Talesnick, M. et al. Géotechnique [https://doi.org/10.1680/jgeot.19.P.

019]

K0 determination using improved experimental technique

MARK TALESNICK
, SHAY NACHUM
and SAM FRYDMAN†

Lateral soil pressures under at-rest conditions are commonly estimated as K0 times vertical normal
pressure, where K0 is obtained from generally accepted empirical formulae, related to the soil’s friction
angle and the stress history. Results of a programme of laterally, rigidly confined, vertical compression
and unloading tests on sand and clay are presented, in which two significant experimental techniques
were employed: (a) lateral pressures were measured using null pressure gauges, preventing gauge
diaphragm deflection; (b) side-wall friction was measured, and minimised using special friction
reduction measures. Comparative tests in which one or both of these measures were not employed
indicate the significant errors induced in both the vertical and the horizontal pressures used for K0
determination. The bank of experimental data, which formed the basis for the commonly accepted
empirical formulae, did not take account of these issues, and they are shown to be inconsistent with
values measured in the present study.

KEYWORDS: earth pressure; friction; laboratory equipment; laboratory tests

INTRODUCTION issues, as well as to issues important to soil pressure

The assignment of values for the magnitude of horizontal measurement and laboratory testing procedures in general.
earth pressure at rest, which is relevant to many problems in
geotechnical engineering, is often based upon the magnitude
of the coefficient of lateral earth pressure at rest, K0, defined
here as the ratio between the effective horizontal pressure, σh′ , BACKGROUND
and the effective vertical pressure, σv′ , for conditions of zero Despite the common conception, there is no ‘correct/
lateral movement. A more basic definition requires that the acceptable’ closed-form solution for the determination/
vertical and horizontal directions be principal directions calculation of K0.
(e.g. Mesri & Hayat, 1993; Terzaghi et al., 1996), but K0 is Jaky (1944), assuming the soil to be in a plastic, limit
often used to define the stress ratio next to retaining state, developed a relationship between K0 during virgin
structures where there is friction along the structure–soil (normally consolidated) loading, (K0)N/C, and the angle of
interface (e.g. Bowles, 1968; Kezdi, 1975; Wu, 1975). internal friction, ϕ′, which he simplified to the following
Unfortunately, there is no correct theoretical formulation approximation
for the determination of the magnitude of lateral earth ðK0 ÞN=C ¼ ð1  sin ϕ′Þ ð1Þ
pressures at rest, or for the magnitude of the coefficient K0
(e.g. Michalowski, 2005). Furthermore, to utilise K0 properly, Although (K0)N/C is obviously not relevant to a plastic,
a representative value for σv′ must be known. This may be limit state in the soil, and consequently Jaky’s theoretical
trivial when considering horizontal, free-field situations. development leading to equation (1) is invalid, it has been
However, it is not simple when considering the conditions almost universally adopted. Numerous researchers have
adjacent to a non-yielding structural element – for example, claimed good experimental agreement with laboratory
in the field next to a retaining structure, in a laboratory measurements under conditions where horizontal/lateral
specimen next to the wall of an oedometer, or within a scaled movements were presumably restricted (e.g. Hendron, 1963;
centrifuge model. In these situations, frictional shear stresses Brooker & Ireland, 1965; Wroth, 1972). Alpan (1967), while
develop along the vertical interfaces, where the vertical soil recommending use of equation (1) for sands, evaluated data
pressure is, consequently, usually unknown. presented by Kenney (1959) for clays and suggested the
This paper reconsiders these issues by performing and following expression
analysing one-dimensional load/unload laboratory tests on
specimens of sand and clay. Testing was performed within a ðK0 ÞN=C ¼ 019 þ 0233  logðIp Þ ð2Þ
specially designed cell, which incorporates measures to where Ip is the plasticity index.
evaluate and reduce side-wall friction and tools to properly For unloading, Schmidt (1966) suggested that K0 be
and rigorously measure horizontal soil pressure. The out- related to (K0)N/C and the overconsolidation ratio, OCR, by
comes have relevance to practical, full-scale engineering an expression of the following form
K0 ¼ ðK0 ÞN=C  OCRα ð3Þ

On the basis of limited experimental evidence, Alpan

(1967: p. 34) suggested, for sands, a decrease in α with increas-
Manuscript received 15 January 2019; revised manuscript accepted
ing friction angle, as shown in Fig. 1, stating that ‘this trend is
12 March 2020.
Discussion on this paper is welcomed by the editor. not surprising since one should expect the effect of over-

Faculty of Civil & Environmental Engineering, Technion, Haifa, consolidation to be less marked in the case of denser sands’.
Israel. For clays, he suggested that α is better related to the plasticity
† Faculty of Civil & Environmental Engineering, Technion, Haifa, of the soil, and recommended equation (4), indicating a
Israel (Orcid:0000-0002-7842-6680). decrease in α with increasing plasticity index, Ip.

Downloaded by [ Technion - Israel Institute of Technology] on [21/06/20]. Copyright © ICE Publishing, all rights reserved.
 2 TALESNICK, NACHUM AND FRYDMAN
0·50 free-field conditions, questioning the applicability of
equation (5). Both factors noted above have the potential
0·48 to induce the development of erroneous hysteresis, as will be
0·46
demonstrated below.

