You are on page 1of 16

281

A critical reappraisal of ‘‘preconsolidation


pressure’’ interpretations using the oedometer
test
Storer J. Boone

Abstract: Over the past 70 years, a number of methods have been proposed for the interpretation of the characteristic ver-
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIV WINDSOR on 11/11/14

tical effective yield stress, or ‘‘preconsolidation pressure,’’ of compressible soils as defined using the standard oedometer
test. The concept of the ‘‘preconsolidation pressure’’ has been extraordinarily useful in geotechnical engineering for analy-
sing and predicting settlement behaviour and for normalizing other engineering parameters for comparative purposes. De-
fining this characteristic stress, however, is often problematic and relies heavily on graphical techniques that are subject to
some uncertainty, particularly when using semi-log plots for this purpose. This paper illustrates fundamental difficulties
with the interpretation of oedometer test and presents an alternative technique for defining unambiguous values of the ver-
tical effective yield stress.
Key words: oedometer, preconsolidation pressure, settlement.
Résumé : Dans les derniers 70 ans, plusieurs méthodes ont été proposées pour l’interprétation de la contrainte d’écou-
lement verticale effective caractéristique, ou « pression de pré-consolidation », de sols compressibles telle qu’obtenue à
partir de l’essai odométrique standard. Le concept de « pression de pré-consolidation » s’est avéré très utile en géotech-
nique pour l’analyse et la prédiction du comportement en tassement et pour la normalisation d’autres paramètres d’ingé-
For personal use only.

nierie dans un but comparatif. La définition de cette contrainte caractéristique est cependant souvent problématique et se
base beaucoup sur des techniques graphiques qui comportent des incertitudes, surtout lorsque des échelles semi logarithmi-
ques sont utilisées. Cet article illustre des difficultés fondamentales lors de l’interprétation d’essais odométriques et pré-
sente une technique alternative servant à définir les valeurs ambigües de la contrainte d’écoulement verticale effective.
Mots-clés : odomètre, pression de pré-consolidation, tassement.
[Traduit par la Rédaction]

Introduction loading behaviour after sampling and unloading, or recom-


Over the past 70 years, a number of methods have been pression, up to the point at which it reaches the maximum
proposed for the interpretation of the vertical effective yield stress it has experienced in the past, and then exhibits
0
stress, s vy 0
, of compressible soils in one-dimensional com- ‘‘virgin’’ compression beyond s vy . An example of such
pression as defined using the standard oedometer test (e.g., test data and mechanical stress history is illustrated in
Casagrande 1936; Becker et al. 1987; ASTM 2004). This Fig. 1. Because this characteristic yield stress may be the
yield stress has been termed the ‘‘preconsolidation pres- result of many factors including past depositional and
sure,’’ s p0 , or the past maximum stress. During one-dimensional stress history, weathering, cementation, and other minera-
compression in the oedometer, very soft and compressible logical and structural factors, the term ‘‘vertical effective
soils commonly exhibit nearly bilinear stress–displacement yield stress’’ is considered more fundamentally correct, yet
behaviour when the data are plotted on a semi-logarithmic the attractively simple concept and term of ‘‘preconsoli-
graph. Below s vy 0
, the displacements are relatively small; dation pressure’’ remains within common usage. Within
and above the characteristic s vy 0
, the soil yields and the ra- this paper, s p0 and s vy
0
are considered equivalent terms de-
tio of displacement to the applied stress difference in- scribing the same behaviour, and s p0 is used in keeping
creases during the subsequent phases of loading, then with familiar practice. Regardless of the causes underlying
decreases again as the soil structure becomes more difficult the yielding behaviour of soils, the ‘‘preconsolidation pres-
0 sure’’ has been an extraordinarily useful concept in geo-
to compress. The characteristic value of s vy has been de-
scribed as the ‘‘preconsolidation pressure,’’ based on the technical engineering for analysing and predicting
mechanical concept that the soil specimen undergoes re- settlement behaviour and for normalizing other geotechni-
cal engineering parameters for comparative purposes (e.g.,
Ladd and Foott 1974; Mesri 1975). Defining this character-
Received 29 July 2008. Accepted 5 August 2009. Published on istic stress, however, is often problematic and relies heav-
the NRC Research Press Web site at cgj.nrc.ca on 17 February ily on graphical techniques that are subject to some
2010. uncertainty, particularly when using semi-log plots for this
S.J. Boone. Golder Associates, Ltd., 309 Exeter Road, Unit 1, purpose (Becker et al. 1987; Grozic et al. 2003, 2005).
London, ON N6L 1C1, Canada (e-mail: sboone@golder.com). Subsequent interpretations of correlated parameters, such

