniques have been performed on
specimens tested under plane strain conditions. High resolution 3-D density images
Khalid A. Alshibli', Member, ASCE, Susan N. Batiste12, Roy A. Swanson3, Stein
can be obtained by computed tomography. The first applications of tomography
Sture4, Member, ASCE, Nicholas C. Costes5, Fellow, ASCE, and Mark R. Lankton6
were in radio astronomy (Bracewell 1956) and today it is widely used in the areas of
medical imaging, seismology, petroleum engineering, acoustic imaging, soil science,
and powder industry (e.g., Coshell et al. 1994, Amos et al. 1996, Zeng et al. 1996,
A bstract: A series o f displacem ent-controlled, conventional, drained Denison and Carlson 1997, Denison et al. 1997, Phillips and Lannutti 1997).
axisymmetric (triaxial) experiments were conducted on dry Ottawa sand specimens
at very low effective confining stresses in a microgravity environment aboard the Computed Tomography (CT) is a technique in which an incident beam of
Space Shuttle during the NASA STS-89 mission. Post-flight analysis included electromagnetic radiation (x-rays or gamma rays) passes through an object and is
studying the internal fabric and failure patterns of these specimens using Computed collected with an array of detectors; the object is rotated such that the beam probes
Tomography (CT). The CT scans of three specimens subjected to different from several angles to collect attenuation data and produce the equivalent o f a
compression levels (uncompressed specimen, a specimen compressed to 3.3% “slice” through the region of interest. The attenuation information collected during
nominal axial strain (ea), and a specimen compressed to 25% ea) are presented to the rotation is then used by a computer program to reconstruct a slice image o f the
investigate the evolution of instability patterns and to quantify void ratio variation. internal structure. It is a powerful nondestructive probing technique for mapping the
The progress of failure is described and discussed. Also, specimens’ densities were internal structure of objects. High energy CT systems can penetrate large objects
calibrated using standard ASTM procedures and void ratio spatial variation was built from a wide variety of materials. The CT images provide more clarity than
calculated and represented by contour maps and histograms. The CT technique conventional radiographs such as x-ray images. Three-dimensional reconstructions
demonstrated good ability to detect specimen inhomogeneities, localization patterns, can be made by stacking the CT slices and attenuation is measured on individual
and quantifying void ratio variation within sand specimens. planes whereas conventional radiographs comprise attenuation data from all planes
within an object superimposed onto one plane.

In this paper, the results of CT analysis performed on sand specimens tested

under triaxial conditions at very low effective confining pressures in a microgravity
1 Assistant Research Professor, Dept, of Civil & Environmental Engineering, Technology Hall, S201, environment will be presented. This includes explaining the density calibration
University of Alabama in Huntsville, Huntsville, AL 35899, Tel. (256) 544-3051, Email: techniques, showing and describing the failure patterns, and quantifying void ratio
[email protected]. variations within the specimens.
2 Graduate student, Dept, of Civil, Environmental, & Architectural Engineering, Campus Box 428.
Univ. of Colorado at Boulder, Boulder, CO 80309-0428.
3 Research Associate, Dept, of Civil, Environmental, & Architectural Engineering, Campus Box 428,
Univ. of Colorado at Boulder, Boulder, CO 80309-0428.
Materials and Methods
4 Professor,, Dept, of Civil, Environmental, & Architectural Engineering, Campus Box 428, Univ. of A series o f displacement-controlled quasi-static and cyclic, conventional
Colorado at Boulder, Boulder, CO 80309-0428. triaxial compression experiments were performed in the SPACEHAB module o f the
5 Adjunct Professor, Dept, of Civil, Environmental, & Architectural Engineering, University of Shuttle Orbiter during the STS-89 mission to the Russian space station Mir in
Colorado at Boulder, 4216 Huntington Rd SE, Huntsville, AL 35802.
January, 1998. The experiments were conducted on cylindrical dry specimens (75
6 Program Manager, Laboratory for Atmospheric and Space Physics, Campus Box 590, Univ. of
Colorado at Boulder, Boulder, CO 80309-0590.

30
32 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 33

nun in diameter and 150 mm long) at effective confining pressures of 0.05, 0.52 and CT Scanning
1.30 kPa. Displacements were controlled through stiff, highly polished tungsten
carbide end-platens which were intended to provide very low friction (the measured In principle, CT consists of measuring the attenuation of an electromagnetic
radiation beam for a number of different paths through the tested object and
angle o f friction with F-75 Ottawa sand is in the range of 4.5°) at the end-platen to
mathematically extracting the density of each point in a plane of the object from the
specimen interface. A constant confining water pressure was transmitted through a
attenuation data recorded by an array of detectors. This 2-D density map constitutes
thin, highly flexible latex rubber membrane surrounding the cylindrical surface
between the end-platens. Detailed recording of data was made in all experiments, a CT slice which gives very accurate information within the plane. By creating
several slices, a true 3-D volumetric image of the object can be built up. More
including axial load, axial displacement, confining pressure, bulk volumetric
information about CT can be found in Bossi et al. (1990).
changes, ambient temperature, and environmental acceleration levels. Video
cameras and a recorder provided 360° coverage of the specimens and monitored
The CT scans reported in this study were performed at the NASA/ Kennedy
overall behavior to track the onset of formation of shear bands or other forms of
instability. Space Center (KSC), Industrial Computed Tomography System (ICTS) using CITA-
201 scanner, manufactured by Scientific Measurement Systems, Inc. A Cobalt-60
gamma ray source was used, with an energy level of approximately 1.25 MeV and
Ottawa sand was used to prepare the specimens. It is a fine-grained, uniform, source spot size of 2.5 mm. The source aperture was set as 36 degree by 4 mm high.
sub-rounded to rounded silica (quartz) sand (commonly denoted as F-75 banding The 125 detector apertures were set at 2 mm wide by 4 mm high, which samples
sand) with mean particle size of 0.22 mm, maximum porosity of 0.446, minimum from approximately a 2-mm thick CT data region. The data object acquisition circle
porosity of 0.327, and specific density of sand particles of 2.65. The specimens was 600 mm at 1581.85 mm from source. Ray spacing data acquisition was every
were prepared in a terrestrial laboratory by air pluviation (raining) of the sand into a 0.5 mm. Second generation data acquisition was used, with time per datum of 0.2
mold at a certain intensity and velocity controlled, respectively, by the opening of seconds, thus taking approximately 1 hour to acquire data per complete slice. Slices
the funnel from which the sand is poured and the distance between the funnel and were computed at 1 mm intervals over the height of each specimen. For further
the mold. This ensured uniform density at 65 ± 1% relative density. A thin latex information on the tomography system the reader is referred to Engel (1998).
membrane (0.30 nun-thick) was used to encase and isolate sand from the
surrounding water medium. The latex membrane was placed along the inside of a 75
mm in diameter by 150 mm long split mold and held aligned by vacuum pressure. Calibration of Computed Tomography Data
The mold was then attached to the bottom pedestal of the triaxial test cell, the
Aside from qualitative observation, the CT data is intended to be used for two
membrane stretched over the bottom end-platen and sealed using two o-rings The
purposes: measuring internal structure geometry and density (or void ratio). The
sand was rained into the mold, then the top end platen was attached by stretching the
data has been calibrated for both, and measures of accuracy are presented below.
membrane over the end platen and sealing it using two o-rings. The vacuum was
then removed from the membrane-mold interface and applied to the inside of the
specimen to prevent its collapse as the mold was split and removed. The external A length scaling calibration o f 0.387 mm/pixel was found through
triaxial test cell was then assembled around the specimen, filled with deionized examination of known features in CT data, namely hardware geometry of the triaxial
water, and pressurized to 103.5 kPa. The internal vacuum was then removed and the test cell. In addition spatial resolution was determined using ASTM standards
specimen pore space vented to atmosphere. The specimens were then launched on a (ASTM-E1695, 1995). The interface between an aluminum rod and air was
Space Shuttle Orbiter to achieve a microgravity environment. While in orbit, the examined and it was determined that data shows 50% modulation at 0.275 line pair
confining pressures were lowered to the desired test pressures and specimens were per millimeter (1pm) and 10% at 0.53 1pm. In other words, objects separated by 1.8
compressed to the prescribed axial strain level. The confining pressures were then mm can be identified with 50% modulation and objects 0.94 mm apart can be
restored to 103.5 kPa for Orbiter re-entry and landing. Specimens were identified with 10% modulation. This affects data interpretation both in geometry
subsequently retrieved and subjected to an extensive post-flight analysis including and void ratio measurements. One cannot resolve below 3 pixels (0.94 nun/0.387
computed tomography, profile measurements, epoxy stabilization, and thin- mm/pixel = 2.4 pixel) at 10% modulation. Considering the average grain size of the
sectioning. More details about the experiments hardware and results can be found in material is 0.22 mm, one cannot resolve below 5.3 grain diameters at 10%
Alshibli el al. (1996) and Sture el al. (1998). modulation, or 8 grain size diameters at 50% modulation.

Calibration o f density was also performed according to ASTM standards

(ASTM-E1935, 1997). Water and Lexan regions included in scans of the specimens
were used as standards. An aluminum standard was also scanned by ICTS and the
34 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 35

data were used in the calibration. Figure 1 shows the relationship between the CT another cannot be directly related. A discussion concerning the quantitative void
Numbers (CTN) o f standard samples and their individual attenuation coefficients ratio relation between specimens is presented later in this paper.
(Ca); which were obtained from the National Institute of Standards and Technology
(NIST) web site (Hubbell and Seltzer, 1996). Coefficient o f attenuation for Silica The CT images of three specimens compressed to different axial strain levels
(S i02) is related to density through the mass coefficient of attenuation, which equals will be presented. They are; U-specimen (Uncompressed specimen), C3.3 specimen
0.05687 cm2/g for dry silica sand regardless of specimen density. To verify the (a specimen compressed to 3.3% axial strain (ea) under 1.30 kPa confining pressure
calibration, the average CTN in a specimen was compared to its bulk density for 11 in a microgravity environment), and C25 specimen (a specimen compressed to 25%
different specimens. The average CTN was measured using all CTN from each axial strain under 1.30 kPa confining pressure in a microgravity environment).
plane over the entire height and diameter of a specimen. The bulk density was Table 1 indicates beginning and end densities of the three specimens. Figure 2
calculated from initial mass and volume measurements, and volume change during illustrates location of internal specimen views in Figures 3 through 5.
testing. For dry specimens, void ratio = (sand particles density/specimen bulk
density) - 1. The void ratio standard deviation was found to be less than 0.017.
TABLE 1. Void Ratios of specimens before and after Compression

Specimen U C3.3 C25

Initial Void Ratio 0.598 0.594

0.606
Final Void Ratio 0.789 0.662

CTN
Figure 1. Relationship between CT Numbers and Coefficients of Attenuation
Figure 2. Generalized Locations of 3-D Volume Rendering and Two Orthogonal
Axial Sections
Results and Analysis
Specimen Deformation Patterns
For the C25 specimen (Figure 3), a large number of radial discrete shear bands
Three-dimensional reconstructions of the specimens using CT data have been
of lower density (darker areas), similar in shape to turbine blades mounted on a
prepared using the image processing software IDL (Interactive Data Language).
central hub, extend outward from the boundary o f the central shear cones. When
Sides of volume renderings can be “cut away” to view the internal structure. Also,
examining cross-sections parallel to the axis of compression, extensive areas o f
axial planes through the center of the specimen can be generated and displayed. The
generally uniform density are seen outside o f shear zones. The rounded shear cones
pixel intensity of the images is linearly related to the density o f the material.
are seen extending at large angles from each end, although the cone on the stationary
However, it is cautioned that data sets for each specimen have been scaled
(bottom) end platen is more clearly developed. The cones are clearly defined by the
individually for optimal viewing. As a result, the color scales from one specimen to
36 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 37

specimen-end-platen interface friction and the restraint posed by the stretched latex shear cone, visible in both vertical planes, is at a very low angle. In contrast, the
membrane in the external contact region. Outside of the cones, several inclined lines shear cones seen in the C25 specimen are steep. This indicates that at the beginning
of low density are identified as the radial shear bands. The shear bands were formed of an experiment low end-friction is present, and probably increases as the specimen
earlier in compression, probably around 8-10% axial strain. Further deformation of bulges. The lack of defined shear bands in the C3.3 specimen indicates that the
the specimen (as the specimen was compressed to 25% axial strain) was numerous shear bands visible in the C25 specimen develop at axial compression
accompanied by substantial deformations, including bulging, as seen by the profile greater than 3.3%.
of the axial sections in Figure 3. As a result, there is some distortion that is seen the
CT scans: that is, when the specimen containing shear bands and a sharp shear cone
is compressed and deformation occurs, these features will distort somewhat, leaving
initially straight features such as shear bands or cones rotated or curved.

shear band, shear cone shear band

radial view

Figure 5. 3-D Volume Rendering and Two Orthogonal Axial Sections

Figure 3. 3-D Volume Rendering and Two Orthogonal Axial Sections of the U-Specimen
of the C25-Specimen
Vertical areas o f slightly different color (density) are present in the
uncompressed specimen (Figure 5), which is probably a result o f the specimen
preparation (air pluviation) technique and the CT noise. Overall, the specimen is
nearly homogenous and no distinct shear pattern is visible. This indicates that the
specimens were well prepared, and that shear bands are not CT artifacts, but real
physical features.

Void Ratio Variation

W hile 3-D volumes and axial sections are useful for qualitatively
understanding the internal structure of a triaxial specimen, other methods arc used to
examine the features quantitatively. The relationship between the CTN and
specimens’ densities, described earlier, was used to obtain quantitative observations.
The data are displayed as contour maps (Figures 6 through 8) and void ratio
Figure 4. 3-D Volume Rendering and Two Orthogonal Axial Sections histograms (Figure 9). The contour maps were produced from a grid of average void
of the C3.3-Specimen ratios. The void ratios were averaged over small annuli within the specimen,
centered around the central long axis. The annuli were at evenly spaced radii and
heights throughout the specimen: This method reports the average void ratio in a
The majority of the C3.3 specimen (Figure 4) shows a slight amount of circumferential region, appropriate to axisymmetric shearing structures. Two
bulging, but no visible internal shear bands. The specimen is much more features o f the contour maps should be noted. First, the shear bands illustrated
homogenous than the C25 specimen although small shear cones are visible. The earlier (Figure 2) are not seen in contour mapping, as both shear bands and areas
outside o f shear bands were combined in each average void ratio calculation.
38 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 39

Second, the apparent increase in void ratio seen at the outer edge of the specimen is ter-circle areas of low void ratio at both ends of the specimen are indicative of
not a physical property of the specimen, but is related to the edge effects of the CT f Uy developed shear cones. On the border of these shear cones is a sharp transition
scan at the specimen-water interface. to higher void ratios. In the area outside of the shear cones, shear banding is present,
and indicated by the large void ratios throughout the remainder of the specimen. At
the smaller radii, the void ratio is high where several shear bands cross and the shear
cones merge. Images of this area (Figure 3) does not reveal individual shear bands,
but a mass of low density material.

Radius (mm)
Figure 6. Void Ratio Distribution of U-Specimen Represented by Contour Map
' ' ' I ...............
0 10 20 30 40 50
In comparing Figures 6 through 8, one can see a progressive internal structure Radius (mm)
forming. For the uncompressed specimen (Figure 6) there is no evidence of definite Figure 7. Void Ratio Distribution of C3.3 Specimen Represented by Contour Map
internal structure (pattern). The void ratio in the center of the specimen is slightly
larger than toward the outside, which may be a result o f the raining technique.
Figure 7 (the C3.3 specimen) shows signs of internal structure, primarily near the Histograms o f the entire data set are produced and normalized for an area o f
end platens where lower void ratio triangular areas (conical in 3-dimensions) are unity under the curve. This does not average out the shear bands and the areas
seen. Also, a higher void ratio area appears to be forming in the central regions of outside the shear bands, as is the case in the contour plots
the specimen. Shear bands are fully developed in the C25 specimen (Figure 8). The
40 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 41

V oid Ratio

Figure 9. Void Ratio Distribution Histograms

Conclusions

The CT technique dem onstrated the ability to detect specim ens’

inhomogeneities, localization patterns, and quantify void ratio variation within sand
specimens. Specimen deformations localize into two major shear cones, affected by
the specimen lestraints at the interface with the two end platens, at early
compression stages (less than 3.3% axial strain). Then, multiple symmetrical radial
shear bands develop at larger axial strain which extend from the conical surfaces in
the radial direction towards the specimen perimeter. Also, there is nearly no spatial
variation in void ratio values for the undeformed specimen and a slight change in
As seen from Figure 9, very little variation in the void ratio distribution occurs void ratio value is detected in the 3.3% axial-strain specimen, whereas a sharp
between ea= 0 and 3.3%. Although shear cones are beginning to form at the very increase in void ratio is observed in the specimen compressed to 25% axial strain
ends of the ea=3.3% specimen (Figure 4), the void ratio in the central portion is especially across the shear bands.
almost unaffected. When the C25 specimen is compressed to 25% axial strain,
however, large changes in the void ratio value and distribution are obseived. lhe It is apparent that conventional triaxial compression specimens remain
curve o f the void ratio distribution widens and shifts to the right, indicating an relatively well into the deformation process. This indicates that the conventional
increase in void ratio. This is due to the shear banding that takes place duiing triaxial experiment is a very appropriate experiment to obtain constitutive properties
compression, which results in areas of relatively high void ratio. for soils, as opposed to an experiment such as plain strain. Based on these findings
and earlier observations and analysis of plane strain experiments (e.g., Lee 1970,
Marachi el al. 1981, Han and Vardoulakis 1991, Han and Drescher 1993, Peric' el al.
1992) it is quite clear that the deformation processes and stability behavior are quite
different for axisymmetry and plane strain. This also raises questions about our
42 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 43

understanding of soil behavior under these two basic but very different modes of Hubbell, J. H., and Seltzer, S. M. (1996). “Tables of X-Ray Mass Attenuation
loading. Coefficients and Mass Energy-Absorption Coefficients”, N IST web Site
(http://physics.nist.gov/PhysRefData/XrayMassCoef/cover.html).
Acknowledgments Lee, K. (1970). “Comparison of Plane Strain and Triaxial Tests on Sand,” Journal o f
Soil Mechanics and Foundations division, ASCE, 96 (3), 901-923.
The authors gratefully acknowledge the financial support provided by NASA/
Marachi, N., Duncan, J., Chan, C., and Seed, H. (1981). “Plane-Strain Testing of
George Marshall Space Flight Center under contract No. NAS8-38779. Thanks are
Sand,” Laboratory Shear Strength o f Soils, ASTM, STP 740, Yong and
also due to Peter Angels of NASA/ Kennedy Space Center for help in performing
Townsend, (editors), 294-302.
the CT scans.
Peric', D., Runesson, K., and Sture, S. (1992). “Evaluation of Plastic Bifurcation for
References Plane Strain versus Axisymmetry,” Journal o f Engineering Mechanics, ASCE,
118 (3), 512-524.
Alshibli, K., Costes, N., and Porter, R. (1996). “Mechanics of Granular Materials,” Phillips, D. IT, and Lannutti, J. J. (1997). “Measuring Physical Density with X-ray
Space Processing of Materials, International Symposium on Optical Science, Computed Tomography,” N D T & E International, 30 (6), 339-350.
Engineering, and Instrumentation, SPIE-The International Society fo r Optical
Engineering, Denver, CO, Vol. 2809, Aug 1996, 303-310. Roscoe, K. (1970). “The Influence of Strains in Soil Mechanics,” Geotechnique, 20
(2), 129-170.
Amos, C. L., Sutherland, T. F., Radzijewski, B., and Doucette, M. (1996). “A Rapid
Technique to Determine Bulk Density of Fine-grained Sediments by X-ray Sture, S., Costes, N., Batiste, S., Lankton, M., Alshibli, K., Jeremic, B, Swanson, R.,
and Frank, M. (1998). "Mechanics of Granular Materials at Low Effective
Computed Tomography,” Journal o f Sedimentary Research, 66 (5), 1023-1039.
Stresses, " ASCE, Journal o f Aerospace Engineering, 11 (3), 67-72.
Bossi, R. H., Friddel, K. D., and Lowrey, A. R. (1990). “Computed Tomography,”
Vardoulakis, I. and Graf, B. (1985). “Calibration of Constitutive Models for
N on-D estructive Testing o f Fibre-R einforced Plastic Com posites, J.
Granular Materials Using Data from Biaxial Experiments,” Geotechnique, 35(3),
Summerscales (Editor), Elsevier Applied Science, NY, 2, 201-252.
299-317.
Bracewell, R. N. (1956). “Strip Integration in Radio Astronomy,” Australian Journal
Vardoulakis, 1., Graf, B., and Hettler, A. (1985). “Shear Band Formation in a Fine
o f Physics, 9, 198-217.
Sand,” Proceedings o f the 5th International Conference on Numerical Methods in
Coshell, L., Mclver, R. G., and Chang, R. (1994). “X-Ray Computed Tomography Geomechanics, Nagoya, 517-522.
o f Australian Oil Shales: Non-D estructive V isualization and Density
Zeng, Y., Gantzer, C. J., Payton, R. L., and Anderson, S. H. (1996) “Fractal
Determination,” Fuel, 73 (8), 1317-1321.
Dimension and Lacunarity of Bulk Density Determined with X-ray Computed
Denison, C., Carlson, W. D., and Ketcham, R. A. (1997). “Three-dimensional Tomography,” Soil Science Society o f America Journal, 60, 1718-1724.
Quantitative Texture Analysis of Metamorphic Rocks Using Fligh-Resolution
Computed X-ray Tomography: Part I. Methods and Techniques,” Journal o f
Metamorphic Geology’, 15(1), 29-44.
Denison, C., and Carlson, W. D. (1997). “Three-dimensional Quantitative Texture
Analysis o f Metamorphic Rocks Using High-Resolution Computed X-ray
Tomography: Part II. Application to Natural Materials, ” Journal o f Metamorphic
Geology, 15 (1), 45-57.
Engel, P. (1998). "NASA/KSC Industrial Computed Tomography (CT) System",
Unpublished Report.
Han, C. and Vardoulakis, I. (1991). “Plane Strain Compression Experiments on
Water-Saturated Fine-Grained Sand,” Geotechnique, 41 (1), 49-78.
Han, C. and Drescher, A. (1993). “Shear Bands in Biaxial Tests on Dry Coarse
Sand,” Soils and Foundations, JSSMFE, 33 (1), 118-132.
 GEOTECHNICAL MEASUREMENTS 45

experimentally by plate bearing tests. Such testing has been discussed by Hansen
(1961), Levy and Morton (1974), Anderson and Slenhamar (1982).

Load-Deformation Performance of Peal Soil under Large Concrete Plates

Mark Vakher1

Photograph 1. Constructing concrete plate road

Abstract
on peat soil.
This paper presents the results of a series of investigations performed with
Field Experiments
peat soils to determine the modulus of deformation, E, and the modulus o f subgrade
reaction, K, using an elastic foundation model. The experiments were performed
In the experiments, the top of the concrete plate (3tn by lm by 0.15m) was
both in situ and under laboratory conditions using concrete plates placed over peat.
covered with a rubber layer. A 328mm diameter steel bearing plate w-as placed on
Experiments were earned out under both static and dynamic loads. Based on the
top of the mbber layer on the center of the concrete plate. A monotonic static load
results of the investigation, empirical equations for the modulus o f subgrade reaction
was applied to the circular plate in stages by truck crane at a moderately rapid rate in
for sustained and resilient deformation are proposed and related to peat soil
nearly uniform increments: llk N , 36kN, 48,5kN, 73.5kN. The maximum load
properties.
corresponded to a maximum pressure p = 0.025 MPa. Under this small loading,
there w'as no primary consolidation. After each loading increment, settlements were
measured at 0, 2, 8, 15, 30 min., and the last reading was taken at 1 hour. At the end
Introduction
of 1 hour, base of peat soil had conditional stabilization of settlement. Unloading
was carried out in stages. In each test, settlement during unloading cycle was also
Construction of airal roads on swampy soils using prefabricated reinforced
measured.
concrete plates (3m by lm by 0.15m) can eliminate the need excavating large
deposits of peaty soil with subsequent replacement using mineral soil. Such
Deflection of the concrete plate was determined by 10 dial gauges, which
construction has been performed in Switzerland (Route Beton, 1964), Germany
were mounted on a reference beam spanning the long axis of the concrete plate. Dial
(Roadloff, 1965), and Belarus (Vakher, 1979), and is shown in Photograph 1. Under
gauges were graduated in units of 0.01mm and capable of recording a maximum
the effect of load from car or truck wheels, the concrete plates flex and deflect
deflection of 25mm (blocks were inserted when settlement exceeded 25mm). The
downward. The deflection of the plates, as well as deformation resistance of peat
load testing arrangement is shown in Photograph 2.
can be calculated using the parameters of modulus of deformation, E, or modulus of
subgrade reaction, K.
The experiments were made in situ on a layer of natural peat soil 1.2m thick,
which was underlain by thick, dense fine sand. Peat soils consisted mostly of Seadge
There are two main methods used to model the behavior o f plates on elastic
peat with Hemic fiber content (15-40 percent), Medium Ash content (8-10 percent),
foundations. The first is the elastic model (the Winkler hypothesis); the second is the
and water content o f 800-950 per.cent (highly absorbent). Dry density of peat soil in
elastic halfspace model, which considers soil to be a linearly deforming medium.
situ ranged from 104kg/nr to 1 ISkg/nr.
The parameters involved in analysis of plates (F. or K) can be determined

'Technician. Haley & Aldrich, Inc., 465 Medford Sireel, Suite 2200, Boston, MA 02129-1400.

44
46 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 47

directly on peat soil. However, tests revealed that the ultimate bearing capacity
increase is insignificant.

1.0 %
J---------
MAV 3
k
’ 0.8 it ^ '
\ 'N
>L o
0.6
\
&\
S/S0 (

0.4
Photograph 2. Loading technique and settlement 0 20 40 60 80
measurement devices. Thickness of Sand (cm)
Figure 1. Decreasing settlement o f plate and maximum bending
The modulus of deformation, E, was defined by using the equation lor elastic
moment by increasing thickness o f sand fill on peat soil
settlement from Gorbunov-Possadov (1973) as:
1.3 - Under load P = 0.025 MPa
2.4 - Under load P = 0.0125 MPa
E=l . » V l * ,i)
S - Settlement with sand fill
Sav ^|\ So - Settlement without sand fill
M - Moment in concrete plate with sand fill
And modulus subgrade reaction, K, defined as bearing stress per unit area of the Mo - Moment in concrete plate without sand fill
contact surface between plate and the subgrade on which it rests from Hough (1957)
and Henry (1986) as:
The modulus o f subgrade reaction for the 3m by lm concrete plate was
K. ------------ ( 2) determined from standard plate bearing tests. This value was used as a basis to
A„ S... compare and normalize the results of the other tests. The modulus of subgrade
reaction was based on the settlement of the 328mm diameter plate bearing directly
where P = applied load, N on the peat soil, and was then scaled up to the size of the concrete plate, by using the
Sav = average bending settlement of concrete plate calculated as the area following equation:
of the differential (mm) settlement o f the plate divided by the
length of plate
// = Poisson’s ratio, for peat soil /u = 0.31 (3)
S rp D bp cop
Ap = supporting area of concrete plate (with Ap = 3m")
= form factor, = 0.83, from Gorbunov-Possadov (1973)
where S b.p = settlement of bearing plate, in mm
Db.f. = Diameter of the bearing plate (D b p =328mm)
The same method was used to determine the benefit of having a layer of sand
between the plates and the peat soil. Tests were performed to measure the settlement Equation (3) was derived from equation (4) (Klepikov, 1967)
of plates bearing on different thickness of sand fill placed over the surface of peat
soil. The decrease of plate settlement with increasing thickness o f sand fill, as well
as decrease of maximum bending moment is shown in Figure 1. Sand fill decreases
settlement of plate by as much as one-half of the settlement ot the plate bearing
48 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 49

Where the modulus of deformation, E, in the test with 328mm diameter bearing
plate, was calculated as:

(1 - M 2) (5)

Figure 2 shows the relation of modulus of subgrade reaction for the

foundation on peat soil to increasing thickness of sand fill. Lines 1 and 4 on Figure 2
show the increase in modulus of subgrade reaction with increased thickness of sand
fill for the full size concrete plate. Lines 2 and 3 show the change of modulus of
subgrade reaction under loading by only the bearing plate.

The results of the analysis shown in Figure 2 indicate that the subgrade
reaction of the peat soil under large concrete plate increases by as much as 50%
when the sand fill increased to be nearly as thick as the concrete plate was wide. The
rate of increase in K shown in Lines 2 and 3 is much higher because the bearing area
of circular bearing plate compared with the rectangular concrete plate is much
smaller. When the thickness of sand fill increased, the depth of peat engaged under
the bearing plate decreased. It was concluded that load tests with bearing plate
reflected the behavior of only the top level of foundation peat soil.

Labora tory Experi men (s

Similar experiments were made under laboratory conditions. The modulus of

subgrade reaction was determined under a scaled-down model of the concrete plate
(300mm by 100mm) using both static and repetitive dynamic loads. The model
concrete plate was made of steel 6mm thick. The laboratory experiments were made
with special apparatus designed for the study. The laboratory device for repetitive
dynamic loads allowed changes in load, speed, and intensity of load. Settlement of
the base of soil in the laboratory was measured with a transducer that was connected
to the monitor.
Figure 2. Modulus o f subgrade reaction K as a function o f the ratio o f the
thickness o f sand fill on the peat soil to width o f plate
In the laboratory, we used different undisturbed peat soils (specimen size
0.5m by 0.5m by 0.3m) with the following ranges in properties: water content 463%-
940%; dry density ranging from 94kg/m3 to 185kg/nr; fiber content 15%-45%; and 1,4 — ©— Concrete plate (3m x lm x 0.15m)
ash content 4%-20%. The model loading plate was placed directly on the peat soil. 2.3 — O— Bearing plate D = 328mm
Tests were also made on sand fill over the peat soil. The applied contact pressure 5 — • — Model o f plate (300m x 100mm) in the laboratory
was the same as in the field tests. For each test, settlement is plotted versus contact 1,2 - Under load P = 0.0125 MPa
pressure, as shown in Figure 3 for static loading and in Figure 4 for repeated 3.4 - Under load P = 0.025 MPa
dynamic loading.
50 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS

Contact Pressure, P(MPa)

Contact Pressure, P(MPa)

0
0.005 0.010 0,015 0.020 Q.Q25

Average Plate Settlement. Sav(mm)

Average Plate Settlement, Sav(mm)

1
Plate Bearing on Undisturbed Peat Soil
Plate Bearing on Undisturbed Peat Soil
1 l-F=40-45%, w=463%, 7=184kg/m3
2 [ - F=40-45%, w=470%, 7=184kg/m3
2 - F=45-50%, w=565%, 7=158kg/m 3
! - F=35-40%, w=650%, 7=135kg/m 3
3 - F=10-15%, w=638%, 7=135kg/m 3
3 1- F=30-35%, w=707%, 7=124kg/m 3
4 -F= 15-20%, w=900%, 7=94kg/m3
2\ 1- F=15%, w=940%, 7=94kg/m3
4\
4
3 '
6 L

Contact Pressure, P(MPa)

Contact Pressure, P(MPa)

Average Plate Settlement, Sav(mm)

Average Plate Settlement, Sav(mm)

Plate on Sand Fill over the Peat Soil

Plate on Sand Fill over the Peat Soil
1 . F=40-45%, w=463%, 7=184kg/m3 T=0.75Wp
1 - F=40-45%,w=470%,7=184kg/m3 T=0.5Wp
2 - F=35-40%, w=786%, 7= 114kg/m3 T=0.75Wp
r 2 - F=20%, w=745%,7=l 17kg/m3 T=0.75Wp
3 - F=35-40%, w=614%, 7=145kg/m3, T=0.25Wp
----0 ^ 3 - F=20%, w=745%,7=l 17kg/m3 T=0.25Wp
V 4 - F=15%, w=745%,7=94kg/m3 T=0.25Wp
4 - F= 15-20%, w=900%, 7=94kg/m3 T=0.25Wp

4S

Figure 4. Average settlement model o f plate as a function o f

Figure 3. Average settlement model o f plate as a function o f
repetitive dynamic plate contact pressure, (P)
nonrepetitive static plate contact pressure, (P)

F - Fiber content o f peat soil F - Fiber content o f peat soil

w - Water content o f peat soil w - Water content o f peat soil
T - The ratio o f the thickness o f sand fill on T - The ratio o f the thickness o f sand fill on
the peat soil to width to the plate Wp the peat soil to width to the plate Wp
7 - Dry density o f peat soil 7 - Dry density o f peat soil
Wp - Width o f plate Wp - Width o f plate
52 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 53

Under repetitive dynamic loading using small pressures of 0.025MPa, the

Modulus of Subgrade Reaction, K(MPa/m)

resilient deformation of foundation soil was measured and found to be smaller than 1.
under static loading of the same pressure. The duration of each cycle of the
repetitive load was 0.5 second. Both repetitive and static load tests found settlement
to vary linearly with the applied load. The laboratory experiments enable definition 1.2
of numerical values of the modulus of subgrade reaction (sustained and resilient
deformation) for foundation of the plate for different peal soils by the following
formula which is analogous to equation (3): 0.!

