Re-appraisal of the Dilatometer for In-situ Assessment of Geotechnical

Properties of Swedish Glacio-Marine Clays

Tara Wood
Department of civil and environmental engineering, Gothenburg, Sweden. E-mail: [email protected]
NCC Construction Sweden AB, Gothenburg, Sweden E-mail: [email protected]

Keywords: ground profile, small strain stiffness, undrained shear strength, in-situ stresses, degradation

ABSTRACT: This paper compares the results of in-situ field and high quality laboratory tests on Swedish soft
highly structured glacio-marine clays. The applicability of SDMT measurements for both soil profiling and
determination of soil properties are considered. Seismic dilatometer (SDMT) tests were found to define the
ground profile as well as piezocone penetration tests (CPTU). Furthermore, it was found that in-situ stress
state can be determined using existing correlations. However new correlations were required to define soft
clay anisotropy in undrained shearing and clay stiffness, pre and post yield, consistent with high quality
laboratory test results. Determination of small strain stiffness (G0) and degradation (G/Go) determined with
the SDMT probe are compared with high quality triaxial tests and show reasonable agreement.

1 INTRODUCTION specified by Marchetti (1980) were found to be

unreliable for Swedish normally to lightly over
Field testing with the dilatometer (DMT) was first consolidated soft clays. The information obtained
introduced to Sweden in the late 1980’s. Initially from the dilatometer using the Swedish correlations
DMT testing was used extensively within research for clays were soil type, density, earth pressure at
and industry for clays. In total 10 DMT blades are rest coefficient, K0, over consolidation ratio (OCR)
registered within Sweden. However, soon after the and undrained shear strength, su. The dilatometer
introduction of the dilatometer the popularity of modulus, MDMT, was found to lie in between the
cone penetration testing (CPTU) increased due to its reloading modulus and post yield (plastic) modulus
speed and ease of use in the field. This led to a thus was not deemed applicable. Other correlations
significant reduction in the use of the dilatometer in for soil parameter determination from dilatometer
Sweden. Currently only 2 of the 10 blades are still tests in soft clays have been specified among others
used. Among Swedish geotechnical practitioners by Chang (1991) and Lunne et al. (1989). Various
there is a general consensus that DMT testing is methods of interpretation of DMT tests for soft clays
unsuitable for clays and that other methods, such as are considered in this paper and compared.
shear vane and CPTU are more reliable for soil
profiling and determination of soil parameters. 2.2 Information required for FEA analysis
The initial purpose of the SDMT testing was to Numerical FEA models for soft clays are discussed
find the in-situ small strain stiffness (Go) and allow by Karstunen (2013) and Olsson (2013). The
comparisons with laboratory determined values. It Gothenburg clays are structured, anisotropic, non-
was found however that SDMT measurements could linear, rate dependent, viscous materials. Both an
also be used to give initial estimations of some adequate characterisation of the ground profile is
parameters used in finite element analyses (FEA). required in addition to input parameters for FEA
constitutive model. Preliminary assessments of both
based on SDMT tests are investigated in this paper.
2 BACKGROUND Soil properties of particular interest for advanced
FEA analysis are: earth pressure at rest (K0), over
2.1 DMT revisited consolidation ratio (OCR) and unit weight (γ) for
Swedish practice for interpretation of DMT tests in determination of initial stress state in the ground.
clay was presented by Larsson & Eskilson (1989). The failure criterion is normally related to drained
Swedish correlations were developed as those shear strength (c’, ’, M), however comparison to
undrained strengths is also useful when validating
drained model parameters. The deformation
parameters required depends on constitutive model
but potentially includes small strain stiffness (G0),
shear modulus degradation (G/G0), pre-yield
parameter (Cs, κ, E50’,E’ur, M0), post-yield parameter
(Cc, λ, Eoed, ML), and creep parameter (rs or μ*). No
attempt of creep from SDMT tests is made in this
paper. The other model parameters are discussed.