Can. Geotech. J. 47: 281–296 (2010) doi:10.1139/T09-093 Published by NRC Research Press
282 Can. Geotech. J. Vol. 47, 2010

Fig. 1. Example one-dimensional stress – voids ratio relationship (Casagrande 1936). He then proposed a method for defining
showing paths of (A) sedimentation and past stress compression, the preconsolidation pressure based on an empirical graphi-
(B) unloading to in situ stress, (C) unloading and swelling due to cal approach, now well-known as the ‘‘Casagrande’’ con-
sampling, test preparation, and saturation, (D) recompression to struction in which the consolidation test data are plotted
past maximum stress, (E) ‘‘virgin’’ compression to maximum stress, with the voids ratio, e, on a vertical linear axis, and the ef-
(F) final unloading, and (G) potential influence of disturbance to fective vertical stress, s’, is plotted on a horizontal logarith-
compression curve. The Casagrande (1936) construction for inter- mic axis (frequently noted as e-logs’ curves) as shown in
pretation of preconsolidation pressure is illustrated in the inset. e0, Fig. 1. The purpose of this interpretation was to create a bi-
initial voids ratio. linear approximation of nonlinear compressibility behaviour
of soils that readily lent itself to calculation of primary con-
solidation settlement, S, using the now-familiar equations
 0 
Cc s þ Ds 0
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIV WINDSOR on 11/11/14

½1a S¼ H log v0 0
1 þ e0 s v0
for normally consolidated soils with a layer height (thick-
ness) of H and an initial voids ratio of e0, where the increase
in stress, Ds’, is greater than the in situ vertical effective
0
stress s v0 (identical with s p0 in this case), and the settlement
is governed by the slope of the virgin compression portion
of the curve (compression index), Cc;
 0 
Cr s v0 þ Ds 0
½1b S¼ H log 0
1 þ e0 s v0
0
for overconsolidated soils where s v0 þ Ds 0 > s p0 and the
settlement is governed by the slope of the recompression
For personal use only.