1
(6) 0>
10Apm.

where Sav = average settlement of model plate without taking flexure into
account (mm) (
Ap m = supporting area of plate model Ap.m= 30,000mm' Dry Density of Peat Soil, 7(Kg/m3)
P = load applied (kg) 2.4
P/Sav = was derived from Figures 3 and 4.

Line 5 in Figure 2 reflects the sustained modulus of subgrade reaction for the

Modulus of Subgrade Reaction, K(MPa/m)

2.0
peat soil under the model plate. The character of Line 5 shows that relationship
between thickness of sand fill and peat soil modulus in laboratory modeling is not
strongly influenced by thickness of sand fill. The magnitudes of sustained and 1.6
resilient modulus of subgrade reaction, K, for each kind of peat sample were
determined. They are plotted relative to the dry density of peat soil in Figure 5.
Line 1 in Figure 5 is based on the results of the plate resting directly on the
1.2
foundation of peal soil. Lines 2, 3 and 4 correspond to the thickness of sand fill
overlying the base of peat equal to 0.25, 0.5, 0.75 times w-idth of plate, respectively.

Discussion 0.8

An empirical relationship w'as developed to approximate the curves plotted in

Figure 5. Equation 7 shows the dependency of both the sustained modulus of 0.4
subgrade reaction and the resilient deformation modulus on the dry density of peat
soil, as follows:
0
90 110 130 150 170 190
Dry Density of Peat Soil, 7(Kg/m3)
K = A |io o r (7) Figure 5. Modulus of subgrade reaction sustained (A) and resilient (B)
deformation versus the dry density of peat soil from laboratory
where K = modulus of subgrade reaction for either sustained or resilient model tests
deformation, MPa/in 1 o - Without sand fill
A = parameter numerically equal to modulus of subgrade reaction for 2 x - Sand fill on peat soil 0.25 Wp
peat soil with dry density y = 100 kg/nr (applied for either 3 • - Sand fill on peat soil 0.50 Wp
sustained or resilient deformation) 4 a - Sand fill on peat soil 0.75 Wp
Wp - Width of plate
54 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 55

n = an empirical exponent characterizing the variation of modulus of 3 Resilient deformation of peat soil under dynamic loading has a linear
subgrade reaction (sustained or resilient deformation) when the dry relationship with load. It is also smaller than the deformation which occurs
density of peat soil changed under sustained loads.

The parameters of equation (7), defined by least-squares method, are 4. Empirical relationships have been developed to determine the modulus of
represented in Table 1. These are based on laboratory data. subgrade reaction for sustained and resilient deformation based on the dry
density of peat soil.
Table 1. Parameters A and n in equation 7.
Acknowledgements
Ratio of Value of Parameters for Computing Modulus of Subgrade
Thickness of Reaction The author is indebted to Dr. Jim Lambrechts of Haley&Aldrich, Inc., for his
Sand Fill on Sustained Deformation Resilient Deformation valuable assistance with the preparation of this paper.
Peat Soil to
Width of Plate A n A n References
0.00 0.215 2.50 0.33 2.00
0.25 0.325 2.00 0.45 1.75 Anderson, K.H. and Stenhamar, P. (1982), "Static Plate Loading Tests on
0.50 0.440 1.60 0.61 1.60 Overconsolidated Clay," ASCE, Vol 1, 108, pp. 918-934.
0.75 0.540 1.40 0.90 1.30
Gorbunov-Possadov, M.I., and T.A. Malikova (1973), Calculation o f Structures on
Comparison of the field and laboratory data shows their distinguishing Elastic Foundation, Gosstroiizdat, Moscow, 627 pp.
difference. The field tests with non-repetitive static load, as well as repetitive
dynamic load under a plate (3m by lm by 0.15m), was used as a standard of Hansen, J.B. (1961), "The Bearing Capacity of Sand, Tested by Loading Circular
comparison for analogous laboratory tests. The approximation to the laboratory tests Plates," in Proceedings o f the 5th International Conference on Soil
is not exact. This difference was taken into account by a correction factor. The Mechanics and Foundation Engineering, Vol. 1, pp. 659-664.
resulting correction to equation (7) to make it applicable to field loading directly on
peat soil is: Henry, F.D.C. (1986), The Design and Construction o f Engineering Foundations,
Chapman and Hall, London, 1,089 pp.

K = 3 A (— )" ( 8) Hough, B.K. (1957), Basic Soils Engineering, Ronald Press Company, NY, 513 pp.
100
Klepikov, S.N. (1967), Calculation o f Structures on Elastic Foundation, Budivelnik,
When there is sand fill between the loaded plate and the peat soil, the equation Kiev, Ukraine, 583 pp.
becomes:
Levy, J.F. and K. Morton (1975), “Loading tests and settlement observations on
K = 4 A (-£ -)" (9) granular soils,” Conference o f British Geotechnical Society, Settlement of
100 Structure, Pentech Press, London, pp. 43-52.

Conclusions Roadloff, W. (1965), “Koonnen Spurbahnen aus Betonfertigteilen eine

Standardbasses sein,” Wasser and Boden, No. 3.
1. Increasing the thickness of sand fill over the peat soil decreases both the
settlements of the plate and the maximum bending moment. Route Beton (1964), “Chemins en Geton pour Ameliorations Foncieres en
2. The load tests performed with the bearing plates effected only upper level of Switzerland,” Route Beton, No. 62.
the foundation soil to a depth generally not deeper than the foundation’s
width. Vakher, M.Y. (1979), Investigation o f Peat Soil as a Foundation fo r Road
Construction, Ph.D. Thesis for Belarus Water Research Institute, Minsk,
Belarus, 211 pp.
 GEOTECHNICAL MEASUREMENTS 57

backing the use of glass ballotini, Taylor-Schneebeli model etc.), it is possible to

analyse the process of grain crushing using various granular materials. The author used
dry granular silica gel (grain size 2-4 mm) which does not aggregate at higher
pressures, clearly disintegrates and, due to the considerable angularity, restricts the
grain rearrangement.
Naturally bonded materials often possess high heterogeneity. Thus, for instance, the
author measured with specimens of the same undisturbed sample of Neogene clay
highly differing intensity of bonding (Feda, 1995). Bonding was studied, therefore, by
(weak) portland cement-sand mixtures. Even in this case, at the lower limit (2 % of
cement), some heterogeneity of the cement distribution has been found. Cement-
stabilized soils are a common technology with transportation engineering (see e.g.
Modeling grain crushing and debonding of soils Mitchell, 1976), but the experimental results are seldom used for the study of various
aspects of soil bonding.

Jaroslav Feda1
Grain crushing

Fig. 1 presents the scheme of grain crushing consisting of two phases. In the first one,
the weakest grains (dotted) are crushed. Since the grain fragments do not participate on
Abstract
the load transfer, increased (effective) porosity results accompanied by strain softening.
In the second phase, owing to the compression rearrangement, grain fragments are
The effects of two structural factors - grain crushing and bonding - were investigated activated (incorporated into the bearing grain skeleton) and strain hardening follows.
using model materials - granulated silica gel (for the compression) and cement-sand Figs. 2 and 3 depict those effects as reflected in the compressibility and shear
mixture (for the shear resistance). Grain crushing correlates well with kinking of the resistance, m, n2 ... indicate the increasing initial porosities. Due to the grain crushing,
(oedometric) compression curves. It is suggested that this is the reason for a garlandlike compression curve acquires a garlandlike form (in the arithmetic scale - Fig. 2) and
shape of compression and creep curves. Ratio of compression indices Ca/Cc is a shear strength envelope becomes nonlinear (Fig. 3). Grain crushing resembles
constant. Bonding finds place within a specific bonding surface where the behavior is loosening of the soil: with increasing load (effective) porosity increases. Figs. 2 and 3
brittle. Outside of this surface, debonded material (sand) behaves in a standard manner, depict this process.
e.g. as ductile (if loose). Both structural factors - grain crushing and debonding - deeply
affect the constitutive behavior of geomaterials.

Introduction =>

Constitutive behavior of soils (geomaterials) reflects the soil structure variation under
the effect of various state parameters (load, time, water etc.). Engineering practice asks
for the quantification of those effects. Geomaterial research attempts to explore their Fig. I Two phases of the process of grain crushing
possible causes, occurrence, limits etc. Two examples of such an analysis will be
treated: grain crushing and debonding (destruction of cementation). In the past, grain The effect of grain crushing will be illustrated by the oedometric compression curve of
(or particle cluster) crushing (disintegration) has been mostly investigated in connection silica gel (grain size 2-4 mm, initial porosity of about 70 %). Fig. 4 shows three
with grinding technology, and rockfill and problematic soils (e.g. volcanic, biogenic, oedometric compression curves in a semilogarithmic plot. They are piecemeal (multi)
calcareous sands) behavior (see e.g. Feda, 1982). Real geomaterials were used. Under linear, horizontal arrows indicating kinks. Compression index Cc (= Ae/Alog o; e, o
the tacit assumption that all particulate materials (i.e. materials consisting of solid void ratio and stress) can be defined but within each linear part. If transformed into the
particles in mutual contact) behave in the same manner (the same philosophy is arithmetic scale, each multilinear compression curve represents a garlandlike curve
(garlandlike curve means a curve consisting of a series of concave secondary
Institute of Theoretical and Applied Mechanics, Academy of Sciences of the Czech Republic, compression curves as depicted by dotted lines in Fig. 2).
Prosccka 76, 190 (K) Praha 9, Czech Republic, e-mail [email protected]

56
 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 59
58
LO AD
'0 0 1 000 kPa 10 000

Fig. 4 Three oedometer compression curves of dry silica gel (2-4 mm).
Arrows indicate kinks
Fig. 2 Origin of the garlandlike compression curve by grain caishing
(n are the initial porosities, m the smallest one) Fig. 5 compares the values of the compression index Cc (ranging between 0.5-1.3) and
that of the gradient of crushing (relative increase of the amount of < 2 mm grains).
There is a clear correlation, maximum gradient corresponding with the max Cc (i.e.
with the largest compressibility). This is believed to prove that the garlandlike form of
the compression curves of silica gel results from grain crushing. The convexity of the
diagram in Fig. 5 corresponds best with the dashed compression curve in Fig. 4.
Symbols in Fig. 5 (upper part) correspond to compression curves in Fig. 4.

Fig. 3 Origin of the nonlinear shear strength envelope by grain crushing

(n are the initial porosities, m the smallest one)
AXIAL S T R E S S

The intensity of crushing used to be defined by the variation of the grain diameter dio, Fig. 5 Compression index Cc (upper part) and gradient of crushing (full circles, lower
the amount of fines (e.g. < 0.075 Urn) or the ratio of the initial and final areas occupied part) as depending on the mean oedometric pressure
by the granulometrical curves and horizontal and vertical axes of the granulometrical
plot (area criterion). In the present study, the extent of grain size < 2 mm was chosen This conclusion does not lack the generality. Garlandlike compression was found e.g.
(initial grain size of silica > 2 mm). It was found that this criterion is congruent with with sand (see e.g. Feda, 1982). Fig. 6 depicts a garlandlike creep of another model
material used called "antiperle" (AP). It consists of 5 mm spherical grains, like
the area criterion.
ballotini, but made of sugar instead of glass (sugar is much more crushable). It is
suggested that even in this case the grain crushing is responsible for the effect produced
within the pattern in Fig. 1.
m

60 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS

TIME (min.) value depends on the grain size. On the horizontal axes of Figs. 5 and 7, the axial stress
is divided by the grain resistance to obtain a dimensionless value of stress (ranging
within 0-0.4) thus enabling the comparison with other granular materials of different
grain resistances. Since the grain resistance is statistically variable, a deeper analysis
shows that grain crushing is a progressive process. Progrcssivity depends on grain
roundness, being substantial with angular grains and is absent with spherical grains of
silica gel. Certainly an important role is played by the granulometry (which was in case
of silica gel 2-4 mm of grain size).

Bonding curve

Bonding has been investigated using a medium alluvial sand with portland cement
admixture (2 % to 8 % by weight). The mixture was tested either dry or water
Fig. 6 Garlandlike creep of antiperle AP at 1.5 MPa (oedometer) inundated (after 5 days of curing with 8 % of water). Shear testing was carried out by a
shear box 6x6 cm with specimen thickness of about 3.2 cm. Loose and dense
For the tested silica gel, Fig. 7 presents the value of the coefficient of secondary specimens were prepared with the mean porosity of about 50 % and 55 %.
compression (creep) Co (= Ae/A]0g t; t time) as depending on the stress level Fig. 8 shows the results of four shear box tests of cemented sand and (open circles).
(oedometer). The similarity in the course of Ca (Fig. 7) and Cc (Fig. 5) enables to They may be interpreted in two ways: either by linear interpolation, using a straight line
propose similar mechanism of structural changes, destruction oi grains, as already (dashed line; this is usually done in the literature resulting in the statement that cement
accepted in case of AP. The ratio of C JC c amounts to about 0.02. stabilization affects but the cohesion) or by a composite envelope (full line). Fig. 9
presents another picture of a composite envelope (open circles: debonded sand). Dense
saturated sand of Fig. 8 is replaced in Fig. 9 by a dry loose sand, with the same
outcome. Fig. 10 generalizes the above pictures.

AXIAL STRESS

Fig. 7 Coefficient of secondary compression Co of four compression tests

(Nos. 4, 6, 8, 9) in an oedometer (2 means an average of two values)

The extent of crushing depends on the grain resistance. This was measured to be
11 MPa (mean value of a series of unconfined compression tests of single grains) for Fig. 8 Shear strength envelope of a saturated dense sand with 4 % of cement,
the angular silica gel grains tested. If rounded, this value rose to 24.5 MPa (in inundated, (r is coefficient of correlation; H, Z - dense, inundated)
accordance with the effect of roundness found by Hess, 1980). For comparison, the
same value of dry clayey grains was 4.6 MPa, that of loessial grains 0.8 MPa. The
 GEOTECHNICAL MEASUREMENTS 63
GEOTECITNICAL MEASUREMENTS
62

loading path (= of the amount of deformation work). Finally always the peak shear
resistance is reached.

Fig. 9 Bonding surface of a dry loose sand with 4 % admixture ot cement

Fig. 11 Loading paths of loose dry sand of two initial porosities

(1: 50.4 %; 2-54.8 %) with 2 % of cement

Composite shear strength envelopes were proposed already by Taylor (1948) who
their initial nonlinear part ascribed to cohesion. In fact, the composite nature follows
from the transition of the initial structural configuration (inborn structure) to the
induced (deformation) structure.

Conclusion

The amount of grain crushing may be studied by a model material (silica gel). One
Fie. 10 Generalized picture of the bonding surface may thus explore the effect of stress level, angularity and grain resistance. The process
of grain destruction takes place in two phases - crushing and mobilization of the
The following interpretation is suggested. At high stress level (loading path 1-1) bearing function of destroyed grains. Owing to this complex phenomenon, garlandlike
debonding (destruction of cementation bonds) occurs already at zero shear stress by compression curve and nonlinear shear envelope result.
the effect of axial loading (point D on the normal stress axis). At lower stress level Garlandlike compression curve is the result of grain crushing. The same effect as with
(2-2) debonding occurs before the shear strength is reached at some nonzero shear compression (state parameter: load) can be recorded with creep (state parameter: time).
stress level The behavior is (for loose sand) ductile (or generally that of the standard Therefore the ratio Co/Cc is approximately a constant (0.02) irrespective of the stress
debonded specimen). At small stress level (3-3) the nonlinear envelope is firstly level.
reached and then the shear strength drops to the residual value R - the behavior is Due to bonding (cementation), the shear strength envelopes of bonded soils are
brittle. The (de)bonding surface D-D is thus defined. It is a kind of yield surface. This composite. Cement-sand mixtures represent an appropriate tool to explore bonding­
interpretation is more logical then by assuming the linear (simple) shear strength debonding effects.
envelope and throws the proper light on the process of debonding. It is also supported
by the fact that the behavior of bonded samples is brittle and that oi debonded sample.
Acknowledgement
Fi° 11 depicts various experimental loading paths within the bonding suiface (full
circles at 54.8 % porosity, open circles for 50.4 % porosity are given for comparison). The research reported was partially supported by the grant agency of the Academy of
They all show that the process of debonding takes place at the peak strength and Sciences of the Czech Republic No. A 207 1803 which is gratefully acknowledged.
debonding before peak failure is negligible, irrespective of the length ot the traversed
64 GEOTECHNICAL MEASUREMENTS

Appendix. References

Feda, J. (1982)."Mechanics of particulate materials - the principles". Elsevier-

Academia, pp. 448.
Feda, J. (1995). "Behaviour of a cemented clay". Can. Geot. J. 32, 5, pp. 8 ;y-yU4.
Hess, W. (1980). "Einfluss der Schubbeanspruchung und des Verformungsverhalten
bei der Druckverkleinerung von Kugeln und kleinen Partikeln". PhD. Thesis, TH
Karlsruhe, pp. 162.
Mitchell, J. K. (1976). "The properties of cement-stabilized soils . Proc. Work. Leura,
P ractical A pplications of T ime D omain
Aus., pp. 365-404.
Taylor, D. W. (1948)."Fundamentals of soil mechanics". Wiley & Sons, pp. 71X). Reflectometry (TDR) to Monitor and A nalyze
Soil And Rock Slopes

Neil Anderson,1 M. ASCE

David Welch,12

ABSTRACT: Time domain reflectometry (TDR) has been used for many years in
the telecommunication industry to locate damage or faulting in communication
cables. More recently TDR has been applied in the geotechnical/geologic field in
order to detect movement and locate shear planes in rock or soil slopes by grouting
coaxial cables in boreholes. TDR can provide significant cost savings over
traditional inclinometers in many applications as well as be remotely accessed. This
paper presents five case histories in which TDR has been successfully used to
monitor and analyze unstable rock slopes and soil embankments. Case histories
cover a range of applications which include an open pit rock slope, a small rock
slope, two embankment/levees over soft/loose soil, and a native soil slope. In
addition to monitoring for soil movement, the last case history also presents the
innovative employment of TDR to measure changing groundwater levels in an
economical remote sensing configuration. This paper also summarizes testing
equipment, cable selection, and cable installation methods for each case history.

T dr Background
Time Domain Reflectometry (TDR) is a technology that has it roots back in
the 1930s with the development of radar and later in the 1950s to locate cable faults
in the telecommunication industry. In the 1970s, TDR began to be applied to the
fields of geotechnical engineering and geology to detect earth and rock movement
and locate where this movement was occurring. In geotechnical and geologic
applications, a coaxial cable is placed and grouted in a bore hole or trench (vertical,

1Principle and Geotechnical Engineer, Neil O. Anderson & Assoc., Inc., 22 Houston Ln, Lodi,
CA 95240, [email protected].
2 Senior Geologist, Neil O. Anderson & Assoc., Inc.

65
66 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 67

inclined, or horizontal). If the cable becomes stressed and deformed by earth or rock faulting, forming a “horst” structure in the mine pit. Due to this structure, weak
movement, this movement is detected and the exact point of deformation is planes developed along the fracture zones.
determined using the TDR reader. To maintain dewatering o f the open pit, 60,000 to 70,000 liters per minute of
groundwater is pumped from sixteen de-watering wells. The high-capacity
T dr Principle submergible pumps used for dc-watering cost upwards of $350,000. One o f the de­
The best way to understand how TDR works is to think oi it as "cable radar". watering wells (DW-7) exists on a bench within a weak zone of fractured rock inside
TDR operates by sending a wavefront of energy into a cable from one end. This the open pit. The well is 60 cm in diameter and 340 m deep. The pit-wall slope
wavefront of energy travels along the cable at nearly the speed of light. As the angle in this area is approximately 25°. Slope movement occurred that crimped the
wavefront encounters "variations" in the cable's physical characteristics, part of the de-watering well casing. When an attempt was made to extract the pump, the
wavefront is reflected back to the source. Cable variations or permutations (and S250,000 pump motor housing broke off at the crimp and was lost down the well.
hence the amount of reflection) is affected by two factors: 1) the physical distance A smaller diameter pump was then installed in the well and at the same time a TDR
between the two conductors, and 2) the insulating material between the conductors, cable was installed 8 m away in a 90 m bore hole. For a summary of the cable type
refered to as the dielectric. All two-conductor cables have an associated "impedance' and installation see Table 1. The purpose o f the TDR cable was to locate the shear
or resistance (expressed in units of ohms), which is a function of these two factors. plane(s) and to monitor and chart the subsurface rock movement of the slope so that
Under normal conditions, cable impedance is fairly constant. When physical sufficient “warning” could be given to retrieve the new pump before additional and
distortion or "damage" of the cable occurs (pinching, stretching or shearing), the excessive crimping of the casing took place. See Figure 1 for pit geologic profile and
impedance of the cable changes at the point of damage. Thus, when a wavefront of TDR cable location.
energy encounters a point of impedance change (physical distortion), energy is
reflected back to the source. TDR can measure two parameters of the returning
(reflected) wavefront: 1) the time elapsed from the original, outgoing wavefront, and
2) the amplitude of the reflected wave. These are plotted as a "spike" of information
(cable impedance on the vertical scale) verses time (distance along the cable) in the
horizontal scale (ref. A&L Instruments, 1999).
By comparing successive cable readings or "signatures", the difference
between the readings w'ill show that damage has occured to the cable and where that
damage is located. The magnitude of the signature spike is also an indication of the
degree of cable damage. With experience, the magnitude can also be a rough
indication o f the amount of movement that has caused the damage. For a
FauH Fault
comprehensive overview of the application of TDR in the earth science field, see the
work of O’Connor and Dowding (1999). i igure 1. General Cross Section of the Open-Pit Mine, Battle Mountain, Navada.
Five case histories are presented in which TDR has been sucessfully used to
monitor soil and rock slopes. These case histories illustrate many of the advantages The baseline or first TDR reading was made in June of 1998. Additional
and limitations of using TDR with slope stability problems. readings to monitor movement occured on a monthly basis. A presentation o f major
cable signature changes is presented in Figure 2
Case History #1, E cho Bay M ine Initial readings indicated that shear planes had developed at the 37 and 42 m
The Cove pit, operated by Echo Bay Minerals near Battle Mountain, Nevada, depths (see the 12/03/98 reading, Figure 2). Successive readings indicated that
is a large open-pit gold and silver mine. The pit has a maximum crest elevation of sufficient movement was occuring to shear the 1.3 cm diameter cable and cause
1707 m and a pit bottom elevation of 1184 m. Side slopes consist o f benched contact between the inner and outer conductors. Eventually, the TDR cable broke
and no monitoring below the break was possible.
terraces ranging from 25 to 45°.
The mineral deposit consists of a sediment hosted gold and silver deposits After seven months the pump was pulled prematurely due to changes in the
with a granodiorite intruded into the Augusta Mountain Sequence, consisting of200- de-water scheme o f the mine and the need for the pump elsewhere, but the
million-year-old limestone, sandstone, and conglomerate. Covering the older rocks inexpensive approach ofTDR gave the mining engineers and geologists warning that
are 10 to 35-million-year-old volcanic tuff. The area has since undergone normal deformation of the cable, and consequently the well casing, was occurring at a
measurable and steady rate.
68 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 69

For this application, I'DR provided an economical way to check for The slope is composed of a highly weathered fine-grained and less weathered
movement, with cable shearing indicating at least 1 cm of movement. The work of blocky metagraywacke rock that is part of the Jurassic Age Mariposa Formation.
O’Connor and Dowding (1999) has shown that when comparing cable signature The bedding is vertical, the strike is northwest, and the slope face is north-south. A
“spike” growth with measured degree of movement from adjacent inclinometers, it zone of jointed and thinly bedded rock containing weakly cemented silt-sized and
is possible to develop a correlation "ratio" for a given cable and grout combination. sand-sized grains appears to be easily susceptible to weathering. This zone is fairly
This ratio is used to more accurately quantify the degree o f movement with TDR well defined and appears to be associated with the reported rock movement on the
signature spike growth. upper slope and in line with several circuitous cracks in the asphalt of the road
surface above the area of loose material.
Following the initial examination, the primary focus was to determine if
potential deep movement, possibly associated with the road crack pattern could be
detected. A 12 m deep TDR coaxial cable was installed behind the zone of
weathering and within the road crack pattern in an attempt to intersect any shear
plane or other movement in the rock. In 6 months of monitoring of the TDR cable
through the winter period, no change in the cable signature was found, indicating no
subsurface movement o f the slope. This was substantiated by no observed changes
in the road cracking patem or widths. Trenching of the road cracks did, however,
indicate that they are potential shear zones. The rock movement appears to be
associated with one particularly weathered bed of vertically inclined rock. The
project is currently in the recommendation stage. Annual monitoring of the TDR
cable will continue to be performed as needed.

CASE HISTORY #3, DISCOVERY BAY RESIDENCE

The Discovery Bay residential development was constructed in phases in the
1980s at the edge of the Sacramento/San Joaquin river delta area in northern
California. The development consists of a large network of channel excavations and
peninsula embankment fills which were made in order to produce a development in
which each single-family residence has a land-side driveway and a water-side boat
dock.
The “delta” is reclaimed land from a former inland saltwater/freshwater
Meters Below Ground Surface marsh created in Holocene times with climate flucutations in sea levels. The general
soil profile consists of unconsolidated sands, silts, and clays overlain by organic
Figure 2. TDR signatures of the Echo Bay Project. Note: The cable signatures deposits of peat and additional channel deposits of sands, sills, and clays. In the
have been vertically separated for clarity of presentation. early 1900s an extensive levee system (2,700 km) was constructed in order to reduce
seasonal flooding of outlying areas and reclaim land for agricultural purposes. Since
C ase H istory #2, J ackson diner construction of levees, much of the reclaimed land has subsided below sea level due
The community of Jackson is located within the California gold belt in the to oxidation of the peat and consolidation of the loose sediments.
Sierra Nevada foothills ofnorthem California. The city engineering department was Since the Discovery Bay development is located on the western edge of the
concerned about recent rock and soil sloughing from a 12 m high cut slope within the delta, 3 to 4.5 m embankment fills were required to elevate the streets and house lots
city limits. A popular diner is located along the toe of the slope and a single-lane above sea and flood levels. Fill material was generated from channel excavations
street is located along the crest. 'The slope-face angle ranges from 40 to 50°. It is and consisted o f a mixture o f organic peat, silty clays, and sandy clays. In 1997 a
believed that the cut slope is at least 50 years old. In the last few years the public private residence, overhanging deck, and boat dock were constructed on a vacant, 10-
works department has covered the slope with plastic during the winter in an effort to year-old lot. Soon after completion of the house and deck there was evidence of deck
reduce increasing erosion and sloughing. and slope movement. The slope failed at the end of July 1997. Initial observations
70 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 71

o f the failed slope were made on July 29, 1997. The failure appeared to be a
rotational failure with the head scarp running roughly parallel to the slope and
varying from 1.2 to 1.8 m from the edge of the original embankment. The head scarp
had dropped approximately Ü.8 m and the lailure had caused the lower rows of TEKTRONIX 1502C TDR READER
timber piles supporting the deck to rotate and drop, causing the deck to drop and pull
away from the house (see Figure3). Cursor -0.040 ft
On July 30, 1997, a 10 cm diameter, hand-augured test hole was excavated Distance/Div 5 ft/div
in the middle of the slide to a depth of 5.03 m. The top o f the test hole was located Vertical Scale 133 mp/div
1.8 m below the top of the embankment. Embankment fill was encountered to a Vp 0.68
Noise Filter 4 avg
depth of 2.7 m. At a depth of 3.7 m some apparent slicken sides were encountered, power ac
indicating a possible shear plane. In order to definitively determine the location of
the shear plane a TDR cable was installed in the test hole (see Tabic 1 for summary
o f installation). The cable extended to the 5.0 m depth. Figure 4. TDR signature of the Discovery Bay Residence, Hole H-l.

CASE HISTORY #4, CONTRA COSTA CANAL LEVEE

Pumping plant # 1 of the Contra Costa County Water District has an intake
canal that draws from the fresh water o f the Sacramento/San Joaquin delta. The
intake canal has embankment levees on either side, built up from material excavated
from the canal. In 1996, a 110 m long section of the south levee had slumped into
the canal. Due to the width o f the canal, the failure did not significantly restrict water
flow in the canal, but it did threaten the support of two power poles that ran along the
crest of the embankment. The general soil profile was 4 m of embankment fill,
consisting o f a mixture ol loose to very loose silty sand and medium stiff sandy/silty
clay, overlying a 1.5 m stratum o f loose native silty sand. The sand was underlain
by very stiff sandy clay.
The failed section of the embankment was over-excavated and replaced, but
due to tidal action in the canal, the depth of over-excavation was limited to just the
embankment fill material. Consequently, three layers of geotextile and a toe trench
Figure 3. Cross section of Discovery Bay residence embankment and cable of rip-rap was placed at the base of the reconstructed embankment.
installation location.
Eight months after reconstruction of the embankment some tension cracks
opened up along a 50 m section, parallel to and just below the crest of the
A return visit was made a week later, on August 7, 1997, to make an initial
embankment. Crack widths varied from 1 to 20 cm. The cracks were filled in and
TDR reading. When the reader was first attached, the instrument showed a
it was elected to install two TDR cables at the base of the embankment, immediately
substantial cable spike at the 3.75 m depth, indicating a large defonnation ot the
above the rip-rap protection (sec Table 1 for installation summary). The purpose of
cable. Coincidentally, while the equipment was being readied to record the cable
the TDR cables was to monitor for additional movement and locate a shear plane if
signature the subsequent reading showed that the cable had separated at the this depth
movement occurred. Two TDR cables were installed to a depth of 4.9 m,
(see Figure 4). This confirmed that the slide was still active (a surface marker
approximately 30 m apart (see Figure 5).
showed the head scarp had dropped an additional 5 cm) and that the shear plane was
located at the 3.75 m depth. This also showed that the slide originated in the weak
foundation soils and not the embankment fill, as originally thought due to previous
problems within the development. This had significant bearing on the lailure
analysis and the recommended repair.
 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 73
72

An interesting item of note is the overall drop o f the subsequent cable

signatures (from the 12/24/98 baseline signature) starting at the 0.5 m depth. This
shows a change in the cable impedance which is indicative of a change in the
insulating property o f the cable dielectric (insulation between the two conductors).
Levee Shear Plane Head Scarp This change is likely being caused by water which has infiltrated into the cable
'TDR Cable 5.06 m Below Ground dielectric. The 0.5 m depth is close to the high tide elevation of water in the channel.
/ Recompacted Fill\ Because only a small diameter hole was necessary, powered hand-equipment
Sandy Silt, Native was able be utilized on the 2:1 (11:V) slope, eliminating the need for more intrusive
and expensive drilling techniques. Cost of installation for both cables was under
Cable Break at 3.11 m $600.