α 0·44
Experimental considerations
0·42 Most of the laboratory-based data published in the
literature aimed at measuring horizontal soil pressures
0·40
under at-rest conditions, and defining K0, were based upon
0·38 one of five different experimental approaches: flexible
34 36 38 40 42 44 46
ring oedometer tests (e.g. Komornik & Zeitlen, 1965;
Angle of internal friction, φ': degrees
Bellotti et al., 1975; Lee et al., 2013), rigid ring oedometer
tests (e.g. Bellotti et al., 1975; Gareau et al., 2006; Monroy
Fig. 1. Relationship between α and angle of internal friction, ϕ′, after et al., 2014), null oedometers (e.g. Hendron, 1963; Ofer,
Alpan (1967) 1981; Mesri & Hayat, 1993), testing within soil pressure
vessels (e.g. Taylor, 1945; Weiler, 1979; Selig, 1980; Selig
et al., 1990) and triaxial cell manipulations (e.g. Bishop,
Ip ¼ 281  log ð185αÞ ð4Þ 1958; Al-Hussaini & Townsend, 1975; Menzies et al., 1977;
Santana & Candeias, 2015). In order to obtain meaningful
Meyerhof (1976) suggested that α be taken as 0·5 for all soils, data, the testing methodology must include the capacities to:
close to Alpan’s upper limit for sands. Mayne & Kulhawy (a) properly apply and maintain at-rest soil conditions;
(1982), from a study of an extensive collection of experimen- (b) properly and accurately measure soil pressure; and
tal data, recommended that α be taken as sin ϕ′, resulting in (c) correctly and accurately account for a representative
the equation value of σv′ .
The difficulty in obtaining meaningful data of horizontal
K0 ¼ ð1  sin ϕ′Þ  OCRsin ϕ′ ð5Þ soil pressure in the at-rest condition is in the definition of the
Note that contrary to Alpan’s suggestion, equation (5) at-rest condition itself. Since there is no correct theoretical
indicates that α is expected to increase, rather than decrease, approach for the determination of the horizontal earth
with increasing sand density. pressure at rest, it is impossible to verify experimental data
Mayne and Kulhawy’s equation (5) has become commonly against theory. Therefore, to best approach the problem from
accepted, and has been adopted in national standards and the experimental front, it is important to uphold the most
handbooks (e.g. CGS, 2006; ISI, 2008). The only soil basic of the boundary conditions. The at-rest condition
parameter required, in order to predict the development of means zero horizontal movement and zero horizontal soil
σh′ as a function of σv′ , is the angle of internal friction (ϕ′), strain. Even when using fluid-filled sensors, deflection is
provided that the pre-consolidation pressure is known. Fig. 2 generated because of (a) the compressibility of the fluid,
presents σh′ as predicted by equation (5) for a load–unload (b) deflection of the sensing membrane and (c) volume
vertical stress history of a granular soil with ϕ′ = 32°. Note the changes of the internal fluid due to temperature variations.
significant hysteresis displayed in horizontal pressure in Daigle & Zhao (2004) illustrated that the parasitic soil
response to unloading of vertical pressure. Lambe & pressures generated due to volume change as a result of
Whitman (1969) suggested that part of the hysteretic temperature change are significant.
response may be due to locked-in horizontal stresses that It is important to consider the rationale of the different
develop within the soil. However, a significant portion of this experimental approaches used for the determination of K0
hysteresis may be erroneous, resulting, as demonstrated by with respect to the required capabilities noted above.
Talesnick (2012), and later in the paper, from other factors Furthermore, since the definition of K0 includes the
not necessarily to do with the soil at all, namely: (a) the soil magnitude of σv′ , it is imperative that it be correctly defined
pressure measurement technique and (b) side-wall friction, in any experimental setting aimed at determining K0.
which controls the representative value of the σv′ . Talesnick In flexible ring oedometers, a soil specimen is confined in a
(2012) presented data that illustrated significantly smaller thin-walled ring with foil strain gauges bonded to its outer
hysteresis of horizontal pressure during unloading in surface and calibrated to known inner pressure. The thinner
the ring, the higher and cleaner the output signal, but the
greater the global lateral strain, making the situation less
compatible with at-rest conditions. The inherent problem
50 with flexible ring oedometers is that they are flexible and so
do not uphold at-rest conditions.
Horizontal pressure, σ'h

40 Furthermore, in any oedometer test, frictional stresses

develop on the ring–specimen interface. During loading the
30 specimen undergoes vertical compression, accompanied by
the development of upward-acting frictional shear forces
20 on the soil along the oedometer’s inner wall. These forces
φ' = 32°
Load reduce the vertical pressure increment at all levels of the
10 specimen (Sivrikaya & Togrol, 2005; Teerachaikupanich
Unload
et al., 2007; Lodahl et al., 2016), so that, over the specimen
0 height, the actual vertical pressure is smaller than that
0 20 40 60 80 100 applied to the specimen top. During unloading, the specimen
Vertical pressure, σ'v rebounds, the direction of the parasitic frictional stresses is
reversed and the vertical pressure at the specimen base may
Fig. 2. Representation of the Mayne & Kulhawy (1982) formulation become greater than that applied at the top. These parasitic
for K0 frictional forces are not usually measured, and their

Downloaded by [ Technion - Israel Institute of Technology] on [21/06/20]. Copyright © ICE Publishing, all rights reserved.
 K0 DETERMINATION USING IMPROVED EXPERIMENTAL TECHNIQUE 3
magnitude is unknown. Consequently, a representative internal deformation in response to the deflection incurred
magnitude of the effective vertical pressure is unknown, during loading. This mismatch results in over-registration
making a reliable determination of K0 difficult. of lateral pressure during unloading, and extraneous
In rigid ring oedometers, soil pressure sensors are hysteresis. There is no method to differentiate between the
embedded flush with the inner surface of the oedometer actual soil response, which may include a hysteretic com-
and respond to local lateral soil pressure. Ring wall friction ponent, and the parasitic hysteresis, which is inherent to
is commonly ignored and problematic, also, in these the measurement technique. The issue to consider is how
oedometers. Furthermore, unless transducer diaphragm much lateral deformation may be allowed without having a
deflection is prevented, spurious pressure measurements significant negative effect on the recorded results. Currently,
result, as discussed earlier. The associated hysteresis in the answer to this question is: none. No lateral deflection
sensor output is sensor and soil dependent and is not easily should be allowed. There has been no credible evidence
calibrated (Zhu et al., 2009). presented in the literature to answer this question in any
A limited number of rigid ring oedometers, which other way.
integrate pressure balance (null type) soil pressure sensors
as part of the oedometer body, are described in the literature
(e.g. Gareau et al., 2006; Monroy et al., 2014). Construction Effects of side-wall friction
of the system is not trivial and the issue of interaction of the The issue of side-wall friction in oedometer testing has not
sensor output with the deformation of the ring itself requires been given a great deal of attention in the literature. The
consideration. Unfortunately, only limited test data are American Society for Testing and Materials (ASTM)
available using these systems, and there has been no reference standard for oedometer testing (ASTM, 2011) does not
to the effect of side-wall friction. provide significant guidance with reference to quantifying
Null oedometers (e.g. Hendron, 1963; Brooker & Ireland, side-wall friction, but does stipulate that the side wall of the
1965; Ofer, 1981; Mesri & Hayat, 1993) attempt to eliminate cell should be highly polished or coated with a low-friction
global lateral cell deflection by applying a controlled ‘null’ material. Sivrikaya & Togrol (2005) and Lodahl et al. (2016)
fluid pressure in the annulus between the sensing element of a illustrated the importance of accounting for side-wall friction
flexible ring oedometer and a rigid outer shell. If the system is in the determination of material properties. When attempting
designed correctly, the ‘null’ fluid pressure is equal to the to examine the development of K0 correctly, it is imperative
internal horizontal soil pressure. If properly designed, these to monitor the development of both σh′ and σv′ at the same
systems can produce reliable results when testing uniform and location. It may not be appropriate to, by default, assume
transversely isotropic materials. However, side-wall friction that the magnitude of the applied vertical pressure at the
needs to be taken into account. sample top is relevant for the determination of K0. In the
When testing within soil pressure vessels, sensors are presence of side-wall friction, the representative value for σv′
embedded in a soil mass in different orientations so as to (say at the sample mid-height) is lower than the vertical
define a representative stress condition in space and sub- pressure applied to the top of the sample during loading, and
sequently the value of K0. The main challenge in this may become greater than it during unloading. Correct choice
approach is correct measurement of soil pressure, considering of the representative vertical pressure is of no less importance
the problems mentioned earlier. than the representative value of the horizontal pressure in the
In triaxial testing (e.g. Bishop, 1958; Menzies et al., 1977; determination of K0.
Al-Hussaini & Townsend, 1975; Santana & Candeias, 2015),
the coefficient K0 is typically determined by adjusting either
the axial load and/or cell pressure to maintain at-rest TEST SET-UP
conditions, employing a feedback loop which constantly Experiments were performed in a prismatic aluminium cell
monitors the specimen diameter (Al-Hussaini & Townsend, with specimen plan of 60 mm by 60 mm (Fig. 3). Two main
1975) or perimeter (Bishop, 1958) at the sample mid-height. issues were considered in the design of the cell: (a) the ability
A second option involves performing feedback on the pore to monitor the vertical loads and/or the vertical pressure
volume of saturated specimens (Santana & Candeias, 2015). reaching the specimen base; (b) the method for measuring
A significant drawback of these approaches is that the horizontal soil pressure at the specimen boundaries.
geometry of triaxial specimens does not usually remain The square shape was selected to allow inclusion of planar
uniform as axial strains develop. The specimen ends are soil pressure measurement gauges flush with the test cell
typically laterally restrained by friction; barrelling or thin- walls. The test cell is constructed of four interchangeable side
ning of the specimen central section occurs as its length walls, each 60 mm high, 82 mm wide and 22 mm thick.
changes. Even when frictionless ends are used (e.g. Rowe & Lateral soil pressures were measured using two null soil
Barden, 1964), only finite segments of the specimen are pressure gauges (Talesnick, 2005; Talesnick et al., 2014)
actually known to be under at-rest conditions, while the mounted and fixed flush with the inner surfaces of two
remaining segments of the specimen may undergo radial opposing cell side walls. These gauges maintain zero
deformation. membrane deflection over their entire area, including the
sensing area, and have been shown to eliminate the issues of
soil arching and associated particle rearrangement described
Effects of diaphragm deflection on soil pressure measurement earlier. The gauges used have a sensing diameter of 13 mm,
The small deflections required to measure soil pressures in outer diameter of 25 mm, thickness of 4 mm and a nominal
the methods described above introduce at least two problems diaphragm thickness of 0·3 mm. In the test configuration
when measuring the development of lateral soil pressure. used, full-scale pressure is limited to 700 kPa. They operated
During the initial loading phase, the parasitic deflection independently, allowing comparison of results obtained
results in a departure from local, at-rest conditions and to an from each.
unknown reduction of the lateral pressure relative to that had The applied vertical force was monitored by a load cell
the at-rest condition been locally maintained. During mounted at the specimen top. The force divided by the
unloading, the sensing element strives to return, elastically, contact area was taken as the average vertical pressure applied
to its original condition. However, the specimen has main- to the sample, hereafter referred to as the applied vertical
tained contact with the sensor and will have developed pressure.