portion of the curve (recompression index), Cr; and


 0 !
0
Cr sp Cc s v0 þ Ds 0
½1c S¼ H log 0 þ H log
1 þ e0 s v0 1 þ e0 s p0
0
for overconsolidated soils where s v0 þ Ds 0 > s p0 .
Since 1936, a number of alternative approaches have been
suggested, most of which are based on similar empirical ob-
servations regarding the stress and voids ratio patterns ex-
hibited by different soils in the oedometer test. Some of
these alternative approaches can be highly subjective and
prove problematic for soils that do not exhibit a well-defined
as undrained shear strength using the ‘‘stress history and change in the curvature of the e-stress curves (i.e., more
normalized soil engineering properties’’ (SHANSEP) (Ladd ‘‘rounded’’ plots). A summary of alternative methods is
and Foott 1974) or Cam-clay (Mayne 1980) methods, for provided in Table 1 and Grozic et al. (2003, 2005) and
example, that use the interpreted preconsolidation pressure Clementino (2005) provided a detailed and graphical sum-
and ‘‘overconsolidation ratio’’ (OCR), s p0 =s v0
0 0
(where s v0 is mary of many of these approaches. A review of bilogarith-
the in situ vertical effective stress), as a basis may be mic approaches (e.g Butterfield 1979; Oikawa 1987;
questionable if there is significant uncertainty in defining Onitsuka et al. 1995) was completed by Onitsuka et al.
the value of s p0 . This paper presents an alternative ap- (1995), in which he concluded that his approach provided
proach to defining this characteristic stress, compares this the most consistent method for rapidly interpreting the pre-
approach with three other methods as applied to soils rang- consolidation pressure and was favourably comparable to
ing from very soft sensitive clays to low-plasticity silty the ‘‘work’’ method of Becker et al. (1987) that was based
clay, and compares back-analyses of four settlement case on theoretical mechanics. Grozic et al. (2003) completed
histories to evaluate the validity of the results. an evaluation of multiple methods of interpretation for a
number of natural soils and concluded that all methods
Background were subject to uncertainty and some were more cumber-
some than others in application. They also completed a
In 1936, Casagrande (1936) described the relationship be- comparison of these methods using low plasticity clay
tween the voids ratio and applied pressure for soils in terms samples (plasticity index PI = 10) consolidated from slurry
of ‘‘virgin’’ compression, rebound, and recompression. For that were subsequently unloaded such that a known me-
natural soils, he proposed that ‘‘. . . it should be possible to chanical preconsolidation pressure was created. An exam-
estimate the load p0 (s p0 in this paper), under which the soil ple of one of these tests is illustrated in Fig. 2 (curve
was consolidated in the ground, the so-called preconsolida- ‘‘f’’). Using these tests on slurry-formed samples, the Casa-
tion load, from a properly conducted consolidation test’’ grande (1936), Janbu (1969), Becker et al. (1987), Burland

Published by NRC Research Press


Boone 283

Table 1. Summary of methods for interpretation of preconsolidation pressure using oedometer tests.

Author(s) Method summary Basis


Casagrande (1936) Select maximum curvature point on plot of e (linear vertical axis) Empirical observations regarding the re-
and s’ (logarithmic horizontal axis), draw line tangent to curve sponse of soils in the oedometer test sub-
at this point, draw second horizontal line from this point, bisect ject to repeated loading and unloading
this angle, and the point at which this bisector intersects a third cycles and represents a point at which the
line drawn tangent to ‘‘virgin’’ compression curve represents the loading portion of the test would join the
‘‘preconsolidation stress,’’ s p0 . virgin compression line were it not for
sample disturbance.
Janbu (1969) Plot constrained modulus, m, versus s’ using linear scales, and the Empirical observations interpreted to repre-
pressure at which m exhibits a marked drop reflects s p0 . sent the structural breakdown of the soil,
graphical construction.
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIV WINDSOR on 11/11/14

Pacheco Silva Plot data as per Casagrande (1936), draw horizontal line at e0, Empirical observation, graphical construc-
(1970) where a line drawn tangent to the virgin compression curve in- tion.
tersects this horizontal line draw vertical line to the compression
curve and a second line to meet the tangent to the virgin com-
pression curve, the stress at this intersection defines s p0 .
Butterfield (1979) Intersection of linear fit lines of ln(1+e) and lns p0 , plotted on linear Empirical observation of test plots, graphical
scales. construction.
Becker et al. (1987) Define the strain energy density (‘‘work per unit volume’’), DWoed, Fundamentally new approach that relates the
for each successive loading increment of the oedometer test, i energy used to compress the sample to the
0
and i+1, etc., using , DWoed = 0:5ðs iþ1 þ s i0 Þ/ (ei + 1 – ei), an ar- applied stress, graphical construction.
ithmetic plot of cumulative DWoed and applied stress, will define
a generally bilinear plot and the intersection of these lines will
0
define the vertical effective yield stress, s vy (equal to s p0 ).
Oikawa (1987) Intersection of linear fit lines of log(1+e) and plotted on a linear Empirical observation of test plots, graphical
scale and stress plotted on a logarithmic (base 10) scale. construction.
For personal use only.