Figure 5. Cross section of the levee showing the location of the IDR cable and CASE HISTORY #5, VACAVILLE HILLS RESIDENCE
shear plane. A custom home was constructed at the base of the Vacaville hills of northern
California. Local geology consists ofTertiary and Cretaceous beds of sandstone and
The cables have been monitored for six months. Readings indicate that some shale interbedded with claystones and mudstones. The claystone is described as
additional movement has occurred and that the shear plane, as suspected, is located highly susceptible to slope failures. The home backs up to the base of a 30+ m high
within the loose sand stratum at the 3.1m depth. A summary of major cable signature natural slope consisting of interbedded sandstones and claystones.
changes in Hole #2 is presented in Figure 6. Movement, however, has not been The base of the slope, which had been trimmed slightly during original home
sufficient yet to damage or shear the cable of Hole # 1. construction, failed in 1994. The failure was 7.5 m high and encompassed
approximately 300 m3 of material. The slide took out a small (1.2 m) two-tiered
175 retaining wall located at the base and extended out into the backyard, within 3 m of
INITIAL ¿ABLE SIGNATURE — the house. A forensic investigation concluded that the slide was caused by an
172.5 improperly installed and improperly functioning french drain coupled with excessive
irrigation on the slope, which caused the slope to become saturated and the subsoils
170 weak.
Three years alter repair of the slide, which consisted of benching in new fill
167.5 and construction o f a new single retaining wall, the slide reactivated, severely
o13
stressing the new wall. At this point it was suspected that a heretofore unknown
MOVEMENT ONLY SUFFICIENT TO
165 CAUSE A SHORT BETWEEN INNER
source o f seasonal subsurface water was saturating and weakening the slope and
AND OUTER CONDUCTOR OF 0.6 CM triggering the slide. This was supported by the fact that water surfaced at the base
DIAMETER CABLE CAUSING
162.5 ✓ " N IMPEDANCE TO DROP.
of the retaining wall during and immediatly after rainstorms. A test hole and test pits
SUFFICIENT MOVEMENT OCCURS excavated in the slide during the summer months did not encounter any subsurface
160 TO COMPLETELY SHEAR CABLE
__ CAUSING INFINITE IMPEDANCE.
water. Consequently, it was elected to cover the slope through the winter and install
instrumentation to monitor for additional slope movement and for the presence of
157.5
.4 subsurface water through a winter cycle. Since the water only surfaced during and
after a rainstorm it was desirous to obtain readings immediately after storm events.
155 Date of Acquisition Three monitoring wells and one grouted TDR cable were installed within the
e n t ir e c a b l e s ig n a t u r e s h o w s ------ 12/24/98 slide mass. Monitor wells consisted of a 1.9 cm perforated pvc pipe placed in a 5 cm
152.5 IMPEDANCE DROP INDICATING ------ 02/16/99
WATER HAS LIKELY INFILTRATED ------ 04/01/99 dia. bore holes. The pipe was backfilled with pea gravel and a 3 foot bentonite plug
CABLE DIELECTRIC.
at the surface. A cross section showing monitoring well locations and depths is
150 presented in Figure 7.
0.6 1.2 1.8 2.4 3 3.6 4.2 4 8 5.4
Meters Below Ground Surface

Figure 6. Contra Costa Canal Levee, Hole #2 cable signatures indicating soil
movement between 12/98 and 4/99.
 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 75
74

Figure 7. Cross section of Vacaville Hills residence slide showing monitoring

well locations and maximum measured depth of groundwater.

initially, a perforated and hollow 16 mm dia. coaxial cable (air Idled

dielectric) was inserted in each of the monitor wells to record ground water depths.
In theory, as water rises and fills the dielectric air space in the cable (acting as a
standpipe) it changes the impedance. This impedance change should be easily
detected with the TDR reader and the depth of ground water recorded. In the lab this
concept worked well. However, in the field the air filled coaxial cables proved
unreliable due to moisture condensation within the cable. A larger diameter cable
would likely work better. The coaxial cables were replaced with common 2- Figure 8. Vacaville Hills Residence cable signature readings indicating
conductor lamp cord wire (18 ga.). The TDR signature of the lamp cord is not the fluctuations of groudwater depths in monitoring well P-3.
"flatter" and smoother plot that is achieved with coaxial cable, but the key is the
"change" in the signature due to water surrounding the cord's two conductors and Cable readings from the single grouted TDR hole indicated no additional
changing its characteristic impedance. The lamp cord signatures proved to be very movement of the slide. The TDR cables detected groundwater in each of the three
reliable for detecting and measuring groundwater heights and correlated well with monitoring wells and the maximum recorded groudwater height is shown on Figure
the manual water level readings which were taken periodically to check the TDR 7.
readings. The remotely accessed TDR proved very useful and economical for
The four instrumented holes were wired to a central location within the instrumenting the subject project and collecting the desired information. Total cost
laundry room of the house with RG-59 coaxial cable. Average length of connecting of installing the cablcs/monitoring wells, setting up the remote site instrumentation,
cable was 25 m. The cables were connected to a new multichannel TDR 5600, A&L and obtaining the readings (discounting the cost for the PC, TDR reader, and the trips
Instruments reader which in turn was connected to and controlled by a personal to manually measure the ground water depths) was approximately $2500.
computer (PC). The PC was connected to the house phone line and the computer
program PC Anywhere was used to access the computer by modem. Readings were Cable Installation and E quipment Summary
remotely made from our office and the signature files immediately downloaded after Table 1 presents a summary of the TDR cable installations for each o f the
major storm events through the rainy season of early 1999. Figure 8 presents a case histories presented in this paper.
summary of several cable signature changes from monitoring well P-3. Manually
measured groundwater depths within the pvc standpipe have been noted on the
applicable signature readings.
 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 77
76

Initially we utilized the Tektronix 1502C reader, which is a very good

Table 1, TDR Cable Installation Summary instrument and versatile in working with many different cable types of different
impedance. However, it is expensive (@ $8,000) and is more awkward when
H ole G r o u tin g R e a d in g
C ase K x c av atiu n D rillin g
D ia.
D epth C a b le
G r o u t M ix
M eth o d In s tru m e n t
working with longer cables in the field because they require multiple readings.
E q u ip m e n t M eth o d (m ) T ype
H is to ry (cm ) Although there is software available that will automatically combine multiple
Flexwell 5 part readings to produce a continuous signature (Huang, 1993), in order to get the
Ingersol Poured
Rand Reverse SLC-12- cement, 1 TDR preferred resolution (approx. 1 data point every 2.5 cm), multiple readings must still
15 90 in from
1 Circulation 50J, 1.3 part water by surface* 5600 be obtained and saved in the field for longer cables since the unit can only save 256
TH1000
Drill Rig cm dia. weight data points at a time. We have now switched to A& L Instrument's TDR 5600
Hollow
reader, which was specifically designed for this application. It is more economical
5 part
Mobil B53 Stem Poured
TDR (@$2,500) and captures the entire cable signature at the desired accuracy (approx.
Beldin cement, 1
Explorer Contiuous 15 40 in from 1 data point every 2.0 cm, regardless of cable length). Due to its presets, however
2 RG-58 part water by 5600
Flight surface* it should be used with either 50 or 75 ohm cable.
Drill Rig weight
Auger
5 part CONCLUSIONS
cement, 3 Poured The following Table 2 presents a summary of the application and use o f TDR
Beldin Tektronix
Hand 10 5.0 part sand, 1 in from
3 - RG-59 1502C for each of the case histories presented in this paper:
Auger part water by surface*
weight
fable 2, TDR Case History Application Summary
5 part
2 Man, 5 cement, 3 Poured Case History
Continuous Bledin TDR TDR Application
hp, Post 5 5.0 part sand, 1 in from
4 Flight RG-58 5600 #1, Echo Bay
Hole Auger part water by surface* Locale shear plane in a rock slope and wam of the degree of movement
Digger weight Mine occuring in order to protect equipment.

Beldin Prototype ft2, Jackson Diner Establish lack of movement in a rock slope.
5 part TDR
2 Man. 5 RG-58 cement, 3 Poured #3, Discovery Bay
Continuous and 2- 5600, Quickly located shear plane in an active slope failure of a manmade soil
hp, Post 3.0 to part sand, 1 in from Residence embankment.
5 Flight 5 multiple
Hole 4.9 cond., part water by surface*
Auger 18ga. channel #4, Contra Costa
Digger weight Detect movement and locate shear plane in a repaired levee.
wire reader Canal Levee
* If equipment is available a more prudent method for grout placement would be to pump or tremie
the grout form the bottom up, however, for eaeh of these applieations this method appeared to work. #5, Vacaville Hills Establish that there was no additional movement of a reactivated slide of
Residence a soil slope and remotely measured groundwater elevations within the
For critical applications (with minor movement or multiple shear planes) or slide mass through a winter cycle to help in the design of a subsurface
depths exceeding 50 feet it is recommended that the higher quality Flexwell cable be drainage system.
used (@ $6.00/m). This cable produces a signature with less "noise" and minor
damage shows up quickly. More econmical RG-59 cable (@ S0.35/m) can be used Following are additional observations and conclusions regarding the use of
for less critical or shallow applications where it may be desired to simply locate the TDR for each of the case histories:
shear plane of a slide.
Our experience is that a hard Portland cement grout mix will make the cable #1, Echo Bay - Allowed for quick and easy assesment of movement at
sensitive to smaller movements and will even work in soft soil conditions when greater depth and accurately located multiple shear planes for help in
simple shear plane location is desired. O’Connor and Dowding ( 1999) have done analysis. Not able to accurately determine degree of movement along shear
significant work on various grout and cable combinations and it matching them to the plane, but for the purpose involved it w'as sufficient. See work o f O'Conner
soil or rock conditions. This is particularly important if it is desired to utilized TDR and Dowding ( 1999) for using TDR to "quantify" degree of movement. Also
to help quantify the degree of movement and the authors recommend that their work it grcatei accuracy is desired for other reasons, additional instrumentation
be referred to.
 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 79
78

should be utilized such as an inclinometer or possibly a fixed and remotely Very economical to install, particularly in difficult locations
accessed down-hole inclinometer. If time and resources permit, the initial Ease o f reading, reading(s) can be made in several minutes by a
use of TDR would help in determining the optimal placement ol fixed down­ single individual, regardless of the length of the cable and depth of
hole.
hole inclinometers.
Cabling of multiple holes can be brought to a central location or out
#2, Jackson Diner - Proved quick and economical for establishing lack of of harms way when traffic is involved.
movement. Since no instrumentation is placed in hole or left in field, cable ► Can be remotely accessed.
can be reliably checked in future years for long term monitoring. Cost of Suitable for long term monitoring since no instrumentation in hole or
installation and additional readings is minimal which is appealing to a public left at site.
works department of a small city. If movement is ever detected, then a more " Can be used to monitor ground water depths.
intensive investigation can be initiated.
We are a small to mid-size geotechnical firm and have had 5 projects, as
#3, Discovery Bay Residence — Rapid installation of TDR in a small illustrated, in the past 2 years in which TDR has been used. We feel that it has
diameter hand-hole allowed installation in an otherwise difficult location and proven to be an effective tool and support its continued use. It can be applied to any
the necessary information (location of shear plane) to be achieved in a timely situation that would damage or chage the impedance characteristics of a two
manner. More expensive instrumentation and installation costs would have conductor cable.
been beyond the resources of the project.

#4, Contra Costa Canal Levee — Again, installation by hand equipment on appendix - R eferences
a slope allowed economical installation in a difficult location. Since TDR did
indicate that additional movement occurred, would liked to have had a more A&L Instruments, L.L.C., (Lodi CA), TDR 5600 Product User’s Manual, 1999.
accurate assessment of degree of movement. Dual cable installations gave
a range of where movement was occurring since one cable sheared olf and Huang, F.-C., O'Connor, K.M., Yurchak, D.M., and Dowding, C.H., NUMOD and
the other did not. By using TDR in conjuntion with inclinometers, the exact NUTS A, 1993. Software for Interactive Acquisition and Analysis o f Time
degree of movement can be obtained. On a larger slide with a greater Domain Reflectometry Measurements, BuMines IC 9346, 42 pp.
investigative budget, multiple installations of economical TDR cable can be
used to better map and understand the slide and/or reduce the number of O Connor, K.M., and Dowding, C.H., GeoMeasurements by Pulsing TDR Cables
inclinometers that might otherwise be installed. and Probes, CRC Press L.L.C., Boca Raton, 1999.

#5, Vacaville Hills Residence — Same advantage as 3 and 4 regaiding

economics of installation for a small budget. Due to the electronic nature of
TDR, cabling allowed all instrumented holes to be brought to a centra
location in order to set up a single remote access station. Major advantage
for this application was the remote accessing of groundwater readings which
were able to be obtained from our office immediately after storm events.

Overall, we have found TDR to be a valuable tool in accessing and analyzing

soil and rock slopes. It does have its limitations, most notably its present limitation
in detecting the exact degree of movement. Work had been done and research is
presently under way to improve this capability. The benefits TDR appear to be
significant, including the following:
 GEOTECHNICAL MEASUREMENTS 81

CABLE TESTER

r\
V7 330
PROTECTIVE
V
\
COVER 50-OHM
CONNECTOR CABLE

Comparison of TDR and Inclinometers for Slope Monitoring

Charles H. Dowding1, Member, and Kevin M. O 'C onnor, Member

EXPANSIVE
Abstract CEMENT
CRIMP
GROUT
IN C A B L E
As TDR technology grows in acceptance, its use stimulates further innovative
applications and comparison with slope inclinometer measurements. This paper
presents cases in which the opportunity arose to compare these two technologies to
detect and measure subsurface deformation in slopes. Among the main points
addressed are (1) the comparison of TDR reflection magnitude and inclinometer Figure 1.-Schematic of cable installation and monitoring.
incremental displacement to help quantify deformation with TDR technology, and (2)
the comparison of the accuracy of the two technologies in detecting and measuring deformities producing these reflections. As shown in Figure 1, a metallic coaxial
shear deformation in localized versus general shear. Case histories are presented that cable can be placed in a drill hole and anchored to the walls by tremie placement of
involve monitoring movement in soil and rock slopes and embankments as well as an expansive cement grout. When localized shear movements in rock or soil are
retrofitting deformed inclinometer casing with coaxial cables. This paper describes sufficient to fracture the grout, cable deformation occurs and can be detected using
installation details. When monitoring to detect narrow shear zones in soils, it is best a TDR cable tester which launches a voltage pulse along the cable. At each location
to use small ratios of hole-to-cable diameter, and prudent use requires that larger where deformation has occurred, a portion of the voltage is reflected back to the TDR
diameter, solid, metallic coaxial cables be installed in separate holes. Grout strength unit which displays the reflections. Travel time of each reflection distinguishes the
should be (1) low enough to fail before bearing capacity of the surrounding soil is locations where cable deformation is occurring, and differences in the reflected signal
reached, and (2) high enough to deform the cable it encapsulates. It is recommended magnitudes can be employed to quantify the magnitude of cable deformation
that other users publish cases in which theses two technologies are compared in order (O’Connor and Dowding, 1999). When a cable is crimped prior to placement in the
to expedite continued assessment. hole as shown in Figure 1, a reflection from each crimp serves as a distance reference
marker in the TDR record.
Coaxial Cable Geometry used fo r TDR Monitoring
Strip Mine Highwall Slope (Case 1)
TDR is analogous to radar in a coaxial cable. Consequently, it is possible to
display all reflections along a cable and identify the type and location of cable The example shown in Figure 2 involved installation of coaxial cable in the
highwall slope of an oil sands mine. Details of the installation are compared with the
other cases in Tables 1 and 2. The bituminous sands contain numerous thin
'Department of Civil Engineering. Northwestern University. Evanston. Illinois 60208, consolidated clay layers that cause highwall slope instability. Consequently, slope
cdow<Jing@northwestern .edu.
movement is an operational problem and many kilometers of inclinometer casing have
-’GeoTDR, Inc., 297 Pinewood Drive, Apple Valley, Minnesota 55124. [email protected].

80
 GEOTECHNICAL MEASUREMENTS 83

Shear Zone

TDR Cable
260

Inclinometer TDR Cable

Displacement, mm Reflection Mag., mrho

* Reference Crimp
a End of Cable
Figure 2.- Installation of coaxial cable and inclinometer in Case 1;
A, location; B, inclinometer incremental displacement and TDR
waveform for one reading

been installed to continuously monitor this movement in the vicinity of the

multimillion dollar dragline. Given the large commitment to man-hours and hardware
required for monitoring, the mining company investigated the potential of TDR for
remote monitoring.

At three locations in the mine, a coaxial cable was grouted into a hole located
10 m from an inclinometer and the results from one comparison are shown in Figure
2. Note the reflection which developed at a depth of 18 m. It is consistent with both
Table 1 - TDR Cable Details

the shear zone location and the inclinometer incremental displacement profile. When
the TDR reflection magnitude of 6 mrho (1 mrho = 1/1000th of the launched voltage)
is compared with the inclinometer incremental displacement (28 mm over the
inclinometer probe wheel base of 2 ft), the linear correlation is 6 mrho/28 mm or
approximately 0.2 mrho/mm. As will be seen in Case 2, this sensitivity is low and can
be enhanced by reducing both the hole diameter and grout strength.

Landfill Slope Defonnation (Case 2)

Deformation of an industrial landfill presented another opportunity to compare

inclinometer and TDR response. The landfill rests on 1 m of silt and sand, underlain
84 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 85

TOR W AVEFORM S
TDR WAVEFORMS INCLINOMETER
o o
00
\ \
0 6 /2 4 /9 8 0 8 /0 7 /9 8
o (X)
\10 10/13/99
0 7 /2 4 /9 8 1 0 /2 5 /9 8 0 7 /2 7 /9 8
00 co
0

1Ò

20
>

60

40
( b a s e l in e - D rrm )
Increm ental cfsp. (m m )
since 1/ 2 9 / 9 8
20 mmo (baseline - 0 mm) Figure 4.-TDR waveforms and inclinometer profile for Case 3.

Figure 3.-TDR waveforms and inclinometer profile for Case 2. resulted from the decreased hole diameter and grout strength. The magnitude of the
larger TDR reflection (23 mrho at a depth of 31 m) is a minimal value as it exceeded
by 9 to 12m of soft clay, which in turn is underlain by stiffer clay. In accordance with the maximum range setting used for data acquisition.
the current standard of practice, inclinometers and piezometers were installed to
define the extent of the deformation and monitor effective stress changes. As a field Road Distress and Retrofit o f Inclinometer Casing (Case 3)
trial of TDR technology to detect and quantify shear within soft clays, coaxial cables
were installed in boreholes adjacent to two of the inclinometers. Distress in a limestone causeway supporting a major highway presented yet
another opportunity to compare TDR and inclinometer response. Movements had
As shown by the lower bulge at the right of Figure 3, inclinometer incremental been occurring since December of 1997, and an instrumentation program was
displacements indicate that subsurface deformation occurred within a shear zone at formulated to determine the cause of movement. Among the instruments installed to
a depth of approximately 30 m within the soft clay. As shown by the 5 mrho monitor subsurface movement were inclinometers and coaxial cables (Figure 1).
reflection at a depth of 22 m, deformation of the coaxial cable first occurred on
7/10/98 at this depth which is the contact between the fill material and underlying Comparison between TDR and inclinometer measurements in Figure 4 shows
layer of sill and sand. On 8/19/99. a spike of 8 mrho appeared at a depth of 28 m and consistent response at a depth of 76.5 m. Movement at this depth occurred within a
a second spike of 23 mrho at a depth of 31 m. At both depths, the inclinometer zone of greater fracture density in the limestone. The reflections at depths of 113 m
incremental displacement was 2.6 mm. These responses yield a sensitivity of 8 and 119 m were caused by movement along dolomite-shale contacts. The TDR
mrho/2.6 mm = 2.7 mrho/mm and 23 mrho/3 mm = 7.7 mrho/mm. As shown by the reflections at a depth of 76.5 m between 6/24/98 and 8/7/98 indicate that the
summary in Tables 1 and 2, the increased sensitivity compared with Case 1 may have reflection grew by 58 mrhos. The growth in inclinometer incremental displacement
86 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 87

TDR WAVEFORMS INCLINOMETER

TDR WAVEFORMS INCLINOMETER

0 9 /1 0 /9 8 0 8 /0 5 /9 9 0 8 /1 4 /9 8
0 9 /1 5 /9 9 05/04/99 1 0 /1 6 /9 9 10/16/99

0.5 0 0.5
20 mrho Incremental disp. (mm)
crim p since 4 / 2 7 / 9 9
(baseline = 0 mm)

Figure 6.- TDR waveforms and inclinometer profile for installation in Case 4.
Figure 5.- TDR waveforms and inclinometer profile for retrofit of inclinometer casing in an
inclined hole. compare inclinometer and TDR response. The approaches of two highway bridges
were constructed by building embankments over existing soils on the west abutment
at this depth, over the same time period, was 12.4 mm so the sensitivity was 58
and over rock on the east abutment. The embankment fill consists of stiff silty clay
mrho/12.4 mm or approximately 5 mrho/mm. When the reading was taken on
and rock fragments and, on the west abutment, it overlies native alluvial soft clay and
8/19/98, it was found that the cable had been sheared off at this depth.
loose sand. The underlying rock on both abutments is shale and limestone. The shale
is soft and erodible while the limestone is conspicuous as ledges in exposures along
Kinking of the inclinometer casing provided an opportunity to demonstrate the
the highway.
use of TDR technology to extend the useful life of this casing. On 3/2/99 it was not
possible to lower a probe down past a depth of 75.6 m in the adjacent inclinometer
Inclinometer casing and coaxial cables were installed in separate holes on both
hole, and on 6/25/99 this inclinometer casing was retrofitted with a grouted coaxial
the east and west abutment slopes. The TDR waveforms and inclinometer
cable to continue monitoring. Ultimately, five of the eight original inclinometer
incremental displacement profile shown in Figure 6 were acquired on the east
casings were retrofitted with grouted coaxial cables, and TDR waveforms for one of
abutment where the embankment was constructed over rock. Both the coaxial cable
these are shown in Figure 5. This hole is inclined at 30 degrees from vertical and on
and inclinometer casing were deformed at depths of 12 m and 19 m. However,
8/14/99 it was not possible to get the inclinometer probe past a downhole distance of
deformation of the coaxial cable was also detected at depths of 11, 13, 16, and 20 m.
86.6 m which corresponds with an actual depth of 75 m. The TDR waveforms
The fill extends to a depth of 13 m where it rests on the soft shale which is underlain
acquired since 9/10/98 show that it has been possible to continue monitoring
by limestone at a depth of 20 m. The water table is located at the top of the shale
movement at this depth.
layer. The largest TDR reflections correspond to depths at which contacts exist— 29
mrho at a depth of 12 m and 26 mrho at 13 m at the contact between fill and shale,
Abutment Embankment Deformation (Case 4)
and 21 mrho at a depth of 19 m at the contact between shale and limestone. By
comparison, the inclinometer incremental displacement was only 0.2 mm at depths of
Suspected deformation of bridge abutments provided another opportunity to
12 m and 19 m. This implies a sensitivity which varies from 21 mrho/0.2 mm to 29
mrho/0.2 mm or 105-145 mrho/mm. These values are considered to be unusual, and
reasons for such a large difference in response between TDR and inclinometers are
88 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 89

discussed below. respond to abrupt changes in shear strains at the boundaries of thick shear zones. The
TDR reflections at depths of 28 m and 31 m in Figure 3 define the upper and lower
Deformation o f Coaxial Cables versus Inclinometer Casing boundaries of the shear zone within the soft clay. Peterson’s laboratory tests indicate
that in this case, the fractured grout was confined by the soil in the shear zone and
Plastic inclinometer casing and solid metallic coaxial cables deform differently localized shearing of the cable occurred.
when subjected to very localized shearing. Metallic coaxial cable deforms easily when
subjected to highly localized shear and has been found useful in rock where Installation Details
deformation occurs along joints, bedding planes, and fractures. On the other hand,
inclinometer probes are sensitive to gradual changes in inclination of the inclinometer Details of the coaxial cable installations in these four cases varied considerably
casing. Localized shearing of inclinometer casing causes kinking such that a probe due to local practices for drilling and installation of slope inclinometers. These details
cannot be moved through the deformed casing. The thinner the localized shear zone, are summarized in Tales 1 and 2 that describe (1) coaxial cables and (2) the grout and
the greater the TDR response and the smaller the slope inclinometer response soil/rock properties. Local drilling practices varied from rotary wash drilling of
(O'Connor and Dowding, 1999). Thus, in situations involving both general shear and unlined 75-106 mm diameter holes in medium clay to 200 mm diameter hollow stem
localized shear, it should be expected that the two technologies will respond auger holes in miscellaneous rock and soil fill. It is important to keep the ratio of hole
differently as illustrated by the case histories summarized in this paper. diameter to cable diameter as small as possible to ensure that the grout fractures with
the least relative deformation. Coaxial cables will not be deformed until the grout
Shear Strain versus Incremental Displacement fractures which is one reason that TDR technology responds poorly to gradual
deformation of soil over large shear zones (general shear).
There are two methods commonly used to plot slope inclinometer
measurements—cumulative displacement (change in profile of the casing) and All the grouts used in these cases were tremied into place with the drilling rig
incremental displacement (local changes in inclination of the casing). While reported water pump - even the thickest and strongest. However, grout strengths varied
as displacement, slope inclinometer incremental displacement (SIID) is also the considerably as shown in Table 2. Early installations (i.e., Case 1) employed very
inclination of an inclinometer probe, and can be interpreted as local shear strain (over strong grouts that were standard mixes established for inclinometer installation at that
the base length between wheels of the probe). site. More recent installations (e.g., Case 2) have employed cement-bentonite grouts
with unconfined compressive strengths that are only 2% to 3% of those used in Case
As a consequence of the difference in response of TDR and inclinometer 1. Case 4 encountered loss of the low viscosity grout, and it was necessary to resort
technologies, SIID under-represents localized shear while TDR under-represents to a thicker grout mix to minimize loss through rock fractures and sand lenses.
general shear. This difference results from the span over which relative displacement
is measured. SIID is the change in angle over a base length of 600 mm (or 24 in.), the Grout strengths must be large enough to deform the cable; yet they must be
wheel-base of the standard probe. Thus a SIID of 1 mm over 600 mm (or 0.04 in. weak enough to fracture before the bearing capacity of the soil outside the shear zone
over 24 in.) is a shear strain of 0.17%. This shear strain is averaged over the base is exceeded (Cole, 1999). The metallic coaxial cable must be installed in its own hole
length which is fairly large when compared to localized shearing along rock fractures and the grout must fracture early so that the cable can be deformed as movement
(that may be less than 2 cm in width), or along thin shear zones in soil. occurs within the surrounding soil (Pierce, 1998). For installation in rock, this
consideration is not as critical due to the relatively high strength and stiffness of rock.
It has been shown that cable-grout composites respond poorly when shear In order to maximize cable/grout composite sensitivity in soil, it is hypothesized that
occurs across large shear zones. Peterson (1993) found that sensitivity declined by the shear capacity of the grout should be less than the bearing capacity of the soil just
a factor of 2 when the distance between clamps used in a laboratory direct shear test outside the localized shear plane.
was increased from 1 mm to 40 mm. Sensitivity declined by a factor of 20 for a clamp
spacing greater than 80 mm. These results indicate that cable-grout composites Strength and stiffness of the cable, grout, and soil/rock for the four sites
respond optimally to localized shear across zones less than 40 mm wide (e.g., shear described in this paper are summarized in Tables 1 and 2. Based on laboratory tests,
strain of 1 mm/40 mm or 2.5%). grouted 22-mm-diameter CommScope solid aluminum coaxial cable (P3-75-875CA)
has a stiffness of 88 MPa and a shear strength of 1.6 MPa. This cable has performed
Most importantly, data from Case 2 indicate that TDR technology may satisfactorily at sites where the deformation occurs in fractured rock. Case 2 indicates
90 GEOTECHNICAL MEASUREMENTS

that this solid aluminum cable installed in small diameter (100 mm and smaller) holes
with weak grout may respond satisfactorily even when installed in soft clay.

Summary

Comparisons of slope inclinometer and TDR responses for the cases presented
in this paper indicate that both technologies provide useful information. TDR
technology is especially sensitive to localized shear so it is most responsive to
concentrated shear strain. On the other hand, slope inclinometers are especially
sensitive to gradual changes in inclination so they are most responsive in soils PERFORMANCE MONITORING FOR A CRITICAL STRUCTURE
undergoing general shear. TDR technology will also respond to abrupt changes in BUILT WITHIN A LANDSLIDE
shear strain at the boundaries of a thick shear band.
Barry K. Myers1, M. ASCE, L. Radley Squier, F. ASCE2,
These differences do not imply that either technology is more correct; rather, Mark P. Biever3, and Rex K.H. Wong, M. ASCE4
the two techniques respond optimally under different conditions. The real challenge
is to provide a more precise and rational explanation for these different íesponses. In
order to expedite this process it is recommended that other users publish case
ABSTRACT
histories in which the two technologies are compared.

Finally, solid aluminum coaxial cables can be installed in deformed The Washington Park Station is a key element of the Tri-County
Metropolitan Transportation District’s (Tri-Met’s) Westside Light Rail Transit
inclinometer casing to allow continued monitoring. The results of installing and
System in Portland, Oregon. The Station platforms are located 79 meters (260 feet)
monitoring coaxial cables installed in deformed inclinometer casing indicate that this
below the ground surface, with sole access provided through two 9.5 meter
will be effective whether the casing has been installed in rock or in soil. Such
(31 foot) diameter concrete-lined vertical shafts. Within theses shafts are high­
retrofitting allows continued monitoring of subsurface deformation without the need
speed elevators for transferring passengers between the surface and the Station,
to drill additional holes. fixed stairways, ventilation ducts, electric power conduits, signal and
communication system conduits, and a fire suppression standpipe system. The
References lower 55 meters (180 feet) of each of the shafts is constructed in bedrock and the
upper portions are within a large ancient landslide. In addition, the drilled pier
Cole, R.G. Compliant TDR Cable Grout Composites to Measure Localized Soil foundation elements for the headhouse structures extend through the landslide and
Deformation. M.S. Thesis, Department of Civil Engineering, Northwestern into the underlying bedrock.
University, Evanston, Illinois, December, 1999, 90pp.
O'Connor, K.M. and C.H. Dowding. GeoMeasurements by Pulsing TDR Cables and The discovery of creep movements within the ancient landslide resulted in
Probes. CRC Press, Boca Raton, 1999, 420pp. the need to incorporate special design features in the shafts. These design features
Pierce, C.E. A Compliant Coaxial Cable-Grout Composite for Time Domain allow for a total of 5 centimeters (2 inches) of horizontal shear movement within
Reflectometry Measurements of Localized Soil Deformation. Ph.D. the shafts, at the location of the basal failure plane of the landslide. Five
Dissertation, Dept of Civil Engineering, Northwestern University, Evanston, centimeters (2 inches) was selected as the design criteria based on an evaluation of
Illinois, December, 1998, 233pp. the creep movements measured in the inclinometer instrumentation and the
Peterson, D. Personal communication, Syncrude Canada Ltd., Edmonton, Albeita, extrapolation of these movements over the 50-year design life of the structure. The
April, 1993. beneficial impacts of the remediation measures installed during construction of the
Station were also considered in the selected 5-centimeter (2-inch) design criteria.

