Download - geng166-0031
-
8/12/2019 geng166-0031
1/18
Proceedings of the Institution of Civil Engineers
Geotechnical Engineering 166 February 2013 Issue GE1
Pages 3148 http://dx.doi.org/10.1680/geng.9.00080
Paper 900080
Received 30/09/2009 Accepted 10/01/2011
Published online 09/02/2012
Keywords: embankments/field testing & monitoring/geotechnical
engineering
ICE Publishing: All rights reserved
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg
Stability for a highembankment founded onsulfide clayj1 Rasmus Muller MSc, CEng
Senior Geotechnical Engineer, Tyrens AB, Borlange, Sweden
j2 Stefan Larsson PhD, CEngProfessor, Royal Institute of Technology, Stockholm, Sweden
j3 Bo Westerberg PhD, CEngAssistant Professor, LuleaUniversity of Technology, Lulea, Sweden
j1 j2 j3
During staged construction of embankments on clay foundations, the undrained shear strength su increases due to
consolidation during the construction process. The increase is usually related to the pre-consolidation pressure 9p by
way of the ratio su=9p and is important when assessing the stability of an embankment. Properties controlling the
increase are usually associated with various unknowns that can be difficult to predict before construction. A case
involving a large embankment built on vertically drained sulfide clay is presented. Empirical knowledge and
experience of similar constructions on sulfide clay were limited, therefore there were uncertainties of the soildraininteraction and how the soil would behave under the embankment load. A trial embankment was built to gather
knowledge and experience of this particular soil and the observational method was adopted. The study presents
how embankment stability was predicted at design stage and controlled during construction. It highlights the
importance and usefulness of obtaining measurement data from different types of measurements. Laboratory tests
and a large number of in situ tests at different stages during construction were performed to assess su=9p ratios in
the sulfide clay. The mean su=9pratio was estimated by means of statistical analysis to 0 .25 for a large stress interval.
Notationa constant dependent on soil type and failure mode
b constant dependent on soil type
Cc compression indexcv coefficient of consolidation in the vertical direction
C creep index
CPTu cone penetration test with pore pressure
measurement
CRS constant rate of strain oedometer test
CUK0DSS K0 consolidated undrained direct simple shear test
FC fall cone test
FCOED fall cone test on oedometer consolidated samples
FS total factor of safety
FVS field vane shear test
g unit weight
kv vertical hydraulic conductivity
OCR overconsolidation ratio
PVD prefabricated vertical drain
qt cone resistance measured by way of CPTu
SHANSEP stress history and normalised soil engineering
properties
su corrected undrained shear strength
su,FC=FVS undrained shear strength measured via fall cone or
field vane test
USA undrained strength analysisu pore water pressure
wL liquid limit
wm natural water content
9p pre-consolidation pressure
v total vertical stress
9v effective vertical stress
j9 angle of shearing resistance
1. IntroductionDuring construction of a staged embankment founded on clay the
drainage conditions can be classified as partly drained (Ladd,
1991). The weight of the embankment induces positive excess
pore pressures, and consolidation (increase in effective stresses)
is ongoing. If a stability analysis is performed under these
conditions, an analysis in terms of total stresses is usually
preferable (more simple) over an effective stress analysis (Lacasse
et al., 1977; Lambe and Whitman, 1979; Leroueil et al., 2001;
31
-
8/12/2019 geng166-0031
2/18
Wood, 1990). An effective stress analysis requires accuratepredictions of the pore pressure development in the design phase
and very careful control of the development during construction.
Furthermore, an effective stress analysis may overpredict the total
factor of safety (FS). Several recently published studies have
adopted total stress analyses to evaluate stability during staged
construction (e.g. Shibuya et al., 2008; Watabe et al., 2002). A
total stress analysis based on the initial undrained shear strength
su0 in the clay would in most cases be conservative (i.e. lead to
more expensive measures to achieve the required degree of
stability). It is therefore preferable to utilise the increase in su
that occurs owing to consolidation and effective stress increase
during construction, which can be done by way of an undrained
strength analysis (USA) (Ladd, 1991). The basic components of a
USA involve the following principal steps
(a) evaluation of the stress history of the soil in terms of initial
in situ effective stress 9v0 and pre-consolidation stress 9p0
(b) estimation of the changes in stresses imposed by the proposed
loadings by way of consolidation analyses
(c) laboratory tests to evaluate su at different consolidation
stressessu=9p(d) stability analyses using su profiles calculated in step (c).
In order to fulfil step (c) above, the use of K0 consolidated
undrained active and passive triaxial tests and direct simple sheartests is suggested (Ladd, 1991). These tests should be conducted
by way of either the recompression technique (Bjerrum, 1973) or
the stress history and normalised soil engineering properties
(SHANSEP) technique (Jamiolkowski et al., 1985; Ladd and
Foott, 1974). The recompression technique means consolidating
the samples to 9v0 before shearing. The SHANSEP technique
essentially requires consolidating the soil samples to stresses well
beyond9p0 and then unloading them to various overconsolidation
ratios (OCR 9p=9v) before shearing. The limited applicability
of the SHANSEP technique, owing to alteration of the soil
structure, has been the subject of discussion in several studies
(Ladd, 1991; Mesri, 1975; Tavenas and Leroueil, 1980). InSweden the recompression technique is normally adopted in order
to avoid destructuring of the clay (Lofroth, 2008; Westerberg,
1999).
It is common to relate su to 9v and OCR by using the following
constitutive relationship (Ladd and Foott, 1974; Roscoe and
Burland, 1968;Schofield and Wroth, 1968)
su=9v a OCRb1:
where a is dependent on the type of soil and the failure mode
(triaxial compression, direct shear or triaxial extension) and b is
dependent on the type of soil. Typical values of b range from 0.7
to 0.9 with 0.8 being a reasonable estimate for most soils
(Larssonet al., 2007a). Any loading and subsequent consolidation
leading to a stress state where 9v > 9p0 results in a normally
consolidated stress state in the clay where 9v 9p: Equation 1can then be rewritten as
su=9p a2:
For classic limit equilibrium analysis of short-term stability of
embankments, direct simple shear (DSS) tests are suggested as the
most relevant test; these simulate the behaviour around a postulated
failure surface most closely (Wroth, 1984). Several researchers
state that peak strengths from DSS tests provide reasonable
estimates of the average undrained strength ratio (su=9p) along
potential failure surfaces (Ladd, 1991; Larsson, 1980). This
assumption implies a constant value ofa for a certain soil. Values of
aranging from 0.15 to 0.50 have been suggested for different types
of soils in numerous studies (e.g.Chung et al., 2007;Jardine and
Hight, 1987;Karlsson and Viberg, 1967;Kim et al., 2009;Ladd,
1991; Larsson, 1980; Mayne, 1980; Mayne and Mitchell, 1988;
Nicholson and Jardine, 1981; Shibuya et al., 2008; Suzuki and
Yasuhara, 2007; Tanaka, 1994; Watabe et al., 2002). From these
studies it can be concluded that the su=9pratio varies with soil type.
There is usually scatter in obtained values from either in situ tests or
laboratory tests and different test methods in field or laboratory
might render different values. An average value of a 0:22 is
suggested for inorganic clays (Jardine and Hight, 1987;Ladd, 1991;
Larsson, 1980, Mayne, 1980) and a 0:25 for organic clays andsilts (Jardine and Hight, 1987; Ladd, 1991; Larsson, 1980).
Stability analyses, according to the USA methodology used for
designing embankments built using staged construction methods,
require reliable predictions of su at the various stages. The
predictions involve estimations of su=9p, 9v and 9p based on
consolidation analyses. These predictions and consolidation ana-
lyses are, however, associated with uncertainties, thus results
from investigations (i.e. field measurements and laboratory tests)
at a specific site should preferably be supported by experience
from similar projects involving similar ground conditions and
empirical knowledge about the behaviour of the soil in question.
