INTERNATIONAL SOCIETY FOR

SOIL MECHANICS AND

GEOTECHNICAL ENGINEERING

This paper was downloaded from the Online Library of

the International Society for Soil Mechanics and
Geotechnical Engineering (ISSMGE). The library is
available here:

https://www.issmge.org/publications/online-library

This is an open-access database that archives thousands

of papers published under the Auspices of the ISSMGE and
maintained by the Innovation and Development
Committee of ISSMGE.
Proceedings of the XVII ECSMGE-2019
Geotechnical Engineering foundation of the future
ISBN 978-9935-9436-1-3
© The authors and IGS: All rights reserved, 2019
doi: 10.32075/17ECSMGE-2019-0014

Collection of high-quality samples in liquefiable soils

using new sampling techniques
Collecte d'échantillons de haute qualité dans des sols liquéfiables à
l'aide de nouvelles techniques d'échantillonnage
A. Viana da Fonseca, C. Ferreira, F. Molina-Gómez & C. Ramos
CONSTRUCT-GEO, Universidade do Porto, Faculdade de Engenharia, Porto, Portugal

ABSTRACT: Due to the difficulty in collecting undisturbed samples of cohesionless loose soils, the mechanical
characterisation of liquefiable soils is routinely performed in advanced laboratory tests on reconstituted
specimens. Following recent advancements in sampling, namely with the development of the Dames & Moore
and the Gel-Push samplers, high-quality samples of loose sands to silty sands from Benavente (Portugal) have
been collected. Documents about the 1909 earthquake demonstrate the occurrence of liquefaction phenomena in
these soils, which are currently being studied as part of two research projects ongoing in the CONSTRUCT-GEO
centre of FEUP. This paper introduces the sampling processes in the field with both techniques, as well as the
laboratory preparation and setup of the samples for element testing. Preliminary assessment of the sampling
quality of the collected samples has been made through the comparison of field and laboratory measurements of
shear wave velocity, obtained by SCPTu and bender-element bench tests, respectively.

RÉSUMÉ: En raison de la difficulté à collecter des échantillons de sols meubles sans cohésion et non perturbés,
la caractérisation mécanique des sols liquéfiables est systématiquement effectuée lors d'essais de laboratoire
avancés sur des échantillons reconstitués. Suite aux progrès récents de l'échantillonnage, notamment avec le
développement des échantillonneurs Dames & Moore et Gel-Push, des échantillons de haute qualité de sables
meubles à sables limoneux de Benavente (Portugal) ont été récupérés. Des documents sur le tremblement de terre
de 1909 démontrent l’apparition de phénomènes de liquéfaction dans ces sols, qui sont actuellement à l’étude
dans le cadre de deux projets de recherche en cours au centre CONSTRUCT-GEO de FEUP. Ce document
présente les processus d'échantillonnage sur le terrain avec les deux techniques, ainsi que la préparation en
laboratoire et la configuration des échantillons pour le test des éléments. L'évaluation pré-liminaire de la qualité
d'échantillonnage des échantillons prélevés a été effectuée en comparant les mesures de la vitesse des ondes de
cisaillement sur le terrain et en laboratoire, par des essais SCPTu et à éléments bender en banc, de manière
spectrale.

Keywords: Dames and Moore sampler, Gel-push sampler, high-quality sampling, liquefiable sands, shear wave
velocity.

1 INTRODUCTION water pressure. Furthermore, loose to medium

sands under undrained conditions are the most
One of the most complex phenomena occurring susceptible for liquefaction triggering (Ramos et
in granular materials is liquefaction. Such al., 2015). The assessment of liquefaction
phenomenon induces the loss of strength and susceptibility can be carried out by means of field
stiffness due to the generation of excess pore