, Principal Engineer, Squier Associates, 4260 Galewood Street, Lake Oswego, OR 97035
‘ Senior Consultant, Squier Associates, 4260 Galewood Street, Lake Oswego, OR 97035
ASenior Staff Geologist, Squier Associates, 4260 Galewood Street, Lake Oswego, OR 97035
Resident Engineer, Tri-Met, 710 N.E. Holliday, Portland, OR 97232

91
92 GEOTECHNICAL MEASUREMENTS
GEOTECHNICAL MEASUREMENTS
93

Additional design considerations included the guide rail supports for the high-speed
elevator: they can tolerate only 0.64 centimeters (0.25 inches) of movement Surface Elevators

without adjustment to maintain safe operation. With these design criteria, both Tri- Ground Surface
Met and the City of Portland Building Department considered it essential to adopt a Sev=213-216m

long-term monitoring program. If performance monitoring showed rates of Headhouse

movement in the landslide that would result in the design criteria being exceeded, Excavations
Elev=206m (west)
then additional landslide remediation measures would need to be implemented. Elev=201m (east)

To achieve the long-term performance monitoring objectives, a monitoring

system was designed and installed. The system provides real-time monitoring of
the elevator shaft movements, and the movements and ground water conditions Slide Plane
within the landslide. The monitoring system consists of: 1) instrumentation located at Shafts
Elev=192m (west)
at the slip joints within the shaft structures and in the landslide mass near the Elev=197m (east)

Station, 2) both manual and automated data acquisition for monitoring the
instrumentation, and 3) an integrated warning system and associated response plan
if detected movements exceed predetermined threshold levels.

INTRODUCTION

The Westside Light Rail Project provides commuter train service from
downtown Portland, Oregon to the west-side suburb communities. The Washington
Park Station accesses the rail corridor at a key location 3.2 kilometers (2 miles)
west of downtown Portland. The Station is located adjacent to the Portland
Metropolitan Zoo, the Forestry Center, the old Oregon Museum of Science and
Industry Building, and the popular Washington Park. Because of the West Hills,
located adjacent to downtown Portland, deep tunnels were required. At the location
Figure 1 Elevation View of Washington Park Station
of the Washington Park Station, the rail tracks had to be located 79 meters
(260 feet) below the ground surface. The Station is the deepest train Station in Historical Landslide Activity
North America. Figure 1 presents an elevation view of the Station and associated
Development has occurred over the years, at times triggering local
elevator shafts.
mstabthty a, the toe of the ancient slide. In this regard, in the late 1950 s ml!
Geology I tL n Z o n l l i h COnn, c,c,io" wilh 8radi"S for a new Portland Zoo. and ¡he
widening of U S. Highway 26 at the toe of the slope. Ground movements occurred
The Washington Park Station is located in the Central West Hills, west of nd a major scarp, about 579 meters (1900 feet) long, appeared in the slope some
downtown Portland, Oregon, within an ancient landslide mass. The West Hills is . . to 183 meters (400 to 600 feet) above the Highway, (refer to Figure 7)
an anticline consisting predominantly of layered basalt flows of the Miocene Age * measures consisled of horizontal drain arrays on the west and a de~ep
Columbia River Basalt Group. Stratigraphically, a thick mantle of loess (wind vertical drainage curtain above the head scatp segment on the east as shown i!,
blown silt), known as Portland Hills Silt covers most of West Hills. Locally, a red igurc 2. Movements of the slide continued on the east over the next 10 years with
to varicolored plastic clay deposit correlated with the Sandy River Mudstone of cumulative totals of up to 0.46 meters (1.5 feet). Consequently the l e g o n
Miocene to Pleistocene age underlines the Portland Hills Silt, and overlies the
S i moreT active*east portion
stabilize the 1 (0 D 0 Tof’ the
inS'a" ed 3 r° Ck but,ress in 1970
slide. ordef to
basalt. The upper basalt is highly weathered and, at depth, weak interflow zones
exist. The basal failure plane of the ancient landslide is in the red plastic clay
and/or weathered basalt. The approximate extent of the ancient landslide is shown
on Figure 2. As shown, the Washington Park Station is at about the mid-section of
the ancient landslide. The ancient slide encompasses about 485,600 square meters
(120 acres).
94 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 95

Hence, at the time of siting the Washington Park Station, the 1958-1959
major slide event had been essentially stabilized. In addition, the instrumentation
along the toe of the slope revealed creep-like movements varying between
0.08 centimeters (0.03 inches) per year just above the buttress, to less than
0.25 centimeters (0.1 inches) per year to the east of the buttress below the Zoo.

< M HORIZONTAL DRAINS S g

M e»»®
4ÏC..-4 V \\> 3 s r -i >■
m o n it o r in g p r o g r a m f o r c o n s t r u c t io n

Initial Instrumentation
'.. V /!:{; ' . v - . c j ç ^ v£ W * 7 in s t a l l e d d u r in g v P t t r Æ ' Æ ' % WASHINGTON '-
M r PARK
f r h i 'A STATION In connection with the final design phase for the Westside Light Rail Tunnel
segment and the Washington Park Station, three inclinometers were installed. The
inclinometers were installed in October 1992 in order to investigate potential creep
movements in the upper reaches of the ancient landslide. Inclinometer B-585 was
J' ■jJV EAST installed about 18 meters (60 feet) north of the Station to a depth of 24 meters
^ B -5 8 4 'i.^ 'ì EXCAVATION
■ : DRAINS . 0' V ‘S k ^ W <T (78 feet). Inclinometers B-584 and B-583 were installed 50 and 143 meters
: WEST
(165 and 470 feet), respectively, south of the Station to depths of 34 and 37 meters
(112 and 122 feet). The locations of the instruments are shown on Figure 2. Four
sets of readings were taken in the inclinometers during the winter months,
November 1992 to March 1993. No movements were observed. An additional
reading set was taken in April 1994, which revealed very small movements at
E SCARP ( 19gg) depth. More readings were collected in June 1994 that confirmed the April
readings. The resultant magnitude of movement observed in the three inclinometers
between November 1992 and June 1994, some 19 months, is shown in Table 1
below, as is the annual creep rate.

Table 1. Total Movement and Creep Rates between

November 1992 and June 1994
Inclinometers Inclinometer Total Movement Annual Creep Rate
centimeters centimeters/year
Figure 2 Washington Park Landslide (inches) (inches/year)
B-585 0.28 (0.11) 0.18 (0.07)
B-584 0.25(0.10) 0.15(0.06)
An ODOT inclinometer installed some 18 meters (60 feet) upslope of the B-583 0.13(0.05) 0.08(0.03)
rock buttress, first read in 1975, revealed a decreasing rate of ground movement of
about 0.56 centimeters (0.22 inches) per year between 1975 and 1976, to about As shown in Table 1, the creep movements were apparently higher in the
0.08 centimeters (0.03 inches) per year between 1987 and 1991, when the upper reaches of the ancient slide, and lower towards the toe. Five additional
instrument became inoperable. Other inclinometers, some monitored by ODOT on reading sets were obtained between June 1994 and December 1994, and during this
the west and others monitored by Squier Associates on the east below the Zoo, period no further movement was observed. The lack of movement during the 1994
revealed similar creep-like movements in the mid-1990’s on the order of less than dry summer season suggested that the creep movements observed were most likely
0.25 centimeters (0.1 inches) per year (average). related to increases in ground water levels during the preceding wet, winter months.
96 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL. MEASUREMENTS 97

Ground Movements During Construction

Additional instrumentation, consisting of eight inclinometers and six open

standpipe piezometers, was installed close to the perimeter of the proposed Station
Headhouse excavations in the Fall 1994. Two excavations were made, one for the
west shaft location and one for the east. The excavation work began in early
January 1995. Shortly after commencement of the excavations, ground movements
were observed at depth in the inclinometers. The movements were most
pronounced in inclinometers IN-26 and IN-28 located immediately upslope of the
Station as shown on Figure 2. Figure 3 shows the relationship of the west and east
excavation levels for the Headhouse construction, and the acceleration of ground
movement in the two inclinometers IN-26 and IN-28 as the depth of excavation
increased.

The cumulative precipitation for the 1994/1995-winter season is also shown

on Figure 3. Squier Associates recommended that the east excavation be halted
some 7.6 meters (25 feet) above design grade, because of concern of eminent
failure of the excavation. The elevation of the failure plane on the east was at
196.6 meters, which was some 4.6 meters higher than on the west. Further, it was
recommended that an array of eleven horizontal drains, from 91 to 244 meters
(300 to 800 feet) in length be installed from within the west excavation as shown on
Figure 2. The installation of the horizontal drains was completed in April, 1995.
The drains had a pronounced impact on lowering ground water levels and slowing Figure 3 Relationship of Excavation Levels for Construction of the
the ground movements. Staged excavations were adopted, with excavations first to Headhouse to Measured Ground Movements and Precipitation
final grade on the west followed later by excavations to final grade on the east.
More highly reinforced drilled pier foundation elements assisted in controlling
ground movements on the east. The Washington Park Station was substantially LONG-TERM MONITORING PROGRAM
completed and backfilled in the summer of 1996.
The primary objective of the long-term monitoring program is to confirm
Presently, the degree of stability of the ancient landslide mass near the that landslide creep movement rates remain at or below an average threshold limit
Station is greater than before because of the continued drawdown of ground water of 0.10 centimeters (0.04 inches) per year. This limit is based on the allowable
levels by the horizontal drains installed during construction. Continued slower maximum ground movement of 5 centimeters (2 inches) over the 50-year design
creep rates near the toe in B-583, as compared to the higher creep rates above the life of the structure. If the rate of movement exceeds this threshold level for a
Station, provide some assurance of movements slowing with time, because of the significant period of time, then additional landslide remediation measures will most
buttressing effect provided by the more slowly moving toe portion of the ancient likely need to be implemented. Further objectives of the program are to confirm the
slide. extent of the ground movements in the vicinity of the Station (refer to Figure 4),
and the ground water conditions that may be of significance in evaluating these
movements. Because the elevator shafts are critical to the Station operations, the
long-term monitoring system also needs to provide advanced warning of any
sudden movements that would cause an immediate concern for the safe operation of
the elevators.

To accomplish these objectives, a long-term monitoring system has been

implemented that includes: 1) a number of different instrument types; 2) both
manual and automated data collection; 3) the use of spreadsheet tools for data
evaluation and management; and 4) an integrated advanced warning system and
98 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 99

associated response plan. The following sections discuss each of these items in Table 2. Summary of Long-term monitoring Instrumentation
detail.
Instrument Type Monitoring Purpose Sensor Type Reading
Instrumentation (Quantity) Frequency
Inclinometers Movement in the Biaxial inclinometer 5 Quarterly
The instruments selected for long-term monitoring include inclinometers, (10) landslide mass probe (0.61 meter 5 Semi-annually
tilt meters installed across slip joints in the elevator shafts, piezometers, and intervals)
outflow monitoring from the horizontal drains. The locations of the selected
Inclinometers (3) Movement at the slide Electrolytic type in- Every 4 Hours
instruments are shown on Figure 4. plane place inclinometer
Tilt Meters (2) Movement across the Two uniaxial vibrating Evaluated every
slip joint in the elevator wire tilt meters 60 seconds and
shafts configured to provide logged for
biaxial monitoring historical data
every 4 hours
Open Standpipe Piezometric head near Water level read with 3 Quarterly
Piezometers (11) the midpoint and base electronic well sounder 8 Semi-annually
of the landslide mass
Vibrating Wire Piezometric head near Vibrating wire pressure 3 Every 4 Hours
Piezometers (11) the midpoint and base transducers 3 Quarterly
of the landslide mass 5 Semi-annually
Horizontal Drain Flow from the Multistage orifice Every 6 Hours
Outflow Flow horizontal drain outfall standpipe flow meter
meter (1) pipe with vibrating wire
pressure transducer

Inclinometers

All of the inclinometer casing installed, with the exception of one

replacement instrument, is 6.99 centimeter (2.75 inch) outside diameter (O.D.) PVC
casing manufactured by the Slope Indicator Company of Bothell, Washington. The
one exception is a 8.48 centimeter (3.34 inch) O.D. replacement casing which was
installed to achieve a longer design life at a location that has historically
♦ Long Term In-Place Inclinometer
experienced larger rates of movement and offset. All of the inclinometer
© Long Term Open Standpipe Piezometer
installations were placed in 10.2 centimeter (4 inch) diameter boreholes and
A Long Term Vibrating Wire Piezometer
backfilled to the surface with cement grout. Since all of the installations are highly
m Row Meter with Single Channel Vibrating
visible, they were protected with steel or aluminum flush-to-grade monuments. The
Wire Datalogger

Bl Remote Monitoring Unit (RMU) depths of the installations range from 24.5 to 46 meters (80 to 150 feet) with the
deeper installations located upslope of the Station in the deepest portions of the
landslide.
Figure 4 Long-term monitoring Instrumentation Plan
The three inclinometers that are considered to be the most critical for early
A summary of the different instrument types, their purpose, and the detection of movements that could be damaging to the Station are outfitted with
associated reading frequencies are presented in Table 2. Model 56804020 electrolytic vertical In-place-Inclinometers (IPI’s) manufactured
by the Slope Indicator Company of Bothell, Washington. The IPI’s are electronic
100 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 101

sensors that measure tilt at one location within the inclinometer casing. The IPI’s elevations. Discharge from the horizontal drain outflow pipe flows into the
were positioned in the casing at the depth of the identified zone of movement. 20.3 centimeter (8 inch) diameter standpipe, and the water head in the standpipe is
measured. The flow rate from the outflow pipe is then calculated using the
Tilt Meters measured water head and an empirical orifice-flow equation. The orifice-flow
equation relates the water head above the center of the orifice hole to the flow rate
One set of tilt meters was installed across a precast slip joint in each of the through the orifice.
elevator shafts. The slip joint is a design feature for the elevator shafts to
compensate for any shear movement along the failure plane of the ancient slide. Multiple orifice holes were used to allow for measurement of flow rates in
The tilt meters provide immediate detection of total movements that exceed the range of 0.095 to 3.5 liters per second (1.5 to 55 gallons per minute). When the
0.64 centimeters (0.25 inches). They also measure smaller ongoing shear water head rises above the lowest orifice hole, then the flow rate is calculated as the
displacements to observe for trends that could indicate a developing rate of sum of the flow from each of the orifice holes. Therefore, for large flow rates near
movement of concern. The locations of the tilt meters are shown on the elevation 3.5 liters per second, all three orifice holes are used.
view of the Station, Figure 1. The tilt meters span the joint by means of a
51 centimeter (20 inch) long stainless steel bar with pivot joints on either end to The water levels within the 20.3 centimeter (8 inch) diameter standpipe are
allow the units to tilt in any given direction. The tilt meter assemblies consist of measured with a Model 4500AL, 17.2 kilopascal (2.5 pounds per square inch),
two Model 6350 uniaxial vibrating wire tilt meters that are welded together at vibrating wire pressure transducer manufactured by Geokon Inc. of Lebanon, New
90 degrees to provide biaxial measurements. Geokon Inc, of Lebanon, New Hampshire. The pressure transducer is buried in pea gravel near the base of a
Hampshire manufactured the tilt meters and assembly. perforated 10.2 centimeters (4 inch) diameter PVC standpipe that is placed within
the 20.3 centimeter standpipe. The 20.3 centimeter diameter standpipe is also
Piezometers centralized within a 30.5 centimeter (12 inch) diameter PVC standpipe that is set in
concrete in the ground. The outer standpipe provides protection for the flow meter
Two different types of piezometers were used for long-term monitoring of and allows for the water discharged through the orifice holes to be collected and
the piezometric head within the landslide mass. These included open standpipe discharged further downstream. A 15.2 centimeter (6 inch) diameter outfall pipe is
piezometers and vibrating wire pressure transducers. Dual instrument installations connected at the base of the 30.5 centimeter diameter standpipe to convey the water
were used to monitor the piezometric head at both the midpoint and base of the to the desired discharge point located 152 meters (500 feet) downstream.
landslide mass at three locations as shown on Figure 4. At these locations the
vibrating wire transducer was placed at the base of the landslide and the open The multistage orifice standpipe flow meter was selected rather than a weir
standpipe was used to monitor the piezometric head at the midpoint of the landslide. plate or flume because of the site geometry constraints. The discharge from the
horizontal drain outflow pipe exited on the side slope of a drainage ravine.
The open standpipe piezometers consist of two 2.5 centimeter (1 inch) Consequently, providing a level bench to install an open channel-type flow meter
diameter PVC standpipes that were stacked, two per installation, one deep and one would have been difficult. In addition, access to the site was difficult and,
shallow with a bentonite seal between the two. The zones identified for monitoring therefore, adopting a design that would minimize the amount of heavy construction
were screened and backfilled with sand to provide a 4.6 meter (15 feet) screened materials, such as concrete and steel, was advantageous.
and sanded monitoring zone. The vibrating wire transducers were installed in a
similar configuration with the transducers wrapped in a protective cover, lowered Data Collection
into the boring, and buried directly near the base of a 3 to 4.6 meter (10 to 15 foot)
thick zone of sand backfill. A bentonite seal was placed between the monitoring Data collection for the long-term monitoring program includes both manual
zones and the borehole was then backfilled with grout to the ground surface. The readings and automated data acquisition. As previously discussed, future
pressure transducers used were Model 4500 vibrating wire pressure transducers movements within the vicinity of the Station are of greatest concern and require
manufactured by Geokon Inc. of Lebanon, New Hampshire. close monitoring. Other areas within the ancient landslide mass are also of interest
to aid in the evaluation of any future movements that may adversely impact the
Horizontal Drain Outflow Flow Meter Station. Therefore, long-term monitoring includes automated data collection at
frequent intervals for those critical instruments in the vicinity of the Station. Less
The horizontal drain outflow pipe is equipped with a multistage orifice frequent data collection by manual methods is provided for those instruments that
standpipe flow meter. The flow meter consists of a 20.3 centimeter (8 inch) are valuable in maintaining a good historical record of movements within the
diameter PVC standpipe with different sized orifice holes at three different overall landslide mass. In addition, real-time monitoring is provided for the tilt
102 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 103

meters to allow for immediate detection of any critical shear movements impacting especially for the real-time monitoring of the tilt meters. Of importance, a separate
the elevator shaft structures. RMU was provided for each of the tilt meter assemblies for system redundancy.

Manual Readings The three RMU’s located outside of the Station are configured to function
primarily as dataloggers. The units collect data on the EPI’s and vibrating wire
The manually read instruments include inclinometers, open standpipe piezometers every 4 hours. The data is then downloaded periodically for
piezometers, and vibrating wire piezometers. The reading frequencies for these evaluation. The two RMU’s within the Station are configured as data acquisition
instruments vary from quarterly to semi-annually as shown in Table 2. and control units that provide continuous monitoring for alarm conditions along
with historical data collection functions. The units are integrated with the Station’s
The inclinometers are read at 0.61 meter (2 foot) intervals using a Slope Supervisory Control and Data Acquisition (SCADA) system so that if alarm
Indicator Company accelerometer-type inclinometer probe. The data are stored in a conditions are detected, the operations personnel can be notified automatically.
portable Datamate readout unit for later download to a PC. Slope Indicator Real-time interrogation of the tilt meters and periodic downloading of historical
Company also manufactures the Datamate. data from a remote location is provided through the use of modem communications.
The tilt meters are read and evaluated for alarm conditions every 60 seconds, and
Water level data are collected from the open standpipe piezometers using an data is logged for historical purposes every 4 hours.
electronic well sounder with a graduated tape. The vibrating wire pressure
transducers are read electronically by connecting the transducer leads to a vibrating One Model LC-1 single channel datalogger, manufactured by Geokon Inc.,
wire readout box. The readout box energizes the sensor and reads its response as a is used to collect water level data from the vibrating wire pressure transducer
frequency measurement. Since the frequency response of the sensor is directly installed in the horizontal drain outflow flow meter. The single channel data logger
related to pressure, the water pressure at the elevation of the sensor can be is a programmable unit that can collect and store readings on a predetermined time
determined. The piezometric head is then calculated by adding the pressure, interval for later retrieval. The unit is concealed within the 30.5 centimeter
expressed in meters of water, to the elevation of the transducer. Geokon Inc diameter standpipe and provides reliable transducer readings every 6 hours. The
manufactures the readout box used to read the transducers. single channel datalogger was selected for this application as an inexpensive way to
collect reliable data on a frequent basis. Data is periodically downloaded from the
Automated Data Acquisition unit for evaluation.

Automated data acquisition is accomplished by the use of both intelligent Data Evaluation/Management
remote monitoring units (RMU’s) and a single channel datalogger. The locations of
the RMU’s are shown on Figure 4. The three locations around the Station each Data from the long-term monitoring system is evaluated on a quarterly
monitor one LPI and a shallow and deep vibrating wire piezometer. The two basis. This evaluation consists of reducing the data and plotting it for comparison
locations within the Station monitor the tilt meters in the west and east elevator with the historical data. The comparison is primarily used to make observations
shafts, respectively. The single channel datalogger is used to collect data from the regarding the rate of movement and the corresponding conditions within the slide
horizontal drain outflow flow meter. mass that are influencing the rate of movement. Trends of increasing rates of
movement are of special interest.
The RMU’s consist of either a Model CR-10 or CR-10X automated
datalogger as manufactured by Campbell Scientific Inc. of Logan, Utah. The CR- The manually collected inclinometer data is downloaded from the Datamate
10X is a newer model with enhanced capabilities that was used for the real-time into a database software tool called DMM®. Another software tool called
monitoring of the tilt meters, and one of the inclinometers and piezometer DigiPro® is then used to plot and evaluate the inclinometer data. Slope Indicator
instrument groups. The Model CR-10 and CR-10X are a microprocessor controlled Company manufactures both of these software tools. The plots that are generated
data acquisition units that can be programmed to collect and evaluate data include cumulative deflections verses depth, and time-history plots of the
independently of the other RMU’s, and without direction from a central control movements at specific depth intervals. The cumulative deflection plots indicate the
unit. This distributed intelligence architecture provides significant advantages. depth at which the movement is occurring and the time-history plots are used to
With the intelligence distributed throughout the monitoring network, data can be evaluate the rates of movement, and magnitudes of movement.
collected and alarm notification initiated at multiple locations independent of each
other. Therefore, if one or more of the RMU’s cease to function, the remainder of
the system continues to operate. This is important in providing a reliable system,
104 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 105

Data from the IPI’s, tilt meters, piezometers, and horizontal drain outflow
Secure Wiring Enclosure
flow meter are entered into a Microsoft Excel® spreadsheet for reduction and
evaluation. Time-history plots are generated for each of the instruments. The time-
history plots for the IPI's and tilt meters are used to evaluate rates of movement
trends and total accumulated deflections for the tilt meters. The horizontal drain
and piezometer data are evaluated along with rainfall data to identify trends of
increasing piezometric head and changes in flow rates under different climatic
conditions.

The results of the data accumulation and evaluation are presented in a

quarterly report that is provided to Tri-Met operations personnel for review and
assessment

Integrated Warning System and Alarm Response Plan

The purpose of the integrated warning system is to provide advanced

warning of shear movements along the elevator shaft slip joints that exceed
0.64 centimeters (0.25 inch). The threshold of 0.64 centimeters was based on the
design tolerance for the guide rail supports for the high-speed elevators. If the total
cumulative movement at the slip joint exceeds 0.64 centimeters, then the guide rail Figure 5 Monitoring System Schematic
supports need to be adjusted to maintain safe operation of the elevators.
Alarm Level Detection
System Design
In the event that the tilt meter readings indicate a cumulative total vector
The architecture for the advanced warning system is presented on Figure 5.
movement of 0.64 centimeters (0.25 inch), then a portion of the CRIOX's operating
The tilt meter assembly is located in the stairwell portion of the elevator shafts and
program becomes activated and begins an alarm subroutine. The alarm subroutine
extends across the slip joint as shown. Transient protection is provided for the tilt
averages the next five sequential readings and verifies that the average value is
meters at a junction box that is located in the stairwell. A roughly 70 meter
equal to or greater than 0.64 centimeters. Based upon the averaging routine, the
(230 foot) long shielded signal cable connects the junction box with the RMU,
CR10X will screen out erroneous or anomalous data points to prevent the
which is located in the lower Headhouse level of the Station as shown on Figure 1.
occurrence of a false alarm. When the CR10X confirms that the alarm is real then it
The signal cable is protected inside a 5 centimeter (2 inch) diameter metallic
will latch the relay and personnel at the primary operations center will be notified
conduit. The junction box at the RMU location provides transient protection for the
by the SCADA system. The CR10X also begins logging the data every 60 seconds.
RMU. Components for the RMU include the CR10X, a 12VDC battery power
The data may be remotely viewed by personnel through the modem communication
supply with 110VAC trickle charger, a vibrating wire interface unit that reads the
link. If the average value does not exceed the alarm level, then the CR10X returns
tilt meters, and the modem communication equipment. The SCADA system
to its normal operating procedures.
interface consists of a latching relay that is energized by the CR10X when an alarm
condition occurs, and an alarm box that provides visual confirmation of the alarm
Response Plan
condition and a manual reset button to unlatch the relay. The relay provides a
closed contact that is recognized by the SCADA system as an alarm. The SCADA
When personnel at the operations center receive notification from the
system then provides a visual and audible notification to the personnel located at
SCADA system, a response plan is initiated. The response plan is in the form of a
the primary operations center 24 kilometers (15 miles) away. The primary
flow chart that is kept at the operations center. The plan includes the following
operations center is manned 24 hours a day.
action steps in the event that an alarm condition is detected by the system.

If an alarm condition is detected by the tilt meters in both elevator shafts,

then the operator immediately:
106 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 107

1) notifies the appropriate Tri-Met personnel, Two design functions that have proven to be especially beneficial are the
2) dispatches someone to the site to inspect the structures for signs of SCADA interface and the dial-up modem communications link. These two
movement or damage, and functions have enhanced the performance of the system for the following reasons:
3) downloads the data from the RMU’s for evaluation.
Simplicity
If an alarm condition is detected by only one of the two RMU’s, then the
operator: As is typical for a large capital construction project by a local transit agency,
many of the engineering staff who had intimate knowledge of the system left the
1) dispatches someone to the site to inspect the elevator shaft and project in 1998. The rail maintenance-of-way (MOW) staff then assumed the
appropriate tilt meter and RMU for signs of damage, responsibility for operations and maintenance. The challenge was to provide an
2) notifies the appropriate Tri-Met personnel immediately if the structure easy-to-understand interface for the MOW staff who are not design engineers. The
shows signs of movement or damage and downloads the data for addition of a simple alarm notification through the SCADA interface, and a clear
evaluation, response plan allows the staff to feel more comfortable in their ability to recognize
3) notifies the appropriate personnel if the warning system equipment is and respond to an alarm condition at the Washington Park Station. Although an
damaged or malfunctioning, or alarm event has not yet occurred at the Station, the SCADA interface has been
4) resets the alarm relay at the RMU if a false alarm has been verified. tested and proven to perform very well for alerting the MOW staff.

Alarm Acknowledgement Procedure Remote Access to the Elevator Shaft RMU’s

When the latching relay is energized by the CR10X, the green light on the As the Station construction was nearing completion, there were concerns
alarm box is turned off and the red light is turned on to indicate an alarm condition. about the ability to monitor the tilt meters from multiple remote locations to allow
The relay interface was intentionally designed to be reset manually at the alarm box for rapid response to an alarm condition, and to minimize the cost of regular data
location to ensure that a visual inspection of the system is made following any collection. The addition of modems for the two RMU locations along with two
alarm detection event. The relay must be reset by manual action of the reset button dedicated telephone lines made remote communication possible.
on the alarm box panel and cannot be reset remotely using the modem
communication link. When the relay has been successfully reset, the red light is With the modem communication link, the geotechnical consultants, as well
turned off and the green light is turned back on. as Tri-Met staff, can access the deflection data off-site from their office desktop
computers using the PC208W® software by Campbell Scientific Inc.. Having the
SYSTEM PERFORMANCE SINCE INSTALLATION ability to access the deflection data off-site reduces the cost of data collection by
minimizing the number of required site visits to download data. More importantly,
The system has performed well since its completion in 1998. No alarms however, is the ability for the Tri-Met engineering staff to immediately connect to
have been detected to date by the warning system and the rates of movement the RMU’s and evaluate the data when an alarm condition is detected by the
indicated by the instrumentation surrounding the Station have been consistent with operations center.
the predicted values, as indicated below in Table 3.
CONCLUSIONS
Table 3
Total Post-construction Movement and Creep Rates For some projects, moving to a site with ideal geotechncial engineering
Between July 1996 and December 1999 conditions is not an option. The Washington Park Station is a good example of a
project where difficult geologic conditions had to be incorporated into the design
Inclinometer Total Movement Annual Creep Rate and operation of the facility. The long-term performance monitoring provides a
centimeters centimeters/year level of assurance that the Station will perform in a safe manner, and that if
(inches) (inches/year) necessary, a timely response will be available to allow emergency actions. Further,
B-585 0.20 (0.08) 0.05 (0.02) the accumulated data would allow for the design and implementation of remedial
measures to enhance stability of the landslide mass, if ever needed in the future.
B-584 0.15(0.06) 0.05 (0.02)
B-583 0.15 (0.06) 0.05 (0.02)
108 GEOTECHNICAL MEASUREMENTS

ACKNOW LEDGMENTS

The design and construction of the Westside Light Rail Transit project was
accomplished under the direction of Tri-Met. Parsons Brinckerhoff Quade &
Douglas (PB) was the lead Civil Design firm: Squier Associates was the prime
geotechnical consultant under contract with PB. The authors acknowledge the
contributions to the monitoring portion of the project of Gary Peterson, Project
Manager, Squier Associates; and David McAllister, Tunnel Design Manager, and
Phillip Rice, Lead Geotechnical Engineer, of PB. The Project Director of Tri-Met W h arf Movements Due To Tidal Fluctuations
was Tuck Wilson; the Civil Engineering Director was Robert Chow, and the
Resident Engineer was Joe Gildner.
Jonathan E. Thomas, P.E.1, M.ASCE
Wilbur J. Morin, P.E.12, F.ASCE

Abstract

Movements o f a pile-supported wharf due to tidal fluctuations acting on unstable

sediment mounds are described and analyzed. A 244 meter long by 61.5 meter
wide general cargo wharf (Berth 14) at Port Klang, Malaysia has moved
approximately 135 millimeters horizontally since its construction in 1963. Results
of a detailed investigation are summarized, along with an analysis conducted for the
structure. A detailed geotechnical study was completed to characterize the soil,
both within and underlying the sediment mounds down to firm bearing horizons for
deep foundations. Additionally, a short-term tidal fluctuation movement­
monitoring program was completed to measure the movements of the wharf during
the peak tidal cycles. The gaps at the expansion joint between the access trestle and
the wharf were measured at high and low tides for correlation purposes. The data
gathered in the field confirmed movement o f the wharf and supporting piles.
Secondly, it allowed determination of maximum permissible dredge depth for the
berth. Thirdly, it provided a determination of the cause of the movement. A
correlation between the measured movements in the berth and empirical
relationships for lateral load on piles was established using the procedure
documented by C. Vigianni (1981) and a finite element model of the berth.