3 FIELD TESTING

Four new SDMT test sites were studied in

Gothenburg indicated in Fig. 1 located where deep Fig.1. Location of test sites around Gothenburg
excavations (>10 m) are planned for construction. At
site 4 two tests were carried out to verify the
repeatability of the SDMT. Other field tests carried 4 LABORATORY TESTING
out at these sites included; shear vanes, CPTU,
piezometric measurements and sampling with the Calibration of DMT correlations were done using
Swedish fixed piston sampler (STII). Field testing results from high quality fixed piston samples taken
and sampling was done with a Geotech 504 boring at site 1. Following extraction samples were
rig. For dilatometer tests care was taken to keep the immediately taken to the laboratory for testing.
expansion of the membrane at a constant rate as Index tests and CRS oedometer tests were carried
work by Smith (1989) showed that rate of expansion out within 1 hour of extraction. Four of the eleven
can affect the P0 and P1 pressure measurements, triaxial tests were started within 2 hours of
where P0 and P1 relate to the pressures to inflate the extraction and all but two triaxial samples were
membrane 0.05 mm and 1.1 mm respectively. For tested within 7 days. These later samples were tested
seismic testing the shear wave was produced using a within 1 month (45m CkoUE and 55m CkoUC).
10 kg hammer hitting a reinforced timber beam, Stepwise (IL) oedometer tests for 10, 18 and 27 m
similar to the arrangement described by Marchetti et were carried out after 4 days, whereas IL tests at 35
al. (2008). The boring rig was used to provide m, 45 m and 55 m were carried out after
reaction force on the shear beam and efficient approximately 1 month.
energy transfer from the hammer to the ground. The The Swedish STII piston sampler provides 3
procedure was repeated at least 3 times and shear samples of height 170 mm and diameter 50 mm. The
wave velocity, Vs, assessed using the method quality of the samples taken from the middle and
outlined by Marchetti et al. (2008). If the variability lower tubes from 10 m, 18 m, 27 m, 45 m and 55 m
coefficient of Vs exceeded 1% further tests were was assessed to be very good to excellent for
performed, although this was rarely necessary. The samples tested within 7 days based on Lunne et al.
seismic probe used to determine Vs consists of a (1997) and Landon et al. (2007). Samples taken
cylindrical probe placed above the DMT blade and from 35 m were disturbed during extraction.
contained within the pushing rods. The probe is Assessment of these samples ranged from good to
equipped with two mono-axial geophone receivers poor, as did the 1 month old triaxial samples.
compliant with the ASTM standards. The receivers
were spaced 0.5 m apart and the signal was
amplified and digitized in the probe. 5 GROUND PROFILE
The location of earlier measurements of in-situ Vs
in Gothenburg by Andreasson (1979) using The area of Gothenburg is characterised by the
downhole and crosshole methods is presented in Fig. crystalline bedrock sculpted by the effects of
1 as Site 5. This site was also used to determine glaciation. The deep gorges in the rock have been
DMT correlations specified by Larsson (1989). Also filled with sediments after recession of the glaciers
shown in Fig. 1 is the location of Site 6 where Vs and in the area of central Gothenburg these
was determined with multichannel analysis of sediments are principally clays. The varying
surface waves (MASW) using the method described sedimentation conditions in the glacio-marine
in Donohue et al. (2004) and is included for environment during clay deposition are significant
comparison with SDMT Vs values in Fig.5.
as they gave rise to different clay structures KD and ED (MPa) vs (m/s)
principally due to the different ionic strength of the 0 2 4 6 8 50 100 150 200
pore water, but also influenced by the speed of 0 0
sediment transport, water depth, landslides, ice
rafting activity and bacteria. 5 5
Zone 1A
The Gothenburg clays have a plasticity index
Ip≈40. They are predominantly illitic but plasticity is 10 Zone 1B 10
Zone 1C
also influenced by the silt fraction (≈30%), pore 15 15

Depth (m)