This paper presents how stability during construction was pre-
dicted and controlled in a project involving a large embankment
built on vertically drained sulfide clay in Sweden. USAs based on
predictedsu profiles were used to calculate FS at critical stages of
construction of the embankment. Previous experience of construc-
tion of large embankments on sulfide clay and empirical know-
ledge related to the interaction between vertical drains and this
type of soil was limited. These shortcomings were compensated
for within the design and construction phases by using information
gathered from a trial embankment and the use of the observational
method. Predictions of su used in stability calculations in the
design phase were compared with measurements during construc-
tion. Furthermore, the paper aims at increasing understanding of
the behaviour of sulfide clay with the help of the relationship
su=9p obtained from laboratory tests and evaluated from in situ
measurements ofsu and pore water pressure u during construction.
32
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg
-
8/12/2019 geng166-0031
3/18
2. Description of the projectThe Veda embankment was built to navigate over an approximately
300 m wide valley situated near the river Angermanalven, 55 km
north of the city of Sundsvall in Sweden (Figure 1(a)). This section
is a part of the Adalsbanan railway line which connects the East
coast railway line with the recently built Botnia line along the north-
eastern coast of Sweden. The subsoil strata in the valley generally
consist of soft sediments of silt, clay (partly sulfide clay) and sand.
The embankment was made of crushed rock fill (mainly granite and
gneiss) derived from a nearby tunnelling project, and rises to aheight 16 m above the original ground surface (Figure 1(b)).
Sulfide soils mostly occur in tropical areas around the world.
The occurrence of sulfide soils in Europe is limited, apart from
the coastal areas around the Gulf of Bothnia, where the
embankment is situated (Figure 1(a)). Common to all sulfide
soils is the presence of sulfide in the form of pyrite (FeS 2) and
iron monosulfide (FeS). The sulfide soils that are found in
Sweden around the Gulf of Bothnia usually have a relatively
high organic content as they were formed as organic matter
degraded in fine-grained sediments under anaerobic conditions
(Larsson et al., 2007b). Typical ranges for properties of these
sulfide soils are total sulfur content of 0.12%, total iron
content of 25%, and organic content of 27% of dry weight
(Westerberg and Andersson, 2009). If the sulfide soil is sub-
jected to oxygen (e.g. excavated), there can be negative environ-
mental impacts such as lowering of pH in ground- or surface
water and leaching of metal ions (Macsik, 1999). In Sweden
there are restrictions regarding the handling of sulfide soils, and
excavation should be avoided.
The embankment load imposes a considerable stress increase in
the soil, which in turn leads to deformation and stability
problems. In order to address the stability issues, support berms
consisting of moraine fill were constructed and the embankmentwas built using staged construction, making it possible to account
for the increase in su in the clay due to consolidation effects.
Prefabricated vertical drains (PVDs) were installed in order to
accelerate consolidation. Recently published studies present re-
sults from in situ measurements of the su increase in vertically
drained clay during staged construction of embankments and fills
during land reclamation (Bergado et al., 2002; Bo et al., 2007;
Loet al., 2008;Long and ORiordan, 2001;Suzuki and Yasuhara,
2007). However, as experience of embankments of this particular
magnitude on sulfide clay is limited, uncertainty about strength,
deformation and consolidation characteristics of the subsoil
prevailed. Generally, pilot tests are recommended owing to thecomplexity and difficulty in predicting the parameters required to
make an accurate design of ground improvement works involving
PVDs (Hansbo, 1997). In the present case a trial embankment
was built and instrumented 1 year before the beginning of the
main embankment works. The design and construction processes
were performed according to the observational method (Peck,
1969). Various parameters were monitored during construction of
the embankment, such as su from field vane shear (FVS) tests
and from cone penetration tests (CPTu), and compared with
design assumptions and predictions.
The geotechnical design requirements issued by the client
concerning stability stated that the FS was to exceed 1.5 in both
the short term (during construction) and the long term. In the
present case the critical phase was the short-term situation or, in
other words, the partially drained situation (i.e. partially consoli-
dated), when the consolidation process was ongoing.
Norw
ay
Swed
en
KalixLocation ofsulfide soils
Finlan
dUme
Both
nia
nBay
VedaEmbankment
Sundsvall
Stockholm
Gteborg
0 250 km 500 km
(a)
25
20
15
10
5
0
5
10
15
Elevation:m
0 100 200 300 400
Length: m(b)
Crestofembankment
Original ground surfaceSectionA
SectionB
SectionC
Silty clay
Sulfide clay
Glacial clay
Sand
Moraine
Figure 1.Veda area: (a) location plan (partly fromSchwab, 1976);
(b) elevation through centre of embankment
33
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg
-
8/12/2019 geng166-0031
4/18
3. Initial ground conditionsThe ground surface in the valley was situated about 4 4.5 m
above sea level (Angermanalven river) near the centre of the
embankment. A deposit of clay, silt and sand up to 16 m thick
was present in the bottom of the valley. The sediments overlaid a
firm moraine (Figure 1(b)). The upper 12 m consisted of silty
clay, followed by an approximately 3.5 m thick layer of post-
glacial sulfide clay, which in turn overlaid 35 m of glacial clay
with silt and sand layers. Near hydrostatic pore-water pressures
prevailed corresponding to a water table situated near the ground
surface.
Site investigations in the area consisted of probing, mainly by
way of CPTu, sampling and pore pressure measurements. Labora-
tory tests on undisturbed samples of clay taken from various
depths were also carried out at six locations. Extractions were
made using the Swedish standard piston sampler (SIS, 2007),
pushing the clay sample into 170 mm long plastic tubes with an
inner diameter of 50 mm. Index properties such as liquid limit
wL, natural water content wn and total unit weight were
determined and su was evaluated by fall cone (FC) tests.
Sedimentation analyses were also performed on samples of the
sulfide soil, indicating clay contents (by total weight) of 2732%
and silt contents of 6065%. Total sulfur, iron and organic
contents (by total dry weight) were determined as 0.3 0.4%,
3.1 3.7% and 2.8 2.9% respectively. These ranges fall well intothose presented as typical ranges for sulfide soils, and the soil
was defined as sulfide clay. Furthermore, from samples taken at
two locations, compression index Cc, creep index C, vertical
hydraulic conductivity kv, coefficient of consolidation in the
vertical direction cv and 9p of the clay were determined by way
of constant rate of strain (CRS) consolidation tests and standard
incremental load oedometer tests. Evaluations of su at different
9p were made by FC tests, performed on samples which had
undergone incremental load oedometer tests (FCOED). The
undrained shear strength in the clay was also determined in situ
by way of FVS and evaluated from CPTu.
According to Swedish practice, values of undrained shear strength
in clays evaluated from FC and FVS tests su,FC=FVS or CPTu
should be corrected by empirical relations when used for stability
calculations.
The corrected value of su in overconsolidated clay classified as
normal inorganic Swedish clay is calculated from FC or FVS
tests as (Larsson and Ahnberg, 2005;Larssonet al., 2007a)
su su,FC=FVS43
wL
0:45OCR
1:3
0:15
3:
From CPTu, su is calculated as (Larsson and Ahnberg, 2005;
Larssonet al., 2007a)
su qt v0
13:4 0:0665wLOCR
1:3
0:24:
When the clay is classified as sulfide clay, the corrected value su
is calculated from FC or FVS tests independent of consolidation
status as (Larssonet al., 2007b)
su su,FC=FVS 3 0:655:
From CPTu, su is calculated as (Larssonet al., 2007b)
su q t v0
20
OCR
1:3
0:2
6:
For clays with OCR, 1.3, the last expression in parentheses in
Equations 3, 4 and 6 should be set equal to 1.0. In this study
clays above elevation +4 m and below 1.5 m related to the sea
level, are classified as normal inorganic Swedish clay andbetween +2 m and1.5 m as sulfide clay (Figure 1(b)).