Introduction

This paper describes the movement of a general cargo berth due to tidal fluctuations
and unstable sediment mounds that have accreted under the structure. A detailed
study was completed regarding the stability and serviceability for all of the port
facilities (Berths 12 through 21) at North Port, in Port Klang, Malaysia. The
facilities at Port Klang are located approximately 50 kilometers west of Kuala
Lumpur in the State of Selangor on the West Coast of Peninsular Malaysia. Direct

1Moffatt & Nichol Engineers, 2209 Century Drive, Suite 500, Raleigh, NC 27612 (919)781-4626
2Consulting Geotechnical Engineer, 6805 Lemon Road, McLean, VA 22101 (703)790-0125

109
I 10 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 1 11

access to the facilities is provided from the Straits of Malacca, which is one of the slope stability analyses for different conditions, and confirmed the root cause of the
busiest shipping lanes in the world. This paper focuses on Berth 14 at the site due movements. Additionally, a correlation between the measured movements in the
to space limitations and the severity of the movement problem at that berth. berth and empirical relationships for lateral load on piles was established.
Significant movement of the different berths at North Port was noted and studied in
1986 with a final conclusion that the accreted sediment mounds would not have a Sedimentation History
detrimental impact in the future. However, in March 1997, Berth 14 translated
enough, approximately 120 millimeters, that the trestle access bridge (Bridge 7) at Prior to the construction of the berths, a natural hydrodynamic pattern existed in the
the north end of Berth 14 fell off of its bearings. A detailed study was executed in Klang Delta, which encompasses the North Port site. Coode & Partners (1986)
1998 to detennine the cause of the movement(s) in the berths, the magnitude of the commissioned by the owners of the facilities, the Kelang Port Authority (KPA)
problem, and to identify alternatives to provide a long-term solution. present the physics of this existing condition in a thorough and detailed report. The
report concludes that the Klang Delta is in unbalanced equilibrium and suspended
The facilities at North Port were conceived over a 55-year period from 1907 until a solids coupled with periodically low-currents lead to a slow build up of sediment
contract was signed in 1960 to construct the current Berths 11-14. The wharves throughout the delta. Clearly, the construction of the pile-supported berths at the
proposed were to consist of 760 meters of pile-supported structure, four transit North Port has changed the sedimentation regime. The first berths (Berths 11 - 14)
sheds, and approximately 10 hectares of land reclamation with rail access. Opening were constructed at the North Port in 1963 and, as a consequence, the local
ceremonies were held for the new facilities in December 1963 according to Din hydrodynamic system was altered. A direct result was the slowing of water
(1981). The other berths (15-21) at North Port were constructed at different times velocities in the vicinity of the piled structures and where the velocity dropped
from 1964 to 1990. Figure 1 provides an overall plan view of the North Port and below critical levels, sediment accumulation occurred. Figure 2 provides a profile
the different berths. Berth 14 at North Port is 244 meters long and 61.5 meters of the sediment mounds that accumulated under Berth 14 from 1947 to 1998. A
wide. It is a general cargo berth that is capable of handling vessels up to 30,000 review shows that over 10 meters of accretion has occurred under the structure.
Dead Weight Tons (DWT). The deck is a reinforced concrete one-way slab
supported on transverse and longitudinal reinforced concrete beams with cast-in­
6U9m
place reinforced concrete pile caps and spun precast concrete cylinder piles.

Lend Facilities a&55m(ACPl, ......

MHXS +5.30m (ACOi [V T/-Q~-y—u—nj jLrr~u lttuttj Injun. ittiit~u in vru u~ar~vToT~u u~~ur-ij tF'vnl

r 1 1 °t~ ~ ï t \ 1 T_JT— 1— T ; T °-"r~'T — T---- \cPlj L'T ? T T T ' -----1-----r~ 1 1 1---- »----,—/ / ^ — T
■fri i ~ ‘r i 1 “ it
Bridge ¿10 f Bridge ¿9 J J Bridge ¿8 f S r ijg e #7 | ff
B ridçe i '. 2 { Bridge # ’ 1 t
L a n d s id e X- , Landside,- _ X■ S ^ Landside Lands.de / / s
V 77777//M
y / / / / / / '/ \ V77777A V 7777m
Der In 18 1 Be th 17 1 B e rth 16 1 E e 'lh 15 i B e rth U ! g t . , n , i ! = „ , t r ., .

1 *
N orth Keiorrç S tro lls
~n

O 50m 100m A

Surface Profile - Circa 1SB6

—Surfaco Profile - July 1998 (Source; Coodo and Partners)
Figure 1: North Port Site Plan - Port Klang, Malaysia (Source: KPM) Surface Profile - C'rca 19S4
Surface Profie - Circa 1997 (Source: Drg. No. 3, Klang STR)-
(Source: Coode end Portncrs)
Surface Profile - C'rca 1947
Moffatt & Nichol Engineers and DPI Konsult Sdn. Bhd. were retained to conduct a (Source: HMS Sharpshooter Survey) -

detailed study into the siltation, stability, and serviceability of the North Port Berths
in 1998. A variety of concerns about their structural integrity had arisen in recent Figure 2: Seabed Surface Profiles (1999) - Berth 14 North Port
years due to obsen/ed damage and significant movements of the wharves. In order
to address the concerns and to provide the information necessary to make the Geotechnical Information
technical evaluations, a series of programs were developed and executed. These
included a formal geotechnical investigation, berth movement monitoring, A formal geotechnical program was completed at the site from May until August
structural condition inspections and analyses, and development of remedial 1998. The study included a field and laboratory program to characterize the soil,
measures and improvements. The information gathered through the work programs both within the sediment mounds and also the underlying soils down to firm
confirmed movement of the wharf and supporting piles, allowed completion of bearing horizons for deep foundations. The complete program for the North Port
112 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 113

included a total of 33 soil borings along with 30 deep soundings using a piezocone the soils at various depths. The inclinometers indicated movement of the soils and
penetrometer (CPT). The borings and field-testing locations were configured to are discussed later in this paper.
provide longitudinal profiles in the front, center, and rear of all berths.
Additionally, the configuration provided transverse sections perpendicular to the
profiles. The laboratory testing provided index properties o f the various strata
along with strength and consolidation properties of the soil.

The geotechnical investigation confirms that the sediment in the mounds is

predominantly very soft silty clay, as indicated by the zero-blow count (N values)
obtained during Standard Penetration Tests (SPT). The sand content is generally
less than 10% and the remaining soil content consists of silt and clay (and colloids).
Texturally, silt predominates, but the plasticity produces a classification of high
plasticity silty clay (CH). The soil stratigraphy at the site, shown in Figure 3, can
be described as the relatively recently accumulated very soft silty clay overlying
sands. The combined thickness of the sediment mounds and the underlying silty
clay vary from 18 to 25 meters. Typically, below a depth of 10 to 15 meters, some
measurable blow-count values are recorded, but these are usually less than 5
indicating soft to very soft soil. Sand underlies the silty clay sediment mounds and
it varies from medium to very dense.

Figure 4: Bulk Density vs. Depth - Figure 5: Plasticity Chart-

BRIDGE 7 LAND Berths 12-14 Berths 12-14

Figure 3: Soil Stratigraphy (1999) - Cross Section Berth 14 North Port

Laboratory testing provided information regarding the bulk density, shear strength,
gradation, Atterberg limits, and relative density of the soils at the site. Figures 4
through 6 provide laboratory test results. Additional field testing was completed
including the installation of five inclinometers to detect differential movement of Figure 6: Field Vane Shear Results vs. Depth - Berth 14
I 14 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 115

Stability of Sediment Mounds Inspection of the surface profiles shows that the slope on the face of the sediment
mound tends to change with time. This is an indication that the sediment mounds
One of the main goals of the study was to ascertain the stability of the existing are not completely stable under the current conditions. The stability of the mound
sediment mound under the berths. The May 199S seabed surface profile shown in at existing slope-angles was examined using STABL. The results of the analysis
Figure 2 was used as the baseline case for the slope stability analyses of Berths 12 - indicate that failures manifest themselves in the form of localized surface
14. For compulation of mound stability, it was assumed that there is no added soil sloughing. The sediment mounds are significantly unstable with factors of safety
support offered by the piles. While the piles do provide support, the 6.09 meter near or below one. More restrictive searches, such as restricting the exit points
bent spacing of the piles is considered too great to offer anything more than further away from the face of the slope results in deeper failures, albeit with higher
localized support. In any case, disregarding the piles is considered a more factors of safety indicating there is reduced or no risk of deep-seated failure.
conservative approach. Janbu and Sliding Block slope stability analyses using
STABL were performed for Berths 12-14. Remolded vane shear values for the
berths were applied to depths of 8 meters and total stress sliding block analyses NORTH PORT KLANG
performed. This is considered appropriate for undrained conditions, where the soils N.P.K 13 N.P.K. 13
Conc.'iti siob
are stressed rapidly compared to their drainage. The shear strength of the soil
appears to be between the remolded and the peak undrained strength. Remolded
values to depths of 8 meters were selected because movements to that depth were
indicated by the inclinometers. The analyses indicated that an unstable slope exists
with shallow slides imposing loads on all rows of the piles as shown in Figure 7.

INCR.D!$P.‘.A;{rrjn
I

Figure 7: Slope Stability Analysis Results for Berth 14 Figure 8: Inclinometer Results - Berths 12-14, North Port
116 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 117

The lateral loads on the piles from the outward movement of the soils in the the slope oi the lines is positive. A review of past measurement programs
sediment mounds are considered the root cause of the movement of the berths at undertaken at the North Port provides insight into the permanent net movement.
North Port. Ting and Tan (1997) first noted the effect of'tidal action on unstable Table 1 summarizes the available data lor Bridge 7 and shows a net seaward
slopes behind wharf structures’ in their study o f the movement of a wharf structure movement.
in East Malaysia. Using inclinometers and strain gauges, it was discovered that at
low (tidal) water levels, the unstable slope had a tendency to move outward, Table 1
Bridge 7
imposing a loads on piles. At high water levels, the movement reversed. However,
Date Bridge 7 Berth 14
it was observed that some 'residual' or net outward movement remained over time.
South Edge | North Edizc
The reason behind the residual displacement of the wharf lies in the inability of the Jun-86 Not visible - Asphalt
supporting soil mass to recover to its original state. While the structure is capable Jun-95 Not visible - Asphalt
of elastic movement, this is not the case in the soil and the residual movement 03/21/97

O
'T

o
translates to the structure over a long period of time resulting in permanent berth 06/06/97 Repaired after Bearine Failure
movement. The tidal action on the upper sediments results in a continuous cyclic 10/13/97 20 25
movement of the soil, which in turn reduces the soil's strength due to remolding. 07/17/98 25 25
Evidence of similar movement is apparent at the North Port and can be seen in the 09/25/98 to 11/18/98 37 30
inclinometer plots. Significant movement has occurred in the upper 3 to 4 meters
and can be detected to depths of 8 to 10 meters as shown in Figure 8. The findings
of Ting and Tan appear to have validity with respect to the berths at North Port. Causes of Movement

Short-Term Tidal Fluctuation Movement Monitoring Based on the history o f the expansion gap measurements and the dislodging of
Bridge 7, it is clear that significant net permanent movement has taken place in the
As part of the project, the consultants took measurements of the gaps at the access 34 years since Berth 14 was built. Three potential causes for the movement are
bridges to the berths, and noted variations occurring during differing tidal cycles. identified: (1) Translational movement of the piles; (2) Deep-seated failure o f the
The data indicated that tidal-impacts on the sediment mounds and the berths should sediment mounds and underlying strata; and (3) Deflection of the piles due to
be given serious consideration. The maximum tidal difference between Low and lateral loads from the sediment mounds. Figure 10 provides the failure modes for
High Water is approximately 5.5 meters. As such, a controlled monitoring program restrained piles taken from Broms (1964).
was initiated in September 1998 that involved measuring the gaps between the
access bridges and the rear of the berths. The program confirmed that the gaps
were varying during tidal cycles. These movements are short-term (i.c. daily with
the tidal cycle) and it is theorized that it results in a long-term net movement of the
structure seaward as indicated by the failure o f Bridge 7. Figure 9 shows the results
of the monitoring for the repaired Bridge 7 at Berth 14.

The measurement program immediately corroborated movements in the berth

during tidal cycles. The data for Bridge 7 shows that the maximum seaward
movement is 8.5 millimeters and the maximum landward movement 1 millimeter.
Those values represent the maximum change in the expansion joint gap between
High and Low Tides. It is the maximum tidal cycle movement in any single tide
cycle measured at the joint. It should be noted that the values do not represent a
permanent net movement of the berths, since the data clearly shows that the berths
move seaward and landward with the tidal cycle. The data also show that the gap at
each of the berths is directly related to tidal cycle with the widest gap measurement Figure 10: Failure Modes for Restrained Piles (taken from Broms)
during the largest tidal level fluctuation. This supports the theory advanced by Ting
and Tan. The overall short period o f observations also confirmed indications of The inclinometer results support the loading of the piles from the sediment mounds.
permanent measurable residual movements, albeit on the order of only 2 to 3 First, a translational stability analysis was carried out using STABL and a sliding
millimeters. The trend lines indicate that the expansion gaps tend to widen since block mode. This type of failure would require that the piles "walk" (similar to
I 18 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 1 19

Figure 10c) seaward from a deep-seated failure plane. The slope stability results
show high factors of safety against such failure planes even when applying very
minor penetration of the pile tips into the dense sand layer. Secondly, there is no
evidence, in our opinion, from the 1997 and current seabed profile that deep-seated
failures have occurred within the sediment mounds or underlying soils. If this had
occurred, a scarp would have formed in the seabed above the failure surface and a
bulge would have formed at the toe of the slope. The surface profile shown in
Figure 3 does not indicate any obvious evidence of such changes. It is possible that
a slump could have filled in with sediments and the bulge at the toe eroded, leaving
no evidence of the failure, however, it is considered unlikely that both these actions
could have occurred simultaneously. Furthermore, deep-seated failures could over­
stress the outer row of piles and there is no visible structural deterioration (i.e.
plastic hinge development at pile head, pile rotation, etc.) evidence that this has
occurred. Finally, the findings from the geotechnical investigation, field
measurements, and structural modeling support that the cause of the berth
movement is most attributable to a combination of actions that act concurrently and
result in the imposition of lateral loads on the piles.

The combination of actions are sliding soil mass due to unstable slope, soil
movement caused by tidal fluctuations, and regular shallow soil failures. All three
of these actions are probable and contribute to loading of the piles. It is important
to note that the degree of net permanent movement observed at the berths is not the
effects of one action or a single event, but is due to a combination of the factors
listed above, acting over time. The existing situation is the result of many years of
continuous movement and to completely understand the phenomenon, it is required
to break the time-cycle down. It is believed that as the piles deflect seaward, due to
the respective soil loads, the surrounding soils, also moving, tend to fill the voids
that are left by the movement of the piles. When the soil loads on the piles are
relieved by a high tide, the piles have a tendency to return to their original position,
due to their elastic properties. However, the voids left by the original seaward
movement are partially filled by the soil and consequently, the piles have to exert a
force to displace the in-filled soil and thereby return to their original position. As
time passes, the piles are unable to return completely to their original position due
to the in-filling process and the subsequent build-up of soil mass, which continues
to exert a load on the piles. This phenomenon occurs over a long period of time.

Empirical Determination of Lateral Soil Loads

Based on the characteristics of the sediment mounds, it is generally considered that

the upper soil stratum is exerting lateral load on the piles as outlined in the
procedure by C. Vigianni (1981). The magnitude of the lateral force acting on the
pile due to soil flow can be estimated using the empirical procedure from Vigianni.
The Vigianni procedure extends the work completed by Broms (1964) and provides
Figure 9: Expansion Joint Gap Measurements (1999) - Bridge 7 Berth 14 the ability to predict the ultimate lateral load on a pile. In the procedure, the
interaction between a cohesive sliding soil mass and a pile is controlled by the
following factors:
 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 121
120

• Length and diameter of the pile(s) • Thickness of the sliding soil mass calculated using undisturbed strength, is most likely to form the upper bound value
• Yield moment of the pile section • Undrained shear strength of both the (e.g. 154 kN) as this condition is present just before slope failure occurs. It is
stable and sliding soil masses therefore considered that the actual loads acting on the piles will lie in the upper
range of the loads shown in Cases 2 and 4, respectively.
For the purposes of this analysis, it is assumed that the lateral loading acting on the
Structural Modelling of Berths and Correlation of Lateral Soil Loads
piles will be the same in all rows that lie within the sediment mound, front to rear.
Figure 2 shows the piles at Berth 14 that are in the sediment mound. A summary of
the respective parameters used to calculate soil loads are provided in Table 2. As part of the study, there was a requirement to perform a structural evaluation of
the existing North Port berths to determine current structural capacities of the
Table 2: Parameters Used in Lateral Soil Load Analysis for Bert l 14 respective berths and also identify any limitations that may be required for future
improvements. To enable the evaluation, finite element models of each berth type
Length of Pile (m) m 2S.0
(Berths 1 2 -2 1 ) were developed. Part of the modeling process was to calibrate the
Diameter of Pile (m) mm 864 models based on the measured movements of the berths. The observed translation
Thickness of Sliding Soil Mass m 3.50 of the berths complicates the analysis, as these loads must somehow be accounted
kN-m 576 for in the models. In order to account for the overall movement of the structures, a
Yield Moment of Pile Section
global model of each berth type was developed.
Undrained Strength of Sliding Soil Mass kPa 1.6
Undrained Strength of Stable Soil Mass kPa 30.2
The model was developed using the finite clement structural software program
STAAD-III and consists of three and four-node slab elements, beams, and pile
Using the data in Table 2, the ultimate lateral loads acting on the piles have been members. The model is three-dimensional and accurately reflects the bent spacing
estimated using the Vigianni method. The loads shown in Table 3 represent the and the layout of the supporting beams and deck slab. It incorporates the full
estimated load acting on an individual pile. The total load acting on the berth is structure from expansion joint to expansion joint in the berth along with accounting
therefore determined by multiplying the (per-pile) soil load by the number of piles for the non-linear soil-structure interaction between the piles and the soils (i.e. soil
springs). The soil springs were developed using a program (LTBASE) that
in each berth.
conducts a nonlinear analysis of laterally loaded piles taking into account the
Table 3: Estimated Lateral Loads Acting On Piles per Vigianni engineering properties of the soil including the modulus of subgradc reaction, shear
Estimated strength, and unit weight. Figure 11, on the following page, shows an isometric and
Case
Lateral Load side view of the Berths 12-14 model. Results of the expansion joint measurement
Description per Pile (kN) program and empirically derived lateral earth loads on the piles were input into the
Soil failure where rigid pile movement develops using 19 model to calibrate and check the magnitude of soil loads acting on the structures.
1 remolded shear strength of soil. The results of the analysis indicate that the resulting loads for Berth 14 falls within
Soil failure where rigid pile movement develops using 67
2 the estimated range determined by the Vigianni approach. This supports the
undisturbed shear strength of soil.
Soil failure where a plastic hinge is allowed using remolded
assumption that the shear strength of the sediment mounds lies between the
94
3 shear strength of soil. remolded value and the undisturbed value. Table 4 summarizes the results of the
Soil failure where a plastic hinge is allowed using 154 analysis for Berth 14.
4 undisturbed shear strength of soil.
Table 4: Earth Loads Per Pile for Berth 14
In the absence of any significant visual signs of structural distress during the site Load/Pile Determined by Structural Model and
116 kN
inspections, it is reasonable to assume that the plastic hinge-type failure (Cases 3 Expansion Joint Measurements
and 4) is not occurring. While there is movement of the upper soil strata evident, as Load/Pile (Lower Limit) Estimated by Vigianni
67 kN
Approach (Case 2)
shown in the inclinometer results, the amount of movement is not considered
Load/Pile (Upper Limit) Estimated by Vigianni
extreme. Therefore, due to the minimal amount of soil movement it is assumed that 154 kN
Approach (Case 4)
the in-situ shear strength of this soil layer most likely lies somewhere between the
remolded and undisturbed strengths. The estimation of lateral lorce on the pile,
122 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 123

because of sloughing of the mounds, thus requiring considerable maintenance

dredging.

Dredging the sediment mounds completely from underneath and behind the berths
will effectively remove any excess lateral loads since it eliminates the mechanism
that causes the movement. Conventional dredging methods, unfortunately, cannot
be used to clear the sediment beneath the berths because of access problems. From
a geotechnical perspective it is desired to dredge the existing sediment mounds to a
lv:3h slope from the front of the berths so as to maintain slope stability. However,
the lv:3h dredge slopes docs not eliminate the earth loads being imposed on the
piles and other structural solutions would still be required to prevent movement of
the structures.

Lateral loads on a piled structure will cause that structure to move in the direction
of the applied load. The existing structures are composed entirely of plumb piles
that rely on the moment interaction between the piles and the bent cap to resist the
lateral loading and to restore the structure to its' original position. Under original
design conditions (i.e. no lateral earth loads) Berth 14 would be adequate to resist
lateral berthing and mooring loads. However, development of the sediment mounds
under and behind the berths has created excessive lateral loads on the piles and has
resulted in permanent long-term movement of the structures. A common and
effective approach used to resist excessive lateral movements in marine structures is
to install batter (raked) piles. It was concluded that installation of batter piles in this
situation was a practical option for constraining the berth movement and subsequent
structural analyses were performed to determine pile size and structural
configuration. The general configuration determined is such that two piles are
installed either side of the bent caps, at the rear of the berth. Piles are angled both
seaward and landward, to account for two-way movement, and attached to the
existing structure by way of a reinforced concrete cap. Based on information from
Figure 9: Three-Dimensional STAAD-III Model - Berths 12-14 construction contractors, it is proposed that the new batter piles would be installed
at an angle of lh:4v, which is considered to be the maximum practical angle at
Remedial Measures
which the piles can be installed. This is a non-displacement method o f pile-
installation with no increase in pore water pressure that may cause slope failure of
The most critical problem at North Port relates to the long-term, net movement of
the surrounding soil. The proposed connection and batter pile members were
the berth structures due to the earth loads from the sediment mounds induced by
analyzed via the global structural model of Berths 12-14. The results indicate that
tidal fluctuations. Eventually, if the movement is not curtailed, the piles will all
this solution is very effective. Table 5 summarizes the results of the analysis and
start to show signs of distress (cracking, spalling, etc.) due to movement of the pile indicates that joint movements are considerably reduced after installation of the
head relative to the point of fixity.
battered piles.
The sediment mounds are the main cause of several problems at North Port. The Table 5: Batter Pile Solution - Structural Modeling Results for Berths 12-14
development of the mounds underneath and behind the berths has primarily caused
Original Joint Retrofitted Model Maximum
movement of the berths. Secondly, the increased soil depth has reduced the Berth Movement Joint Movement Compressive Pile
effective pile length on the back rows of piles, which results in an increased relative A (mm) A (mm) Load Pnm(kN)
stiffness of the piles. This causes the rear rows of piles to carry a larger proportion 12-14 10.00 1.44 2571
of the lateral loads since they are the stiffest. Additionally, the design dredge
depths of the different berths have generally been very difficult to maintain,
124 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 125

Conclusions References

The movement of the berths at North Port in Port Klang is a serious problem. It has Broms, B. B., “Lateral Resistance of Piles in Cohesive Soils”, Journal of Soil Mechanics and
Foundations Division, American Society of Civil Engineers, Vol. 90, New York, 1964 pp. 27-63.
resulted in a catastrophic failure at the bearing for one access bridge and the
movement continues. The accretion of sediment mounds under the berth has led to
Coodc and Partners, "Study of Siltation and Ancillary Problems in the Navigation Waterways of the
inherent instability in the mounds, movement of the soils in the mound from the Klang Delta", 1986. Report provided by the Owner’s of Berths 12-21 at North Port
large tidal fluctuation at the site, and subsequent lateral loading on the piles causing
fluctuating movement(s) of the berths. The net permanent movement is seaward. Kelang Port Management (KPM) Sdn. Bhd.
This paper discusses the data gathering and engineering effort that went into a 1998 Tingkat 2, Blok A, Bangunan Pentadbiran
Off Jalan Parang, Pelabuhan Utara,
study that addressed the overall stability, serviceability, and development of 42005 Pelabuhan Klang
remedial solutions for the problems facing the berths. The slope stability analyses Attn: En. Mohd. Yussof Bakar
completed showed that the sediment mounds are unstable and need to be cut to a
lv:3h slope from the front face of each of the berths. Additionally, an empirical Din, Harun, Barlow, P.G.R. and Wong, K.H., "Port Kelang - The First 80 Years", SEATEC III
determination from Vigianni (1981) was utilized to establish the lateral loads being Seminar, Asian Ports and Development, March 1981. pp. 1-13.
imparted on the supporting piles for the berths by the sediment mounds. Detailed
Ting, W.H. and Tan, Y.K., "The Movement of a Wharf Structure Subject to Fluctuation of Water
structural modeling was utilized to address the stability and serviceability of the Level", 14th International Conference on Soil Mechanics and Foundation Engineering, Vol. 1,
berth structures. Part of this included calibrating the models based on measured Hamburg, 1997 pp. 903-906.
expansion joint gap movements during tidal cycles and comparing the results with
the Vigianni loads. The results indicate that the loads fall somewhere between the Vigianni, C., "Ultimate Lateral Load on Piles Used to Stabilize Landslides", 10lh International
Conference on Soil Mechanics and Foundation Engineering, Vol. 3, Stockholm, 1981. pp. 555-560.
remolded shear strength values and the undisturbed values which was anticipated.

The recommended solution to the movement problem was to install batter piles on
the back face o f the berths. The analysis and modeling shows that the retrofit
should be very effective in arresting the landward and seaward movement and it is
relatively inexpensive when compared to replacing the berths. Part of the
recommendations for this project were to install steel plates and strain gauges at the
expansion joints between the access bridges and the back of the berths to measure
the actual loads causing the berth movements. This information would prove to be
invaluable in confirming and designing the retrofit to the berths.

The data and information available regarding this subject is very limited and further
research and data gathering would prove to be invaluable. Future projects could
include the following:

Methods to dredge accreted material from under pile supported

structures.
- Installation of strain gauges to measure the actual loads exerted on the
structures.
Development of other solutions to the movement problem

Acknowledgements

The authors would like to acknowledge the assistance provided by the staff of
Kelang Port Management (KPM) for this project. Without their patience and
professionalism the success of the project and ultimately this paper would not have
been possible.
 GEOTECHNICAL MEASUREMENTS 127

monitor the performance of remediated landslides. Once inclinometers have been

installed, the geotechnical consultant is sometime requested to determine the
effectiveness of the mitigated landslide in a limited time frame However, analysis of
the monitoring data may result in erroneous interpretation of slope movements, when
the mitigated landslide is performing adequately.

Where remediation involves engineered fill, the engineer should recognize that fills
consolidate and settle under their own weight. Movements are also seen from changing
development conditions, hydrocompaction, and the weight of significant structural
Inclinometer Data Analysis for Remediated Landslides loading. In addition, the earth material below the fill may consolidate or heave due to
the changing subsurface environment, such as increased moisture exposure due to
William R. Stevens, Member ASCE1and Bill E. Zehrbach, Member, ASCE12 landscape irrigation The resulting movement may occur over both short and long
periods of time Additionally, engineered fills placed upon hillsides will be subjected
to differential settlement consistent with fill thickness. A complicating factor is surficial
creep due to the shrink and swell behavior of expansive soil.
Abstract
An example of some of these conditions was encountered in Northern California where
Landslides are frequently remediated by constructing engineered fills, improving landslides were stabilized using engineered fill. Misinterpretation of the gathered data
drainage, or other more specialized construction methods. Monitoring instruments, could have occurred; however, the use of certain analyses and methods of data
such as surface monuments and inclinometers, are sometimes installed to evaluate the presentation clearly showed that the mitigated landslide was performing within
performance of the remedial measures. Where remediation involves engineered fills, expected limits. A large graded parcel in the Bay Area of California was instrumented
the engineer should recognize that compacted fills undergo an equilibration process that with inclinometers, survey monuments and piezometers in several remediated
can take years. This process can involve heave caused by expansive soil and landslides. Instrumentation data from Area 1-5 in this development is presented.
consolidation due to the weight of the fill and imposed structural loads. Additionally,
fills placed upon hillsides can be subjected to differential settlement consistent with fill Site History
thickness and development changes, surface creep, and lateral extension. Monitoring
data can indicate various subsurface movements which are a product o f settlement of Area 1-5 is a mitigated landslide within a 1,200-acre residential and golf course
the fill mass and lateral extension, and not related to movement of the remediated development that was investigated in the 1980's. The geotechnical reports for the
landslide. Some of these conditions were encountered at a site in Northern California development provided recommendations for the remediation of several landslides,
Misinterpretation of the gathered data could easily have occurred, possibly initiating including Area 1-5, situated within the project. Area 1-5 was found to be underlain by
unwarranted remedial measures or preventing development However, the use of an approximate 12.2 meters (40 foot) thick landslide complex. Figures 1 and 2 show
certain analytical and data presentation methods clearly showed that the remediated a plan view and cross section, respectively, of the mitigated landslide. The landslide
landslide was performing as designed was located on a southeasterly facing slope and was about 100 meters wide and 228
meters long (330 feet wide and 750 feet long). Remediating the landslide included the
Introduction complete removal of the upper portion of the landslide and the reconstruction of this
area with engineered fill, keying and benching into bedrock, overexcavation of soft
Landslide remediation frequently utilizes engineered fills and improved drainage material, and the installation of keyway and bench subdrains. The central portion of the
techniques, as well as other, more specialized construction methods. Instrumentation landslide complex was removed by cut grading, creating a fairly level area at the toe of
is sometimes installed to evaluate the effectiveness of the mitigation measures the engineered fill. The lower portion of the landslide was left unmitigated. Grading
Inclinometer casing and survey monuments are some methods that can be utilized to of Area 1-5 occurred in 1988 During grading of a nearby site, several thousand cubic
meters of soil about 4.5 meters (15 feet) thick were temporarily stockpiled at the top
1Senior Engineer, Consolidated Engineering Laboratories, 7060 Koll Center Parkway, of Area 1-5 in the fall of 1995.
Suite 300, Pleasanton, California, 94566
2 President, Earth Systems Consultants Northern California, 47853 Warm Springs
Boulevard, Fremont, California, 94539

126
128 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 129

The study area is underlain by the Panoche Formation, which contains claystone,
mudstone, siltstone, and shale (ASMI, 1988). The development site is located in a
long, moderately wide valley with northwest-southeast trending ridges on both sides
Local areas of diorite and silica carbonate occupy the higher, more resistant ridges.
Tectonics have sheared the Panoche Formation, creating zones of weak, sheared rock.
The soils on the hillsides formed from the Panoche Formation consist of silty and sandy
clay Typically, the material within the landslide masses are clayey soils with a
Plasticity Index on the order of 24 to 32 with Liquid Limits o f 40 to 50. Clay gouge
along shear planes possess Plasticity Indices in the range of 21 to 63 and Liquid Limit
values from 41 to 84. The in-situ water content of the landslide mass material varied
from 17 to 24 percent. The engineered fill that was utilized to remediate the landslide
was derived from bedrock materials excavated from other parts of the site The
resulting fill material, compacted to a minimum 93 percent relative compaction (ASTM
D 1557-91), is classified as a clayey gravel to gravelly clay.

Instrumentation

Two sets of instrument clusters consisting of an inclinometer casing, open pipe

piezometer, and a survey monument were installed at Area 1-5 in December 1991 at
Sites 1 and 2 of Area 1-5 Site 1 is located in a level portion of a 6.1 meter (20-foot)
deep cut at the toe of the engineered fill as shown on Figures 1 and 2. Site 2 is located
Approximate Scale in Meters within 10 feet of the top of the 9.1 meter (30-foot) high engineered fill slope The
inclinometer casings were installed to a depth of 8.2 and 18 6 meters (27 and 61 feet),
Figure 1 Plan view of mitigated landslide. The shaded areas are mapped landslide for Sites 1 and 2, respectively. Survey monuments were installed on top of 15 cm (6-
deposits and the hatched area show limits of Area 1-5 engineered fill inch) diameter, concrete piers reinforced lull depth by a #5 bar and extended about 1.5
meters (5 feet) into the ground Baseline readings for the inclinometer casings and
piezometers were obtained in December 1991 and for the survey monuments in May
1992.

Instrumentation Data

A total of 8 surveys and 15 inclinometer and piezometer readings were performed for
Area 1-5 between September 1992 and November 1998. Electronic distance
measurement surveys were taken on the survey monuments to determine the easting,
northing, and elevation movement of the monument. The monitoring program at the
development also included rainfall measurements collected daily by the golf course
superintendent, subjective subdrain flow information once or twice a year, and annual
geologic reconnaissances. The information contained in this paper discusses
inclinometer, survey, and rainfall data.