Depth (m)
water and other clay minerals. Different clay
Zone 2D
structures and sedimentary boundaries due to 20 20
varying conditions should be identifiable within the
Vs profile. To help identify if this is possible the 25 25
sedimentary geology classifications of Alte et al. Zone 3aD
30
(1989) and Bergsten (1991) have been amalgamated 30

in Table 1. The results from two SDMT tests taken 35 35

3 m apart at site 4 are shown in Fig. 3 together with Site 4.1 ED Site 4.1
these boundaries. The dilatometer parameters ED and Site 4.2 ED Site 4.2
KD are defined as ED=34.7(P1-P0) and Site 4.1 KD
KD=(P0-u0/σ’vo). The repeatability of SDMT tests is Site 4.2 KD

excellent, particularly for the seismic. Clearly small Fig. 3. Results of SDMT tests at Site 4
local variations exist but the clay appears
homogeneous. A sand layer was identified at 16 m layer at around 16 m depth. Based on Fig. 4 CPTU
in both SDMT profiles and confirmed by CPTU and DMT tests provide similar evaluations of the
tests. This layer could be significant for the planned ground profile.
cut and cover tunnel at this site as the layer lies just The usefulness of Vs as both a profiling tool and
below the planned excavation depth. Bergsten for understanding the effects of stress history can
(1991) noted fissures in samples below 23m due to also be seen in Figure 5 where Vs profiles from 5
erosion, this boundary appears be identified in the Vs sites are compared. The 5 sites were all subjected to
profile. Further erosion and increased sedimentation loading in the 1800’s due to land reclamation. At
events are apparent below this level in the Vs profile this time excavations for a dock at site 3 and canal at
and ED profile but not in the KD profile. site 4 were carried out. This dock was later refilled
The classification of the ground profile at site 4 in 1934. The Vs values in the zone 1 clays are
using different methods is presented in Figure 4. greater at site 3 and 4 due to recent stress history but
Assessment (a) from DMT uses Marchetti and are most prevalent at site 3 where additional loads
Crapps (1981), while (b) uses the chart given by were applied. SDMT measurements at site 2 and 3
Larsson (1989), (c) uses CPTU tests from Larsson were done during a cold period (< -10°C), which
(2007) and (d) is based on all measurements. clearly caused very high measurements of Vs in the
Assessment (a) erroneously identified the dry crust upper 5 m of the ground profile. Below the Zone 1
as silt otherwise it is very similar to (c) and (d). clays (12 m) profiles for all the sites are very similar
Method (b) correctly identified the stiffer dry crust confirming the homogeneity of these clays and the
but failed to identify the very soft clay within the ability of the Vs profile to identify changes at the
zone 1B clays. All methods identified a frictional expected geological boundaries.
Table 1. Geological profile of Gothenburg with Zone 1
(1A, 1B, 1C), Zone 2 (2D) and Zone 3 (3aD, 3bD) clays 0 made ground
Strata Age Base of strata 5 dry crust
(years) (±1m) 10 silt/sand
Made Ground ≈ 150 ≈ 2m
(Site 3 ≈ 7.5m) 15 very soft clay/mud
Post Glacial 1A clay 8000 5.5 m 20
soft clay
Post Glacial 1B clay 9000 8m 25
Post Glacial 1C clay 10000 12 m firm clay
30
Post Glacial 2D clay 10600 21 m stiff clay
35
Glacial 3aD clay 12000 42 m
Depth (m)
(a) (b) (c) (d)
Glacial 3bD clay 13000 57 to 100 m
Fig.4. Ground profile with different assessment methods
 Vs (m/s) 6.1.2 Determination of horizontal stresses
50 100 150 200 Horizontal stresses can be determined if the in-
5 situ earth pressure at rest coefficient, K0 is known.
Site 1
0 At site 1 in-situ K0 was assessed using the
Site 2 relationship derived by Schmidt (1966). Values of
‐5 OCR and ’ were taken from laboratory tests. The
Site 3
‐10 value of K0nc was estimated from the ratio of
Elevation (m)