A selection of soil properties along with a typical result from a
CPTu in the area is presented in Figure 2. Profiles showing the
measured initial pore water pressure u0, evaluated 9v0 and 9p0
are also shown. The initial overconsolidation ratio OCR0 was
around 5 at the top of the clay deposit and around 1.3 at the
bottom. In the sulfide clay, Cc, C, wL and wn were significantly
higher while and kv were lower compared to the upper silty
clay and the bottom glacial clay. As seen in Figure 2, the ratio
C=Cc was in the order of 0.060.08 in the sulfide clay. Previous
studies (Mesri and Godlewski, 1977; Mesri et al., 1994) havepresented ratios in the order of 0.050.06. Hence, the evaluated
ratios were a little higher in the sulfide clay at Veda than the
findings for other organic clays presented in the aforementioned
studies.
Typical values of cv were in the range 0.951.4 m2/year over the
relevant stress range (50250 kPa).
There was a relatively large scatter in su obtained from laboratory
FC tests and in situ FVS and CPTu tests. The coefficient of
variation was 2030% in the sulfide clay, in line with the findings
ofLarssonet al. (2007b), and 2070% in the upper silty clay and
the bottom glacial clay. The large scatter in the upper silty clay
and the bottom glacial clay presumably arose from the inhomo-
geneities due to silt and sand layers. Unfortunately, more
sophisticated studies of su, for example by way of triaxial tests or
DSS tests, were not carried out at this stage of the project.
34
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg
-
8/12/2019 geng166-0031
5/18
4. The trial embankment
4.1 Purpose of the trial embankment
Construction of the trial embankment served the purpose of
increasing understanding of strength and deformation properties
in the soil concerning different aspects, such as the ratio su=9pand the consolidation rate in the different clay layers. Further-
more, the trial embankment enabled testing of the equipment and
techniques proposed for measurements of the planned Veda
embankment.
4.2 Description of the trial embankment
In November 2005, PVDs in a triangular pattern with 1.07 m
spacing and instrumentation for the trial embankment were
installed. Drains of the brand Membradrain MD 7007, with a
cross-section of 100 mm 3 3 mm, filter permittivity of 0.3 s1
and a discharge capacity larger than 3 m3/day, were used. The
crest of the trial embankment was situated about 30 m beside the
crest of the Veda embankment, made of crushed rock material
and constructed in two stages. The length and height were 50 m
and 6.7 m respectively (Figure 3).
4
3
2
1 0
1
2
3
4
Elevation:m
4
8
12
Coreresistance,
:MPa
qc
Silty
clay
Sulfide
clay
Glacial
clay
0
10
20
30
40
su0:kPa
FVSFCCPTMean
Design2
0
40
60
80
w
w
n
L
a
nd
:%
wnwLMeanwn
MeanwL
15
17
19
21
:kN/m3 M
ean
0
10
20
30
40
50
kv
10
:m/s(1
)
kvMeankv
0
2040
6080100
u0
v0
and
and
p0:kPa
pu0v0
p0
0
0204060810
Cc
CcMeanCc
00020
040
060
080
10
C
C
MeanC
Figure
2.
Soilp
rofile
and
selected
soilproperties
atthes
ite
before
construction
started
Piezometers
Concrete well
Settlement plate
Inclinometers
T1 T2
0 10 20
Length: m
Section1
Section2
Section3
(a)
Piezo-meters
Berm
Trialembankment
Concretewell Settlement plates
Inclinometers
Berm
PVDs c/c 107 m
15
10
5
0
5
10
15
Elevation:m
20 0 20
Length from centreline: m
(b)
Moraine
Sand
Glacial clay
Sulfide clay
Silty clay
Figure 3.Trial embankment: (a) layout plan; (b) cross-section
35
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg
-
8/12/2019 geng166-0031
6/18
4.3 Measurements and results from the trialembankment
Measurements of settlements, horizontal deformation and pore
pressure development were performed in three sections and
continued for 1 year, until the works with the Veda embankment
started in December 2006.Figure 4illustrates the staged construc-
tion sequence and corresponding measured excess pore pressures
u-excess at different elevations in the soil. Corresponding meas-
urements from different control sections showed good agreement,
therefore mean values are presented. Pore pressure measurements
were conducted with the use of BAT IS piezometers.
Two concrete wells (1.0 m in diameter), T1 and T2, were installed
through the embankment to enable measurements of su during
construction. In situ tests (FVS and CPTu) were made on two
occasions, after 50 days and 140 days. After 220 days, undis-
turbed sampling of the clay using the Swedish standard piston
sampler (SIS, 2007) and subsequent K0 consolidated undrained
direct simple shear tests (CUK0DSS) were performed. The load
from the Veda embankment was expected to induce normally
consolidated stress states at the stages when stability was to be
analysed, that is 9v . 9p0. Hence, evaluation of su at stresses
above 9p0 and with OCR 1 was of interest. Table 1 presents
the evaluated in situ vertical effective stress at the time of
sampling 9v,220, the consolidation stresses used in the tests
9p,DSS, the evaluatedsu values and the obtainedsu=9p ratios. Thesu profiles shown inFigure 5(a)are based on measurements made
50 and 140 days after the beginning of the filling operations,
together with the evaluated mean values from initial tests at the
location of the trial embankment. The profiles were evaluated
based on mean values from tests in wells T1 and T2.Figures 5(b)
and 5(c) show measured u, evaluated 9v and 9p profiles
representing the conditions after 50, 140 and 220 days respec-
tively. The 9v profiles were calculated based on initial stress
conditions, measured u profiles and the total stresses v induced
by the embankment fill ( 18 kN/m3). It can be seen that parts
of the curves representing 9v after 50 and 140 days indicate areas
where the clay was still somewhat overconsolidated. When su=9pratios were evaluated this was taken into account, assigning
b 0:8 in Equation 1 according to the recommendations given
by Larsson et al. (2007a). Evaluated su=9p ratios obtained at
different times and from different test methods ranged from about
0.2 t o 0.45 (Figure 6). Mean values in the sulfide clay were
approximately 0.29 and in the rest of the clay approximately0.33. The coefficient of variation ranged from 0.09 to 0.25 in the
sulfide clay and 0.1 4 t o 0.20 in the other clay layers. Values
chosen for design of the Veda embankment are also presented.
Measurements of ground surface settlements near the crest of the
trial embankment are presented in Figure 7(a). The settlement
plates used were placed on a sand blanket and made of square
steel plates (0.5 m 3 0.5 m) with a steel rod welded to the centre
of the plates. Polyvinyl chloride (PVC) tubes were used to protect
the steel rods from the rock fill. The tips of the rods were
monitored over time by surveying with geodetic total station. In
Figure 7(b), horizontal displacements at different elevations for aselection of time steps are shown. After 1 year, when measure-
ments were interrupted, the crest settlements were in the order of
1.0 1.2 m. Measurement of horizontal displacements was inter-
rupted after about 200 days. At that time, the maximum
horizontal displacement was about 0.17 m and located in the
sulfide clay layer.