IGS 1 ECSMGE-2019 - Proceedings

 A.2 - Investigation by in situ tests

and laboratory procedures. Nevertheless, there is valuable information regarding soil behaviour
some variability of the results obtained by type and the location in depth of the sandy layers.
different techniques, due to the soil spatial In addition, shear wave velocity (VS) values
uncertainty and the disturbance of samples used obtained from in situ and bender-element (BE)
for element testing in the laboratory. bench tests in the laboratory were used to
Obtaining undisturbed samples of loose to evaluate the quality of the samples. Finally, a
medium-dense sands is a challenge, since these comparative analysis is presented regarding the
soils are often susceptible to experiment quality of the samples retrieved by both samplers.
significant volume changes and collapse its
structure during sampling. Besides, any change in 2 MATERIALS AND METHODS
relative density or fabric will affect the
liquefaction resistance of the soil, as the sample 2.1 Site description
becomes denser and thus less representative of
the in situ conditions. Disturbance of soil samples An experimental campaign to retrieve high-
is mainly attributed to the excessive friction quality samples of liquefiable soils was carried
generated during penetration of the sampler into out in a pilot site located in the municipality of
the ground (Chen et al., 2014). In order to solve Benavente, in Lower Tagus River Valley (LTV)
this issue, the Dames & Moore (D&M) and the region, south of Portugal. According to Eurocode
Gel-push (GP) samplers have been developed. 8, the South of Portugal is the region with the
Such samplers are known as advanced sampling highest seismic risk of the country due to the
devices, due to the incorporation of novel proximity of the boundary of African and
materials to reduce the friction between the Eurasian plates. Besides, such area is affected by
sample and the walls of the liner in which the soil the occurrence of large magnitude (>8) distant
is collected. The D&M sampler is a hydraulic earthquakes and of medium magnitude (>6) near
fixed-piston device, similar to the Osterberg, with earthquakes (Azevedo et al., 2010).
a shorter length to reduce wall friction and brass The selection of the location of the pilot site
tubes to generate lower friction in the interface. was based on historical records of liquefaction
On the other hand, the GP sampler uses a viscous during the earthquake of 23rd April 1909 in
polymer gel to substantially reduce wall friction Benavente (Teves-Costa and Batlló, 2011; Cabral
during field extraction and sample extrusion in et al., 2013). In addition, from extensive
the laboratory. Both devices allow retrieving geological and geotechnical data reported by
relatively “undisturbed” samples of medium- Saldanha et al. (2018), the presence of thick
dense sands, silty sands, silts, compressible silty profiles of recent alluvial sandy deposits was
sands, silty clays and clays. These samplers have identified in this region. The selected site is
already been implemented with success for located near the shore of the Tagus River and its
laboratory characterization of liquefiable soils in position coordinates are 39°1'0.77"N,
Adapazari, Turkey (Bray and Sancio, 2006), 8°50'25.89"W (Figure 1). Prior to sampling, the
Dhaka, Bangladesh (Ishihara et al., 2016) and site was investigated with SCPTu in order to
Christchurch, New Zealand (Markham et al., identify consistent soil layers suitable for
2016; Bray et al., 2017). sampling. Vs values were measured at each 1 m
This paper deals with the description of the depth. Figure 2 presents the SCPTu test results,
collection of high-quality samples of liquefiable which were interpreted according to Robertson
soils from a pilot site in Benavente (south of (2009). Based on site characterisation, sandy
Portugal) by means of D&M and GP sampling. layers were identified at 5 to 12 m depth. This
The selection of the depths to retrieve the samples work focuses on the study of the soil samples
was based on SCPTu test results, as these provide collected at such depths.

ECSMGE-2019 – Proceedings 2 IGS

 Collection of high-quality samples in liquefiable soils using new sampling techniques

insertion, which allows recovering samples with

45-50cm length, preventing sample from falling
inside the borehole (Viana da Fonseca and
Pineda, 2017). The key component to ensure the
vacuum is a neoprene skirt seal, located in the
transition of the pressure cylinder and the liner.
This seal prevents the entrance of disturbed soil
into the liner from the bottom of the borehole and
also acts as a check valve for the driving pressure.
Figure 3 shows a scheme of the components of
Figure 1. Pilot site location (adapt Google Earth). the D&M sampler.