The inclinometer data were analyzed utilizing DigiPro, an inclinometer data analysis
computer program, by SINCO, which can create plots of Cumulative Displacement,
Cumulative Deviation, Incremental Deviation, Incremental Displacement, Incremental
Deviation, and other useful graphs. (It should be noted that the DigiPro program could
Figure 2. Cross section AA’ through Area 1-5 landslide and engineered fill. not generate metric scale graphs, Figures 3 through 5, when the inclinometer equipment
130 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 131

has been calibrated and constructed to English units) A spreadsheet program was movement may indicate that the engineered fill may be going through a period of
utilized to analyse the inclinometer and survey data to create Horizontal Vector adjustment (Al-Homoud, et al, 1995 and Day, 1991) The readings obtained between
Movement, Resultant Movement, and Rate of Movement graphs. Each of these graphs November 1993 to July 1995 are fairly stable, without significant movement The
shows different aspects in the interpretation of instrumentation data second episode of movement is the readings between July and October 1995. After
October 1995, the readings are again stable.
Cumulative Displacement - Figures 3 and 4 present the cumulative displacements of the
inclinometer casings from the baseline readings for Sites 1 and 2, respectively. The
base of the casings are assumed to be fixed. Displacement indicated in the data is
assumed to take place in the overlying materials. Note that the displacement scales for
both figures are the same, which can be important when comparing inclinometer data
from several sites.

!
Î

' i
11/ 04/ 94;
. !)09/ 17/92
■305/ 19/93
J11/ 03/93
i 03/ 23/94
. 01/ 20/95
a04/ 19/95:
)07/ 03/95
110/ 28/ 95;
302/ 28/96!
- 08/ 1-3/ 96- 50-
701/ 31/97
}06/ 05/97
: 07/ 08/ 9B
-.11/ 23/98 50
Figure 4 Cumulative Displacement inclinometer graph of Area 1-5, Site 2.

70----- ‘---------------------j 7-00.4

----- ‘------------------
0.4 12 Incremental Displacement - Figure 5 shows the Incremental Displacement graph for
-04 0.4 1.2
A - Axis, Inch B - Axis, Inch Site 2. Incremental Displacement is the amount of variation for each 60 cm (2-foot)
interval reading obtained from the inclinometer probe relative to the baseline reading.
Figure 3 Cumulative Displacement inclinometer graph of Area 1-5, Site 1. Movement zones can be detected where the graph of the reading being analyzed
deviates from the baseline reading (zero point). The A-axis plot shows movement at
The Cumulative Displacement graphs are initially utilized to determine if movement depths of 12.5, 10.7, 9.5, 8.8, and 8 2 meters (41, 35, 31, 29, and 27 feet). Minor
indicated in the data is significant. Since less than about 1/4 inch (considered to be the movement is occurring in the zone above a depth of 7.6 meters (25 feet). Surficial
approximate accuracy of the inclinometer results) of movement was measured for Site creep movement can be seen at a depth of 30 cm (1-foot). The B-axis plot shows
1 at the toe of the engineered fill, further data analysis of this site was considered movement at depths of 15.5 and 10.7 meters (51 and 35 feet), with minor movement
unnecessary. However, at the top of the slope at Site 2, the monitoring showed between 12.5 and 8.2 meters (41 and 27).
movement in the upper 12.8 meters (42 feet ).
Horizontal Vector Movement - The Horizontal Vector Movement graph on Figure 6
The Cumulative Displacement graph for Site 2 shows two episodes of movement. The shows the magnitude and direction of movement. The graph shows the resultant
first episode consists of the initial three readings through November 1993. This magnitude of movement for the A and B axes of the inclinometer and the easting and
132 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 133

northing survey readings. The survey data is in the horizontal plane, and does not 4.0
include the elevation readings. Figure 6 contains the Horizontal Vector Movement
graph for the 3-foot depth inclinometer and survey monument readings for Site 2. The 30
3-foot depth was selected for analysis to minimize the effect of surficial creep. The
arrow shows the downslope direction at Site 2 20

I0

00

-10 --EP isode 1__

- 2.0

Episode
2{ \
-3.0

-40
-4.0 -3 0 -2 0 -1 0 00 10 2.0 3.0 40

-O - Inclinometer, 1 meter depth —0 —Survey Monument |

North is Up, Arrow Shows Downhill Direction. Graph a in cm

Figure 6. Horizontal Vector Movement graph of inclinometer and survey monument

data.

Figure 5. Incremental Displacement inclinometer graph of Area 1-5, Site 2.

The movements plotted in the Horizontal Vector Movement graph are generally
downslope towards the south-southeast. Two episodes of movement are seen in the
inclinometer plot on this graph that correspond to the cumulative displacement graph,
Figure 4. The first episodic movement to the south are seen in the initial three readings.
The next several readings are grouped around 18 to 2.0 cm (0.7 to 0.8 inches) of
southward movement, but show a small amount of movement eastward The
movements (second episode) then jump to 3.3 cm (1.3 inches) southward, the last
several readings are stable. The survey monument data show a similar pattern of
movement Figure 7. Resultant Movement graph of inclinometer and survey data.

Resultant Movement - Figure 7 is a chronological graph of the magnitude of the The initial readings of inclinometer movement through November 1993 shows about
resultant movement for the survey data and the 3-foot depth inclinometer reading. This 1.8 cm (0.7 inches) of movement, which slowly increases to about 2 cm (0.8 inches)
graph also shows two episodes of movement similar to those seen in the Cumulative through July 1995. The inclinometer readings then show an acceleration of movement
Displacement and Horizontal Vector Movement graphs to 3.3 cm (1 3 inches) in October 1995 Significant movement has not occurred since
then The survey data graph shows similar, episodic movement
134 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 135

Figure 7 also shows the seasonal, 90-day, cumulative rainfall totals for each rainy The survey data shows similar movement indications. The rate of movement indicates
season from 1991 to 1999. Correlations between the 3, 5, 10, 20, 30, 60, and 90 day that in about 30 years, the area around Site 2 would have moved 30 cm (11 8 inches),
cumulative rainfall and the onset of slope movement were examined However which may not be within acceptable limits for some structures constructed at the top
quantitative correlation between the amount of rainfall and hillside movement is not of this slope However, after about 3-1/2 years of data collection, the amount of
possible due to the low frequency of displacement data collection. If survey and movement appears to have stabilized at an average of about 0.3 cm (0.12 inches) per
inclinometer data were collected on a weekly or bi-weekly basis, correlations may have year This equates to about 15 cm (5.9 inches) in 30 years.
been possible
A soil stockpile was placed at the top of the slope in the fall o f 1995 for a short period
Rate of Movement - The inclinometer movement rate for Site 2 at the 3 foot depth for of time. The October 1995 inclinometer and survey monument reading showed an
the inclinometer is presented in Figure 8 The short term rate of movement is increase in the amount of movement soon thereafter. The inclinometer data showed
calculated by taking difference in the amount of movement from the reading being about 3.3 cm (1.3 inches) of total movement after the stockpile was placed at the top
analyzed to the previous reading, and then dividing by the duration between the two of the slope, which is an increase of about 1 3 cm (0.5 inches) since initial stabilization
readings. This graph also shows the two episodes of movement seen in the other in 1993 Likewise, the rate of movement increased from 1.1 cm to 3.7 cm (0.43 to 1.46
graphs. The movement rate was about 1.1 cm (0.43 inches) per year during the first inches) per year. Since the fall of 1995, the inclinometer data has remained fairly
two years. Between about October 1993 and August 1994, the rate of movement was constant at about 3.5 cm (1.4 inches) of movement and the rate o f movement has
cyclical, but averaged about 0.5 cm (0.20 inches) per year. The rate of movement remained low.
increased in the second episode of movement to approximately 3.7 cm (1.46 inches) per Figure 7 also plots a hypothetical extension of the inclinometer resultant movement if
year in the fall of 1995. Since then, the rate has decreased to about 0.3 cm (0 12 the October 1995 reading was deleted. This would remove the movement seen in the
inches) per year Additionally, the rate of movement appears to be cyclical, with more inclinometer data caused by the placement of the temporary soil stockpile. The
movement occurring during the winter months than in the summer. engineered fill then appears to have stabilized at a cumulative amount of movement of
about 2.3 cm (0.9 inches).

The process of reconstructing a landslide mass involves making a cut into the hillside,
which could produce stress relaxation in the underlying rock mass and in the material
being cut. The corresponding strains from stress relaxation can occur relatively rapidly,
and may continue for an extended period of time after the cut has been made. The
stress relaxation can be due to many factors: stress relief from overburden removal,
chemical and mechanical weathering of the exposed bedrock after being exposed to the
elements during cut grading, and increased subsurface water flow (Anderson and
Richards, 1987 and Bowles, 1979). High, locked-in stresses could be present in the
rock mass from tectonics, rapid denudation, and topographic effects (Jaeger and Cook,
1979, and Franklin and Dussealt, 1989).

Other forms of movement that may be seen in the instrumentation data is adjustment
of the engineered fill to changing environmental conditions (Lawton, et al, 1991, Al-
- o - Short Term Rate o f Movement
Hamoud,etal, 1995, Day, 1991, and Sherard, et al, 1963). This can include expansion,
hydrocompaction, lateral deformation towards an open face, and surface creep The
Figure 8. Short Term Rate of Movement graph of inclinometer data. fill material is placed at a certain moisture content and compaction effort, usually above
optimum moisture content and generally between 90 and 95 percent relative
Conclusions compaction After construction, the fill material may undergo changes to adjust to a
new state of stress and water content that is different than the conditions during
The first two years of inclinometer measurements may indicate that the mitigated construction. This phenomena can be seen in the two episodes of movement shown in
landslide may not be performing as anticipated. The data from the inclinometer at the Figures 4, 6, 7, and 8 The first episodic movement could be attributed to the
top of the slope, Site 2, shows about 2 cm (0.8 inches) of movement and a short term stabilization of the engineered fill after construction. The second episode is the
rate of movement of approximately 1 cm (0.39 inches) per year during this time period. movement in the engineered fill caused by placement of the temporary soil stockpile at
136 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 137

the top o f the slope. Development of a site usually involves the introduction of Brandon, Thomas. L , Duncan, J. Michael, and Gardner, William. S., 1989,
landscape irrigation, which may increase the water content of portions of the fill “Hydrocompression Settlement of Deep Fill,”, Journal of Geotechnical Engineering,
material, increasing the in-situ unit weight and possibly inducing hydrocompaction ASCE, 116(10), 1536-1548

It is concluded that the instrumentation data analyzed for Area 1-5, Site 2 indicates that Day, Robert W., 1991, “Swell versus Saturation for Compacted Clay,” Journal of
movements are most likely the result of fill equilibration and that the engineered fill is Geotechnical Engineering, ASCE, 118(8), 1272-1278
performing as designed. No movement was observed in the underlying Panoche
Formation. Initial movement of about 1.8 cm (0.7 inches) in the first two years is Franklin, John A. and Dusseault, Maurice B , 1989, Rock Engineering McGraw-Hill
believed to be movement caused by fill equilibration. An additional 1.3 cm (0.5 inches) Publishing Company
of movement occurred during the placement of a 6-meter (20-foot) high soil stockpile
at the top of the slope; movement stopped when the stockpile was removed. Jaeger., J. C. and Cook, N. G. W., 1979, Fundamentals of Rock Mechanics. Third
Edition, Chapman and Hall.
Future Investigations and Research
Lawton, Evert C., Fragaszy, Richard J., and Hetherington, Mark D , 1991, Journal of
Once development is completed and landscape irrigation commences, additional Geotechnical Engineering, ASCE, 118(9), 1376-1394.
movements are anticipated due to the wetting effects of irrigation (Brandon, et al,
1989). Monitoring of Area 1-5 should continue to assess the performance of the Sherard, James L., Woodward, Richard J., Gizienski, Stanley T , an dCleveneger,
engineered fill after development is completed. Further research is needed to more William A., 1963, Earth and Earth-Rock Dams. John Wiley and Sons, Inc.
accurately determine the effects of stress relaxation and fill equilibration in hillside
grading. Research should include installing instrumentation in ungraded hillsides as Slope Indicator Company, 1993, DataMate and DigiPro software programs, Bothello,
well as in hillsides affected by grading, preferably installed prior to commencing grading Washington.
operations. The research should document each aspect of grading, from the excavation
of the back-cut slopes to the placement of fill Data collection and analysis should
continue until full development of the graded hillside.

Acknowledgments

The support and comments from Mr. Howard McMaster of Earth Systems Consultants,
Dr Jak Daemen of the University of Nevada, Reno, College ofMines, Mr Larry Lewis
of SINCO, and Mr. Murray Levish assisted in the compilation of this paper.

References

Al-Hamoud, A S , Basma, A. A., Husein Malkawi, A. I., and Al Bashabsheh, M. A.,

1995, “Cyclic Swelling Behaviour of Clays,” Journal of Geotechnical Engineering,
ASCE, 121(7), 562-565.

Anderson, M G and Richards, K. S., editors, 1987, Slope Stability, Geotechnical

Engineering and Geomorphology. John Wiley and Sons Ltd

ASMT, 1988, “Geotechnical Investigation Report,” Sky Valley Development

Bowles, Joseph E , 1979, Physical and Geotechnical Properties of Soils, McGraw-Hill

Book Company.
 GEOTECHNICAL MEASUREMENTS 139

square and 0.6 meters thick were then constructed for each tieback on the rock
and clay fill slopes below the homes. When tested and locked off these tied back
anchor blocks formed the primary line o f defense to mitigate the shallow building
foundation and slope improvement failures. This paper will discuss the design of
the tieback restraint system, the difficult construction constraints encountered on
the steep slopes, and the current successful results o f slope and building
monitoring.

Development Histoi^ and Background

The 172-unit Stone Point Sweetwater single-family residential development was

constructed in 1987. Mass grading o f hills and ravines, prepared the site for the
maximum number o f homes that could fit, resulting in 6 meter wide streets and 15
STABILIZATION OF FILL SLOPES SOLVED USING
meter wide lots. Many o f these lots where built on cut and fill transitions with
DEEP-HIGH CAPACITY TIEBACKS
slopes as high as 18 meters A
Chester A. Carville, C.E., G.E., M. ASCE1 typical slope is shown in Photo 1.
Ronald J. Tanenbaum, PhD., P.E., G.E., F. ASCE2 With a potential for movement, the
slabs were post tensioned.
Abstract
The developer was allowed to over­
The Stonepoint Sweetwater Reconstruction Project consists o f a 172 unit steepen the slopes to 1-1/4
residential development o f single family homes on shallow concrete slab horizontal to 1 vertical. San Diego
foundations built in 1987 near San Diego, California. The project has experienced County required the construction of
various levels o f distress and damage from geotechnical related causes attributed a rock boulder buttress about 6
to unstable fill slopes and associated lateral fill slope extension, volcanic rock fills meters thick on the outer fill slope
Photo 1 - Typical Site Slope
with voids and inadequately compacted clayey fill soils, and soil creep on fill slopes face with rocks not be exceed 1.8
as high as 18 meters and as steep as 1 1/4 to 1. Differential building settlements of cm. in diameter. No provision was made to separate the finer grained fill behind
up to 20 cm and horizontal movements of building and slope improvements of the buttress or topsoil on top of the buttress from penetrating the voids.
several centimeters forced immediate action and the expenditure of several million
dollars to solve the problem over a 5 year design/construct period from 1995 to Construction of a clean boulder buttress did not occur. Rather, the buttress was
2000. composed o f a mixture o f boulders, some over 0.9 meters in diameter, and fill soil
consisting predominantly o f clay. This mixture of clay and boulders was inherently
Considerable investigation and evaluation o f numerous options resulted in the unstable at 1-1/4 to 1 slope.
selection o f a unique but economic solution o f tied back individual anchor blocks
as the primary slope restraint system to maintain the homes from further The end result o f this construction was a combination of ailments:
movement. Over 100 wire tendon tiebacks were installed and pressure grouted • Settlement due to inadequate compaction;
into the volcanic bedrock using down hole percussion drilling techniques. The • Soil loss into the voids of relatively clean nested boulders in direct
tiebacks were drilled to depths as deep as 61 meters and tested to capacities of response to surface water infiltration;
1423 kN. Concrete and steel anchor blocks measuring as much as 2.4 meters • Severe lateral slope extension and slope creep; and
• Limited soil/clay expansion.

'Geotechnical Consultant, Carville Engineering & Construction, P.O. Box 1030, Cardiff, CA

Associate and Senior Geotechnical Consultant, URS Dames & Moore, 1615 Murray Canyon
Road, Suite 1000, San Diego, CA 92108

138
140 GEOTECHNICAL. MEASUREMENTS GEOTECHNICAL MEASUREMENTS 141

Structural Response (Dames& Moore, 1995a). Beneath the fills was either the Sweetwater Formation,
consisting o f angular conglomerates, siltstones, mudstones and sandstones, or the
As the slopes began to move laterally and downslope and the surface settle, a Santiago Peak Volcanics, which is a complex assemblage o f volcanic (andesitea),
number o f events occurred. Homes on the top o f the fill slopes settled and creeped volcaniclastic (breccias, tuffs and tuff breccias) and sedimentary rocks (sandstones,
towards the slope. Distortions in the floor slabs of 5.1 cm. were not uncommon shales and conglomerates).
with several homes approaching 10.2 to 24.4 cm. out of level and tilting.
Separation between the driveway slab and the garage slab approached 5.1 cm. in Anchor Load Testing Program
some residences. Significant interior wall and ceiling cracks developed, doors and
windows jammed, floor tiles cracked, and kitchen cabinets and counters separated Prior to detailed engineering analysis a full scale tieback testing program was
from the walls. Some of the homes settled toward each other; and several water performed on two typical slopes within
lines ruptured. the project area (Dames & Moore,
1995b). Four test anchors were drilled
The Homeowners Association (HOA) common areas also experienced distress. using air percussion equipment
The Association pool and deck cracked and was repaired twice. A concrete block suspended by crane on a drill platform.
and glass sound wall on the top o f the highest slope broke, common area retaining (Photos 2 & 3) Test anchors were
walls cracked and tilted, and surface drainage systems were disrupted. installed to depths varying from 21 to
37.5 meters in both the Sweetwater
Actions Taken Formation and the Santiago Peak
Volcanics using epoxy coated Williams
As expected, a lengthy legal process ensued. The experts for the HOA estimated all thread high strength steel bar. The
that total damage amounted to about $35 million. The developer declared Photo 2 - Test Anchor Site bars were installed in 12.7 cm. drill
bankruptcy and the insurance company could only provide about $10 million. This holes and pressure grouted with bond
left the HOA with about $7.5 million to effect repairs after attorney and expert lengths varying from 0.9 to 9 meters. GeoKon vibrating wire strain gauges were
expenses. Recognizing that the homes could not be effectively repaired until the installed at various locations
slope movements where reduced and minimized, the HOA board engaged Dames throughout the length o f each bar.
& Moore to design repairs to the common area slopes, retaining and sound walls, Each tieback was tested to failure
and pool area. The work performed included the following: and readings recorded for stress-
• review o f historic documents; strain relationships o f the anchor
• visual and manometer survey o f representative homes; plate and the anchor bar.
• limited geotechnical investigation;
• alternative evaluation to reduce slope deformations and repair other The results o f the testing program
common area problems; provided extremely valuable
• plans, specifications and cost estimate preparation for selected information for the engineering
alternatives; analysis discussed below, and for
• contractor selection, negotiation and permitting; the competing contractors in the
phased construction program. Photo 3 - Drilling Test Anchor
• construction management and supervision; and
• monitoring performance of stabilization efforts.
Engineering Analyses
Geologic Setting
Fill soils consisting o f reddish brown silty clay, sandy clay with cobbles and various In our effort to identify the most cost-effective means of stabilizing the fill slopes,
amounts of boulders were encountered to thicknesses exceeding 4.6 meters we found it necessary to go beyond conventional, limit equilibrium, slope stability
142 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 143

analysis. Using the geotechnical engineering properties o f the materials developed The question to be answered was how much of the slope movements could be
during the investigation phases for both the legal action and the slope attributed to lateral slope deformation versus fill settlement. The latter could have
reconstruction design, the limit equilibrium slope stability analyses o f the slopes, developed through consolidation, soil loss into voids, and/or hydro-compaction in
performed using Bishop’s modified method of analysis with the PCSTABL5M response to surface water (rainfall and/or irrigation) seeping into the soil. While these
computer program (Itasca, 1995), indicated the slopes to be statically and mechanisms could be expected to cause similar horizontal and vertical displacements,
pseudostatically stable under existing conditions. Since the method of analysis the choice of mitigation would be significantly different for lateral slope deformation
cannot address the deformations associated with lateral slope extension and creep,
versus consolidation-related problems. Providing lateral support of the fill slope, such
a two-dimensional explicit finite difference computer model was applied.
as anchors or retaining walls, may mitigate lateral slope deformation and creep.
The applied model is known as FLAC (Fast Lagrangian Analysis o f Continua) However, if fill settlement were the culprit, lateral support measures would not be very
developed by Itasca Consulting Group, Inc (Itasca, 1995). This program and
earlier versions o f it have been successfully used by Dames & Moore over the last
25 years to solve a large variety of practical and scientific problems in
geomechanics that could not be readily addressed by conventional means.

The particular aspect o f this project that required the use o f FLAC was related to
the nature o f observed movements that have been occurring over a period o f 10
years. Based on our review o f the monitoring data, there was no clear way of
deciding whether these movements were due to fill settlement, soil creep, lateral Figure 1. Model Mesh for FLAC Analysis
slope extension, or a combination thereof. Because the effectiveness and choice of
mitigation measures greatly depend on the causal mechanism o f the problems they effective. Instead, there would be a good chance that, after ten years, this process may
are supposed to rectify, it was necessary to utilize a method o f analysis capable of be approaching its final stages by now. Alternatively, it should be possible to slow it
rationally simulating these different mechanisms. Numerical analysis with down significantly by preventing further water from seeping into the ground through
programs like FLAC is able to overcome the inherent limitations o f conventional
improved surface drainage and carefully managed irrigation.
stress-oriented slope stability analysis by modeling soil deformations as well as soil
stresses.
Our analysis approach was based on the likely condition that lateral slope deformation
Figure 1 shows the model mesh that was used to simulate the 12.2 meters high fill would occur concentrated in a narrow layer along the interface between fill and native
slope along Jamacha Boulevard at Photo 1. Because the fill consists of a random soils, where seepage from rainfall and irrigation tends to collect. Given this scenario,
mixture of sandy silty clayey soils with cobbles and boulders, it was not possible to slope deformation could be expected to have caused purely translational (vertical and
reliably characterize its physical properties such as shear strength and compressibility. horizontal) movements o f the fill wedge, with only negligible rotation (=differential
Lacking such data for input in our analysis, we mainly relied on back calculations based settlements) o f the homes. The process o f fill settlement, on the other hand, would
on earth movements observed during the ten years since construction of this fill. The result in both translational as well as rotational deformations. By attributing the entire
most reliable data for this purpose were judged to be the lateral displacements and rotational component of earth movement to fill settlement alone, it became possible to
differential settlements that the homes had undergone to date. The former were separate the two contributing factors by a 2-step analysis as described below.
measured as roughly 5.1 cm. cracks between driveway pavements and garage The first step of the analysis was to simulate the process of fill settlement until the
structures, and the latter had created typical differences in floor elevations of 5.1 to 7.6 observed differential (=rotation) of 5.1 to 7.6 cm. between front and back of the homes
cm. between the front and the back of the homes. was reached. The critical question was whether the simultaneously generated
horizontal displacements could fully account for the 5.1 cm. lateral displacement
144 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 145

observed in the form o f pavement separations at the driveway and garage interface. If Figure 3 shows the results of Step 2 of our analysis, which simulated soil creep in a
this were the case, lateral slope deformation could essentially be discounted as a narrow zone along the contact between fill and natural soils. The soil model utilized
contributing factor and the most effective mitigation would likely be minimizing future was a simple power law with the deformation rate defined by two material parameters.
seepage of water into the ground. However, if the 5.1 cm. could not be fully A series o f analysis runs was performed with different deformation parameters, until a
accounted for by the 10-year simulation produced approximately 2.5 cm. lateral deformation. This amount of
process of fill DIFT.SETT. = <t7.4c«i gap - Z
deformation, when added to the 2.5 cm. gap produced by the process o f fill settlement
settlement, then discussed above, accounted for the 5.1 cm. total gap observed in the form of
mitigation measures garage/driveway separations.
should also address
the problem of
lateral deformation
by providing lateral
support.

Slope Deformation
Analyses

The process of water-induced fill settlement was simulated by utilizing an elastic-plastic

Mohr-Coulomb plasticity model equipped with a yield cap to handle permanent Figure 4 End of Consolidation plus I.ateral Slope Deformation:
volumetric strains (i.e. hydro-compaction). The yield cap was incrementally contracted Mobilized Shear Ratio to Failure
until the differential settlements observed in the field were created. Figure 2 shows the
results of this analysis step in the form of a displacement vector plot. The differential
A contour plot of mobilized shear strength after the two analysis steps o f fill settlement
settlement of 6.4 cm. was accompanied by a lateral displacement (gap) of about 2.5 cm.
(consolidation) and lateral deformation is shown in Figure 4. The presented values are
Since this movement fell short of the average 5.1 cm. wide cracks observed at the site, it
ratios of shear stress in the soil divided by available shear strength. A ratio of unit value
was concluded that lateral slope deformation was a likely contributor to the observed
would signal the beginning o f shear failure.
earth movements.

Proposed Mitigation Measures

The effectiveness of the proposed mitigation by lateral support with soil anchors was
evaluated with Step 3 of our analysis assuming 2 rows of 9-meter-on-center anchors
preloaded to 445 kN each. The effect of these anchors is to reduce the ratio of
mobilized shear strength in the soil by taking up a portion of the down-slope driving
forces that produce lateral deformation. Evaluating the degree of deformation rate
reduction that can be accomplished in this manner is not accessible to reliable
quantitative analysis, but is largely a matter of engineering judgement. Based on our
Figure 3. Lateral Slope Displacement: Incremental Displacement Vectors
(Deformation along Full Interface) experience, even a small reduction in mobilized shear strength translates into a
146 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 147

significant slow down of lateral deformation. For example, a 10-percent reduction of continuous monitoring o f soil
mobilized shear resistance can result in a reduction of deformation rate down to 1 deformation and anchor
forces during the first phase.
After observing the anchors
for 2 to 3 years, a decision
was made on how to expand
this type o f mitigation in a
second phase, in the most
economic manner.
Construction Contract
Photo 4 - Slope Before & After Tieback Instillation
Figure 5. Lateral Slope Deformation with Anchors:
Mobilized Shear Ratio to Failure Time was o f the essence to
implement the Phase I
engineered solution, so it was decided that a qualified foundation contractor would
percent of the original rate. Figure 5 shows computed contours of mobilization shear
be preselected and a negotiated contract executed. Several national foundation
specialty companies participated and observed the test anchor and exploration
program. This provided the contractor knowledge o f the site conditions prior to
the completion o f plans and specifications. Schnable Foundation Company was
selected and a contract was signed without the significant time delays o f a formal
bidding process. Peripheral
construction items were left to the
construction manager to negociate
or contract with other specialized
firms as the need arose. Schnable
Figure 6. Mobilized Shear Resistance: Comparison with and without Anchors also participated in a portion of
the design process in the second
ratios about 10 years after anchor installation, and Figure 6 presents a corresponding
phase to optimize the terrain
plot of mobilized shear resistance along the fill-soil contact. These plots indicate a constraints and economics of
reduction of mobilized shear resistance of about 22 percent. With roughly half of this construction.
reduction being accomplished at the time of anchor installation and preloading, the
deformation rate is expected to slow down immediately by a significant amount. Tieback Construction
Photo 5—Klem Drill Rig

Based on the results obtained from the analytical models, it was concluded that a Initial construction began in
September 1995, with the clearing o f slope vegetation and benching o f the slope
system of rock anchors would offer the best solution to reduce future lateral
for equipment access. This occurred along the steepest 18 meter high slope above
deformation of the slopes. The anchor type was based on cost, site construction
a busy highway with limited access for large equipment. Photo 4 shows a
restrictions, and previous experiences o f the selected contractor and design team. foreground view o f the slope before construction and a background view of
installed tiebacks and anchors. Tieback drilling was conducted using a track-
Because it is not possible to reliably predict the actual performance of the anchors, we mounted Klem drilling rig with the 15.2 cm. bore hole angled downward 30
recommended that installation be accomplished in two construction phases, with degrees from the horizontal. (Photo 5) Because o f the loose nature o f the rock fill
148 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS

soils and difficult geologic rock conditions, During construction various forms o f geotechnical
continuous drill casing was used in conjunction instrumentation were installed on selected anchors to
with an air percussion hammer and a lost crown monitor their long term performance. This
drill bit. Drill hole conditions varied from hole instrumentation consisted of GeoKon vibrating wire
to hole load cells and Tensmeg strain gauges. A typical
installation is shown in Photo 8.

and from one foot to the next between voids, Project Evaluation and Summary
loose soil, ground water under pressure and
very hard volcanic rock. Following Phase I construction in late 1995 and early
Photo 6 - Forming for Concrete 1996, an evaluation period o f over one year followed
Once the bore hole was completed and flushed, Reaction Block
before Phase II final construction o f the remaining
a Dwyidag 5 or 7 wire strand tendon with double corrosion protection was slopes was accomplished. This period o f time allowed
inserted inside the drill casing and pressure grouted in the bond zone as the casing for engineering analysis o f data accumulated during
was incrementally withdrawn. The drill hole and tendons varied in length from 26 construction and from monitoring information o f the
to 61 meters depending on design capacity required and depth to competent instrumentation program. This evaluation period was Photo 8 - Typical
formation. The bond length o f the tendon also varied depending on capacity most valuable in economically optimizing the entire Tieback Load Cell
Installation
required from 10 to 12 meters. Anchor tendons project such that over one million dollars remains with
were load tested to 1423 kN and locked off at 534 the Homeowner Association. This money is available for future slope work should
kN. the ongoing monitoring program show deficiencies in the installed program or on
other slopes not repaired. To date all repaired slopes appear to be functioning
Two different types o f anchor reaction blocks and within anticipated limits.
installation methods were utilized during
construction. The project first started with the use ACKNOWLEDGMENTS
of a cast-in-place concrete reaction blocks 2.4
meters wide by 2.4 meters high by 0.6 meters thick We wish to acknowledge the Stone Point HOA and Board o f Directors for their
as shown in Photos 6 and 7. Later sequences of consideration to contract with us to perform the engineering and construction
construction used shotcrete applied reaction blocks management. We also wish to thank Schnable Foundation Company, Inc. for their
2.7 meters wide by 2.1 meters high by 0.6 meters expertise and special consideration to the project and especially acknowledge Hank
Photo 7 - Completed thick to simplify the installation, save time and Schnable, Ron Chapman and Jim Kirkendahl for their involvement in this project.
Reaction Blocks
money and to keep the top o f the block below the
final ground surface. After installation o f each tieback anchor, extensive testing
was performed to evaluate if the anchor met the project engineering specifications.
The two primary tests for anchor evaluation were performance and proof testing.
Performance testing involved incremental loading and unloading o f 10 percent of
all anchors. Proof testing was performed on the remainder o f the anchors by
applying the load in increments without decreasing the load at each interval. All
testing was performed to 1.33 times the design load. Any anchor that did not meet
an acceptable criteria for elongation, total movement or creep was replaced with a
new anchor.
150 GEOTECHNICAL MEASUREMENTS

APPENDIX I REFERENCES

Dames & Moore ( 1995a). Geotechnical Investigation, Stone Point Sweetwater

Reconstruction Project, for Stone Point Sweetwater Homeowners Association,
Job No. 30049-001-035.