4.4 Discussion and conclusions from the trial
embankment
The dissipation of u-excess, that is the consolidation rate, was
much slower in the sulfide clay than in the rest of the clay deposit
(Figures 4 and5(b)). This had an influence on the design of the
PVDs for the Veda embankment, as discussed by Muller and
Larsson (2008). The difference in consolidation rate highly influ-
enced the increase in 9v (and consequently 9p). Hence, the rate
of increase ofsu in the sulfide clay was expected to be lower than
in the silty clay and glacial clay (c.f. Figure 5(a)). As shown in
7
6
5
4
3
2
1
0
Fillheight:m
140
120
100
80
60
40
20
0
u-excess:kPa
0 50 100 150 200 250 300 350 400
Time: days
Load sequenceu-excess at 1 mu-excess at 0 mu-excess at 1 mu-excess at 3 m
Figure 4.Trial embankment: load sequence and measured
u-excess with respect to time
Elevation:m
9v,220:kPa
9p,DSS:kPa
su:kPa
su=9p,DSS
+1 99 120 51.7 0.43
+1 99 180 64.2 0.36
0 63 80 25.7 0.32
0 63 120 37.0 0.31
0 63 180 41.4 0.23
1 113 130 33.3 0.26
1 113 180 37.7 0.21
3 169 160 61.2 0.38
3 169 220 73.2 0.33
Table 1.K0 consolidated undrained direct simple shear test
(CUK0DSS) results
36
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg
-
8/12/2019 geng166-0031
7/18
Figure 5(a), in tests performed after 50 days in the part of the
clay which was still overconsolidated (i.e. 9p 9p0), the su
increase was negligible. According to the pore pressure measure-
ments after 140 days, the clay between elevations +0.5 and
0.5 m should still be somewhat overconsolidated, in other words
a negligible su increase. However, the measurements of su at this
point in time indicated an increase of around 5 kPa (about 30%).
This raised questions regarding the following.
(a) The accuracy of the pore pressure measurements since there
was uncertainty related to the exact positions of the
instruments relative to the drains. Did the measuredu
represent the average value between adjacent drains?
(b) The accuracy of the in situ measurements ofsu:Were the
techniques suitable and were measurements adequately
corrected? Were these instruments suitably positioned for
measurements of average values?
It was concluded that measurements of representative values from
neither piezometers nor CPTu or FVS tests could be expected
from single measurements. Parallel measurements of the same
quantity (u or su), at the same depth but at different locations,
should provide the basis for evaluations of average values.
Furthermore, evaluations of su from measurements by CPTu and
FVS tests should be conducted in parallel. This would enable
estimations of su profiles during construction of the Veda
embankment from three independent sources, CPTu tests, FVS
tests and from pore pressure measurements together with estima-
0 50 100 150 200
v and
p: kPa
v0
p0
v 50 days
p 50 days
v 140 days
p 140 days
v 220 days
(c)
0 10 20 30 40 50 60
su
: kPa
4
3
2
1
0
1
2
3
4
Elevation:m
Initial meanCPT 50 daysFVS 50 daysCPT 140 daysFVS 140 days
(a)
0 40 80 120 160
u: kPa
u0u 50 daysu 140 daysu 220 days
(b)
Figure 5.Trial embankment, evaluated profiles: (a) su; (b) u;
(c) 9v and 9p
0 01 02 03 04 05
su p/
4
3
2
1
0
1
2
3
4
Elevation:m
FVS
CPTu
CUK0DSS
FCOED
Mean
Design
Figure 6.Trial embankment: evaluated su=9p ratio plotted against
depth
37
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg
-
8/12/2019 geng166-0031
8/18
tions of the induced stresses from the embankment and prognoses
of the su=9p ratios.
The choice of design values of su for the stability analyses of the
Veda embankment was uncertain as the scatter in the evaluated
su=9p ratios was quite large. There were no clear trends betweendifferent testing methods, thus no method was regarded as more
reliable than any other. It should be noted that Larsson et al.
(2007b) suggest CPTu as being more reliable than FVS tests. Due
to the uncertainties and considering the suggestions by Larsson
(1980),Jardine and Hight (1987) andLadd (1991), design values
were not strictly evaluated from the mean values (Figure 6).
Values of su=9p ratios corresponding to a cautious estimate were
employed as design values in the stability analyses performed
during the design phase.
5. Preconstruction design of the Vedaembankment
The Veda embankment was designed to be built in seven stages
according to Figure 8. The seventh stage represented unloading
of the surcharge. PVDs installed in a triangular pattern with a
spacing of 0.8 m and large moraine support berms were planned.
The drains were of the same brand as those used at the trial
embankment.Figure 9presents a layout plan and a typical section
of the embankment.
5.1 Preconstruction stability analysis
Construction works were intended to proceed continuously until
completion of stage 5 (Figure 8). Thereafter a 6-month-long
pause was scheduled before applying the surcharge (stage 6). The
most critical phases regarding stability in this project wereidentified when constructing stage 5 and when applying the
surcharge in stage 6.
The geometrical conditions concerning the berm length in com-
parison with the thickness of the clay deposit (Figure 9), and the
occurrence of the significantly weaker sulfide clay layer compared
to the rest of the soil, rendered non-circular slip surfaces with a
significant horizontal translation being least stable. As a result, a
major part of the critical slip surfaces was located beneath the
berms, that is corresponding to a direct simple shear state. In
areas close to the centre of the embankment, su corresponding to
triaxial compression would have been more appropriate due tothe stress rotation. However, su corresponding to triaxial com-
pression is generally significantly larger than su corresponding to
direct simple shearing (e.g.Ladd, 1991). Use of predicted vertical
effective stresses and values of su=9p corresponding to direct
simple shearing were considered as reasonable and conservative
simplifications in this case. In the stability analyses, su profiles
representing the initial conditions (su0) were used in the unim-
proved part of the soil and su profiles corresponding to predicted
design conditions were used where PVDs were installed. Analys-
ing stage 6, two different su profiles were used in the PVD
improved area as the effective stresses were predicted to be
higher below the embankment compared to the support berms,
given the differences in fill heights. The appearance and location
of the critical slip surface representing analyses for stages 5 and
6 is presented in Figure 10. The rock fill embankment and the
moraine support berms were defined as granular materials with
an angle of shearing resistance, j9 348, corresponding to the
0 100 200 300 400
Time: days
0
200
400
600
800
1000
1200
1400
Crests
ettlement:mm
Section 1
Section 2
Section 3
(a)
4
2
0
2
4
6
8
Elevation:m
Section 1, 11 days
Section 1, 56 daysSection 1, 73 days
Section 1, 197 days
Section 3, 11 days
Section 3, 56 days
Section 3, 73 days
Section 3, 197 days
0 40 80 120 160
Horizontal displacement: mm
(b)
Figure 7.Trial embankment: (a) measured settlements and
(b) measured horizontal displacements
38
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg
-
8/12/2019 geng166-0031
9/18
estimated constant voids values. The unit weights of the rock fill
and berm were 18 kN/m3, according to the design code
(Swedish Rail Administration, 2009).
Stability analyses were carried out as analytical calculations by
way of the MorgernsternPrice limit equilibrium method, in the
computer program Beast (Frimann-Clausen, 2003). USAs were
employed to account for the increase in su owing to consolidation
(Ladd, 1991). The vertical effective stress increases in the soil at
different stages of the construction process were evaluated byconsolidation analyses according to the Swedish design guide for
vertical drains (Swedish Road Administration, 1989), based on
the approach presented by Hansbo (1981). Experiences gained
from measurements at the trial embankment were also consid-
ered. The stability analysis for stage 5 was based on the
predictions ofsu after stage 3 (as stages 4 and 5 were intended to
take place continuously). Similarly, stage 6 was based on the
predictions of su after stage 5 and 6 months consolidation time.