Downloaded by [ Technion - Israel Institute of Technology] on [21/06/20]. Copyright © ICE Publishing, all rights reserved.
 4 TALESNICK, NACHUM AND FRYDMAN

Side walls

Null gauge

Specimen
base

Base
supports

Load cell

(a)

(b) (c)

Fig. 3. Stages in the set-up of the test cell: (a) components; (b) base and cell assemblies; (c) complete assembly

The cell’s walls were not connected to the base, and so it (the base vertical pressure) to be compared to the soil
acted as a floating cell. The specimen base pedestal was pressure measured at the centre of the specimen base. The
placed upon a second load cell, allowing the force reaching difference between average pressure over the sample base and
the specimen bottom to be measured (Fig. 3). This force that measured locally, at the centre of the base, indicated the
divided by the base area is referred to as the base vertical non-uniformity of pressure at the base of the sample, largely
pressure. When a specimen is loaded vertically, the difference resulting from side-wall friction.
between the force applied on the upper surface of the
specimen and that registered at its base represents the vertical
frictional load developed on the side walls of the cell. The
two load cells had measurement capacity of 20 kN. TESTING OF SAND SAMPLES
To examine the effect of wall friction on the development The sand tested is a fine, uniformly graded, quartz dune
of lateral soil pressure and K0, two side wall conditions were sand from Caesarea, classified as SP according to the Unified
considered. The default condition was that of the machine- Soil Classification System (ASTM, 2017). The mean grain
milled aluminium side walls. In the second condition, size (D50) is 0·25 mm (cu = 1·7 and cc = 0·78), the specific
reduced side-wall friction was achieved through the use of gravity of the particles, Gs is 2·66, and the minimum and
friction reduction measures (FRMs) based on Tognon et al. maximum dry unit weights are 14·9 and 17·3 kN/m3,
(1999), which consisted of a single layer of polyethylene respectively. The test specimens were prepared close to
sheeting (0·1 mm thick) placed on a thin layer of graphite minimum dry density by funnelling dry sand grains into
grease. The grease was applied to fix the sheeting to the the test cell from a height of 10 mm (ASTM, 2000). The tests
side-wall surface (Fig. 4), allowing sliding between them. were performed on specimens of initial height 30 mm and
Fig. 4(a) represents the case where FRM is applied to a side unit weights ranging from 15·0 to 15·1 kN/m3. At this unit
wall without a null gauge installed and Fig. 4(b) the case weight, the friction angle, ϕ, of the sand is about 32°
of a side wall fitted with FRM in which a null gauge is (Frydman, 2000). Vertical load was applied to the specimen
installed. Data presented herein obtained through testing top through a rigid aluminium block at a rate of 25 kPa/min,
when utilising FRM relate to experiments performed with and subsequently unloaded at the same rate. On average, a
side-wall treatment as shown in Fig. 4. vertical pressure of 275 kPa was applied to the specimen tops
Specimens of both sand and clay were tested. In several and ultimate vertical compression was of the order of up to
tests performed on the sand, an additional null gauge was 3%. The vertical position of the compressed specimen was
flush mounted at the centre of the specimen baseplate, such that its centre reasonably aligned with the centre of the
allowing the average pressure registered by the lower load cell soil pressure gauges mounted in the side walls.

Downloaded by [ Technion - Israel Institute of Technology] on [21/06/20]. Copyright © ICE Publishing, all rights reserved.
 K0 DETERMINATION USING IMPROVED EXPERIMENTAL TECHNIQUE 5

Polyethylene sheets

(a) (b)

Fig. 4. Side-wall FRMs: (a) without null gauge; (b) with null gauge

Effects of side-wall friction by the soil null gauge fixed at the centre of the specimen base.
Figure 5 shows the development of soil pressure in two In the case of the test with FRM, the pressure registered by
sand specimens as a function of the applied vertical pressure. the base null gauge was equal to the applied vertical pressure.
Relevant information from these tests is summarised in This represents a difference of + 7% from the base vertical
Table 1. pressure as measured by the base load cell. Without FRM,
Figures 5(a) and 5(b) refer to tests on specimens with the base soil null gauge registered 85% of the applied vertical
the default side-wall friction condition and with FRM, pressure, representing a difference of + 13% from the base
respectively. The plots of both figures illustrate near-linear vertical pressure. Clearly, side-wall friction results in non-
development of base vertical pressure and horizontal uniform vertical stresses both along the sample height and
pressure as a function of the applied vertical pressure across its width. The chosen FRMs are efficient in reducing
during initial loading. In both cases, the plots illustrate the side-wall friction, contributing to improved longitudinal and
development of upward-acting frictional forces on the side lateral uniformity of the stress regime within the sample.
walls of the cell. The loads registered by the base load cell During unloading, hysteresis was registered by all sensors.
were 28% and 7% smaller than those applied to its top in Hysteresis in both measures of vertical base pressure (top
the tests performed without and with FRM, respectively. The graphs) was significantly greater in the test performed
same trend is true when considering the pressures registered without FRM and is a result of the reversal in direction of