Site 4
‐15 horizontal and vertical yield stress (σ’ch and σ’cv)
Site 6 (MASW) determined from undrained triaxial stress paths in
‐20
compression and extension. These values were
‐25 confirmed by K0 consolidation tests reported by
‐30 Olsson (2013). The assessment of K0 from the
‐35
dilatometer using Marchetti (1980) and Larsson
(1989) are presented in Figure 6. The K0 correlation
‐40 by Lunne et al. (1989) is almost identical to Larsson
‐45 (1989) thus is not plotted. Further verification is
‐50 provided by field measurement of K0 at site 5
presented by Smith (1989) which includes
‐55 measurement with Glotz cells and self-boring
Fig.5. Comparison of Vs measurements in Gothenburg pressuremeter (SBP). Larsson (1989) appears to be
slightly more consistent with field and laboratory
assessed values at these two sites.
6 ESTIMATION OF DESIGN PARAMETERS

Parameters of particular interest for FEA analysis of K0insitu

soft soils were discussed in Section 2.2. Methods of 0

0 0.5 1 1.5 2 2.5

determining preliminary estimations of all of these

parameters with the exception of creep parameters
are put forward based on SDMT tests in this section. 10

The laboratory high quality samples described in

section 4 were used to assess the validity of existing 20
Site 5 DMT ko SGI

empirical correlations and determine some new Site 5 Glotz cell ko

correlations.
Depth (m)

30 Site 5 SBP ko

Site 5 ko=konc*OCRIL^(1.2sinø')
6.1 Determination of in-situ stresses 40
Site 1 ko=konc*OCRIL^(1.2sinø')

6.1.1 Determination of vertical stress

Site 1 SDMT ko SGI method
The unit weights of clays tested have been 50

evaluated from DMT tests using Marchetti and Site 1 SDMT ko Marchetti

Crapps (1981) and Larsson (1989) and were 60

compared with measured values from samples
extracted at site 1. Measurements of Vs were also Fig.6. Comparison of in-situ K0 estimations
used to assess soil density using the mass density
correlation presented by Mayne et al. (1999). 6.1.3 Determination of over consolidation ratio
Assessment using Marchetti and Crapps (1981)
slightly over estimated densities (<5%), while Four existing correlations of vertical OCR based
Larsson (1989) gave overestimations of 10-15%. on the dilatometer horizontal stress index KD are
The Mayne et al. (1999) correlation gave very presented in Figure 7 together with laboratory
accurate soil densities, within 1.3% of measured determined values. The rate dependency of the clays
values thus appears to provide the best basis for is evident particularly with depth seen by the
determination of vertical total stress. The impact of enhanced yield stress of 1 day CRS oedometer tests
small inaccuracies of unit weight for effective stress when compared to increment load (IL) tests which
determination will be small considering uncertainties took 10 days. The best correlation appears to be
in pore water pressures. obtained using Chang (1991) whose correlation was
 strength, suav is significantly higher (75-85%). The
σ'c/σ'vo (kPa)
difference between suav from site 1 and τmob
0.50 1.50 2.50
0
determined by Larsson et al. (1985) will be due to
Site 1 CRS
the impact of softening, and uncertainties like
5
drainage, rate effects and geometry. Post peak
10
Site 1 CkoUC softening in undrained triaxial test results presented
here was up to 80% thus is of a similar order to the
Depth (m)

15
20 Site 1 IL differences between site 1 test results and Larsson
(1985). The discrepancy between site 1 suav and suav
25
Site 1 SDMT Larsson (1989)1 reported by Larsson et al. (1985) and site 1 suDSS is
30 most likely related to issues of storage effects and
35 Site 1 SDMT Marchetti (1980) sample disturbance.
40 The calculation of undrained strength from DMT
Site 1 SDMT Chang (1991) tests has been determined in three different ways.
45
The most common method uses a critical state soil
50 Site 1 Lunne (1989) mechanics type model such as that proposed by
55 Ladd (1977) where (su/σ’vo) is defined in Eq.1.