The stresses induced by the loading and consolidation process
were expected to render 9v . 9p0 at both stage 5 and 6 through-
out the whole clay deposit. Based on these analyses and the
design su=9p values (Figure 6), predicted su profiles were
obtained (Figure 11). Calculations resulted in FS 1.54 and
FS 1.53 for stages 5 and 6 respectively, the size and shape of
the critical slip surface presented inFigure 10 was valid for both
stages. Corresponding calculations, without accounting for the su
increase, resulted in FS 1.33 and FS 1.02.
A simple parametric study was also performed to check the
sensitivity of the stability to variations in the su=9p values, tan j9
in the rock fill, in the rock fill and the initial su0: The
parametric study was based on normalised values, meaning the
use of the evaluated mean su=9p values (Figure 6),
tan j9 0:675 (348), 18 kN/m3 and the mean su0 profile
(Figure 2). The parameters were varied one at a time by
multipliers of 0.8, 0.9, 1.0, 1.1 and 1.2. Results from the
parametric study are presented in Figure 12, where calculated FS
with different multipliers divided by FSnorm calculated withvalues corresponding to the norm analysis (FS=FSnorm) are
shown. The calculated FS was most sensitive to variations in in
the rock fill (10:8) meaning that an increase (or decrease) in of
10% would render approximately 8% lower (or higher) calculated
FS. The second most influential parameter was the su=9p ratio
(10:6). The stability was least sensitive to variations in tanj9
(10:2) ands u0 (10:2).
6. Construction of the Veda embankment
6.1 Changes in staged construction sequence
Works began in December 2006 when a 0.5 m thick drainage
layer of sand was applied and PVDs installed. Day 0 corresponds
to the start of the actual filling works (stage 1) on 13 March
2007. The load sequence was altered somewhat from the design
(Figure 8). This influenced the stability in mainly two ways: the
higher surcharge (approximately 1 m) introduced a slightly larger
20
18
16
14
12
10
8
6
4
2
0
Fillheight:m
0 100 200 300 400 500 600 700Time: days
200
180
160
140
120
100
80
60
40
20
0
Me
asured
-excess:kPa
u
Stage 6
Stage 5
Stage 4
Stage 3
Stage 2
Stage 1
Load sequence designLoad sequence embankmentLoad sequence bermu-excess at 25u-excess at 10u-excess at 00u-excess at 05u-excess at 20
Figure 8.Veda embankment: loading sequence and measured
u-excess with respect to time
39
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg
-
8/12/2019 geng166-0031
10/18
load, and a rest period (approximately 2 months) initiated
between stages 4 and 5 led to slightly different stress conditions
compared with the intended design. Allowance of a rest period
between stages 4 and 5 enabled some consolidation to take place,
which was accounted for when the stability for stage 5 was
checked during construction.
6.2 Measurements
Regarding the stability of the embankment, the key quantities
were in the rock fill and the su=9p ratio in the clay. An
investigation of existing methods to measure the density or unit
weight in coarse rock fill material was undertaken. However,
owing to time constraints, no suitable method for use in the
present case was found. This quantity was therefore not measured
during the construction works.
Pore pressures were measured continuously throughout the con-
struction process at different levels in the soil (Figure 8).Figures
13(a)13(d) show the evaluated u profiles and 9v profiles under
the embankment and under the berm. The consolidation process
had produced normally consolidated conditions, that is 9v 9p,
for the time steps presented (except after 70 days under the
Trial embankment
100 200 300
A1 B1 C1
A2 B2C2
Piezometers
Concrete well
Settlement plate
Settlement tube
Inclinometers
0 50 100Length: m
(a)
25
20
15
10
5
0
5
10
15
Elevation:m
20 0 20 40 60 80 100 120 140
Length from centreline: m
S.P.: Settlement plateC.W.: Concrete wellINC.: InclinometersS.T.: Settlement tube
Drainage (sand)Original ground surface
Berm (moraine)S.T.S.P.C.W.S.P.
Stage 3Stage 2Stage 1
Embankment (rock fill)
Preloading (rock fill)INC.
C.W.S.P.
S.P.
Moraine
Sand Piezometers
PiezometersSulfide clayGlacial clay
Silty clay
(b)
Stage 6
Stage 5
Stage 4Stage 3
Stage 2
Stage 1
PVD S c/c 08 m
Figure 9.Veda embankment: (a) layout plan; (b) cross-section
40
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg
-
8/12/2019 geng166-0031
11/18
embankment). The predicted (design) u profiles and9v profiles at
stages 5 and 6 are also presented for comparison. For stage 5, it
should be noted that the staged construction sequence was altered
(as described above). The total stresses in the soil were different
when measurements were made (after stage 4) when compared
with the predictions (after stage 3). From the measured u and
calculated v (based on the initial conditions and the prevailing
fill height), profiles of 9v were evaluated. The effect of load
spreading with respect to the limited embankment width and the
limited clay depth was analysed by numerical modelling using a
simple linear-elastic material model, and considered in the
determination ofv.
The undrained shear strength su in the clay was measured by
CPTu and FVS tests in six wells (Figure 9) at the critical
phases during the staged construction process that is, after
130 days (stage 5) and 265 days (stage 6). In addition, FVS
tests were performed after 70 days in the six wells and CPTu
tests were performed after 590 days in the three wells under the
berm (A2, B2 and C2). Evaluated su from tests under the
25
20
15
10
5
0
5
10
15
Elevation:m
20 0 20 40 60 80 100 120 140
Length from centreline: m
Init ial profilesu
Critical slip surface for analysis ofstage 5 and of stage 6
Original ground surfaceand GW surface
34
18 kN/m
3
34
18 kN/m
3
Preloading(rock fill)
Moraine
Sand PVDs c/c 08 m
Sulfide clay
Glacial clay
Silty clay
su profiles withregard to
consolidation
Figure 10.Veda embankment: example of a slip surface analysis,
showing the critical slip surface for stage 5 and stage 6
0 20 40 60 80 100
su: kPa
4
3
2
1
0
1
2
3
4
Elevation:m Initial
Stage 5(embankment and berm)Stage 6 ( )embankmentStage 6 (berm)
Figure 11.Veda embankment: predicted undrained shear
strength su plotted against depth for different stages during
construction
08 09 10 11 12
Multiplier
12
11
10
09
08
FS/FSnorm
su p/
su0
tan
Figure 12.Veda embankment: Calculated FS/FSnorm plotted
against values of the multiplier for different parameters
41
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg
-
8/12/2019 geng166-0031
12/18
embankment and berm are presented in Figures 14(a)14(d).
Calculations of su (su 9v 3 su=9p) based on the evaluated 9vprofiles (Figures 13(b) and 13(d)) and the chosen design su=9pratios (Figure 6) are also presented. In Figures 14(a)14(d) the
predicted (design) su profiles at stages 5 and 6 are also
presented for comparison.
Settlements of the ground surface near the crest of the Veda
0 50 100 150 200 250
u: kPa
4
3
2
1
0
1
2
3
4
Elevation:m
Silty clay
Sulfideclay
Glacial clay
(a) (b)
u0u 70 days,u 130 days (stage 5),Design (stage 5)u 265 days (stage 6),Design (stage 6)
0 50 100 150 200 250
v
: kPa
300
p0
v0
v 70 days
Design (stage 5)
Design (stage 6)
v 130 days (stage 5)
v 265 days (stage 6)
0 50 100 150
u: kPa
4
3
2
1
0
1
2
3
4
Elevation:m
Silty clay
Sulfideclay
Glacial clay
(c) (d)
u0
u, 70 daysu 130 days (stage 5),Design (stage 5)u 265 days (stage 6),Design (stage 6)
0 50 100 150 200 250
v: kPa
p0
v0
v 70 days
Design (stage 5)
Design (stage 6)
v 130 days (stage 5)
v 265 days (stage 6)
u 590 days,
v 590 days
Figure 13.Veda embankment: evaluated measurements of u
and 9v at different stages during construction: (a) uunder
embankment; (b) 9v under embankment; (c) u under berm;
(d) 9v under berm
42
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg
-
8/12/2019 geng166-0031
13/18
embankment are presented in Figure 15(a), and approximately
1.0 1.3 m of settlement was measured after 560 days. InFigures
15(b) and 15(c), horizontal displacements at different elevations
for a selection of time steps are shown. Initially, horizontal
deformations towards the centre of the embankment developed
(negative values inFigures 15(b) and 15(c)). These displacements
occurred due to the construction sequence that is, building the
berms before the embankment in stages 1, 2 and 3 (Figure 8).