300 300
Load base vertical pressure Load base vertical pressure
Unload base vertical pressure Unload base vertical pressure
Vertical pressure: kPa

Vertical pressure: kPa

200 200

100 100
Load base null gauge Load base null gauge
Unload base null gauge Unload base null gauge
0 0
0 100 200 300 0 100 200 300
Applied vertical pressure: kPa Applied vertical pressure: kPa

160 160
Load left NG Load right NG
Horizontal pressure, σ'h: kPa

Horizontal pressure, σ'h: kPa

Unload left NG Unload right NG

120 120

80 80

40 Load right NG 40 Load left NG

Unload right NG Unload left NG
0 0
0 100 200 300 0 100 200 300
Applied vertical pressure: kPa Applied vertical pressure: kPa
(a) (b)

Fig. 5. Development of soil pressure on sand specimens as a function of applied vertical pressure: (a) default side-wall friction; (b) FRMs

Downloaded by [ Technion - Israel Institute of Technology] on [21/06/20]. Copyright © ICE Publishing, all rights reserved.
 6 TALESNICK, NACHUM AND FRYDMAN
Table 1. Data from loading tests on sand

Applied Base vertical Average Base Horizontal (K0)N/C based (K0)N/C based
vertical pressure: kPa vertical NG: pressure: kPa on applied on average
pressure: kPa pressure: kPa kPa pressure pressure

Without FRM 278 190 234 235 102 0·37 0·44

With FRM 278 255 267 277 146 0·53 0·55

vertical frictional loads at the specimen sides as the specimen 160

rebounds during unloading. The greater the friction is, the Without FRM

Horizontal pressure, σ'h: kPa

greater the hysteresis.
The horizontal pressures (σh′ ) registered by the two 120

lateral null gauges (left side–right side) in each experiment

were consistent with one another. Further discussion and 80
analysis will consider the average of the two in each
experiment.
40
Two significant issues regarding the development of
horizontal soil pressures are apparent in Fig. 5.
0
(a) During loading, at any level of applied vertical pressure, 0 100 200 300
the horizontal pressure registered without FRM Average vertical pressure, σ'v: kPa
(Fig. 5(a)) was significantly smaller than with FRM Load Unload
(Fig. 5(b)).
(a)
(b) During unloading, hysteresis in horizontal pressures was
evident in both cases. 160
Horizontal pressure, σ'h: kPa

With FRM
120
The coefficient of lateral earth pressure at rest, K0
The loading stage, (K0)N/C. The magnitude of the
coefficient of lateral earth pressure at rest, (K0)N/C, during 80
the initial loading segment, may be determined for the two
cases shown in Fig. 5. To do so, a relevant value of σv′ must be
40
defined. Fig. 5 has indicated that the vertical pressure varies
along the height of the specimen, and therefore adoption of
the applied pressure, as is customarily done, may not, 0
satisfactorily, represent the value which should be related 0 100 200 300
to the horizontal pressure measured at mid-height for Average vertical pressure, σ'v: kPa
determination of (K0)N/C. The form of the variation of the (b)
vertical pressure along the specimen height and across its
width is unknown. For present purposes it is assumed that σv′ Fig. 6. Horizontal pressure as a function of average vertical pressure:
varies linearly from top to base (i.e. uniform frictional shear (a) default side-wall friction; (b) FRMs
stress) and that the average of the applied and base pressures
is the average (across-width) vertical pressure at specimen
mid-height. The measured horizontal pressures are shown estimating (K0)N/C provides a reasonable value, while this is
plotted against this average vertical pressure and the applied not the case when side-wall friction is not reduced.
vertical pressure in Fig. 6, for default and reduced friction The assumption of uniform friction on the side walls
wall conditions. results in an estimated frictional shear stress of 44 kPa.
The coefficient of earth pressure at rest ((K0)N/C) is The measured horizontal pressure was 104 kPa, and
commonly calculated by relating to the horizontal pressure consequently the interface friction angle is estimated as
as a function of the applied vertical pressure. Doing so for the arctan (44/104) = 23°. A similar value was found from
experiments performed with and without FRM yields values interface shear measurements recorded by Avraham (2014),
(Table 1) of 0·53 and 0·37, respectively. However, when using Alaaldin (2018) and from direct shear tests between the sand
the average vertical pressure at specimen mid-height, values and aluminium performed as part of this research. For the
of 0·55 and 0·44 result. case in which friction reduction was employed, the interface
The average vertical pressure at specimen mid-height, used friction angle is estimated to be about 4°.
above to estimate (K0)N/C, does not account for the variation The value of (K0)N/C = 0·55 obtained from tests in which
of vertical pressure across the specimen width. In the case FRM were employed would, according to equation (1),
where FRM was applied, this variation has been illustrated indicate a ϕ′ value of approximately 27°, significantly lower
to be marginal; however, in experiments performed without than would be expected for this sand (e.g. Frydman, 2000).
FRM, this is not the case and the vertical pressure next to the This is considered further below.
side wall is less than that at the centre. Consequently, the
value of 0·44 determined for (K0)N/C when FRMs were not
applied would be expected to be higher, approaching (and The unloading stage, K0. During unloading, Fig. 5 indi-
possibly equal to) the value of 0·55 found with FRM. Clearly, cates a non-linear and hysteretic response for testing
by significantly reducing side-wall friction, the common performed with and without FRM. Fig. 7 presents loading
practice of adopting the applied vertical pressure for and unloading curves for σh′ plotted against σv′ , both

Downloaded by [ Technion - Israel Institute of Technology] on [21/06/20]. Copyright © ICE Publishing, all rights reserved.
 K0 DETERMINATION USING IMPROVED EXPERIMENTAL TECHNIQUE 7
1·0 1·0
Normalised horizontal pressure

Normalised horizontal pressure

0·8 0·8

0·6 0·6

0·4 Load average σ'v

0·4
Unload average σ'v
0·2
0·2 Load applied σ'v

0 Unload applied σ'v

0 0·2 0·4 0·6 0·8 1·0 0
Normalised applied vertical pressure 0 0·2 0·4 0·6 0·8 1·0
Normalised vertical pressure
Load without FRM Load with FRM

Unload without FRM Unload with FRM Fig. 8. Comparison between normalised load–unload curves of the
test done with FRM
(a)

1·0
Normalised horizontal pressure

Alpan, 1967; Mayne & Kulhawy, 1982), but here also

0·8 ′
depending on (σh′ /σh,max ).
Assuming the validity of equation (3), it may be combined
0·6 with equation (9), to obtain the following expression