Figure 3. Schematic of D&M sampler components

The operation principle of this sampler is

based on the fixed hydraulic piston sampler
proposed by Osterberg (1973), which consists of
three different stages. The first stage refers to the
drilling of an external hole to stabilize the
excavation and prevent the groundwater seepage
through the borehole. In the second stage, the
sampler is inserted in the borehole to a specific
depth. A thin-walled tube or liner is then pushed
into the ground at constant pressure until its
Figure 2. SCPTu results maximum length is reached, by means of the
injection of a fluid (usually water) at 1400 kPa
2.2 Dames and Moore sampler minimum pressure, in order to achieve a constant
Dames and Moore (D&M) sampler is an penetration rate. The soil sample is collected
advanced sampler used to collect mainly fine inside the brass liner. At the completion of liner
sands and silty soils (Bray and Sancio, 2006). advancement, the tube remains stationary for a
This device consists of an Osterberg-type minimum of 1 min. The third stage corresponds
hydraulic activated fixed-piston sampler, that can to the sample recovery. The sampler is retracted
retrieve relatively “undisturbed” soil samples to the surface and the liner is extracted and
(Markham et al., 2016). D&M sampler has a prepared for transport and storage. Figure 4
50 cm long liner, made of smooth brass that illustrates the operation phases of a fixed
effectively minimizes the friction between the hydraulic piston sampler, which is the same
tube walls and the soil. principle of the D&M.
On the other hand, the D&M sampler presents
an innovation related to its sealing system. This
system keeps the sample inside the liner during
the retrieval due to the vacuum generated during

IGS 3 ECSMGE-2019 - Proceedings

 A.2 - Investigation by in situ tests

In this study, the GP-Static (GP-S) was

implemented. This sampler follows the concepts
of fixed-piston sampling (Osterberg, 1973) with
triple core barrel, but makes use of a viscous
polymer gel during drilling. The main
components of the GP-S sampler, which
differentiate this equipment from conventional
Osterberg and D&M samplers, are a cutting shoe
and its three pistons: the stationary piston, the
sampling tube advancing piston, and the core-
catcher activating piston. The first piston is fixed
and the other two are travelling pistons. The outer
tube secures the borehole and keeps the
penetration rod and piston fixed in alignment
during penetration. The advancing piston
contains the gel, ensures the downward
movement of the system and activates the catcher
while it is inserted into the soil. The core-catcher
piston captures the sample inside a metallic liner
tube, with approximate inner/outer diameters of
71/76 mm and 1 m length (Taylor, 2015).
The sampling methodology of the GP-S
includes three operation phases. The first phase
Figure 4. Fixed hydraulic piston sampler operation covers the sampler assembly and preparation of
(ASTM International, 2015) the gel, to a concentration ratio of 1% of polymer
in clean water, which is immediately inserted into
2.3 Gel-push sampler the device. At this phase, a borehole of 150 mm
The Gel-push sampler is an advanced device diameter must be drilled, as in the D&M
capable of recovering high-quality undisturbed operation procedure. In the second phase, the
samples of granular soils, developed by the sampler is lowered into the borehole and the GP-
Japanese geotechnical consulting company Kiso- S is connected to a water pump. Afterwards,
Jiban Consultants. This sampler is composed by clean water is pumped to the sampler at a constant
a triple core barrel and uses a viscous polymer gel pressure of 50 MPa or a penetration rate of 1
as its drilling fluid, hence the name Gel-push m/min. The core barrel starts to advance and the
(GP). Its name is a reference to the operating cutting shoe penetrates the soil. Simultaneously,
principle and the viscous polymer gel which is the hydraulic piston closes a bypass valve and the
used for soil collecting (Mori and Sakai, 2016). fixed piston enables to squeeze gel in the core
The gel lubricates and reduces friction between catcher lubricating the end of the collected soil.
the cut sample and the tube, both during sampling In the third phase, the core barrel advances
as well as during extrusion in the laboratory downward into the soil until 1 m depth (liner
(Taylor et al, 2012). Such innovation is a key length). The remnant gel flows through the liner,
factor in liquefaction assessment since the allowing the sliding of the sample and the blades
rheological properties of the polymer allows of the catcher close, holding the soil. Figure 5
preserving the soil structure (Viana da Fonseca displays the operation phases of the sampler.
and Pineda, 2017).