Fig. 7. Comparison of estimated overconsolidation ratio (su/σ’vo)OC=a*OCR^m (1)

σ’c/σ’vo using DMT with laboratory tests
The DMT correlation uses the horizontal stress
based on soft marine clays thus is recommended for index, KD and estimated σ’vo to define undrained
soft glacio-marine clays and can be used in the shear strength; Marchetti (1980), Lunne et al.
assessment of soil stiffness discussed in 6.3.1. (1989), Chang (1991). The DMT undrained
strengths obtained are essentially corrected field
shear vane strengths and agree well with correlated
6.2 Determination of strength properties shear vanes from site 1 in Fig. 8. There is a
significant variation in the ratio (su/σ’vo) for different
6.2.1 Determination of undrained strength soils as shown by Lutengger (1991) which explains
The Gothenburg clays are highly structured, and why so many “local” correlations exist to determine
exhibit significant anisotropy. This is illustrated in a and m.
Fig. 8 by the differences in the undrained shear Other authors argue it is more appropriate to
strengths of triaxial tests in compression and estimate undrained strength based on a simple
extension. It should also be noted that the results of bearing capacity approach using the inflation
undrained direct simple shear tests lie very close to pressure P1, and estimated σh0; Larsson (1989),
values from triaxial extension tests. Based on work Roque et al. (1988), refer to Eq.2.
covering soft clay anisotropy by Hight (1998),
Lunne et al.(1997) and Karstunen (2013) one would su=(P1-σh0)/Nb (2)
expect direct simple shear strengths to be greater,
more similar to the average value from triaxial For Swedish clays Larsson (1989) suggests a value
compression and extensions tests. The results of for Nb of 10.3. The correlation is based on correlated
correlated in-situ shear vane and fall cone tests are shear vane tests. At site 1 estimations of su
also plotted in Fig 8. The correlation applied is determined with Eq.2 are similar to τu from shear
presented by Larsson et al. (1985) and reduces the vane tests and DMT correlations using KD.
fall cone and shear vane strengths by 15 to 20%. Alternative methods of assessment of su from
This correlation considers mobilized strengths, τmob, G0SDMT, and empirical G0 estimations by Andreasson
calculated from landslides, pile tests, foundations, (1979) and Bråten et al. (2010) have been
and embankments in addition to average laboratory investigated. The values su in the Scandinavian G0
derived shear strengths, suav, (suCKoUC, suDSS, suCKoUE). correlations again relate to τu and give values of su
It is also said to incorporate effects of loading rate that lie close to both τu assessed at site 1 and DMT
and varying plasticity. In Larsson et al. (1985) suDSS estimates using -KD and P1. None of the correlations
of post glacial clays are shown to be similar to suav, discussed so far provide good estimates of undrained
and τmob assessed from failures in the ground and are characteristic shear strengths to help validate FEA
in agreement with correlated fall cone and shear analysis. To depths of around 35 m reasonable
vane tests, τu. Based on the results presented in estimates of strength in extension suCkoUE can be
Fig. 8 the average characteristic undrained
made from Lunne et al. (1989). Below 35m Using Eq. 3 and the OCR using Chang (1991), the
Marchetti (1980) gives a better indication of suCkoUE. critical state friction angle can be assessed if a ratio
of su/σ’vo is assumed. Clearly the ratio su/σ’vo
6.2.2 Determination of drained strength depends on a number of factors such as, direction of
The determination of drained strength from DMT shearing, structure and not just OCR thus even use
tests is generally limited to frictional soils. An of specific ratio’s for compression and extension
attempt has been made to assess the critical state tests were not found to yield reliable estimations of
friction angle ’cs using the critical state soil critical state friction angles based on Wroth (1984)
mechanics concept presented by Wroth (1984) given equation. Use of the average value of su/σ’vo =0.37
in Eq. 3 was investigated where Λ =1-(Cs/Cc). An gave a friction angle of 36° which is similar to
assumption is made that the cohesive intercept c’ is laboratory assessed critical state friction angle in
zero during critical state shearing. extension. However laboratory values of φ’cs in
compression varied between 32° in the post glacial
su/σ’vo=½sin’cs*OCRΛ (3) clays to 30.5° in the glacial clays. This method is
therefore unreliable and not recommended for the
For the samples tested at site 1 the ratio su/σ’vo for determination of ’cs. Reliance should instead be put
laboratory tests in extension, simple shear and upon good quality laboratory tests for determination
compression ranged between 0.25 to 0.35, 0.3 to 0.4 of this parameter.
and 0.45-0.6 respectively with an average value 6.3 Determination of stiffness properties
around 0.37. For the highly structured Gothenburg
clays the ratio (Cs/Cc) is around 0.015-0.025 thus Λ 6.3.1 Medium to large strain stiffness properties
is very close to 1.
The constrained modulus, MDMT, is the confined
drained vertical modulus at σ’vo. This is determined
su lab/field (kPa) from the dilatometer modulus ED, which is
0
0 50 100 150 200 250 effectively a disturbed modulus at strains slightly
greater than many engineering situations calculated
5 using elasticity theory which is then multiplied by an
10 empirical factor RM, (for clays is based on KD).
Correlations for RM have been suggested by
15
Marchetti (1980) and Chang (1991) however as seen
20 in Fig. 9 the assessments of MDMT do not represent
either of the moduli typically used to define pre-
25
yield “elastic” stiffness (M0) or post yield “plastic”
30 stiffness ML (where ML=1/Mv). It is not either the
35
initial “disturbed” modulus found from initial
loading in oedometer tests.
Depth (m)