Dames & Moore (1995b). Anchor Load Testing, Stone Point Sweetwater
Reconstruction Project, Spring Valley, California, for Stone Point Sweetwater
Homeowners Association, Job No. 30049-001-035.
Strain Measurements
of Concrete Piers
Itasca ( 1995). FLAC Users Manual, Version 3.3, published by Itasca Consulting
in Expansive Soils
Group, Inc., Minneapolis, Minnesota.
Thomas A. Chapel, PE1Member, ASCE, and
John D. Nelson, PhD, PE*2 Member, ASCE

Abstract

Foundation systems in expansive soils within the western US commonly

utilize drilled concrete piers and grade beams. A field test site was established to
investigate the behavior of full scale drilled concrete piers. Five concrete piers were
instrumented and installed to depths ranging from 3 to 7.6 m in a known expansive
soils deposit on the Colorado State University campus. Four reinforced concrete
piers were instrumented with vibrating wire strain gauges at 1.8, 3, 4.3 and 5.5 m
depths. One pier was not instrumented. Two of the piers were installed beneath a
simulated slab-on-grade and three were placed in an uncovered area. The uncovered
area was irrigated beginning a year after installation of the piers. A vibrating wire
strain gauge was installed in a laboratory test cylinder. The laboratory cylinder was
used to compare baseline readings with the field cylinders, and to monitor
environmental factors (e.g. temperature and shrinkage) influencing strain. Strain in
the concrete piers was monitored over a two year period and compared with soil
water content changes, free field heave and displacement of the piers at the ground
surface. Performance of the drilled concrete piers was compared with that of a
patented helical pier system installed in the same environment. Skin friction
mobilized in the heaving soil zones has not been sufficient to overcome the
anchoring capacity of the deep piers, but the 3 m pier heaved approximately 30 mm
over the 2 year period. Significant tensile strains have been measured, and have
continued to increase. It is believed that strain in the concrete piers can be correlated
with movement of the wetting front downward through the soil.

'Project Engineer, CTL/Thompson, Inc., 375 E. Horsetooth Rd, Bldg 3, Ste 100,
Ft. Collins, CO 80525

2Professor, Geotechnical Engineering Program, Civil Engineering Department,

Colorado State University, Ft. Collins, CO 80523

151
152 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 153

Introduction
1 i j
1 1 1 1 ! • |
Foundation systems in expansive soils in the western United States commonly j j
160 j ------------j-------------- j - - j -------------j - j
utilize drilled concrete piers and grade beams. Typical pier lengths have been in the 1
i ° A o 9 a*j o j O :
4 to 6 m range, and typical diameters have ranged from 20 to 40 cm. A field test area 1° !
was constructed at the Expansive Soil Field Test Site on the Foothills Campus of j j i
Colorado State University in Fort Collins, Colorado with the following objectives: lbO
i i i i i
! ! : ! ! ! ! !
• To monitor strain in conventional drilled concrete piers. xS i h u l a [ t p o j C o n lc r r -e i: e S jl
• To investigate the performance of a conventional drilled concrete pier 140 _ j: -,_ l ; : ‘ .r ? -- i
foundation system compared with the performance o f a patented helical i o "f o ^ . 'd ^ d ' cs !- O i
pier foundation system in expansive soils. i 4 • ! . / [ * -J 1 LE G E N D
• To investigate the movement of water through expansive soil in the j\ ^ [ ' [ j*
130 -------------- i —o - —i - o - 4 - 0 ^ < i> - < :m , o - - . H -■ © — 1 / r\ S hallow H e lic a l Pier
region adjacent to helical piers and conventional drilled concrete piers. 1 , 1 * 1 -1 d i Deep H e lic a l Pier-
-L - a '■ *i i 1* • C o n c re te P ie r
j
The Expansive Soil Field Test site is located on an expansive soil deposit on ^ A ccess Tuloe
O R L G I N A L __ S C A L E _
the Colorado State University Foothills Campus. Geologically, the project area is 120
near the eastern boundary of the Front Range of the Southern Rocky Mountain m
Physiographic Province. The Expansive Soil Field Test Site is located on a well
documented expansive soil deposit [Porter (1977), Goode (1982), Nelson et al
(1994), Nelson and Miller (1992), Hamberg (1985), and Chapel (1998)]. East of the
project site are the flat lying Cretaceous sediments of the Colorado Piedmont. West Figure 1. Layout of the Helical Pier Test Site (after Chapel, 1998).
of the site Cretaceous and older sediments dip sharply away from the uplifted Front
Range stratigraphy. The project site is underlain by the lowermost part of the Upper ranged from 34 kPa (0.4 tsf) to 718 kPa (7.5 tsf). Liquid Limit, Plasticity Index,
Cretaceous Pierre Formation, a mostly dark gray shale with sandy shale lenses gradation, natural dry density and water content are all generally similar for all the
[Braddock et al (1989)]. Locally the Shale ranges in color from very dark gray to soils encountered in the exploratory borings and pits. The locations of the
buff, with thin multi-colored lenses of bentonitic clay and silty clay. The expansive exploratory borings and excavations, and the general layout of the site are shown on
layers vary in thickness from two to about 20 centimeters. At the project site the Fig. 1.
Pierre Shale is covered by a thin, discontinuous layer of alluvium. Sediments strike
nearly north-south and dip 20 to 25 degrees to the East. Six soil borings and four test Theoretical Considerations and Design Details
excavations were used to examine the subsurface soil profile and collect soil samples
for laboratory testing. Samples were collected by thin walled sampler, generally at Expansive soils research as been conducted on the Foothills Campus of
30 to 60 cm intervals. Two other borings were sampled at greater depths using a Colorado State University since about 1976 [Porter (1977), Goode (1982), Nelson
California sampler. et al (1994), Hamberg (1985), and Chapel (1998)]. The helical pier test site was
located geologically up-strike from the previous test sites so that the geological and
One dimensional swell-consolidation tests, (ASTM D546) natural water geotechnical characteristics of all the sites would be the same. Five full scale
content, Atterberg limits, and Unconfmed Compressive Strength tests were conventional concrete piers and six helical piers were incorporated into a field site.
performed on selected samples. Soil properties are summarized in Table 1. The test Some features are unique to this site, and others are similar to adjacent sites. Small
results illustrate an interesting characteristic of the expansive sediments o f the lower diameter PVC tubes and a down-hole moisture/density gauge were used to measure
Pierre Shale: the highly expansive layers are interbedded with sandy and silty clays water content. Strain gauges and conventional survey monitoring stations were also
of moderate and low swell potential in a variable manner. Highly expansive lenses included in the field test site.
only a few inches thick may be separated by similar thicknesses of clays
characterized by low swell potential. In general, there appear from the field and Site Preparation. The North half of the site is an exact duplicate of the South half of
laboratory data to be three main zones of expansive material: one at 1 to 2 m, another the site in terms of instrumentation. The North half of the site was left in a relatively
at 3.5 to 4.5 m, and a third at around 6 m. As shown in Table 1, swell pressures undisturbed state, and the South half was covered by a simulated concrete slab and
isolated from surface water by a perimeter drain. The drain consists of 51 mm
154 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 155

Table 1. Summary of Soil Properties (after Chapel, 1998). zone.

Natural Percent
Borehole/ Natural Water Liquid Plasticity Passing Volume Swell gfrnin and Strain Gauge Theory. Temperature measurements for each strain gauge
Testpil Depth Material Density Content Limit Index No. 200 Change Pressure
at each depth permit corrections for temperature change. The thermal coefficient for
Loc. (m) Type (kN/m3) (%) (%) (%) Sieve (%)* (kPa)‘
E-212 1.2 Clay 14.4
the strain gauge is approximately 2 x 10'6 less than that of the concrete surrounding
E-212 5.8 Shale it. Therefore, as the temperature of the concrete and the strain gauge increase,
C-212 1.2 Clay 14.7 55 29 tension in the strain gauge increases at a slightly slower rate than in the concrete. To
C-212 2.7 Shale account for this difference, the strain reading must be corrected for temperature
B-1 1.5 Clay 57 35 34
B-1 2.6 Shale 11.0
changes. The correction is made by multiplying the difference in thermal
431
NE 1.7 Shale 17.4 15.6 54 30 99 2.5 216 coefficients by the change in temperature and subtracting their product from the
SE 1.1 Shale 16.0 24.3 60 34 95 0.5 77 apparent strain reading. Thus, when the temperature increases, the true tensile strain
SW 0.9 Shale 17.9 12.4 4.0 287 ~ will be slightly higher than the apparent reading taken directly from the gauge. For
NW 0.6 Clay 17.4 12.1 47 25 81 4.5 297
NW 0.9 Shale 18.6 14.6 54 29 82 6.8
a one degree Celsius temperature increase, the actual tensile strain reading will be
7Ì8
* swelled under 47.9 kPa surcharge 2 x 10‘6 greater than the apparent reading [Slope Indicator Company (1994)).
Mathematically, the relationship is
diameter slotted PVC pipe underlain by Gund-Seal, a 14 mil polyethylene geo­
membrane with a thin layer of bentonite bonded to it. The drain was bedded in 20 ApeCorT=Ap£0-(ACTxAT) ( 1)
mm nominal diameter crushed aggregate and installed at a depth below the ground
surface of approximately 300 mm to intercept surface runoff, yet minimize any effect where Apecorr= corrected strain
on horizontal migration of near surface soil water. The surface of this portion of the Ape0 = apparent strain
site was also sealed with Gund-Seal. Seams in the membrane were overlapped a ACr = difference in thermal coefficients. (For the Slope Indicator gauges,
minimum of 60 cm, ACt = -2 x 1 0 > e /C )
AT = current temperature minus initial temperature (C°).
and holes were sealed. The entire "slab" surface was covered with 15 cm of 20 mm
nominal diameter crushed rock to simulate the weight of a concrete slab while An additional thermal consideration is the actual compressive strain that
maintaining a flexible, impermeable barrier. Flexibility of the simulated slab was exists due to the temperature difference between the ground surface and the soil at
needed to reduce water penetration through cracks that would likely develop in a the base of the piers. Temperature induced strain is given as top and bottom of the
rigid slab. The flexibility also prevented the piers from acting as a group. piers.

Pertinent elements of the design included full scale concrete piers, helical £=CstAT
piers, survey movements, PVC access tubes for water content measurements and an ( 2)
irrigation system. Details regarding the helical piers and water content
measurements may be found in Chapel (1998). Details concerning the concrete piers where £ is strain, CST is the coefficient of thermal expansion of the rebar
and strain measurements are presented below. reinforcing, and AT is the difference in temperature between the areas of interest.

Concrete Piers. Swell pressures ranging from 34 kPa to 718 kPa (0.4 tsf to 7.5 tsf) The temperature induced strain can be quite variable in an area such as
have been measured from soils in this area [Nelson et al (1994)]. Piers were Colorado’s front range where air temperatures may deviate by 50 degrees Celsius or
designed using these swell pressures. A pier diameter of 36 cm was chosen in order more during the course of a year. Soil temperatures were measured by the Colorado
to provide enough area for the the instrumentation and the required reinforcing. State University’s Atmospheric Sciences Department at the weather station at nearby
Christman Field. Surface soil readings were compiled from meteorological data and
In August 1995, five 36 cm diameter concrete piers were installed at the field averaged for comparison purposes.
site. Four concrete piers were installed to a depth of 7.6 m each and reinforced as
described above. One unreinforced concrete pier was installed to a depth of 3 m near The strain caused by thermal expansion of the pier due to the temperature
the northeast comer of the site. This pier was installed to provide pertinent data gradient between the top of the concrete pier and the base was also considered. To
regarding the vertical displacement of concrete piers placed entirely in the active determine the significance of the temperature induced strain, the strain caused by the
156 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 157

average temperature difference between the ground surface and the strain gauge at to those installed in
the 1.8 m depth was calculated. The total force in the pier is the sum of the force in the full scale piers at
the concrete (Pc) and the force in the reinforcing steel (Ps). For a pier with no load the helical pier test
site was embedded
Pc + Ps = 0 (3) in the cylinder. The
gauge was mounted
Also, the strain in the concrete (ec) is equal to the strain in the steel reinforcing (gj parallel to the
longitudinal axis in
Thus, the center of a 15 cm on o
3 o,
diameter by 46 cm
£c = £s (4) o c
and, long m odified B •§
acAT+Pc/AcEc=asAT+Ps/AsEs (5) co n crete te st 2 g
u*
where cylinder. The c
a c, a s = the coefficient of thermal expansion for the concrete or gauge’s electronic 3t-H
steel, lead was routed </5
AT = the temperature difference, horizontally through
Ac, As = the area of the concrete or steel, and the wall of the test
Ec, Es = the modulus of elasticity for the concrete or steel. cylinder 76 mm
from the end o f the
Solving equations (3) and (5) for typical values of material properties and a cylinder. The Laboratory Strain x lO^tens. pos.)
temperature difference of 13°C shows that the strain due to temperature differences attached wiring was
is approximately 1.2 x 10'6. Since the temperature induced strain is very small in then connected to Figure 2. Correlation Between Strain Gauge and
comparison with the strain measured in the piers, no correction was necessary for the same readout Laboratory Strain.
temperature induced strain. device used to
measure the field installed strain gauges. The specimen was field cured for 24 hours,
Strain Gauges. The reinforced concrete piers were each instrumented with strain then transported to a moist curing room at Colorado State University. Strain
gauges mounted in the center of the pier at depths of 1.8, 3.0, 4.3, and 5.5 m readings were taken periodically as the concrete cured.
(measured from the ground surface). The gauges were attached to the reinforcing
cage prior to placement of the cage in the excavation. The test specimen was fitted with a frame designed for measuring strains in
concrete specimens. It consists of two circular end members 200 mm apart
The strain gauges installed in the piers are Slope Indicator VS Embedment connected by a dial indicator gauge. The two end frames were fixed to the concrete
Strain Gauges, which are specifically intended for uses in reinforced and mass specimen with set screws so that the strain gauge embedded in the concrete was
concrete applications. The Slope indicator gauges are 181 mm in length which is centered between the ends of the external strain measurement device.
about seven times the nominal aggregate diameter. Variability of the strain readings
is reduced by using a gauge at least twice the length of the maximum dimension of The cylinder was loaded using a test frame in the Colorado State University
the coarse aggregate in the concrete [Stanton & Blake (1992)]. Each gauge is 11 mm Materials Testing Laboratory. At each increment of load, strain was measured
in diameter, and the flanges at each end are 32 mm in diameter. The end flanges are directly from the gauge embedded in the concrete test cylinder, and the deflection
intended to provide a good embedment in the concrete. The gauges are factory was measured directly from the dial indicator. Measurements were taken during
sealed to reduce the possibility of time dependent creep. Strain readings are loading and unloading at the same intervals. Fig. 2 shows the strain data correlation
transmitted to the surface by 22 gauge, shielded, polyurethane jacketed cable. The between the strain gauge readout and the actual strain. In keeping with the
cables were fixed to the rebar reinforcing cage at 0.6 m intervals to the surface. convention established previously for field strain data as part of this project, tensile
strain is shown as positive and compressive strain is negative. It can be seen from
Laboratory Test Cylinder. In order to calibrate the strain gauges under controlled Fig. 2 the strain gauge reading correlated closely to the direct measurement of strain,
conditions and verify the field strain data, a concrete cylinder was constructed for particularly at the lower range of strains.
testing in the laboratory. A Slope Indicator VS Embedment Strain Gauge identical
158 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 159

Analysis

Strain and temperature

in the concrete piers, soil water
content, precipitation amounts,
elevation of the ground surface,
and movement of the tops of the
concrete and helical piers were
recorded at monthly intervals
throughout the duration of the
project. More frequent readings a). 1.8 m depth
were obtained when indicated
by special conditions. For
example, daily strain and
temperature measurements were
made during the curing of the
concrete piers. Additional data
were collected during pullout
capacity testing, soil boring and
sampling and test pit
excavations. An irrigation b). 3 m depth.
system was used during the
second year o f the project, and
irrigation amounts were
monitored during that time.
Fig. 3 shows the general trend
in soil water migration observed Gravimetric Water Content
on the site over the project
duration.
Figure 3. Vertical water migration under the
Laboratory Strain Readings, covered site, August 1995 to October 1997.
Strain during curing was clearly c). 4.3 m depth.
demonstrated by the laboratory test cylinder. Tensile strain in the laboratory
specimen increased for approximately 14 days while the sample was in the moist
curing room, then began to level off. At approximately 20 days the test cylinder was
removed from the moist curing conditions. Soon after, compressive strains
developed. At approximately 70 days the rate of change of strain began to decrease.
It was previously shown that correction for temperature had a negligible effect.
Using a method presented by the American Concrete Institute (1992) and the strain
data from the laboratory test specimen, the ultimate curing strain of this material was
calculated to be 876 x 10'6 (compression).

Field Installed Strain Gauges. The design of the strain gauges results in an initial
strain reading greater than zero that is different for each gauge. This represents the
initial tension in the steel strip that comprises the strain gauge. The initial reading Figure 4. Strain during curing of concrete piers.
160 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 161

must be subtracted from subsequent readings to determine the strain. Strain gauge
readings in the concrete piers were taken prior to concrete placement, then
immediately after concrete placement, and at hourly, daily, weekly and then monthly
recording intervals. Changes in strain during curing were recorded and plotted for
the first week after casting to show the effect of concrete curing on the strain gauge
readings. The temperature correction for each gauge was approximately 5 x 10'6 or
less. Figure 4 shows the strain during curing for each of the instrumented concrete
piers. The data indicate that these gauges functioned well initially and show
potential for consistent, long term performance.
a). 1.8 m depth
Individually, the strain readings followed a similar pattern. For the first two
days there was a general fluctuation caused by the cement hydration and plastic
shrinkage. This is particularly apparent in the gauges in Pier D. 1-130, at the 3 m, 4.3
m, and 5.5 m levels. After this initial hydration period, the tensile strain increased
towards an equilibrium point. Although this equilibrium point is different for the
different piers, the actual difference is small. The reading after 28 days was taken
as the equilibrium point and is used as the “zero” reading for correcting all
subsequent strain readings.

Fig. 5 shows the measured strain in the concrete piers over the length of the
project. Readings have continued since completion of this project, and show that
tensile strain in the piers has continued to increase.

In Fig. 5 it can be seen that after about 150 days, tensile strain in the piers
began to increase. There appears to be a cycle o f increasing tensile strains
punctuated by decreases in tensile strain on approximate yearly intervals. The
exception to this trend is at the 5.5 m depth, where there is a steady increase in
tensile strain with time.

Each of the four concrete piers also shows a trend o f increasing tensile strain
with time. In general, the tensile strain increased more rapidly at the 1.8 m and 5.5
m levels than at the 3.0 m and 4.3 m levels in the covered and uncovered piers on the
East side of the site. The tensile strain did not increase as much on the West side of
the site, particularly near the top of the piers.

Movement of the ground surface (free field heave) and movement of the tops
of the piers relative to a fixed survey monument were also monitored during the
course o f the investigation. While the ground surface moved as a result o f soil heave
and weather changes, no measurable movement of the deep piers was observed. The
pier placed to a depth of 1.8 m, however, heaved approximately 30 mm during the
project.

d). 5.5 m depth Elapsed Time (Days) Conclusions

Figure 5. Strain in concrete piers. Water migrates vertically as well as laterally through the soil in a time
162 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 163

dependent process. Evapotranspiration and suction processes reduce the water Collins, CO, USA.
content near the surface. However, when evapotranspiration is eliminated or greatly Chen, F.H. (1988). Foundations on Expansive Soils, Elsvier Science Publishers,
reduced, as by construction of a slab-on-grade, the water content begins to increase Amsterdam, The Netherlands.
from the surface downward. The increase in water content may also be influenced Dunnicliff, J. (1993). Geotechnical Instumentation fo r Monitoring Field
by other factors, such as irrigation, leaking water pipes, or inadequate drainage Performance, J. Wiley & Sons, New York.
conditions. As the water content increases and migrates vertically and laterally Goode, J. (1982). “Heave Prediction and Moisture Migration Beneath Slabs on
through the expansive soil deposit, the soil volume increases result in strains within Expansive Soils.” Master o f Science Thesis, Colorado State University, Fort
foundation members. The observed strain is cyclical in nature, increasing at varying Collins, CO, USA.
rates apparently related to the movement of water through the soil and seasonal Hamberg, D.J. (1985). “A Simplified Method for Predicting Heave in Expansive
variations. Soils.” Master o f Science Thesis, Colorado State University, Fort Collins,
CO, USA.
Properly designed pier foundations can effectively withstand uplift from Nelson, J.D., and Miller, D.J. (1992). Expansive Soils: Problems and Practice in
expansive soils. The results of this investigation suggest that the key element of Foundation and Pavement Engineering, J. Wiley and Sons, New York.
effective foundation design in such conditions is the depth of embedment for Nelson, J.D., Durkee, D.B., Reichler, J.D., and Miller, D.J. (1994). “Moisture
concrete piers. If adequate length of embedment is provided, skin friction restraining Movement and Heave Beneath Simulated Foundation Slabs on Expansive
the piers will be greater than that attempting to uplift the piers, and the foundation Soils.” U.S. Army Corps o f Engineers Waterways Experiment Station,
will remain effective. The key drawbacks to this system are that as water continues Vicksburg Mississippi, Contr. No. DACA39-92-C-0097, Geotechnical Eng.
to migrate down the pier shaft, the resisting skin friction will continue to decline. As Program, Civil Engineering Department, Colorado State University, Fort
the pier materials age and become less reliable, the uplift against them will continue Collins, CO, USA.
to increase, thereby increasing the probability for failure. As the depth of Porter, A. A. (1977). “The Mechanics of Swelling In Expansive Clays.” Master o f
conventional concrete piers increases, so do the difficulties associated with drilling Science Thesis, Colorado State University, Fort Collins, CO, USA.
and placing reinforcing and concrete. Rhodes, J.A. (Committee Chair) (1992). “Prediction of Creep, Shrinkage, and
Temperature Effects in Concrete Structures.” Rep. No.209-R, American
The ultimate depth of water migration, or the design life of foundation Concrete Institute.
systems with finite practical depths is part o f the focus for research that is continuing Slope Indicator Company (1994). VS Embedment Strain Gauge Product Literature,
at the Expansive Soils Field Test site. Slope Indicator Company, Seattle, WA.
Stanton, J.F., and Blake, J.B. (1992). “Performance of Vibrating Element Strain
Acknowledgments Gages Embedded in Concrete.” Unpublished Manuscript, University o f
Washington, Seattle, WA.
This project was supported primarily by the A.B. Chance Company and
Colorado State University. Additional support was provided by D & B Drilling and
Gundle Lining Systems, Inc. Michael Ptak, of Colorado State University, performed
helical pier load testing, soil borings and laboratory soil investigations. Additional
soil boring data and laboratory test results were provided by Mark Abshire, of
CTL/Thompson, Inc.

References

Braddock, W.A., Calvert, R.H., O ’Connor, J.T., and Swann G.A. (1989). “Geologic
Map of the Horsetooth Reservoir Quadrangle, Larimer County, Colorado.”
U.S. Geological Survey, G.Q.1625.
Bowles, J.E. (1988). Foundation Analysis and Design, 4th ed., McGraw Hill, Inc.,
New York.
Chapel, T.A. (1998). “Field Investigation of Helical And Concrete Piers in
Expansive Soils.” Master o f Science Thesis, Colorado State University, Fort
 GEOTECHNICAL MEASUREMENTS 165

the pore air and pore water pressure decreases. Once the reservoir is filled the core
will become saturated and the piezometers will measure pore water pressures.
This paper presents a general description of the earth-core rockfill dams with details
provided on the core material properties and construction methods. A general
discussion of the instrumentation program with details on the vibrating wire
piezometers and their installation are also provided. Finally, the pore pressure
Pore Pressures in Partially Saturated Soil measurements are presented.
A Case History
The Diamond Valley Lake Project Introduction
The Diamond Valley Lake Project is owned by the Metropolitan Water District of
Southern California (Metropolitan). Located about 145 kilometers (90 miles)
Michael G. Smith,1 Member, ASCE southeast of Los Angeles and 120 km (75 mi) north of San Diego, it has a design
Russell L. Jernigan,*23 Member, ASCE capacity of 987 million cubic meters (800,000 acre-feet). The purpose of the
Arleen A. Arita,3 Member, ASCE reservoir is to provide adequate surface water storage to meet drought and
emergency supply needs. Additionally, the reservoir will be used as a key
component in a traditional seasonal storage / supply cycle.
Abstract
The three earth-core rockfill dams of the Diamond Valley Lake Project were The Diamond Valley Lake Project is situated in a flat alluvial valley bounded on the
instrumented with vibrating wire piezometers during their construction. The north and south by rugged hills. No permanent natural water source supplies the
vibrating wire piezometers were buried within the central earth core of the dams and reservoir valley and the drainage area is very small compared to the reservoir size.
as additional embankment was constructed the pore pressures within the core The natural topography and three earth-core rockfill dams define the reservoir as
material were recorded. The primary purpose of the piezometers is to monitor pore shown on Figure 1.
water pressures in the core during first filling and long term operation. During Each dam was constructed with an alluvial earth-core, two upstream filter / drainage
construction the piezometers provide baseline readings with which to compare zones, three downstream filter / drainage zones, and the surrounding shells consisting
readings taken during first reservoir filling and long term operation. The of rockfill. The basic physical data for the three dams is summarized in Table 1.
construction data also provided information on the development and dissipation of
pore pressures within partially saturated soil due to fill placement. Table 1 Summary of Physical Data

The earth-core rockfill dams range in height from 58 to 114 meters (190 to 375
feetj.The central earth-core material generally consists of clayey sand and sandy lean West Dam East Dam Saddle Dam
clay excavated from the upstream alluvial valley. Immediately after compaction, the Maximum Structural
material has a degree of saturation of approximately 85 percent. Vibrating wire Height above 114(375) 98(320) 58(190)
Foundation, m (ft)
piezometers with coarse porous tips are installed at regular intervals within the core
material by placing them in shallow trenches and compacting core material directly
around the piezometer. Measurements were taken immediately. Crest length, m (ft) 2,500 (8,300) 3,300(10,900) 700 (2,300)

For this paper the term “pore pressure" refers to the measurements of the coarse
porous tipped vibrating wire piezometers. This type of piezometer measures Embankment
Volume, million m3 50 (65) 33(43) 1.9 (2.5)
approximate pore air pressure in partially saturated soil, and pore water pressure in
(million cy)
saturated soils. Initially after compaction the pore air pressure is about atmospheric
pressure and the pore water pressure is below atmospheric pressure due to capillary
suction. As additional fill is placed and the soil compresses the difference between Both the Saddle and East Dams have bedrock foundations, while the West Dam has a
mixed foundation of bedrock and dense alluvium. All of the foundations were
S en io r P roject E ngineer, U R S G reiner W oodw ard C lyde. 2 020 East First Street. S anta A na. C alifornia 92705
treated with multiple lines of grout curtains. Additionally, selected portions of the
2P roject E ngineer. U R S G rein er W oodw ard C lyde. 4 582 S. U lster S treet S uite 1000, Denver, C o lorado 80237 East Dam and the alluvial portions of the West Dam had plastic concrete cutoff walls
3 A rea P roject E n g in ee r for R eserv o ir D esign. M etropolitan W a ter D istrict o f S outhern C alifornia. 700 N orth A lam eda Street, installed in the foundations.
L os A ngeles, C alifo rn ia 90012

164
166 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS

(1333) NOUVA313
IV

Saddle
Dam
Pumping
Plant

Forebay

Recreation (1333) NOI1DA313

West Area 1
1 8
8 8

-h-rTT-j—
TT ttF ttt-T
Area

WEST DAM - TYPICAL SECTION ON ROCK FOUNDATION

_st Dam

Figure 1 Diamond Valley Lake

Embankment Section
The three embankments have similar design cross-sections. The details of the
individual designs vary slightly to suit specific site conditions. The basic design is
shown on Figure 2, which shows two typical sections of the West Dam. A 0.5H:1V
central earth-core composed of alluvium borrowed immediately upstream of either
the West or East Dams forms the water barrier. The core material is typically clayey
sand, SC and sandy lean clay, CL. Flanking the central earth-core both upstream and
downstream are a series of filters and drainage zones. These are comprised of a
sound quartzite, mined from within the valley, which was crushed and screened to
specification. The upstream and downstream rockfill shell slopes are both 2H:lV.
The rockfill has a maximum size of 0.8 m (30 in) and consists of shot rock from the
within the valley.

Core Material
The cores of the West and East Dams consist of 10.9 million cubic meters (14.2
mcy) and 7.3 million cubic meters (9.6 mcy) of compacted alluvial material
respectively. The core material was required to have a fines content (less than 0.075
(1333) N0UVA313
mm or No. 200 sieve) of between 20 percent and 80 percent, a relative compaction of
at least 98 percent, and a water content between 0 percent and 2 percent above
optimum. Relative compaction and water content difference from optimum water
content were determined in accordance with ASTM D 5080-93 “Rapid
Determination of Percent Compaction”, using a compactive effort of 980 kN-m/m3
(20,250 ft-lbf/ft3) which is intermediate to the Standard and Modified compactive
efforts and equivalent to the compactive effort of the California Department of Water
Resources Standard Test Method S-10. Statistical data for the core materials in the
West and East Dams are presented in Table 2.
Figure 2 Typical Cross-Sections of West Dam
168 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 169

Table 2 Zone 1 Core Material Statistical Summary

West Dam East Dam

Property Average Standard Average Standard
Deviation Deviation
Percent Passing No. 200 Sieve 46.0 7.2 34.7 3.7
Water Content Difference
from Optimum, % +0.9 0.9 + 1.0 0.8
(+ wetZ-dry)
Relative Compaction, % 100.2 1.6 99.9 1.6
Plasticity Index 10.2 2.1 14.2 2.4
Total Density, kg/m3 2,205 45 2,240 37
(pcf) (137.5) (2.8) (139.7) (2.3)
Water Content, % 11.3 1.2 10.7 1.1
Dry Density, kg/m3 1,985 48 (3.0) 2,025 (126.3) 46 (2.9)
(pcf) (123.8)
Specific Gravity 2.73 .02 2.75 .01
Degree of Saturation, % 85 8 85 6
100(1 500 200 100 50 20 111 5 2 1 UJ 0.2 0.1 0.05 0.02 0.01 0.005 0.002 0.001

The gradation envelopes for the core materials (Zone 1) of the West and East Dams G rain Size (m m )

are presented on Figure 3. Zone 1 was required to be placed with a maximum lift Figure 3 Core (Zone 1) 1st through 99th Percentile Particle-Size Envelopes
thickness of 200 mm (8 in) prior to compaction. Caterpillar 825 compactors were
used to obtain the required relative compaction. Instrumentation
Block samples of the core material were collected at regular intervals within the The embankments were extensively instrumented during construction. Types of
West and East Dams. Specimens were then trimmed from the block samples and instrumentation included:
tested for engineering properties. The testing included consolidated undrained • Vibrating Wire Piezometers • Standpipe and Casagrande Piezometers
triaxial compression, hydraulic conductivity and confined compression. The • Settlement Sensors • Extensometers
hydraulic conductivity specimens were carved with the long axis oriented • Survey Monuments • GPS Stations
horizontally. The median value of hydraulic conductivity for the core material of the • Inclinometers • Accelerographs
West and East Dams was lxlO '7 and 9x1 O'8 cm/s, respectively. • V-Notch Weirs
The initial degree of saturation of the core material based on sand-cone density
Instrumentation is generally concentrated along key sections of the dams. The West
testing performed after compaction is summarized in Table 2. Compression of the
and East dams have six and five instrumentation sections respectively. Pore
core materials due to placement of subsequent embankment material increases the
pressures within the cores of the dams are monitored with vibrating wire piezometers
degree of saturation. Using the compaction dry unit weight, water content and
that were either directly buried within or drilled in the core during construction. Pore
specific gravity, the strain to cause degree of saturation of 100 percent was
pressures and water levels within the foundations and drain zones are monitored
calculated. This strain was then used with the results of confined compression
using standpipe and Casagrande piezometers. Casagrande piezometers are installed
testing to estimate the stress required to cause a degree of saturation of 100 percent.
in boreholes and consist of a porous tip connected to a PVC pipe with a vibrating
This stress was then compared with the estimated stresses at the locations of
wire piezometer in the tip. The tip is isolated within a specific zone to measure pore
piezometers in the core. The results indicated that it is unlikely that the core became
pressures in that zone. Internal deformations were measured during construction
saturated at the levels of the piezometers during construction, except perhaps at the
using settlement sensors and extensometers that were installed by burying them
lowest level of piezometers, and then only near the end of construction. This was
within the cores of the dams. Post construction surface deformation of the dams is
confirmed by comparing the strains required to cause 100 percent saturation to the
monitored using survey monuments on the crest and slopes of the dams, GPS
average strains estimated to have occurred in the core during construction based on
stations at the crest, and inclinometers at the downstream toe. Seepage is monitored
the instrumentation data from settlement sensors. This conclusion is also consistent
using v-notch weirs. Earthquake ground motions are recorded by accelerographs at
with analyses performed during design.
the crests, abutments, bedrock outcrops and alluvium.
170 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 171

Vibrating Wire Piezometers

Vibrating wire piezometers consist of a tensioned wire attached to a flexible
diaphragm at one end and a rigid base at the other. Electromagnetic coils located next
to the wire are used to both pluck the wire and to convert the wire vibrations into an
electrical output current with frequency equal to the natural resonant frequency of the
wire. A 50-micron stainless steel porous filter separates the diaphragm from the core
material. In the factory and again in the field prior to installation; a calibration is
performed by applying fluid pressure to the diaphragm in increments. As pressure is
applied to the diaphragm the tension in the wire changes and the resonant frequency
changes. For each pressure there is a corresponding frequency output. Immediately
prior to installation water is used to fill a small cavity between the diaphragm and the
porous filter.
Vibrating wire piezometers were installed in the Zone 1 cores either in instrumentation
trenches or boreholes. Typical layouts of vibrating wire piezometers within the West
and East Dam cores are shown on Figures 4 and 5 respectively. The section on Figure
u/s Distance along Section - meters d/s
4 illustrates the installation of three levels of vibrating wire piezometers in
instrumentation trenches with the instrument cables running in trenches out to the Figure 5. East Dam Instrumentation Section 5 Location of Core Vibrating Wire Piezometers
downstream face of the dam to a monitoring station. Figure 5 illustrates the drilling in
of two vibrating wire piezometers, EA and EB, when the embankment reached the When installed in trenches, piezometers were placed in small side trenches to a main
level of vibrating wire piezometers EC and ED. The remaining vibrating wire instrumentation trench. Zone 1 material was compacted directly around the tip of
piezometers EC through EF were installed in trenches. the vibrating wire piezometers installed in trenches. Gasoline powered hand held
tampers were used for the first meter of fill above piezometers. Piezometers
installed in boreholes were lowered into the borehole to the desired elevation and
then graded clean sand was tremied around the piezometer. Bentonite and grout
were used to seal the boreholes. Borings were shallow, with a maximum depth of
about 25 meters (80 ft) and drilled using hollow stem augers to minimize the
potential for damaging the core. Drilling fluid was not allowed to be used. Within
the core, the main trench carrying the instrumentation cables was first backfilled
with bentonite enriched Zone 1 material immediately above and below the cables,
and Zone 1 was placed and compacted above the bcntonite/Zone 1 mixture. Walk
behind type compactors were used for the first meter of fill above the cables. With a
portable readout box the piezometers were able to be read immediately.