Burland (1990) Determine e100 (e at effective stress of 100 kPa) and Cc, plot voids Empirical comparison of compression beha-
index, Iv = (e – e100)/Cc on a linear scale and applied stress on a viour of remoulded soil to natural soil
logarithmic scale, and s p0 is defined by the intersection of a lin- based on the voids ratio and stress beha-
ear fit of the virgin and recompression segments of the resulting viours in each of these two states, graphi-
curve. cal construction.
Jacobsen (1992) s p0 ¼ 2:5s k , where sk is defined as the point of maximum curva- Empirical observations of Danish clay beha-
ture used within the ‘‘Casagrande construction.’’ viour in oedometer tests, graphical con-
struction.
Onitsuka et al. Intersection of linear fit lines of ln(1+e) plotted on an linear scale Empirical observation of test plots, graphical
(1995) and stress plotted on a logarithmic (base 10) scale. construction.

Fig. 2. Example stress – voids ratio curves for Leda clay (curves a through d; Quigley et al. 1981), Chicago clay (curve e; Finno, R.J. Three
oedometer tests and associated data from site in Evanston, Ill. Personal communication, April 2008.), slurry-formed laboratory sample
(curve f; Grozic et al. 2003), and a highly overconsolidated soil from Åsgard (curve g; Grozic et al. 2005).

(1990), and Onitsuka et al. (1995) methods produced esti- laboratory-prepared specimens within 1% to 3% of the
mates within 1% to 16% of the actual value. In a discus- known value. However, Grozic et al. (2005) demonstrated
sion and reply to Grozic et al. (2003, 2005), Clementino as well that this method also produced some ambiguity for
(2005) illustrated that the Pacheco Silva (1970) method more highly overconsolidated natural soils that exhibited
produced unambiguous interpretations of s p0 for the more rounded test plots, with this variability ranging be-

Published by NRC Research Press


284 Can. Geotech. J. Vol. 47, 2010

tween 3% and 55% with the three authors each interpreting Fig. 3. Example of highly overconsolidated soil showing two pos-
three tests independently. Figure 3 illustrates one example sible interpretations using the method of Pacheco Silva (1970) (data
in which Grozic et al. (2005) fit a line through some of from Fig. 2, curve g; Grozic et al. 2005). Note effects of scale
the data within the ‘‘virgin compression’’ range of a sam- comparing the plot in this figure with Fig. 2, curve g.
ple of highly overconsolidated natural soil subjected to a
controlled rate of the strain oedometer test. An alternative
line is also shown, following the guidance of the Pacheco
Silva (1970) approach in which this line is drawn tangent
to the ‘‘virgin compression curve. These two interpretations
demonstrate a subjective variability of about 65% for two
reasonable examinations of this data. Figure 4 illustrates
the variation in OCR, as indicated by the various attempts
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIV WINDSOR on 11/11/14

made by Grozic et al. (2003) to determine s p0 using the


methods of Casagrande (1936), Janbu (1969), Andresen et
al. (1979), Becker et al. (1987), and Onitsuka et al.
(1995). The potential disparity in estimation of s p0 is
clearly evident in Fig. 4, where the OCR for individual
samples (indicated by the vertical dotted lines) ranges over
an order of magnitude, perhaps due to the ‘‘quirks’’ in the
plotting of data suggested by Grozic et al. (2003).
Pinto (1992) and Clementino (2005) provided a clear il-
lustration of how scale effects (one of the ‘‘quirks’’ in plot-
ting) can influence use of the Casagrande (1936) method as
illustrated by Figs. 2 and 5. In addition to judgements re-
lated to the preconsolidation pressure, perceived ‘‘curvature’’
For personal use only.