ECSMGE-2019 – Proceedings 4 IGS

 Collection of high-quality samples in liquefiable soils using new sampling techniques

Figure 6. Samples collected by GP-S and D&M, and

Figure 5. GP-S operation (Taylor, 2015) Ic from CPTu data

3 EXPERIENCE WITH THE NEW

SAMPLING TECHNIQUES
Three different boreholes separated by 2 m
distance were drilled, with the purpose of
retrieving soil samples with the minimum spatial Figure 7. D&M operation: (a) key components; (b)
variability between investigation points. The device assembled; (c) liner after sampling.
sampling techniques were implemented as
follows: two boreholes with the GP-S and one
with the D&M. In total, 29 samples (17 GP-S and
12 D&M) were collected. This paper focuses on
the performance of both sampling techniques in
collecting liquefiable soils and only the samples
retrieved at 5 to 12 m depth will be analysed. Figure 8. GP-S operation: (a) key components; (b)
Figure 6 illustrates the sampling profile. device assembled; (c) activated core catcher.
During the experimental campaign, it was
recognised that both samplers were successful in After sampling, the samples (inside the liners)
sampling silty sands (5 and 8 m) and silts (9 - 10 were hermetically sealed and transported to the
m). However, clean medium sands (7 and 11 m) laboratory, in the vertical position inside a
were not collected, since the sample dropped wooden box, specifically designed for these
during lifting with both samplers. The recovery liners. Each box includes two horizontal sections
ratio of GP-S ranged between 43% and 88% and to ensure the vertical alignment of the liners.
between 80% and 94% for the D&M. Figures 7 With the purpose of minimising the lateral
and 8 show the operation of both samplers. movement of the liners, foam was placed around
the tubes inside the box. The foam laminae were
also placed at the top and bottom of the boxes in

IGS 5 ECSMGE-2019 - Proceedings

 A.2 - Investigation by in situ tests

an effort to isolate the samples from vibrations. SCPTu and in the laboratory by means of bender-
Figure 9 shows one box, used for transporting the element (BE) bench tests.
samples from the site to the laboratory.

(a) (b) (c)

Figure 9. Transportation box: (a) ready for transport;
(b) detail of liner confinement using foam; (c) foam
isolation at the sides and the base.

4 SAMPLING QUALITY ASSESSMENT

In the laboratory, each sample was unsealed
and the total weight and effective length were
measured, in order to estimate its unit weight (γ).
Such values were compared against the in situ γ Figure 10. Comparison between the unit weight of the
obtained by means of the CPTu correlation samples collected by GP-S and D&M and the
proposed by Robertson and Cabal (2010). This estimated unit weight in depth (from CPTu data)
quick comparison allowed for a preliminary
assessment of the density variation due to the In addition, Ferreira et al. (2011) suggested a
sampling process. This revealed a good fitting classification based on five different quality
between the results, which is indicative of low zones (Table 1).
compression during sampling. Figure 10 presents
the γ comparison between lab and in situ values. Table 1. Sample quality classification based on Vs
The extrusion of samples from the liner was normalised ratio (Ferreira et al., 2011)
performed using a vertical hydraulic piston. Quality Sample
Vs* ratio Sample condition
zone quality
Subsequently, samples were divided into smaller A ≥ 85% Excellent Perfect
specimens for element testing (three specimens B 85% - 70% Very good Undisturbed
per liner on average). After extrusion, each C 70% - 60% Good Fairly undisturbed
specimen was transferred into a PVC tube, D 60% - 50% Fair Fairly disturbed
E < 50% Poor Disturbed
carefully sealed and stored in a wet room under
controlled temperature and humidity until testing.
The assessment of sampling quality was made, Normalisation with respect to soil state was
done using the empirical void ratio function
based on the comparison of field and laboratory
shear wave velocities, as proposed by Ferreira et 𝐹(𝑒) = 𝑒 −1.3 (Lo Presti et al., 1997) and
al. (2011). Such assessment compares the assuming 𝑝´𝑙𝑎𝑏 = 1 𝑘𝑃𝑎. Moreover, BE results
normalized shear wave velocity (Vs*) measured were interpreted according to the procedure
in the laboratory with the corresponding presented in Viana da Fonseca et al. (2009).
normalized in situ values, at the same depth. In Figure 11 shows the summary of the sampling
this work, VS data were obtained in the field by quality assessment results.