CPTu, 590 days
0 20 40 60 80 100
su
: kPa
4
3
2
1
0
1
2
3
4
Elevation:m
Silty clay
Sulfideclay
Glacial clay
(a) (b)
su0FVS, 265 days (stage 6)CPTu, 265 days (stage 6) v us
Design (stage 6)
0 20 40 60 80 100
0 20 60 1004
3
2
1
0
1
2
3
4
Elevation:m
Silty clay
Sulfideclay
Glacial clay
(c) (d)
0 20 40 60 80 100
su
: kPa
su0
FVS, 70 daysFVS, 130 days (stage 5)CPTu, 130 days (stage 5) v us
Design (stage 5)
su: kPa su: kPa
40 80
su0FVS, 265 days (stage 6)
CPTu, 265 days (stage 6) v us
Design (stage 6)
su0
FVS, 70 daysFVS, 130 days (stage 5)CPTu, 130 days (stage 5) v us
Design (stage 5)
p (stage 5) p (stage 6)
p (stage 5) p (stage 6)
Figure 14.Veda embankment: evaluated measurements of su at
different stages during construction; (a) under embankment at
stage 5; (b) under embankment at stage 6; (c) under berm at
stage 5; (d) under berm at stage 6 and after 590 days
43
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg
-
8/12/2019 geng166-0031
14/18
The horizontal displacements changed direction as construction
proceeded, resulting in maximum displacements of approximately
0.05 m and 0.08 m after 560 days. As for the trial embankment,the maximum displacements occurred in the sulfide clay layer.
6.3 Stability calculations
By incorporating the evaluated su profiles, updated calculations
of the total stability were made. These resulted in FS 1.70 and
FS 1.47 for stages 5 and 6 respectively and the stability was
judged as satisfactory at both stages.
7. DiscussionConventional limit equilibrium USAs accounting for strength
increase due to consolidation were performed to assess the
stability of the embankment during construction. As previousexperience and empirical knowledge of these characteristics for
sulfide clay were limited, the observational method was em-
ployed. The stability analyses were most sensitive to variations in
in the rock fill followed by the su=9p ratio in the sulfide clay
(Figure 12). Owing to time constraints, no suitable method of
measuring in the rock fill for use in the present case was
found, and so unfortunately this quantity was not verified during
construction. The value of 18 kN/m3 used was in accordance
with common Swedish practice and should be fairly representa-
tive of actual conditions. This is an issue that is seldom
addressed in embankment projects in Sweden; nevertheless, it
should be the subject of careful consideration in the design and
construction phases of embankment projects. One method for
doing this basically includes excavating a pit in the fill material,
weighing the material, lining the pit with an impermeable sheet
and filling the pit with a known volume of water or sand (e.g.
Bertram, 1973; Gordon and Miller, 1966; ICOLD, 1986; Zeller
and Zeindler, 1957). The increase in su was evaluated from in
situ measurements. An important experience gained from this
project is the value of independent measurement of a certainquantity. The increase in su was estimated from three different
sources and from measurements at six different locations.
Evaluations were made from direct measurements of su (FVS
and CPTu) and indirectly by way of measurements of u in the
clay. All three sources were associated with uncertainties, hence
assumptions and judgements had to be made in order to estimate
the average field conditions. Some of the uncertainties are
presented below.
(a) The positions of the instruments in the drain pattern, as the
vicinity to the drains affected the measured values; this is
also discussed bySuzuki and Yasuhara (2007) andLo et al.(2008). For instance, measurements from a piezometer tip
positioned close to a drain would have indicated a faster
dissipation rate than a tip installed at a larger distance from a
drain. Hence, evaluations ofu from measurements of a single
tip might not have been relevant for a large soil volume.
Furthermore, as a result of consolidation settlements in the
clay, the piezometer tips were to some extent pushed deeper
into the soil. Measurements of the depths of the tips were
made during the consolidation process.
(b) CPTu and FVS tests were considered as tests of relative
values ofsu of the clay. In order to derive design values from
these tests, empirical correction factors were used and the
validity of the design values was therefore dependent on the
validity of these factors. The applicability of correction
factors of clay loaded beyond9p0 is discussed byTavenaset
al. (1978) among others.
(c) Estimations of in the clay and fill material. As seen from
0 100 200 300 400 500 600Time: days
0
200
400
600
800
1000
1200
1400
Crestsettlement:mm
Settlement plate, section A
Settlement plate, section B
Settlement tube, section B
Settlement plate, section C
Settlement tube, section C
(a)
80 40 0 40 80Horizontal displacement: mm
4
2
0
2
6
4
Elevation:m
Section B,41 days
(b)
Section B,76 days
Section B,
254 daysSection B,441 daysSection B,562 days
Section B,140 days
80 40 0 40 80Horizontal displacement: mm
4
2
0
2
6
4
Elevation:m
Section C,41 days
(c)
Section ,76 days
C
Section ,
254 days
C
Section ,441 days
C
Section ,562 days
C
Section ,140 days
C
Figure 15.Veda embankment: (a) measured settlements;
(b) measured horizontal displacements at section B; (c) measured
horizontal displacements at section C
44
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg
-
8/12/2019 geng166-0031
15/18
the simple parametric study, analyses of the stability werestrongly influenced by the assumed value of in the rock fill.
(d) The model used for estimations of load spreading strongly
influenced the assessments of9v, hence the evaluations ofsu
from measurements ofu:
(e) Evaluations of9p0 from CRS tests. For instance, different test
methods or strain rates affect the evaluations (e.g. Olson,
1986).
(f) Estimations of the ratiosu=9p. As seen fromFigure 6and
Figure 16, the scatter in observed values of the ratio were
quite large in this case. It was therefore preferable to compare
measurements with valid empirical relationships.
As seen from Figures 14(a)14(d), evaluations of su from the
three sources agreed fairly well in the sulfide clay. In the silty
clay and in the glacial clay the scatter was larger, presumably
arising from the less homogeneous structure (varves of sand and
silt) of these clays. However, these clay layers had less impact on
the stability. By compiling results from different types of tests
and at different locations, uncertainties in the measurements were
reduced, and a sufficiently reliable assessment of the stability was
enabled.
In this project, su=9p ratios were obtained in the laboratory by
CUK0DSS and FCOED tests. In the field, the su=9p ratios were
obtained by CPTu and FVS tests, and evaluations of 9v bymeasurements of u: Figure 16shows measured su plotted against
evaluated9p in the sulfide clay obtained from the laboratory and
the in situ tests. According to the KolmogorovSmirnov normal-
ity test for the 5% level of significance, the su data for each valueof 9p did not differ significantly from a normal distribution. The
coefficient of variation with respect to the measured su was about
25%, the same order as the coefficient of variation related to the
initial conditions (Figure 2). The uncertainty related to the
positions of the instruments in the drain pattern was probably
reduced by simultaneously measuring at different locations.