of the stress – voids ratio plot and overall mechanical be-


haviour of individual test results may also be influenced by ures of variability. Figure 8 presents the ratio of each at-
scale effects as shown in the comparison between Leda clay, tempt and the Casagrade-determined or average value to
Chicago silty clay, and the Åsgard soils illustrated in Figs. 2 illustrate the broad range of possible estimates that may be
and 3. Scale effects are particularly relevant when computer- made depending on the method used. Figure 8a indicates
ized plotting routines automatically select scales based on that very few attempts (24 of the 149 attempts or 16%)
maximum and minimum values of stress and voids ratio. were within 20% of the Casagrande-determined value. Fur-
The strain energy density (‘‘work’’) approach is also thermore, Fig. 8b indicates that only about a third of the at-
somewhat subjective in the determination of the ‘‘best-fit’’ tempts (64 of 187) were within 20% of the average of the
linear interpretations of the cumulative strain energy density attempts for any given test.
and stress plots. Figure 6 illustrates several possible interpre- These comparisons clearly indicate that graphical linear
tations of a test completed on a high-quality block sample of approximations of the test result curves, regardless of plot-
silty clay obtained from the centre of a deep excavation near
ting form and graphical construction, are inconsistent and
Chicago. Determining which of the points to use in the lin-
ear interpretations of the early part of the plot as well as many depend strongly upon the initial loading curve (Becker
within the ‘‘virgin compression’’ zone of the test requires et al. 1987; Morin 1988). The character of the e-stress plot,
considerable judgement for those tests that exhibit ‘‘curved’’ therefore, has a profound role in the interpretation and the
plots. The same data are plotted in Fig. 7 using the conven- degree of ‘‘accuracy’’ or ambiguity that any of the methods
tional e-stress semi-logarithmic scale where the Casagrande may achieve. Such variation and influences of plotting scale
and Pacheco Silva approaches are both applied. The range are troubling due to the implications for calculations or cor-
of differences among these and the strain energy density ap- relations based on s p0 or OCR.
proach is about 28% and 32%, when compared with the In their original paper, Grozic et al. (2003) concluded that
average and minimum values, respectively, for this particu- ‘‘When interpreting the preconsolidation stress of glaciomar-
lar sample of Chicago silty clay. ine clays, the work method of Becker et al. and the bilogar-
Grozic et al. (2003) presented 187 attempts at estimating ithmic method proposed by Onitsuka should be used. In
s p0 using seven different methods. A total of 149 of these at- addition, pc0 (s p0 in this paper) should be estimated using the
tempts were comparable to values of s p0 determined using traditional Casagrande and Janbu methods as a basis of com-
the Casagrande (1936) method. Given that, for the 33 oed- parison because of the broad experience using these meth-
ometer tests presented by Grozic et al. (2003), a true value ods.’’ Given the level of uncertainty that exists when using
of s p0 remained unknown, the attempts at defining s p0 are these methods, particularly when examining tests of natural
compared as part of this paper to the Casagrande approach, soils, and the difficulty with their use that Grozic et al.
where possible, because of its widespread use. In addition, (2003, 2005) and Clementino (2005) so clearly identified,
the estimated values of s p0 were compared with the average this recommendation may prove frustrating in practice and
of all the attempts for each individual test, simply as meas- lead to no better results. Furthermore, when examining sig-

Published by NRC Research Press


Boone 285

Fig. 4. Comparison of interpreted data from Grozic et al. (2003) showing differences in OCR for individual samples based on use of the
methods of Casagrande (1936), Janbu (1969), Becker et al. (1987), Burland (1990), Jacobsen (1992), and Onitsuka et al. (1995).
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIV WINDSOR on 11/11/14

Fig. 5. Interpretation of s p0 using (a) Pacheco Silva (1970) and (b) Casagrande (1936) methods, and (c) illustration of the effects of scale
when using these methods (after Clementino 2005).
For personal use only.

0
nificant numbers of tests for large projects, such an exercise minimum stress and s v0 is of little importance unless un-
can be impractical. loading and swelling behaviour is critical to the engineer-
ing problem being evaluated.
Proposed interpretation principles and  The initial part of the test curve represents only recom-
method pression of the swelling and disturbance associated with
sampling, specimen preparation, saturation, and test set-
Because of the aforementioned difficulties with graphical up.
approaches to oedometer data interpretation, a new method  The use of e0 derived from the start of the oedometer
is proposed that relies on a few logical steps, has some di- test, at which point the soil sample is entirely unloaded,
rect physical meaning and can be readily incorporated into within the settlement calculations or interpretation of the
spreadsheet calculations or manually calculated, thus pro- test is dubious unless e0 in eqs. [1a] to [1c] is interpreted
ducing an unambiguous single value result. This approach based on in situ water content for saturated soils rather
is based on the following fundamental principles: than e0 derived from the oedometer test.
 Deformation behaviour in the oedometer test between the  Once the applied effective stress is equal to the in situ

Published by NRC Research Press


286 Can. Geotech. J. Vol. 47, 2010

Fig. 6. Application of strain energy density method to the interpretation of s p0 from a block sample of Chicago clay (curve e in Fig. 2). An
alternative interpretation of the early part of the curve completed for this paper is illustrated by dotted lines.
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIV WINDSOR on 11/11/14
For personal use only.