40

Soil modului (kPa)

45
0 20000 40000 60000
0
50 Elastic modulus, Mo from site 6
5
55
10 Initial 'disturbed' Mo oedometer
60 modulus site 1
fall cone 15
CPTU Qt Larsson (2007) Plastic oedometer modulus ML
shear vane 20
Depth m

site 1
suDSS
25
suCkoUC MDMT site 1 Chang (1991)
suCkoUE 30
DMT Marchetti (1980)
DMT Larsson (1989a) 35 MDMT Site 1 Marchetti (1980)
DMT Chang (1991)
DMT Lunne (1989) 40
Su from Go, Bråten et al (1991) Estimated elastic modulus Mo
45
Su from Go, Andreasson (1979) from 5x MDMT Marchetti
Su CkoUC from vs, Long et al. (2013) 50 (1980)
Su CKoUC from average lab. Go/SuCkoUC =373 Estimated plastic modulus from
55 0.125xMDMT Marchetti (1980)
Fig.8. Comparison of estimated undrained shear strength 60
using field and laboratory tests
Fig. 9. Comparison of estimated oedometer moduli.
Instead a reasonable assessment of both “elastic” Go (kPa)

and “plastic” modulus could be found for all the

0 50000 100000
levels studied by applying a factor to the original ‐5
MDMT modulus. This factor was found to be 5 for the
pre yield elastic modulus in the range where OCR Go Bender Element
varies between 1 and 2. The factor for the plastic ‐15
(σ'vo)

modulus ML was found to be 0.125. Therefore, Go Andreasson

similar direct correlations should exist for stiffness (1979)

Elevation (m)
parameters that are more appropriate for FEA such ‐25
Go Bråten et al.
as λ, κ, E’oed, E’ur. However, correlation of (2010)
correlated values is generally inappropriate. The
Go Hardin & black
deviatoric stiffness E’50 from compression triaxial ‐35 (1968)
tests at site 1 were found at axial strains of 0.35 to
0.6% which is less than the strains applied during Go SDMT

inflation of the membrane and determination of ‐45

MDMT. Using elastic theory E’DMT can be found from Go from MDMT
Marchetti et al
MDMT (again if υ’, is known) using eq. 4: (2008)
‐55
E’DMT =F.MDMT((1+υ’)(1-2υ’))/(1-υ’) (4)
The value of υ’ is not a constant and varies during Fig. 10. Comparison of estimated small strain stiffness G0
shearing. Values of υ’ at engineering strains are in
the range of 0.1 to 0.3 but this gives gross under The use of SDMT tests to define stiffness
estimation of compression E’50 (< 35%). If the degradation is discussed by Mayne et al (1999). A
poisons ratio for the clays at failure (υ’=0.42) is used relationship is presented for the normalized shear
and the factor F=5 (as used for the M0) Eq. 4 gives modulus G/G0 based on the degree of mobilized
estimations of E’50 similar to those from CKoUE shear strength. If results of e.g. field shear vane tests
tests for the clay studied. One could therefore make are available. An alternate method can be used based
an approximate estimation of compression E’50 by on Hardin and Drnevich (1972) where the reference
first determining the value for extension E’50 using strain, γr=τmax/G0, can be assessed from Fig. 8. The
Eq. 4 and υ’= 0.42 and then adjusting for anisotropy. modulus degradation is defined using Eq. 6.
G/G0=1/(1+γ/γr) (6)
6.3.2 Small strain stiffness properties and
This hyperbolic function is plotted Fig. 11 together
degradation
with degradation curves determined in the laboratory
The determination of in-situ small strain stiffness and SDMT points. A reasonable fit is achieved,
parameters with downhole seismic measurements there is some under and overestimation of stiffness
was first reported in Sweden by Andreasson (1979). at small and medium strains respectively but these
SDMT field measurements were later reported by will tend to counterbalance each other at typical
Marchetti et al. (2008). The determination of small engineering strains of 10-4 to 10-3.
strain stiffness G0 is found using elastic wave theory
using the relation in Eq. 5: 1.2