ECSMGE-2019 – Proceedings 6 IGS

 Collection of high-quality samples in liquefiable soils using new sampling techniques

be drawn:
• The results demonstrate that the GP-S and
D&M samplers induce minimal to low
disturbance in the fabric and structure in silts to
silty sands. The low friction of these samplers
introduces low compression of the soil during
sampling, providing representative soil samples.
• Key issues, which require specific attention to
preserve the high quality of the samples, also
include transport, handling and storage before
element testing. It is then necessary to consider
specific measures for minimising vibrations
during transport of the samples, namely using
specifically designed boxes, with the samples
Figure 11. Results of sampling quality assessment. positioned vertically and properly insulated.
• The comparison between seismic wave
From the sampling quality assessment, it can velocities measured in situ and in the laboratory
be observed that only two samples were is an excellent method to estimate the quality of
categorized as fair, one retrieved with the GP-S soil samples. However, for the comparison to be
and the other with the D&M. Furthermore, two valid, it is necessary to normalise the respective
D&M samples presented Vs ratios higher than void ratio function and mean effective stress.
100%. Nevertheless, such values are acceptable This method allowed estimating the effect of
since they are attributed to the uncertainties changes in the fabric and structure of the soil
associated with direct field measurements in during the sampling process.
terms of resolution (spaced 1.0 m into the
downhole). Both samplers presented a good 6 ACKNOWLEDGEMENTS
performance in collecting liquefiable soils as
almost all samples were categorized as excellent LIQUEFACT project (Assessment
to very good, which indicates a high-quality and mitigation of liquefaction
sample condition. potential across Europe: a holistic
approach to protect structures / infrastructures for
5 CONCLUSIONS improved resilience to earthquake-induced
liquefaction disasters) has received funding from
This paper described a sampling campaign on the European Union's Horizon 2020 research and
liquefiable soils by means of two advanced innovation programme under grant agreement
sampling techniques, GP-S and D&M. High- GAP-700748. The authors also acknowledge the
quality samples were collected in a pilot site Portuguese Foundation for Science and
located in Benavente, near Lisbon. Sampling Technology (FCT) for the support of the
quality was assessed using a comparison between PTDC/ECM/GEO/1780/2014 project
normalized Vs values measured in the field and in “Liquefaction Assessment Protocols to Protect
the laboratory. A preliminary assessment of the Critical Infrastructures against Earthquake
quality of each sample indicated close results Damage: LIQ2PROEARTH”. The second and
between both samplers. In addition, more than fourth authors also acknowledge the support of
90% of samples studied in this work were FCT through grants SFRH/BPD/120470/2016
classified, based on the Vs criteria, as excellent or and SFRH/BD/120035/2016, respectively.
very good quality. The following conclusions can