0·4 σ h =ðσ h;max Þ ¼ OCRα1 ð10Þ

0·2
Equation (10) has been used to prepare Fig. 9, in which
theoretical curves for various values of α are superimposed
0
on the experimental results shown in Fig. 7(b), where FRMs
0 0·2 0·4 0·6 0·8 1·0
were applied to the side walls. Avalue of α = 0·37 is consistent
with the experimental measurements, significantly lower
Normalised average vertical pressure
than the value of 0·5 suggested by Meyerhof (1976).
(b) According to Mayne and Kulhawy’s suggestion that
α = sinϕ′, this would indicate a ϕ′ value of 21°, unacceptably
Fig. 7. Normalised load–unload curves: (a) function of applied low for sands, while Alpan’s Fig. 1 would indicate a ϕ′ value
vertical pressure; (b) function of average vertical pressure
of above 46°, higher than would be expected.

normalised by their respective maximum values during

Effect of soil pressure transducer deflection
loading, for tests performed with and without friction
The empirical relations shown in equations (1)–(5) for
reduction. In Fig. 7(a), σv′ is represented by the vertical
(K0)N/C and K0 were based on published experimental data,
applied pressure, while in Fig. 7(b), σv′ is taken as the average
and their reliability is obviously related to the reliability
vertical pressure at specimen mid-height. The normalised
of these data. As discussed earlier, both side-wall friction
plots of Fig. 7(a) illustrate greater hysteresis when friction
and soil pressure transducer compliance/interaction affect
reduction is not applied. On the other hand, Fig. 7(b)
this reliability. The importance of side-wall friction has
illustrates that when σv′ is taken as the average mid-height
been demonstrated above. Most published soil pressure
vertical pressure, the plots are almost coincident. Following
the previous observation that, with FRM, the values of
applied and mid-height vertical pressure are almost equal
during loading, Fig. 8 illustrates that this is true also for 1·0
unloading. Considering the normalised, unloading curves, α = 0·37
the secant slope (ratio of vertical to horizontal values) is
α = 0·44
given by 0·8
Normalised horizontal pressure

secant slope α = 0·5

s ¼ ðσ′h =σ′h;max Þ=ðσ′v =σ′v;max Þ

ð6Þ 0·6
¼ ðσ′h =σ′v Þ=ðσ′h;max =σ′v;max Þ

¼ K0 =ðK0 ÞN=C ð7Þ 0·4

rewriting equation (6)

s ¼ ðσ′h =σ′h;max Þ  OCR ð8Þ 0·2
Load with FRM
equating equations (6) and (8) Unload with FRM

K0 ¼ ðK0 ÞN=C  OCR  ðσ′h =σ′h;max Þ ð9Þ 0

0 0·2 0·4 0·6 0·8 1·0
Thus, the value of K0 for sand during unloading is expressed Normalised average vertical pressure
in a form depending on the normally consolidated value and
the OCR, as has been done previously (e.g. Schmidt, 1966; Fig. 9. Correlation of α to the unload segment of a load–unload cycle

Downloaded by [ Technion - Israel Institute of Technology] on [21/06/20]. Copyright © ICE Publishing, all rights reserved.
 8 TALESNICK, NACHUM AND FRYDMAN
measurements were obtained using techniques that the range of values recommended by both Taylor (1945) and
included transducer compliance, resulting in unreliable Weiler & Kulhawy (1982) for reliable cell response. Weiler
estimations of K0. and Kulhawy suggested that, in dense soils, a ratio as high as
Figure 10 shows results of a test performed without FRM 5000 may be necessary. In the present test, co-linearity was
and without nulling the lateral soil pressure cells. Horizontal observed in Fig. 10 up to a measured horizontal stress of
pressure was obtained from the pressure–deflection cali- 38 kPa, at which the diaphragm deflection was 0·0021 mm,
bration of the sensing diaphragm. The measured horizontal corresponding to a diameter-to-deflection ratio of 6200. It
pressure is plotted against average vertical pressure at would appear that diaphragm deflection had an insignificant
specimen mid-height, compared to a parallel test with effect on the registered lateral pressure below this value, but a
nulling. greater effect at larger deflection. This test illustrates the
On loading, coincident straight lines are observed up to a negative effects of transducer compliance and side-wall
measured horizontal pressure of 38 kPa, but then increasing friction on pressure measurement, throwing doubt on
departure is evident. (K0)N/C at maximum stress was 0·36 previously reported K0 values.
without nulling, compared to 0·44 with nulling. The (K0)N/C
value of 0·36 would correspond to a ϕ′ value of 40° according
to equation (1). Fig. 11 shows the normalised horizontal
pressure for the non-nulled test as a function of the normal- TESTING OF CLAY SPECIMENS
ised average vertical pressure at specimen mid-height. Clay specimens were prepared by remoulding a natural
During unloading, an α value of about 0·75 is observed, montmorillonite clay sampled from the Jezreal valley of
corresponding to a ϕ′ value of 49° assuming equation (5), northern Israel. Geotechnical characteristics of the material
significantly higher than the value of 40° predicted by are presented in Table 2. The material was air dried and
equation (1) from the loading curve, and unrealistically crushed to an aggregate size of less than 4·75 mm (mesh #4).
high for this sand. The diaphragm deflection at the registered Test specimens were prepared by mixing a known mass of
lateral pressure (78 kPa) was 0·0044 mm. The corresponding air-dried soil with a pre-determined mass of water, to obtain
diameter-to-deflection ratio is approximately 3000, within a nominal water content of 24%. The material was then cured
for 24 h in a sealed, nylon bag, remixed and placed loosely
into the test cell, which was outfitted with a collar that
extended 30 mm above the cell top. The specimen was
160 compressed in load control, under a vertical pressure, on
Load deflecting average, of 1050 kPa, from an initial height of 60 mm to a
Horizontal pressure, σ'h: kPa

height of 30 mm, corresponding to a nominal dry unit

120 Unload deflecting
weight of 14·4 kN/m3. A typical compression load–unload
result is shown in Fig. 12. Note the very large volumetric
80 strains; specimens typically underwent a void ratio change
from approximately 2·5 to 0·86 during loading, and were then
unloaded after a 20 min wait period. Compression and
40 Load nulled rebound indices, Cc and Cr were found to be 1·14 and 0·03,
Unload nulled respectively. Note that these values are for the unsaturated
0
0 50 100 150 200 250
Average vertical pressure, σ 'v: kPa
Table 2. Indicative properties of clay tested

Fines content (200 mesh) 100%

Fig. 10. Comparison between pressure measurements based on Specific gravity of solids (Gs) 2·72
deflecting membrane and nulled soil pressure transducers Liquid limit (ωl ) 75
Plastic limit (ωp) 21
USCS classification CH
1·0 Maximum dry density, Proctor: kN/m3 14·4
Optimum water content: % 25·7