2
G0=ρVs (5)
1
Where ρ is the mass density and can be determined
from the correlation suggested by Mayne (1999). Hardin Drnevich 0.8
The results of SDMT measurements of G0 are 10m CkoUC

presented in Fig. 10 together with laboratory values 55m CkoUE

G/GoBE (kPa)

0.6
using bender elements and empirical correlations. 35m CkoUC

The similarity of laboratory and field values of G0 is 27m CkoUE

0.4
a clear indication of the quality of the samples SDMT 10m
SDMT 27m
tested. Empirical correlations based on undrained SDMT 35m
0.2
strength and plasticity (PI or LL) gave reasonable SDMT 55m
agreement; Bråten et al. (2010), Andreasson (1979), 0

whereas correlations presented by Hardin & Black 1.00E‐06 1.00E‐05 1.00E‐04 1.00E‐03 1.00E‐02 1.00E‐01

(1968) gave poor agreement. Marchetti et al (2008) γ

reported correlations of G0/MDMT and KD however

this correlation is only in agreement up to 5 m depth. Fig.11. Shear modulus degradation curves from
laboratory and in-situ measurements
7 CONCLUSIONS Karstunen, M. (2013) “Modelling rate-dependent
behaviour of structured clays”, Proc. Installation
The work presented in this paper shows that SDMT Effects in Geotechnical Engineering, 43-50, Taylor &
Francis Group, London.
tests are useful for both soil profiling and Ladd, C. C. (1977). Stress-deformation and strength
determination of initial conditions required for characteristics, State of the Art Report. Proc. of 9th
advanced FEA analyses. The use of Vs as a profiling ISFMFE, 1977, 4, 421-494.
tool and for determination and cross checking of soil Landon, M. DeGroot, D., and Sheahan, T. (2007). “Non-
models and parameters in FEA analysis is very destructive sample quality assessment of shear wave
velocity”, Journal of geotechnical and geo-
useful. environmental engineering, 133 (4), 424-432.
When assessing soil parameters using empirical Larsson, R. (1989).”Dilatometerförsök- En in-situ metod
correlations, it is important to consider what för bestämning av lagerföljd och egenskaper i jord”,
correlations are based upon and if this is relevant. A SGI Information 10, Linköping.
reliable method to find drained strength (M) from Larsson, R. Bergdahl, U. Uriksson, L. (1985).
SDMT tests was not found however a correlation of “Evaluation of shear strength in cohesive soils with
special reference to Swedish practice and experience”,
peak undrained strength from standard K0 SGI Information 3, Linköping.
consolidated triaxial tests appears promising. Larsson, R. (2007). ”CPT test: Equipment, testing,
The greatest contribution of the SDMT test in the evaluation. An in-situ method for determination of
characterization of soft clays for FEA analysis is its stratigraphy and properties in soil profiles”, SGI,
ability to assess the stiffness degradation using small Information 15, Linköping.
Long, M, Quigley, P., O’Connor, P. (2013). “Undrained
and intermediate strain properties in conjunction shear strength and stiffness of Irish glacial till from
with the Hardin-Drnevich backbone curve. Such shear wave velocity”, Ground Engineering, 46.11
curves are difficult and expensive to achieve in the Lunne, T., Berre, T. & Strandvik, S. (1997). ”Sample
laboratory due to problems associated with sampling disturbance effects in soft low plastic Norwegian