IGS 7 ECSMGE-2019 - Proceedings

 A.2 - Investigation by in situ tests

10.1520/GTJ20150244.
7 REFERENCES Mori, K., Sakai, K. 2016. The GP sampler: a new
innovation in core sampling. Geotechnical and
ASTM International 2015. D6519 - Standard Geophysical Site Characterisation 5. Sydney,
Practice for Sampling of Soil Using the Hydraulically Australia: AGS, pp. 99–124.
Operated Stationary Piston Sampler. Osterberg, J.O. 1973. An improved hydraulic
Azevedo, J. Guerreiro, L., Bento, R., Lopes, M., piston sampler, 8th ICSMFE. Moscow, pp. 317–321.
Proença, J. 2010. Seismic vulnerability of lifelines in Lo Presti, D.C.F., Jamiolkowski, M., Pallara, O.,
the greater Lisbon area, Bulletin of Earthquake Cavallaro, A., Pedroni, S. 1997. Shear modulus and
Engineering. Springer Netherlands, 8(1), pp. 157– damping of soils, Géotechnique, 47(3), pp. 603–617.
180. doi: 10.1007/s10518-009-9124-7. doi: 10.1680/geot.1997.47.3.603.
Bray, J.D., Markham, C.S., Cubrinovski, M. 2017. Ramos, C., Viana da Fonseca, A., Vaunat, J. 2015.
Liquefaction assessments at shallow foundation Modeling flow instability of an Algerian sand with the
building sites in the Central Business District of dilatancy rule in CASM, Geomechanics and
Christchurch, New Zealand, Soil Dynamics and Engineering. Techno-Press, 9(6), pp. 729–742. doi:
Earthquake Engineering. Elsevier, 92, pp. 153–164. 10.12989/gae.2015.9.6.729.
doi: 10.1016/J.SOILDYN.2016.09.049. Robertson, P.K. 2009. Interpretation of cone
Bray, J.D., Sancio, R.B. 2006. Assessment of the penetration tests - a unified approach, Canadian
Liquefaction Susceptibility of Fine-Grained Soils, J. geotechnical journal. NRC Research Press, 46(11),
Geotechnical and Geoenvironmental Engineering, pp. 1337–1355. doi: 10.1139/T09-065.
132(9), pp. 1165–1177. doi: 10.1061/(ASCE)1090- Robertson, P.K., Cabal, K.L. 2010. Estimating soil
0241(2006)132:9(1165). unit weight from CPT, in 2nd International
Cabral, J., Moniz, C., Batlló, J., Figueiredo, P., Symposium on Cone Penetration Testing. CA, USA.
Carvalho, J., Matias, L., Teves-Costa, P., Dias, R., Saldanha, A.S., Viana da Fonseca, A., Ferreira, C.
Simão, N. 2013. The 1909 Benavente (Portugal) 2018. Microzonation of the liquefaction susceptibility:
earthquake: search for the source, Natural Hazards, case study in the lower Tagus valley, Geotecnia, 142,
69(2), pp. 1211–1227. doi: 10.1007/s11069-011- pp. 7-34. doi: 10.24849/j.geot.2018.142.01.
0062-8. Taylor, M.L. 2015. The geotechnical
Chen, C.-C., Lee, W.F., Chen, J.-W., Ishihara, K. characterisation of Christchurch sands for advanced
2014. Liquefaction potential of non-plastic silty sand, soil modelling. University of Canterbury.
J. Marine Science and Technology, 22(2), pp. 137– Taylor, M.L., Cubrinovski, M., Haycock, I. 2012.
145. doi: 10.6119/JMST-013-0117-3. Application of new “Gel-push” sampling procedure to
Eurocode 8 2005. Design of structures for obtain high quality laboratory test data for advanced
earthquake resistance, European Committee for geotechnical analyses, in 2012 NZSEE Conference.
Standardization. Christchurch, New Zeland, pp. 1–8.
Ferreira, C., Viana da Fonseca, A., Nash, D. 2011. Teves-Costa, P., Batlló, J. 2011. The 23 April 1909
Shear wave velocities for sample quality assessment Benavente earthquake (Portugal): macroseismic field
on a residual soil, Soils and Foundations, 51(4), pp. revision, J.Seismology, 15(1), pp. 59–70. doi:
683–692. doi: 10.3208/sandf.51.683. 10.1007/s10950-010-9207-6.
Ishihara, K., Harada, K. Lee, W.F., Chan, C.C., Viana da Fonseca, A., Ferreira, C., Fahey, M. 2009.
Safiullah, A.M.M. 2016. Post-liquefaction settlement A Framework Interpreting Bender Element Tests,
analyses based on the volume change characteristics Combining Time-Domain and Frequency-Domain
of undisturbed and reconstituted samples, Soils and Methods, Geotechnical Testing Journal, 32(2), p.
Foundations. Elsevier, 56(3), pp. 533–546. doi: 100974. doi: 10.1520/GTJ100974.
10.1016/J.SANDF.2016.04.019. Viana da Fonseca, A., Pineda, J. 2017. Getting
Markham, C.S., Bray J., Riemer M., Cubrinovski high-quality samples in “sensitive” soils for advanced
M. 2016. Characterization of Shallow Soils in the laboratory tests, Innovative Infrastructure Solutions.
Central Business District of Christchurch, New Springer International Publishing, 2(1), p. 34. doi:
Zealand, Geotechnical Testing Journal. ASTM 10.1007/s41062-017-0086-3.
International, 39(6), p. 20150244. doi:

ECSMGE-2019 – Proceedings 8 IGS