Fig. 7. Application of the Casagrande (1936) and Pacheco Silva Fig. 8. Comparison of interpreted values of s p0 using (a) the com-
(1970) methods to the interpretation of s p0 from the block sample of 0
mon Casagrande method, s pðCasagrandeÞ , and (b) the average of all
0
Chicago clay for this paper (curve e in Fig. 2, and same data as methods identified in Fig. 4, s pðaverageÞ .
used for Fig. 6). wn, natural water content; LL, liquid limit; PL,
plastic limit; Su (field vane), undrained shear strength from field vane;
Su (cpt), undrained shear strength from cone penetration test.

tive value of Cc, accounting for prior recompression of


sampling disturbance — note that for highly sensitive
soils (e.g., Leda clay), the value of Cc will decrease
within the higher stress ranges (concave upward curve),
reflecting a greater degree of particle interaction and
greater resistance to structure compression.
Based on these principles, the value of s p0 can be esti-
vertical effective stress, the voids ratio should be repre- mated by the following steps, also illustrated in Fig. 9:
sentative of the in situ voids ratio, ev0, excluding the po- (1) If the oedometer test does not include a load increment
tential for significant sample disturbance. equal to s v00
, to allow direct determination of the value
 The curvature evident within the plot of the e-stress data of the voids ratio at this stress, ev0, determine the slope
0
at stresses above s v0 is the combined result of natural re- of the line between load increments that passes through
0
compression between s v0 and s p0 and sample disturbance. 0
s v0 , noted as Ccv0, so that the value of ev0 can be mathe-
 The maximum value of Cc should be the most representa- matically interpolated (see insets, Fig. 9).

Published by NRC Research Press


Boone 287

Fig. 9. Example application to determine s p0 using proposed method on (a) Leda clay from Ottawa, Ont. and (b) Chicago clay also shown in
curve e, Fig. 2, with expanded e scale for illustration purposes.
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIV WINDSOR on 11/11/14

(2) Determine the load increment at which Cc is a maxi- lanandan et al. 2008). Even with small assumed values of
mum, Ccmax, where the ordinates of the corresponding Cr, on the order of 0.05Cc to 0.1Cc (e.g., Leonards 1976;
For personal use only.