disturbance and storage. This is a very promising clay”, Proc. Int. Symp. on Recent Developments in
direction for advanced field testing in soft soils. Soil and Pavement Mechanics, pp. 81-102, Rotterdam
: Almeida, Missouri, 259-264.
Lunne, T., S. Lacasse, and N. S. Rad. (1989). "State of
the art report on in situ testing of soils." Proc. XII
8 REFERENCES ICSMFE, Rio de Janeiro. Vol. 4.
Lutenegger, A. J. (2006). Cavity expansion model to
Alte, B., Olsson, T., Sällfors, G. Bergsten, H. (1989) estimate undrained shear strength in soft clay from
“Djupdykning I Göteborgsleran”, Chalmers dilatometer. In Proceedings from the Second
University of Technology. International Flat Dilatometer Conference , 319-326.
Andréasson, B. (1979). “Defromation characteristics of Mayne, Paul W., James A. Schneider, and G. K. Martin.
soft high –plastic clays under dynamic loading”. PhD (1999). "Small-and large-strain soil properties from
Thesis, Chalmers University of Technology. seismic flat dilatometer tests." Proc. 2nd Int. Symp. on
Bergsten, H. (1991). “Late Weichselian-Holocene Pre-Failure Deformation Characteristics of
stratigraphy and environment conditions in the Geomaterials, Torino. Vol. 1.
Göteborg area, south-western Sweden“, PhD Thesis, Marchetti, S. (1980). "In situ tests by flat dilatometer."
Chalmers University of Technology. Journal of the Geotechnical Engineering Division
Bråten, C. Døssland, Å. L. Gjestvang, M., Kaynia, A.M. 106.3, 299-321.
Loe, M. M., Løset, Ø. (2010). ”Dimensionering for Marchetti, S., and Crapps, D. K. (1981). “Flat dilatometer
jordskjelv”. Rådgivande Ingenjørers Forening. manual”. Int Report of GPE Inc.
Chang, M. F. (1991). "Interpretation of overconsolidation Marchetti, S., Monaco, P., Totani, G., & Marchetti, D.
ratio from in situ tests in recent clay deposits in (2008). “In situ tests by seismic dilatometer (SDMT)”.
Singapore and Malaysia." Canadian Geotechnical From research to practice in geotechnical
Journal 28.2, 210-225. engineering, 180, 292-311.
Donohue, S., Long, M., Gavin, K., & O'Connor, P. Olsson, M. (2013). “On rate dependency of Gothenburg
(2004). The use of multichannel analysis of surface Clay”,PhD, Chalmers University of Technology.
waves in determining Gmax for soft clay. In Roque, R., Janbu, N., & Senneset, K. (1988). Basic
Proceedings 2nd. Int. Conf on Geotechnical Site interpretation procedures of flat dilatometer tests. In
Characterisation, ISC (Vol. 2, pp. 459-466). Proceedings of 1st International Symposium on
Hardin, B. O., and Black, W.L. (1968). "Vibration Penetration Testing, Orlando, Florida, Vol. 1.
Modulus of Normally Consolidated Clay," J. J. Soil Schmidt, B. (1966) "Earth pressures at rest related to
Mech. Fdns Div. Am. Sot. Civ. Engrs, (11), 1531- stress history." Canadian Geotechnical Journal 3.4,
1537. 239-242.
Hardin, B. O., & Drnevich, V. P. (1972). Shear modulus Smith, M. (1989). “Dilatometer tests in soft Swedish
and damping in soils. Journal of the Soil Mechanics clays”, MSc Thesis. Chalmers University of
and Foundations Division, 98(7), 667-692.H Technology.
Hight, D. W. (1998). “Soil characterisation: the Wroth, C. P. (1984). "The interpretation of in situ soil
importance of structure, anisotropy and natural tests." Géotechnique 34(4), 449-489.
variability”. 38th Rankine Lecture. Géotechnique.