In Figure 16, a best linear regression curve of su on 9p by the
method of least squares is shown. Furthermore, the confidence
limits of this regression are shown for the 10% level of signifi-
cance. Since the number of data is large, the confidence limits of
the fitted mean value of su are narrow. The confidence limits of
prediction of the whole sample are considerably larger as
illustrated. The design values of su were set by way of the ratio
su=9p 0:25 for the range 50 kPa , 9p , 250 kPa, which corre-
sponds to a straight line inFigure 16passing through the origin.
As illustrated, the deviation from the best linear regression curve
is very small. Furthermore, according to a t-test, the null
hypothesis that su=9p 0:25 and that the curve passes through
the origin cannot be rejected for the 10% level of significance.
Considering that the sulfide clay has a relatively high organic
content, it is worth noting that the ratio su=9p 0:25 agrees well
with the suggestions in Larsson (1980),Jardine and Hight (1987)
andLadd (1991), regarding organic clay.
In Figures 17(a) and 17(b), recordings of maximum horizontal
displacements (occurring in the sulfide clay layer) are plotted
against crest settlements. Tavenas et al. (1979) suggested a linear
80
70
60
50
40
30
20
10
0
su:kPa
0 50 100 150 200 250 300
p: kPa
Data from field measurements by FVS and CPTuData from laboratory CUK0DSSData from laboratory FCOEDLinear regression, best fitLinear regression, through origo / 025su p Confidence interval for the best-fit regression90% limit of prediction
Figure 16.Variation in undrained shear strength su plotted
against evaluated preconsolidation stress 9p
45
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg
-
8/12/2019 geng166-0031
16/18
increase in maximum horizontal displacement dh with the crest
settlement ds, with a mean of 0.16 based on 13 case history
embankments on soft clay sites. Ratios of dh=ds were suggested
as: 0.18 for overconsolidated clays; 1.0 during construction
for normally consolidated clays (undrained distortion); and0.16
during initial consolidation for normally consolidated clays. At
the Veda embankment, after development of the initial negative
horizontal displacements, dh=ds was approximately 0.060.14
and gradually increased to 0.140.19. At the trial embankment
dh=ds was approximately 0.150.18 during the whole construc-
tion process. These findings correspond rather well with the
suggestions by Tavenas et al. (1979) regarding deformation of anormally consolidated clay under the stresses imposed by
embankments.
7. ConclusionMeasurements during construction aimed to compensate for
uncertainties in the design assumptions. Uncertainties were
compensated for by evaluating a given property by way of a
number of different measurement methods and at different
locations.
The utilisation of shear strength increase, according to the ratio
su=9
p, was possible because of the information gained from thetrial embankment and laboratory tests in combination with the
use of the observational method. A controllable design was
possible without too much in-built safety.
The large amount of measurements during construction showed
that the ratio su=9p 0.25 is valid for a large stress interval.
Since the scatter is considerable, use of a single test or only a
few tests can result in misleading conclusions.
AcknowledgementsThe authors wish to express their thanks to the Sven Tyrens
Foundation for making this study possible and to Adalsbanan
Railroad Line for allowing the use of all measurements presented
in this paper. Mrs Tara Wood at the Chalmers University of
Technology is also acknowledged for proofreading and valuable
comments on the text.
REFERENCES
Bergado DT, Balasubramanian AS, Fannin RJ and Holtz RD
(2002) Prefabricated vertical drains (PVDs) in soft Bangkok
clay: a case study of the new Bangkok International Airport
project. Canadian Geotechnical Journal39(2): 304315.
Bertram GE(1973) Field tests for compacted rockfill. In
Embankment Dam Engineering, Casagrande Volume(Hirschfeld RC and Poulos SJ (eds)). Wiley, New York, NY,
pp. 120.
Bjerrum L (1973) Problems of soil mechanics and construction on
soft clays: SOA Report. Proceedings of the 8th International
Conference on Soil Mechanics and Foundation Engineering,
Moscow 3: 111159.
Bo MW, Arulrajah A and Nikraz H (2007) Preloading and
prefabricated vertical drains design for foreshore land
reclamation projects: a case study. Ground Improvement
11(2): 6776.
Chung SG, Kim GJ, Kim MS and Ryu CK (2007) Undrained shear
strength from field vane tests on Busan Clay. Marine
Georesources and Geotechnology 25(34): 167179.
Frimann-Clausen CJ (ed.)(2003)BEAST a Computer Program
for Limit Equilibrium Analysis by the Method of Slices,
revision 4. Roquefort les Pins.
Gordon BB and Miller RK (1966) Control of earth and rock-fill
0 250 500 750 1000 1250 1500
Crest settlement: mm
200
150
100
50
0
50
100
Maxim
um
horizontaldisplacement:mm
Section BSection Cd /d 016*h s
*Tavenas . (1979)et al
(a)
0 250 500 750 1000 1250 1500
Crest settlement: mm
200
150
100
50
0
50
100
Maximum
horizontaldisplacement:mm
Section 1Section 3d /d 016*h s
*Tavenas . (1979)et al
(b)
Figure 17.Veda embankment: measured maximum horizontal
displacement plotted against crest settlements: (a) Veda
embankment; (b) trial embankment.Tavenas et al.(1979) mean
ratio of dh/ds 0.16 is based on 13 case history embankments
on soft clay sites
46
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg
-
8/12/2019 geng166-0031
17/18
for Oroville Dam. Journal of the Soil Mechanics andFoundation Division, ASCE92(3): 123.
Hansbo S(1981) Consolidation of fine-grained soils by
prefabricated vertical drains. Proceedings of the 10th
International Conference on Soil Mechanics and Foundation
Engineering, Stockholm 3: 667682.
Hansbo S(1997) Practical aspects of vertical drain design.
Proceedings of the 14th International Conference on Soil
Mechanics and Foundation Engineering, Hamburg3: 1749
1752.
ICOLD (International Commission on Large Dams) (1986)Quality
Control for Fill Dams. ICOLD, Paris, Bulletin No. 56,
Appendix B.
Jamiolkowski M, Ladd CC, Germaine JT and Lancelotta R (1985)
New developments in field and laboratory testing of soils.
Proceedings of the 11th International Conference on Soil
Mechanics and Foundation Engineering, San Francisco, vol.
1, pp. 57153.
Jardine RJ and Hight DW (1987) The behaviour and analysis of
embankments on soft clays. In Embankments on Soft Clays
(Jardine RJ and Hight DW (eds)). Public Works Research
Centre, Athens, Greece, pp. 159244.
Karlsson R and Viberg L (1967)Ratio c/p9 in Relation to Liquid
Limit and Plasticity Index, with Special Reference to Swedish
Clays. Swedish Geotechnical Institute, Stockholm, Sweden
Preliminary Report, 23, pp. 4347.Kim C, Kim S and Lee J (2009) Estimating clay undrained shear
strength using CPTu results. Proceedings of the Institution
of Civil Engineers Geotechnical Engineering162(2):
119127.
Lacasse SM, Ladd CC and Barsvary AK (1977) Undrained
behaviour of embankments on New Liscard varved clay.
Canadian Geotechnical Journal14(3): 367388.
Ladd CC(1991) Stability evaluation during staged construction.
Journal of Geotechnical Engineering (22nd Terzaghi
Lecture), ASCE117(4): 537 615.
Ladd CC and Foott R(1974) New design procedure for stability
of soft clays. Journal of the Geotechnical EngineeringDivision, ASCE100(7): 763786.
Lambe TW and Whitman RV (eds) (1979)Soil Mechanics, SI
Version. Wiley, New York, NY, USA.
Larsson R(1980) Undrained shear strength in stability calculation
of embankments and foundations on soft clays. Canadian
Geotechnical Journal17(4): 591 602.