Pore Pressure Measurements

Initially after compaction the voids of the soil are filled with water and air. The air is
nearly at atmospheric pressure and the water is below atmospheric pressure due to
capillary suction. The coarse porous tipped vibrating wire piezometers used in these
dams do not measure the less than atmospheric pore pressures of the water. A “high
air entry value” ceramic tip with very fine pores is required to measure these pressures.
As the soil is compressed under the weight of additional fill, the pore air and pore
water pressures increase. Data from embankments where both coarse porous and fine
ceramic tipped piezometers were installed indicate that the difference between pore
Figure 4. West Dam Instrumentation Section 5 Location of Core Vibrating Wire Piezometers water and pore air pressure decreases with increasing fill height (Sherard, 1981). This
is caused by a decrease in capillary suction as the material compresses and the degree
of saturation increases. Sherard concluded that for low plasticity soils, with low initial
172 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 173

capillary suction, the air and water pressures are about the same. It is noted that the
core materials for the West and East dams have low plasticity.
Pore pressure measurements were taken at installation and weekly during construction.
The pore pressure measurements for piezometers at West Dam Instrumentation

Elevation - meters
Section 5 are plotted on Figure 6. Locations of the vibrating wire piezometers are
presented on Figure 4. At this section the West Dam core is approximately 100 meters
(335 ft) high above the foundation which is dense alluvium. For these plots the pore
pressure is converted to equivalent water head and added to the installed elevation of
the piezometer. The elevation of the embankment was measured at centerline. This is
a convenient way to display the measurements and the fill placement history. It is not
intended to imply that the elevations are of a piezometric or phreatic surface.
The core material was not saturated at the time of construction, except perhaps at the
lowest level of piezometers, and then only near the end of construction. The pore 6/19/97 1/5/98 7/24/98 2/9/99 8/28/99
pressures are a result of the core material compressing due to the load of the fill
placed above it. For this reason, the pore pressures are plotted versus fill height
above the piezometer on Figure 7. Also shown on the plots are lines of equal pore
pressure ratio. The pore pressure ratio was calculated as the pore pressure divided by
the overburden pressure. The overburden pressure was calculated as the difference

Elevation -meters
in elevation of the fill at centerline and the elevation of the piezometer multiplied by
the density of the core material. Finite element analyses performed during design
indicated that for a given elevation, the stresses at the center of the core were
typically only about 5 percent higher than at the downstream edge. For this reason it
is felt reasonable to calculate the overburden at the centerline regardless of the
location of the piezometer. It is clear from the plots that the pore pressures are
highest near the center of the dam. In each plot the piezometer nearest the centerline
has the highest pore pressure. To further illustrate this, the pore pressure ratios for
August 1999 are plotted and contoured on the dam section on Figure 8. The data 6/19/97 1/5/98 7/24/98 2/9/99 8/28/99
show that the pore pressure ratios are highest near the center and lower portion of Date
the core and least near the downstream chimney drain and top of the core.
The pore pressure measurements for piezometers at East Dam Instrumentation
Section 5 are plotted on Figure 9. Locations of the vibrating wire piezometers are
presented on Figure 5. At this section the East Dam core is approximately 53 meters

Elevation -meters
(175 ft) high above the foundation, which is rock. RCC was used as backfdl material
in a local depression in the foundation in this area. Pore pressures are plotted versus
fill height above the piezometer on Figure 10. Also shown on the plots are lines of
equal pore pressure ratio. Like the West data, it is clear from the plots that the pore
pressures are highest near the center of the dam. With the exception of the first
portion of measurements from piezometer EA, in each plot the piezometer nearest
the centerline has the highest pore pressure. To further illustrate this, the pore
pressure ratios for August 1999 are plotted and contoured on the dam section on
Figure 11. The data show that the pore pressure ratios are highest near the center of
the core and least near the downstream chimney drain.

Figure 6. W est Dam Instrum entation Section 5 Pore Pressure and Em bankm ent Elevation vs. Tim e
174 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 175

0 10 20 30 40 50 60 70 80 90 100

«ft «*>* Distance along Section - meters <ift

Figure 8. West Dam Instrumentation Section 5 Pore Pressure Ratios August 1999
P o re P re s s u re - m e te rs o f w a te r

Conclusions
A case history of the development and dissipation of pore pressures in a partially
saturated soil due to fill placement is presented. The pore pressures were measured in
the earth cores of two major embankments during construction. The pore pressures were
measured using vibrating wire piezometers with coarse porous tips and, therefore, were
measuring approximate pore air pressures in the partially saturated soils. It is concluded
that due to the low plasticity of the core materials, after some initial loading, the pore air
and pore water pressures are, for practical purposes, the same.
Analyses of the construction data indicate that it is unlikely that the core material at the
0 10 20 30 40 50 60 70 80 90 100
levels of the piezometers became saturated, except perhaps at the lowest level of
piezometers, and then only near the end of construction. The data indicate that the pore
pressure development is negligible during the first 5 to 15 meters of fill placement,
increasing at varying rates after this depending on the proximity to drainage layers. This
is consistent with measurements from other embankment dams (Sherard, 1981), which
P o re P re s s u re - m e te rs o f w a te r

indicated the same trend of pore pressure development versus fill height. The highest
pore pressure ratios measured were on the order of 0.4 to 0.5 near the central and lower
portions of the core and typically 0 to 0.1 near the chimney drain.
The height of fill, compressibility of the fill, the degree of saturation, and the ability to
dissipate pore pressures influence the magnitude of the pore pressure at any location
within the dam core. The factors affecting pore pressure magnitude are complex and
interrelated. All of these factors change during construction. Though complex, it is
hoped that the case history presented will add to the knowledge and understanding of the
development of pore pressures in partially saturated soil.
0 10 20 30 40 50 60 70 80 90 100 Appendix. References
F ill H e ig h t a b o v e P ie z o m e te r - m e te rs
Sherard, J.L.( 1981).’’Piezometers in Earth Dam Impervious Sections.” Recent
Figure 7. W est Dam - Instrum entation Section 5 Pore Pressure vs Fill Height Developments in Geotechnical Engineering For Hydro Projects, ASCE. 125-165.
176 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 177

80

70

P o re P re s s u re - m e te rs o f w a te r
Elevation - meters 60

50

40

30

20
10

0
0 10 20 30 40 50 60 70 80 90 100

Date F ill H e ig h t ab o v e P ie z o m e te r - m eters

80

70

P o re P re s s u re - m e te rs o f w a te r
60
Elevation - meters

50

40

30

20

10

0
0 10 20 30 40 50 60 70 80 90 100
Date Fill H e ig h t ab o v e P ie z o m e te r - m eters

F o re P re s s u re * m e te rs o f w a te r
Elevation - meters

0 10 20 30 40 50 60 70 80 90 100
Date Fill H eight a b o v e P ie z o m e te r - m eters

Figure 9. East Dam Instrumentation Section 5 Pore Pressure and Embankment Elevation vs Time Figure 10. East Dam Instrumentation Section 5 Pore Pressure vs Fill Height
178 GEOTECHNICAL MEASUREMENTS

PERFORMANCE OF AN ANCHORED SHEET-PILE WALL

Shailendra N. Endley1, Wayne A Dunlap12
David M. Knuckey3 and Karun Sreeranta1

Abstract

Port of Freeport has built a new General Cargo Dock and Transit Shed.
The project included a 12m high-anchored sheet-pile wall and a relieving
platform supported by auger cast-in-place piles.

Detailed instrumentation of the dock system was undertaken. The

instrumentation included inclinometers, strain gauges and total earth pressure
cells on the sheet-pile wall and strain gauges on the anchor rods. Measurements
were made over a period of six years to determine deflection, moment and soil
pressure on the sheet-piles and tension on tie backs at various stages.
Figure 11. East Dam Instrumentation Section 5 Pore Pressure Ratios August 1999
The deflections measured in the sheet-pile wall were surprisingly larger
than originally calculated. A parametric study was conducted to determine what
factors might have been responsible for the large and unexpected deflections. It
was concluded that the high deflections were due to the manner in which the
backfill was placed and the reduced section modulus. It was determined that the
manner in which the sheet-pile was installed did not provide the section modulus
as expected. Alternative methods of construction should be examined for future
dock construction, which will eliminate the high construction stresses observed in
the present project.

Introduction

In 1985, the Port of Freeport initiated plans to develop a wharf and dock
facility for the American Rice Co. The final structure, completed in 1987,
included a 183m long dock with a water depth of 9.7m, MLW. Plans included
dredging the harbor to a depth of 11.6m (plus 0.6m of overdredging) in the near
future. The dock consisted o f a sheet-pile bulkhead and a relieving platform, a
common structure along the Texas Gulf Coast (Figure 1). A unique feature of the
dock was the use of auger-cast concrete piles to bear the weight o f the dock loads.
Not only did the auger-cast piles provide significant savings over conventional
driven piles; they eliminated the possibility of vibration and backfill displacement
which might have created additional lateral forces against the sheet-pile bulkhead.

1Chief Engineer, Professional Service Industries, Inc., Houston, Texas

' Professor, Texas A&M University, College Station, Texas
3 Director of Engineering, Port of Freeport, Freeport, Texas

179
180 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 181

Due to the uncertainties concerning lateral pressures in relieving type This analysis was done for the final dredged depth of 12 meters and for a design
platforms, particularly with the use of auger-cast piles, the Port of Freeport stress in steel of 250 M Pa.
authorized a study to evaluate the new dock structure with a view that possible
savings could result with future docks. The study included instrumentation of the Port of Freeport decided to use Larssen's "Vs Section" as shown on Figure
bulkhead and tie rods, measurements during and after construction and analyses 2. The suppliers section modulus was 970 cm ’ per unit length of wall.
of the results before and after harbor dredging. This paper discusses results of
measurements and analysis.
Instrumentation

Soils Data Instrumentation was done on various components of the anchored sheet-pile
wail. The instrumentation included:
A detailed soil investigation was performed. The underwater bank of the
ship channel sloped up gently to land from a maximum channel depth of 9.7m (1) Inclinometers at six locations along the 183 meter long wall, to measure
MLW. The soils consisted of stiff overconsolidated clays and dense sands deflections.
overlain by recent river deposits. Both total stress and effective stress tests were
performed to evaluate the end of construction and long term stability conditions (2) Earth pressure cells at three locations behind the sheet-pile wall.
for the bulkhead design.
(3) Strain gages at three locations along the wall to determine strains in the
The sand used for the backfill behind the sheet-pile bulkhead was sampled sheet-pile wall. At each location seven strain gages were placed equally
and tested before and during backfill operations. Properties shown in Figure 1 spaced between elevations 0 to -21.6 meters.
were used to design the sheet-pile. Ranges of soil properties were used during the
studies. The sand fill used behind the sheet-pile had less than 5 percent fines. A (4) Load cells on three tie rods, two at each location (front and back o f the tie
friction angle of 20 degrees was assumed for the short-term condition as the sand rod). These load cells were calibrated to measure tie rod tension.
being placed in 30 feet of water.
Instrumentation details are published paper presented by Endley, et. al. (1).

Design Considerations
Construction Sequence
Based on the soils data provided, the sheet-pile bulkhead was designed by
the structural engineer using the free earth support method applying Rowe’s Construction o f the sheet-pile bulkhead began in August 1986 with a
moment reduction. Further, beam column analysis was performed using the p-y barge-mounted dragline stripping approximately .6m of soft detritus from the
method. The results called for a sheet-pile section with a section modulus of underwater slope. Immediately thereafter, the sheet-piles were driven. Strain
721cm" per unit length of wall. The sheet-pile extended to an elevation of 21.6m gages on the sheet-piles were checked for continuity upon completion of the
MLW. Results of the beam column analysis are shown below. driving. After the whalers and tie rods were installed, the tie rod load cells were
also checked and then protected from the ensuing backfill operation by encasing
P-Y Curve P-Y Curve P-Y Curve P-Y Curve the load cells in pipe. At this point, some gages on the sheet-piles were found to
Active-Passive Active-Passive API Code API Code be drifting but still operable while all the tie rod lead cells were functional.
Long Term Short Term Long Term Short Term Inclinometer casings were grouted in place and the base readings were made and
Mmax 155 kN-m 103 kN-m 113 kN-m 117 kN-m the earth pressure cells were installed behind the wall with the help of divers.

Y m ax 12.4 cm 5.9 cm 4.3 cm 4.1 cm All the backfill operations were conducted from barges using clamshell
bucket. The process of dropping the sand from a clam shell bucket approximately
T m ax 101 kN-m 73 kN-m 66 kN-m 66 kN-m 3 to 5 m. above the water surface proved to be devastating to several tie rod load
182 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 183

cells as indicated by continuity checks made immediately after the backfilling couple of months (January ’87), they were still much higher than is usually
operation. Most of the damage was to the wiring and it was possible to conduct assumed for granular materials, and they were certainly much higher than the
on-site repairs. At this time readings were also made on the inclinometers and value of 0.33 to 0.5 assumed in the design of the bulkhead. The tie rod tension
sheet-pile strain gages to determine the effect of backfilling operations on the ranged from 71 kN/m to 106 kN/rn.
sheet-piles. Following backfilling operation, the auger-cast piles were installed
on 2m centers. At this stage the depth of water was about 9.7 meters. At this stage, several of the sheet-pile strain gages had become inoperable,
some due to the contractor’s operations, some to moisture penetration. However,
none of the operable gages indicated stresses, which were higher than 154 M Pa.
Instrumentation Results To further determine the behavior of the sheet-pile bulkhead, the displacements of
the bulkhead obtained from the inclinometer data were used to obtain bending
The initial readings of the inclinometers after backfilling (October ’86) moments and stresses. This was accomplished by first determining the radius of
showed rather alarmingly high sheet-pile deflections - on the order of 13cm in the curvature along the sheet-pile, and then calculating bending moments (M) at each
south (toward the water) direction (Figure 3). The deflection readings were about node based on the sheet-pile moment of inertia (I) and modulus o f elasticity (E).
the same in all six inclinometers. In the east-west direction deflections, shown in
Figure 3, averaged about 3cm toward the east along the wall. However, Following are finite difference equation was used to determine moments
evaluation of the strain gage data did not indicate significantly high stresses from the deflection readings obtained from inclinometer:
associated with these high deflections. The corresponding bending moments in
the sheet-pile are shown in Figure 4. It should be realized that this channel at this
stage was 9.7 m deep whereas final design was done for a 12 in depth. _ El ( Dm - Du - D i+i + D i-3_)__ ;

The next set of readings was taken in December 1986, immediately after
the auger-cast piles were completed. On average, there was a slight decrease in where:
deflection on the order of 1-2 cm in the south direction.
M, = Moment at ilh node
In November 1987, after completion of the dock and transit shed, another D, = Deflection of i,h node
set of readings were taken. These readings showed no substantial change from X = length between nodes
the previous readings. Since the performance of the bulkhead indicated no
change owing to the weight of the completed structure, but the deflections were Stresses were obtained by dividing the moment by the manufacturer’s specified
still considered high, it was thought that the deflection measurements might be in section modulus. The results of these calculations for average and stresses are
error. Consequently, a diver made measurements, which extended to the mudline. shown on Figure 6. As can be seen the stresses, as calculated based on bent shape
of sheet-pile, are approaching yield stress of about 235 MPa.
Although the diver’s readings and the inclinometer readings were not a
prefect match, they compared closely enough to convince us that the inclinometer The major problem with the high stresses was whether the planned
readings were indeed correct. deepening of the harbor from existing 9.7m to 12m in front of the bulkhead would
increase the stresses beyond the yield stress. Since the original design utilized the
p-y approach as an aid in determining the stresses, this approach was again taken
Analysis of the Data to understand the cause for the high deflections. Using typical values for K of
0.5, the manufacturer’s pile properties and the stiffness o f the tie rods, several
At this stage, the available data were analyzed in detail to ascertain the cases were run, none of which showed deflections near as high as the observed
cause for the large deflections. The first clue came from the pressure cell deflections. Subsequently, the value of K was changed to 1.0, which is about the
readings, shown in Figure 5 in terms of an earth pressure coefficient based on the average value measured immediately after completion of the backfill. It was
effective stresses. The readings taken immediately after the backfill placement obvious that the tie rod tension had an influence on the computed deflections, and
(November ’86) showed values near 1.0 at a depth of 6m below the top of the several assumptions were made for the tie rod stiffness including one which had
sheet-pile and 0.7 at 9m depth. Although these values decreased somewhat after a some initial slack in the rods before the wall moved enough to start tensioning the
184 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 185

rods. The only assumptions which gave calculated deflections close to the (position C-C). Axes B-B and C-C give pile properties which are fairly close to
measured ones, utilized a moment of inertia which was one half of the those for a single sheet-pile with principle axis at D-D.
manufacturer’s suggested value (the logic for this will be discussed later)
Table 2 shows the results of p-y calculations using the two-section modulii
The high K values measured by the pressure cells persisted for well over believed to be the closest to reality. The section modulus of 1414 cm3 per unit
one year, at which time the wires to the cells were accidentally cut. With time, length of wall, which corresponds to an inclination of the neutral axis of 28
the pressures decreased about 15% but the values still ranged between 0.65 to degrees, seems to provide a deflection, which is closer to the measured deflections
0.80. One possible explanation is that the process of dumping the sand from a before dredging. The calculated stress with this section modulus is about 216
significant height above the water momentarily created a dense semi-fluid, which MPa, well below the yield stress. The results of studies show that the stresses
exerted a high enough pressure to cause the large deflections of the sheet-pile. As after dredging 2.4m would increase to somewhat less than 316 MPa still below
the pore water escaped from the sand and it transformed into a solid mass, the the yield stress of 385 M Pa, even though the deflections would increase to nearly
pressure against the sheet-piles would decrease and the piles would attempt to at 19cm.
least partially spring back. However, as they attempt to do this the sand is now
passively resisting the movement. O f course the passive resistance is much TABLE
higher than the active resistance. Thus, the high earth pressure may actually be
passive pressures rather than active ones. RESULTS OF ANALYSES USING VARIOUS
SECTION MODULUS VALUES
This still does not totally explain the high deflections. In Bowles, 1988<2’,
there is a discussion regarding the location of the neutral axis for sheet-piles as Before Dredging
practiced in the U.S. and Europe (Figure 7). By crimping two sections together,
the manufacturers of the Larssen piles infer that the neutral axis is located along Section Modulus Deflection Bending Moment Stress
the line of the interlocks, which also happens to be the position of maximum, Case (cm) (cm) (N.m) (M.Pa)
shear stress. There has been some doubt that the crimping can resist these shear
stresses and prevent movement between the adjacent piles. If the piles act 1 590 6.5 85633 160
individually, and the neutral axis shifts up or down as shown by Bowles, then the
moment of inertia is also reduced to about one half of the values presented by the II 426 10.7 84176 216
manufacturer. The p-y studies showed that these lower values, in conjunction
with the high earth pressures, would give deflection values near the measured After Dredging
ones.
Section Modulus Deflection Bending Moment Stress
Apparently, the question regarding the location of the principle axis has Case (cm) (cm) (N.m) (M.Pa)
been raised before. The effect of crimping on the Larssen sections has been
discussed by previous researchers, but only from the viewpoint that the crimping III 590 10.7 123913 231
was adequate to resist movement along the interlocks.
IV 426 18.7 122082 316
The problem seems to be more fundamental, however. Figure 8 shows the
results of calculations to obtain the moment of inertia for the sheet-piles. Position
A-A is the principle axis recommended by the manufacturer. In reality, the
principle axis should be located along B-B, at an angle of 28 degrees with the line
Based on these calculations, it was recommended that the dredging be
of the interlocks. This would explain the significant East-West movement
allowed to proceed, although the uncertainties associated with the sheet-pile
measured by the inclinometers (Figure 3). In fact, if the angle o f the principle axis
properties dictated some caution. Consequently, the dredging was performed in
is based on the geometrical relationship between the measured East-West and
two stages (1.2 m at a time) with deflection measurements made after the first
North-South deflections (using average values), then 12 degrees should be used
stage to ensure that sheet-pile movement was within the calculated limits.
 GEOTECHNICAL MEASUREMENTS GEOTECHNICAL MEASUREMENTS 187
186

The deflection measurement showed that neither stage of the channel ACKNOWLEDGEMENTS
deepening produced significant changes in the deflections (Figure 9), justifying
the decision to continue with the dredging. However, there is still some Mr. Carl Frederickson performed the strain gages and tie rod load cells
uncertainty regarding the earth pressures from the sand backfill, particularly installation and Mr. Richard Bartoskewitz supervised the installation of the
whether the passive pressure was actually acting on the bulkhead. If it were, then instruments.
the small additional deflections associated with the dredging could actually have
released some pressure on the sheet-piles. In the extreme case, the backfill
pressures would revert to the active case, with pressures of approximately one REFERENCES
half of those measured initially by the pressure cells. Unfortunately, the loss of
the pressure cells prevented us from knowing whether this was the case. 1. Endley, S.N., Dunlap, Snow, C, Knuckey D., Briaud, J.L., Lowery, Lee
“Performance of an Anchored Sheetpile Wall”, ASCE Texas Section
Figures 10 and 11 show the bending moments and stresses developed in Spring 1991, San Antonio, Texas.
the sheet-pile, calculated from measured deflections of the inclinometer.
2. Bowles, J.E., “Foundation Engineering”, McGraw-Hill Book Co., New
CONCLUSIONS York, 1988.

Further data acquisition and analysis on the project is sill in progress, but
on the basis of the measurements, the following conclusions arc appropriate:

1. High lateral earth pressures resulted from the backfill construction

methods, which in turn, produced large sheet-pile deflections and stresses.
Less aggressive placement methods such as hydraulic dredging or
tremieing would reduce the lateral earth pressures.

2. The installation of piezometers to measure backfill pore pressures during

and after construction would help to evaluate the effect of the construction
method on the lateral pressure in the backfill.

3. The evidence from the measurements indicates that the manufacturer’s

values of section modules and moment of inertia are not correct for the
Larssen vs. sections. Crimping (or even welding) the two sections does
not alter the properties.

4. The question of responsibility for instrumentation on a construction

project is age old, but the contractor must be made to share the
responsibility for instrumentation protection during construction.

5. The section modulus and moment of inertia of the specified sheet-pile

sections should be independently verified, especially when some unique or
not so common sheet-pile sections are used.
 GEOTECHNICAL MEASUREMENTS 189

Figure 1 Com ponents of W harf and Dock No. 3, Port of Freeport

Section Modulus, per foot of Wall = 975 cm3

Figure 2 Properties of Larssen “ Vs Section”

TRANSIT SHED

Figure 3 Measured Deflections (October

1986)

Deflection (mm)
-150 -125 -100 -75 -50 -25 0 25

Depth (meters)
 O
O

Figure 4 Bending Moments Calculated From Deflections, October 1986

400
Inclinometer #1
300

GEOTECHNICAL MEASUREMENTS
- HB- Inclinometer #2
- ♦ - -Inclinometer #3
200
Inclinometer #4
100 - -B -
0 --O -

O) 25 2D
c - l o o 1-
*u
-200

-300

-400
Depth (meters)

Figure 5 Measured Earth Pressures

0 0.5 1 1.5 2
Earth Pressure Coefficient, K
0 1
# November '86
o
O January '87 rr
0
H
3 " m
n
i
n
>
00 • •
B 6
c.
o
Û 9 o •
m
z
H
c/i

12

15
 JL.
1

Figure 6 Sheet Pile Stresses Calculated From Deflections, October 1986

400000
— • — Inclinometer #1

GEOTECHNICAL MEASUREMENTS
I— * — Inclinometer #2
Inclinometer #3
— 0— Inclinometer #4
«■ 100000

25 2D
55 -100000

Depth (meters)

■
(A) NEUTRAL AXIS LOCATION AS COMMONLY USED IS DESIGN

GEOTECHNICAL MEASUREMENTS

(B) NEUTRALAXIS LOCATION ALONG INTERLOCKS AS

COMMONLY USED IN EUROPEAN PRACTICE

Figure 7 Approaches used to Determined Neutral Axis (from Bowles)

vC
U>
 -1

o
■fc.

GEOTECHNICAL MEASUREMENTS
POSITION OF NEUTRAL SECTION MODULUS MOMENT OF INERTIA
AXIS (cm3 per unit length) (cm4 per unit length)

A-A (with crimping) 975 21,933

B-B (with crimping) 426 6,451
C-C (with crimping) 590 11,820
D-D (without crimping) 418 6,700

Figure 8 Larssen “Vs” Sheet-Pile Section Properties

■
Figure 9 Comparison of Deflections Before and After Dredging

GEOTECHNICAL MEASUREMENTS

o
 o'
O

Figure 10 Comparison of Bending Moments

Before and After Dredging

GEOTECHNICAL MEASUREMENTS
I
Figure 11 Comparison of Stresses in Sheet Pile
Before and After Dredging

GEOTECHNICAL MEASUREMENTS

•o
^1
 Subject Index
Page number refers to the first page of paper

Anchors, 138, 179 Monitoring, 91, 109

Axial loads, 1
Peat, 44
Bonding, 56 Piers, 151
Piezometers, 164
Case reports, 65, 80, 164 Piles, 109
Cohesive soils, 1 Plates, 44
Comparative studies, 80 Platforms, 179
Compression, 56 Pore pressure, 1, 164
Computer applications, 30 Pore water pressure, 164
Concrete, 44
Concrete structures, 151 Radiography, 30
Creep, 56 Remedial action, 126
Crushing, 56 Remote sensing, 65, 80
Repeated loading, 44
Dams, rockfill, 164
Data analysis, 126 Sampling designs, 16
Deformation, 44 Sand, 16, 30
Density measurement, 30 Sediment, 109
Differential settlement, 138, 151 Shafts, 91
Docks, 179 Shear deformation, 80
Shear strength, 1
Earth pressure, 179 Slope indicators, 80, 126
Elastic foundations, 44 Slope stability, 65, 91, 109, 126
Expansive soils, 151 Slope stabilization, 138
Soils, unsaturated, 151
Steady state, 16
Field investigations, 151
Strain gages, 151, 179
Field tests, 44
Fills, 138
Test procedures, 1
Tidal currents, 109
Grains, 56
Tieback restraint system s, 138
Torsion, 1
Instrumentation, 65, 80, 91, 164, 179 Triaxial tests, 30

Landslides, 65, 91, 126 Underground structures, 91

Lateral loads, 109
Light rail transit, 91
Void ratio, 30
Liquefaction, 16
Load combinations, 1
Wharves, 109
M odeling, 56
M odulus o f elasticity, 44

199
 Author Index
Page number refers to the first page of paper

Alshibli, Khalid A., 30 Naeini, Seyed Abolhassan, 16

Anderson, N eil, 65 Nelson, John D., P.E., 151
Arita, Arleen A ., 164
O ’Connor, Kevin M., 80
Batiste, Susan N ., 30
Baziar, Mohammad Hassan, 16 Penumadu, Dayakar, 1
Biever, Mark R, 91
Smith, Michael G., 164
Carville, Chester A., 138 Squier, L. Radley, 91
Chapel, Thomas A., P.E., 151 Sreerama, Karun, 179
Costes, N icholas C., 30 Stevens, William R., 126
Sture, Stein, 30
Dowding, Charles H., 80 Swanson, Roy A., 30
Dunlap, Wayne A ., 179
Tanenbaum, Ronald J., P.E., 138
Endley, Shailendra N., 179 Thomas, Jonathan E., P.E., 109

Feda, Jaroslav, 56 Vakher, Mark, 44

Jernigan, Russell L., 164 Welch, David, 65

Wong, Rex K. H„ 91
Knuckey, David M., 179
Zehrbach, Bill E., 126
Lankton, Mark R., 30

M andeville, Douglas, 1
Morin, Wilbur J., RE., 109
Myers, Barry K„ 91

200
 G eotechnical S pecial P ublication

Geotechnical
Measurements
Lab and Field
Edited by W. Allen Marr

GEO
INSTITUTE

N um ber 106
 eotechnical Measurements: Lab and Field provides insight
G into the sta te of the practice in geotechnical m easure­
m ents. Geotechnical m easurem ents include observations of the
physical sta te of subsurface m aterials and how th a t state
responds to changes in the physical environm ent. Both labora­
tory and in situ m easurem ents are essential to geotechnical
practice and have provided the lifeblood for advances in mod­
ern geotechnical engineering.

This proceedings is conveniently divided into two sections.

The first focuses on laboratory m easurem ents and th eir use to
fu rth er understanding and characterizing geotechnical behav­
ior. The second section focuses on the application of m ethods
to m easure actual perform ance of the geotechnical elem ents of
constructed facilities.

These technical papers were presented a t the GeoDenver

2000 Conference, sponsored by the G eo-Institute of the
Am erican Society of Civil Engineers, held in Denver, Colorado,
August 5-8, 2000. The objective of the conference was to foster
an exchange of data on innovative practice based on data from
actual projects.

ISBN 0 -7 ñ 4 4 -0 5 1 fl-2

Am erican Society
o f C ivil Engineers 9 780784 405185