0·8
Normalised horizontal pressure

60

0·6 55
α = 0·3
50
Specimen height: mm

α = 0·5
0·4 45
α = 0·7

α = 0·75 40

0·2
Load deflection 35

Unload deflection 30
0
0 0·2 0·4 0·6 0·8 1·0 25
Normalised average vertical pressure 0 400 800 1200
Applied vertical pressure: kPa
Fig. 11. Correlation of α to unload segment, measurements made
using deflecting membrane transducer Fig. 12. Typical compression load–unload curve of clay specimens

Downloaded by [ Technion - Israel Institute of Technology] on [21/06/20]. Copyright © ICE Publishing, all rights reserved.
 K0 DETERMINATION USING IMPROVED EXPERIMENTAL TECHNIQUE 9
clay, in terms of total stresses, and differ from values relevant with a corresponding, significant increase in Poisson’s ratio.
to the saturated case. This response differs from that of the sand, which underwent
The loading and unloading procedures were performed at very small deformations and demonstrated near-linear
average rates of 40 kPa/min. During loading and unloading, development of lateral pressure with increasing vertical
changes in specimen height, lateral soil pressure and load pressure. During unloading, the change of lateral soil
reaching the specimen base were monitored, as for the sand pressure is hysteretic, yet surprisingly flat. This form of
samples. Specimens were tested with and without FRMs, and response also appeared in the data of Monroy et al. (2014).
with horizontal pressure sensors configured in both null and Figure 13(b) is from a parallel test, without FRM. Three
deflection modes. During the loading and unloading differences can be noted in comparison to Fig. 13(a). During
segments of these tests the soil water potential in the clay loading, the frictional loads are significantly higher and
specimens undoubtably changed. In the current testing represent roughly 27% of the applied vertical load.
configuration, the pressures measured cannot be separated Significant hysteresis is seen in the development of base
into representative soil, air and water components. The vertical pressure, indicating reversal of the frictional forces
pressures measured clearly represent total and not effective along the specimen sides. The hysteresis in the horizontal soil
stresses. pressure is larger and of a different form compared to that in
Figure 13 presents a typical set of load–unload results in Fig. 13(a) and appears to be more curved.
tests where lateral soil pressure measurement was done in null At maximum vertical applied pressure, using the same
mode. During loading with FRM (Fig. 13(a)) the plot of assumptions as presented for the sand tests, the side-wall
base vertical pressure indicated development of upward friction for the test without FRM is estimated to be 151 kPa,
frictional force of approximately 9%. During unloading and the interface friction angle is estimated as
there is a small amount of hysteresis in the measured load tan1(151/503) = 17°. The value measured in a direct shear
reaching the specimen base. The two horizontal pressure test was found to be 21° – that is, slightly higher but of similar
measurements agree well over both the loading and unload- order. In the case of the test with FRM, the interface friction
ing segments; only the average will be considered henceforth. angle was estimated to be 5°, similar to the value of 4°
The data illustrate that the rate of development of lateral soil determined for the sand specimens.
pressure is not constant, but increases with applied vertical When the clay specimens are compressed, the relative
pressure. This observation was also shown by Monroy et al. location of the lateral null gauges along the specimen height
(2014), and can be attributed to the large increase in dry varies. At the start of compression, they are about three-
density that the specimen undergoes during the compression, quarters of the way down the specimen height, whereas at the
end of compression they are at specimen mid-height.
Assuming that the average vertical pressure varies linearly
along the height during compression, the change in their
1000 relative location must be considered when estimating the
vertical pressure acting adjacent to them. This has been done
800
in Fig. 14, which shows a comparison of normalised plots of
Pressure: kPa

horizontal pressure against vertical pressure at lateral null

600
gauge location for three testing combinations – FRM and
nulling, no FRM and nulling, and no FRM and no nulling.
400
In the latter test, the maximum gauge diaphragm deflection
was of the order of 0·025 mm; the ratio of diaphragm
200
diameter to deflection was about 500, considerably smaller
than generally accepted lower limits (e.g. Weiler & Kulhawy,
0
1982). It is observed that all normalised loading branches
0 400 800 1200
effectively coincided. When the soil pressure sensors were
Applied vertical pressure: kPa
nulled, wall friction resulted in only a small difference in the
Load base pressure Load right NG

Unload base pressure Unload right NG

1·0
Load left NG
Load deflection
Unload left NG
Unload deflection
(a) 0·8
Normalised horizontal pressure

1000

800 0·6
Pressure: kPa

600
0·4
400
Load with FRM
200 Unload with FRM
0·2
Load without FRM
0
Unload without FRM
0 400 800 1200
0
Applied vertical pressure: kPa
0 0·2 0·4 0·6 0·8 1·0
(b) Normalised representative vertical pressure

Fig. 13. Development of soil pressure on clay specimens as a function Fig. 14. Comparison between normalised load–unload curves; tests on
of applied vertical pressure: (a) FRMs; (b) default side-wall friction clay specimens

Downloaded by [ Technion - Israel Institute of Technology] on [21/06/20]. Copyright © ICE Publishing, all rights reserved.
 10 TALESNICK, NACHUM AND FRYDMAN
normalised unload curves, both relatively straight. However, 0·4
Load without grease
without nulling, a significant difference is observed; the
Unload without grease
unloading curve is significantly displaced, with large 0·3
Load with grease
hysteresis.

Relative friction (abs)

Unload with grease
Figure 15 shows the unloading portions of Fig. 14, 0·2
together with theoretical curves for different values of α,
based on equation (10). When nulling and FRM were 0·1
employed, the full curve is fitted, excellently, by an α value
of 0·08. When neither nulling nor FRM were employed, 0
an α value of about 0·5 appears most appropriate. Load with FRM
–0·1
Unload with FRM

DISCUSSION AND CONCLUSIONS –0·2

0 400 800 1200
The major contributions of the testing performed in the
present investigation resulted from two aspects of their Vertical pressure applied to specimen top: kPa
performance.
Fig. 16. Side-wall friction in a standard oedometer ring when testing
clay specimens
(a) Side-wall friction was measured, and a technique was
developed and applied to minimise it, enabling the
application of an almost uniform vertical pressure shows the results of three tests performed on identical,
throughout the sample height. compacted clay samples placed in standard consolidometer
(b) Use of null soil pressure sensors, in which deflection of rings and then loaded. In one case, no FRM was applied; in
the gauge diaphragm is eliminated, provided more the second, the ring inner surface was coated with grease; and
reliable measurement of lateral soil pressure. in the third, the FRM used in the present research was
employed.
Parallel measurements of lateral soil pressure on the Load reaching the base of the sample was measured as
boundaries of both sand and clay samples, with and described previously. The figure shows the per cent decrease
without nulling, demonstrated that even tiny diaphragm in vertical pressure from sample top to base in each of
deflections can cause erroneous evaluations. Tests on initially the tests as a function of the applied vertical stress. The
very loose sand showed that a ratio of diaphragm diameter to decrease in vertical stress was about 15%, 10% and 5%,
deflection as high as about 6000 may be required in order respectively, for the first, second and third test conditions.
to ensure reliable measurement over loading segments. The tests performed on sand and clay specimens with
Significantly higher ratios may be required for unloading. FRM in null mode pressure measurement indicate that the
Since evaluation of K0 is, by definition, the ratio between commonly employed empirical formulae for K0 are not valid,
horizontal and vertical effective pressures, its evaluation must having been developed on the basis of erroneous pressure
ensure knowledge of both of these pressures at the same
location. Many past evaluations have been based on testing
on laterally confined soil cylinders in which horizontal 1·2
pressure is measured at the soil–container interface, some-
where along its height, while the vertical pressure is taken as 1·0 Load
that applied at the sample top. It has been shown here that
Unload
without application of special FRMs, significant non-
uniformity of vertical pressure can develop along the K0 0·8
sample, as well as across its width. The friction can be of
significance even in the case of consolidation tests carried out 0·6
on thin samples tested in standard oedometer rings. Fig. 16