Larsson R and Ahnberg H (2005) On the evaluation of undrained
shear strength and preconsolidation pressure from common
field tests in clay. Canadian Geotechnical Journal42(4):
12211231.
Larsson R, Sallfors G, Bengtsson P-E et al. (2007a) Evaluation
of Shear Strength in Cohesive Soils with Special Reference
to Swedish Practice and Experience. Swedish
Geotechnical Institute, Linkoping, Sweden, Information No. 3
(in Swedish).
Larsson R, Westerberg B, Albing D, Knutsson S and Carlsson E
(2007b)Sulphide Soil Geotechnical Classification and
Undrained Shear Strength. Swedish Geotechnical Institute,Linkoping, Sweden, Report No. 69 (in Swedish).
Leroueil S, Demers D and Saihi F (2001) Considerations on
stability of embankments on clay. Soils and Foundations
41(5): 117127.
Lo SR, Mak J, Gnanendran CT, Zhang R and Manivannan G
(2008) Long-term performance of a wide embankment on
soft clay improved with prefabricated vertical drains.
Canadian Geotechnical Journal45(8): 10731091.
Lofroth H (2008) Undrained Shear Strength in Clay Slopes
Influence of Stress Conditions. A Model and Test Study. Swedish
Geotechnical Institute, Linkoping, Sweden, Report No. 71.
Long MM and ORiordan NJ (2001) Field behaviour of very soft
clays at the Athlone embankments. Geotechnique51(4):
293309.
Macsik J(1999) Soil Improvement Based on Environmental
Geotechnics, Environmental and Geotechnical Aspects of
Drainage of Redox-sensitive Soils and Stabilisation of Soils
with By-products. PhD thesis, LuleaUniversity of
Technology Division of Mining and Geotechnical
Engineering, Lulea, Sweden, doctoral thesis No. 1999:09.
Mayne PW(1980) Cam-Clay predictions of undrained strength.
Journal of the Geotechnical Engineering Division, ASCE
106(11): 1219 1242.
Mayne PW and Mitchell JK (1988) Profiling of overconsolidation
ratio in clays with field vane. Canadian Geotechnical Journal25(1): 150157.
Mesri G(1975) New design procedure for stability of soft clays.
Discussion.Journal of the Geotechnical Engineering
Division, ASCE101(4): 409412.
Mesri G and Godlewski PM (1977) Time- and stress-
compressibility interrelationships. Journal of the
Geotechnical Engineering Division, ASCE103(5): 417429.
Mesri G, Kwan Lo DO and Feng TW (1994) Settlement of
Embankments on Soft Clays. American Society of Civil
Engineers, Reston, VI, USA, Geotechnical Special
Publication 40, pp. 856.
Muller R and Larsson S (2008) Veda trial embankment comparison between measured and calculated deformations
and pore pressure development. Proceedings of the 2nd
International Workshop on Geotechnics of Soft Soils: Focus
on Ground Improvement(Karstunen M and Leoni M (eds)).
Taylor and Francis, London, UK, pp. 405 410.
Nicholson DP and Jardine RJ (1981) Performance of vertical
drains at Queenborough bypass. Geotechnique 31(1): 67 90.
Olson RE(1986)State of the Art: Consolidation Testing.
American Society for Testing and Materials, West
Conshohocken, PA, USA, STP 892, pp. 770.
Peck RB(1969) Advantages and limitations of the observational
method in applied soil mechanics (Ninth Rankine Lecture).
Geotechnique 19(2): 171187.
Roscoe KH and Burland JB (1968) On the generalised stress
strain behaviour of wet clay. Proceedings of the Symposium
on Engineering Plasticity. University Press, Cambridge, UK,
pp. 535 609.
47
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg
-
8/12/2019 geng166-0031
18/18
Schofield AN and Wroth CP (eds) (1968)Critical State SoilMechanics. McGraw Hill, London, UK.
Schwab EF(1976) Bearing Capacity, Strength and Deformation
Behaviour of Soft Organic Sulphide Soils. Royal Institute of
Technology Division of Soil and Rock Mechanics,
Stockholm, Sweden.
Shibuya S, Jung M, Chae J and Fujiwara T (2008) Evaluation of
short-term stability for sea-wall structure at Kobe airport.
Journal of Civil Engineering, KSCE12(3): 155163.
SIS (Swedish Standards Institute)(2007) Geotechnical
Investigation and Testing Sampling Methods and
Groundwater Measurements Part 1: Technical Principles
for Execution. SIS, Stockholm, Sweden, SS-EN ISO 22475
1:2006.
Suzuki K and Yasuhara K (2007) Increase in undrained shear
strength of clay with respect to rate of consolidation.Soils
and Foundations 47(2): 303318.
Swedish Rail Administration(2009) TK Geo. Swedish Rail
Administration, Borlange, Sweden, BVH 1585.001 (in
Swedish).
Swedish Road Administration(1989) Vertical Drains Design
Guide. Swedish Road Administration, Borlange, Sweden,
Publication 1987:30 (in Swedish).
Tanaka H(1994) Vane shear strength of a Japanese marine clay
and applicability of Bjerrums correction factor. Soils and
Foundations 34(3): 39 47.Tavenas F and Leroueil S (1980) The behaviour of embankments
on clay foundations. Canadian Geotechnical Journal17(2):
236260.
Tavenas F, Blanchet R, Garneu R and Leroueil S (1978) Thestability of staged-constructed embankments on soft clays.
Canadian Geotechnical Journal15(2): 283305.
Tavenas F, Mieussens C and Bourges F (1979) Lateral
displacements in clay foundations under embankments.
Canadian Geotechnical Journal16(3): 532550.
Watabe Y, Tsuchida T and Adachi K (2002) Undrained shear
strength of Pleistocene Clay in Osaka Bay. Journal of
Geotechnical and Geoenvironmental Engineering, ASCE
128(3): 216226.
Westerberg B(1999) Behaviour and Modelling of a Natural
Soft Clay Triaxial Testing, Constitutive Relations and
Finite Element Modelling. PhD thesis, LuleaUniversity
of Technology Division of Mining and Geotechnical
Engineering, Lulea, Sweden, doctoral thesis no.
1999:13.
Westerberg B and Andersson M(2009) Undrained shear strength
and compression properties of Swedish fine-grained sulphide
soils. Proceedings of the 17th International Conference on
Soil Mechanics and Geotechnical Engineering, Alexandria,
vol. 1, pp. 7275.
Wood DM (ed.) (1990) Soil Behaviour and Critical State Soil
Mechanics. Cambridge University Press, Cambridge, UK.
Wroth CP(1984) The interpretation of in situ soil tests (24th
Rankine Lecture). Geotechnique 34(4): 449489.
Zeller J and Zeindler H (1957) Test fills with coarse shellmaterials for Goschenenalp Dam. Proceedings of the 4th
International Conference on Soil Mechanics and Foundation
Engineering, London 2: 405409.
WHAT DO YOU THINK?
To discuss this paper, please email up to 500 words to the
editor at [email protected]. Your contribution will be
forwarded to the author(s) for a reply and, if considered
appropriate by the editorial panel, will be published as a
discussion in a future issue of the journal.
Proceedingsjournals rely entirely on contributions sent in
by civil engineering professionals, academics and students.
Papers should be 20005000 words long (briefing papersshould be 10002000 words long), with adequate illustra-
tions and references. You can submit your paper online via
www.icevirtuallibrary.com/content/journals, where you
will also find detailed author guidelines.
Geotechnical Engineering
Volume 166 Issue GE1
Stability for a high embankment founded
on sulfide clay
Muller, Larsson and Westerberg