0·4
1·0 0 50 100 150 200 250 300
Unload deflection Average vertical pressure, σv: kPa
(a)
α = 0·08
0·8
α = 0·2 0·60
Normalised horizontal pressure

α = 0·5 0·56
0·6
0·52

Load
K0 0·48
0·4
Unload
With nulling 0·44

Unload with FRM 0·40

0·2
Unload without FRM
0·36
0 200 400 600 800 1000
0 Representative vertical pressure, σv: kPa
0 0·2 0·4 0·6 0·8 1·0 (b)
Normalised representative vertical pressure
Fig. 17. K0 plotted against average/representative vertical pressure:
Fig. 15. Correlation of α to unload segment on clay specimens (a) sand; (b) clay

Downloaded by [ Technion - Israel Institute of Technology] on [21/06/20]. Copyright © ICE Publishing, all rights reserved.
 K0 DETERMINATION USING IMPROVED EXPERIMENTAL TECHNIQUE 11
measurements. However, the empirical equation (3) relating Al-Hussaini, M. M. & Townsend, F. C. (1975). Investigation of
K0 for overconsolidated soil to the normally consolidated K0 testing in cohesionless soils, Technical Report S-75-16,
value, (K0)N/C, appears to fit the experimental evidence Project No. 4A061101A91D. Vicksburg, MS, USA: U.S.
satisfactorily, but, again, the commonly accepted empirical Army Engineer Waterways Experiment Station.
Alpan, I. (1967). The empirical evaluation of the coefficient K0 and
expressions for the exponent, α, are not valid. Fig. 17 shows
K0R. Soils Found. 7, No. 1, 31–40.
K0 as a function of vertical pressure for both the sand and ASTM (2000). D-4254-00; Standard test methods for minimum
clay specimens tested with FRM in null mode. It is noted index density and unit weight of soils and calculation of relative
that, in the case of sand (Fig. 17(a)), K0 is effectively constant density. West Conshohocken, PA, USA: ASTM International.
during loading and increases during unloading, whereas, in ASTM (2011). D2435-11; Standard test methods for one-
the case of clay (Fig. 17(b)), the opposite is observed, with K0 dimensional consolidation properties of soils using incremental
increasing during loading and staying effectively constant loading. West Conshohocken, PA, USA: ASTM International.
during unloading. While the form of the curve for sand is ASTM (2017). D2487-17; Standard practice for classification of
similar to that commonly presented (although α is different), soils for engineering purposes (Unified Soil Classification
the clay curve is significantly different. The increase in K0 System). West Conshohocken, PA, USA: ASTM International.
Avraham, R. (2014). Measurement of pressures acting on and around
during loading of the clay may result from its unsaturated a driven model of a pile in sand. MSc thesis, Technion, Israel
condition, and from the significant compression, densifica- Institute of Technology, Haifa, Israel.
tion and coupled increase in Poisson’s ratio which occurs Bellotti, R., Formigoni, G. & Jamiolkowski, M. B. (1975). Remarks
during loading. The almost constant K0 during unloading on the effect of overconsolidation on the coefficient of earth
corresponds to the value of α = 0·08 (i.e. almost zero) fitted in pressure at rest. In Proceedings of the Istanbul conference on soil
Fig. 15. That figure indicated that, if lateral pressures were mechanics and foundation engineering, pp. 17–25. Istanbul,
measured without nulling (as has usually been done), the Turkey: Istanbul Teknik Universitesi Insaat Facultest Matbaasi.
unload curve would be best fitted with an α value of about Bishop, A. W. (1958). Test requirements for measuring the
0·5, consistent with values commonly reported for clay when coefficient of earth pressure at rest. In Proceedings of the
conference on earth pressure problems, pp. 2–14. Brussels,
FRMs are not used and nulling is not applied. This
Belgium: International Society of Soil Mechanics &
emphasises the error in previous conclusions which have Foundation Engineering.
been commonly accepted. Fig. 17 is based on the data shown Bowles, J. E. (1968). Foundation analysis and design. New York, NY,
in Fig. 13(a), which is typical of all the tests performed on the USA: McGraw Hill.
clay. From the figure it is obvious that locked-in horizontal Brooker, E. N. & Ireland, H. O. (1965). Earth pressures at rest related
pressures do develop as a result of the application of the to stress history. Can. Geotech. J. 11, No. 1, 1–15.
load–unload cycle; the unload branch of the plot consistently CGS (Canadian Geotechnical Society) (2006). Canadian
falls above the loading branch of the plot. The interesting foundation engineering manual, 4th edn. Richmond, BC,
observation that the relationship between the horizontal and Canada: CGS.
representative vertical stress during unloading remains Daigle, L. & Zhao, J. Q. (2004). The influence of temperature
on earth pressure cell readings. Can. Geotech. J. 41, No. 3,
relatively constant is the actual material response when 551–559.
measurement is made without membrane deflection. This Frydman, S. (2000). Shear strength of Israeli soils. Isr. J. Earth Sci.
result, which deserves further investigation, is not inherent to 49, No. 2, 55–64.
the sensor; as noted in Fig. 17(a) the response of the sand is Gareau, L. F., Molenkamp, F. & Sharma, J. (2006). An improved
completely different from that of the clay. oedometer apparatus to measure lateral stress during testing.
Further experimental work, involving reliable, parallel Geotech. Test. J. ASTM 29, No. 3, 200–206.
measurements of horizontal and vertical pressures under Hendron, A. J. (1963). The behavior of sand in one-dimensional
conditions of lateral confinement on different soils, is required compression. PhD thesis, Department of Civil and Environmental
in order to establish more reliable empirical expressions. Engineering, University of Illinois at Urbana-Champaign,
Urban, IL, USA.
ISI (Israel Standards Institute) (2008). Israel Standard 940, Part 1:
Geotechnical design: Geotechnics and foundations for civil
NOTATION engineering. Tel Aviv, Israel: ISI.
Cc compression index Jaky, J. (1944). The coefficient of earth pressure at rest. J. Soc. Hung.
Cr rebound index Archit. Engrs 78, No. 22, 355–358 (in Hungarian).
cc coefficient of curvature Kenney, T. C. (1959). Discussion on proc. paper 1732 (Wu,
cu coefficient of uniformity 1958). J. Soil Mech. Found. Engng Div., ASCE 85, No. SM3,
D50 mean grain size 67–79.
Gs specific gravity of solids Kezdi, A. (1975). Lateral earth pressure. In Foundation engineering
Ip plasticity index handbook (eds H. F. Winterkorn and H. Y. Fang), pp. 197–220.
K0 coefficient of lateral earth pressure at rest New York, NY, USA: Van Nostrand Reinhold.
(K0)N/C coefficient of lateral earth pressure at rest for normally Komornik, A. & Zeitlen, J. G. (1965). An apparatus for measuring
consolidated conditions lateral soil swelling pressure in the laboratory. In Proceedings
α constant of the 6th international conference on soil mechanics and
σh horizontal total pressure foundation engineering, Montreal, Canada, vol. 1, pp. 278–281.
σh′ horizontal effective pressure Toronto, ON, Canada: Toronto University Press.
σh,max maximum horizontal total pressure Lambe, T. W. & Whitman, R. V. (1969). Soil mechanics. New York,
′
σh,max maximum horizontal effective pressure NY, USA: Wiley.
σv′ vertical effective pressure Lee, J., Yun, T. S., Lee, D. & Lee, J. (2013). Assessment of K0
′
σv,max maximum vertical effective pressure correlation to strength for granular materials. Soils Found. 53,
ϕ′ angle of internal friction No. 4, 584–595.
ωl liquid limit Lodahl, M. R., Sørensen, K. K., Mortensen, N. & Trankjær, H.
ωp plastic limit (2016). Oedometer tests with measurement of internal friction
between oedometer ring and clay specimen. In Proceedings of
the 17th Nordic geotechnical meeting: NGM 2016 – challenges in
REFERENCES Nordic geotechnics, pp. 289–298. Reykjavik, Iceland: The
Alaaldin, M. (2018). Effect of wall friction on earth pressures and Icelandic Geotechnical Society.
deformations in geotechnical modeling. MSc thesis, Technion, Mayne, P. W. & Kulhawy, F. H. (1982). K0–OCR relationships in
Israel Institute of Technology, Haifa, Israel. soil. J. Geotech. Div., ASCE 108, No. GT6, 851–872.

Downloaded by [ Technion - Israel Institute of Technology] on [21/06/20]. Copyright © ICE Publishing, all rights reserved.
 12 TALESNICK, NACHUM AND FRYDMAN
Menzies, B. K., Sutton, H. & Davies, R. E. (1977). A new system for Talesnick, M. (2012). A different approach and result to the
automatically simulating K0 consolidation and K0 swelling in the measurement of K0 of granular soils. Géotechnique 62, No. 11,
conventional triaxial test. Géotechnique 27, No. 4, 593–596, 1041 1045, https://doi.org/10.1680/geot.11.P.009.
https://doi.org/10.1680/geot.1977.27.4.593. Talesnick, M., Ringel, M. & Avraham, R. (2014). Measurement of
Mesri, G. & Hayat, T. M. (1993). The coefficient of earth pressure at contact soil pressure in physical modeling of soil–structure
rest. Can. Geotech. J. 30, No. 4, 647–666. interaction. Int. J. Phys. Modelling Geotech. 14, No. 1, 3–12,
Meyerhof, G. G. (1976). Bearing capacity and settlement of pile https://doi.org/10.1680/ijpmg.13.00008.
foundations. 11th Terzaghi lecture. J. Geotech., Div., ASCE 102, Taylor, D. W. (1945). Review of pressure distribution theories earth
No. GT3, 195–228. pressure cell investigations and pressure distribution data,
Michalowski, R. L. (2005). Coefficient of earth pressure at rest. Contract No. W22-053 eng-185. Washington, DC, USA:
J. Geotech. Geoenviron. Engng 131, No. 11, 1429. U.S. Army Corps of Engineers.
Monroy, R., Zdravkovic, L. & Ridley, A. M. (2014). Evaluation of Teerachaikupanich, N., Okumura, S., Matsunaga, K. & Ohta, H.
an active system to measure lateral stresses in unsaturated soils. (2007). Estimation of coefficient of earth pressure at rest using
Geotech. Test. J. 37, No. 1, 71–84. modified oedometer test. Soils Found. 47, No. 2, 349–360.
Ofer, Z. (1981). Laboratory instrument for measuring lateral soil Terzaghi, K., Peck, R. B. & Mesri, G. (1996). Soil mechanics in
pressure and swelling pressure. Geotech. Test. J. 4, No. 4, engineering practice, 3rd edn. New York, NY, USA: John Wiley
177–182. & Sons.
Rowe, P. W. & Barden, L. (1964). Importance of free ends in triaxial Tognon, A. R., Rowe, R. K. & Brachman, R. W. I. (1999).
testing. J. Soil Mech. Found. Div., ASCE 90, No. SMI, 1–27. Evaluation of side wall friction for a buried pipe testing facility.
Santana, T. & Candeias, M. (2015). K0 measurement in a sand Geotext. Geomembr. 17, No. 4, 193– 212.
using back volume change. Soils and Rocks, São Paulo 38, No. 1, Weiler, W. A. (1979). An investigation of the behavior of stress cells in
3–8. soil. MSc thesis, Cornell University, Ithaca, NY, USA.
Schmidt, B. (1966). Discussion of ‘Earth pressures at rest related to Weiler, W. A. & Kulhawy, F. H. (1982). Factors affecting stress cell
stress history’ (Brooker & Ireland, 1965). Can. Geotech. J. 3, measurements in soil. J. Geotech. Found. Div., ASCE 108,
No. 4, 239–242. No. GT12, 1529–1548.
Selig, E. T. (1980). Soil stress gauge calibration. Geotech. Test. J., Wroth, C. P. (1972). General theories of earth pressure and
ASTM 3, No. 4, 153–158. deformations, general report, Session 1. In Proceedings of the
Selig, E., Zhang, J. & Eberson, W. (1990). Evaluation of 5th European conference on soil mechanics and foundation
dynamic earth pressure cells for subgrade. Transp. Res. Rec. engineering, vol. 2, pp. 33–52. Madrid, Spain: Spanish Society
1596, 970391. of Soil Mechanics & Foundation Engineering.
Sivrikaya, O. & Togrol, E. (2005). Measurement of side friction Wu, T. H. (1975). Retaining walls. In Foundation engineering hand-
between specimen and consolidation ring with newly designed book (eds H. F. Winterkorn and H. Y. Fang), pp. 402–417.
oedometer cell. Geotech. Test. J., ASTM 29, No. 1, 87–94. New York, NY, USA: Van Nostrand Reinhold.
Talesnick, M. (2005). Measuring soil contact pressure on a solid Zhu, B., Jardine, R. J. & Foray, P. (2009). The use of miniature soil
boundary and quantifying soil arching. Geotech. Test. J., ASTM stress measuring cells in laboratory applications involving stress
28, No. 2, 171–179. reversals. Soils Found. 49, No. 5, 675–688.

Downloaded by [ Technion - Israel Institute of Technology] on [21/06/20]. Copyright © ICE Publishing, all rights reserved.