0
ending stress and voids ratio are thus noted s vmax and Holtz and Kovacs 1981), this method provides reasonable
emin, respectively. and unambiguous results.
(3) Determine the recompression index, Cr, defined as the As a check on the mathematical determination of s p0 , the
average slope of an unload–reload cycle (e.g., Leonards fundamental principals of this method may also be applied
1976; Vipulanandan et al. 2008), preferably conducted graphically. In this case, a line is drawn coincident with
0
at stresses above and below s v0 and less than s p0 Ccmax, another is drawn parallel to Cr, passing through the
(Fig. 9a). 0
point (s v0 , ev0), and the intersection of these lines defines
(4) Calculate the intercept values along the voids ratio axis 0
s p . One of the advantages of the mathematical approach is
for the compression and recompression index lines (ec that graphical interpolation of s p0 on a logarithmic scale is
and er, respectively) using
avoided.
0
½2 ec ¼ Ccmax logs vmax þ emin
Examination of proposed method
0
½3 er ¼ Cr logs v0 þ ev0
Data from more than 160 oedometer tests were used to
(5) Calculate the intersection of these two lines, the first de- examine the suitability of this approach for general applica-
fined by the Ccmax line and the second defined by a line tion. Many of these tests were completed on low-plasticity
parallel to the Cr line passing through the in situ vertical silty clay from the Great Lakes area of Ontario, Canada.
effective stress, solving for the voids ratio at the precon- However, test results from quick Leda clay, sensitive clays
solidation pressure, ep, by from the New Liskeard and Sault Ste. Marie areas of On-
tario, glacial clays from the Chicago and Detroit areas, and
½4 ep ¼ ðec =Ccmax  er =Cr Þ=ð1=Ccmax  1=Cr Þ the results reported in Becker et al. (1987) and Grozic et al.
(6) The ‘‘preconsolidation pressure’’ may then be found by (2003) were also used as a comparison. The range of pre-
consolidation pressures examined fell between 30 and over
½5 s p0 ¼ 10½ðec ep Þ=Ccmax  600 kPa, and the calculated OCR varied between about 1
and more than 10 (using the proposed method). The plasti-
This approach will tend to result in lower values of s p0 if city index, PI, for the samples ranged between about 4 and
the sample is disturbed, as with other methods of s p0 inter- 60. Where recompression index values were not available,
pretation, as a result of the effect of disturbance on the over- approximations were made using Cr = 0.1Cc and Cr =
all test data (e.g., Soderman and Kim 1970; Brumund et al. 0.05Cc for the low-plasticity and sensitive clays, respectively
1976). Use of a recompression index derived from unload– (e.g., Leonards 1976; Holtz and Kovacs 1981).
reload cycles at high stresses will tend to produce larger val- Examples of this evaluation applied to published oedome-
ues of s p0 as a result of larger Cr values arising from greater ter test curves are illustrated in Figs. 10 and 11 and are sum-
potential changes to soil structure (particularly for sensitive marized in Table 2, where the s p0 values interpreted using
clays) at stresses greater than s p0 (e.g., Leonards 1976; Vipu- the Casagrande (1936) (s pðCÞ0 0
), Becker et al. (1987) (s pðBÞ ),

Published by NRC Research Press


288 Can. Geotech. J. Vol. 47, 2010

Fig. 10. Comparison of interpreted values of s p0 for (a and b) Wallaceburg clay and (c) Beaufort Sea silty clay from Becker et al. (1987)
using proposed method and showing range of s p0 estimated using Casagrande (1936) method and using strain energy method as reported by
Becker et al. (1987), with additional Casagrande (1936) and Pacheco Silva (1970) method estimates completed for this paper also shown.
0 0
s pðC this paperÞ , single estimate of s pðCÞ completed for this paper.
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIV WINDSOR on 11/11/14

Fig. 11. Comparison of interpreted values of s p0 for (a) Port Allen clay from Kaufman and Sherman (1964) and (b) New Liskeard clay from
Quigley and Ogunbadejo (1980) using proposed method and showing reported s p0 using Casagrande (1936) method, with additional Pacheco
For personal use only.

Silva (1970) and strain energy method estimates completed for this paper also shown.

Table 2. Summary of estimated preconsolidation pressure Table 3. Summary of undrained shear strength and preconsolida-
values, Figs. 10 and 11. tion pressure ratio comparisons for 107 samples with available su
data.
0 0 0 0
s pðCÞ s pðPSÞ s pðBÞ s pðPÞ
Test (kPa) (kPa) (kPa) (kPa) 0 0 0 0
su =s pðCÞ su =s pðPSÞ su =s pðBÞ su =s pðPÞ
Wallaceburg (a) 115 94 120 104
Mean 0.29 0.32 0.23 0.26
Wallaceburg (b) 150 173 145 156
Standard deviation 0.13 0.17 0.09 0.09
Beaufort Sea 200 298 220 272
Port Allen 250 300 300 283 Note: Undrained shear strength determined using either Nilcon push-in
New Liskeard 299 336 338 322 field vane shear, borehole vane shear or adjacent cone penetration test ca-
librated to nearby on-site field vane shear testing.
0 0
Note: Interpretation methods s pðCÞ , Casagrande (1936); s pðPSÞ ,
0 0
Pacheco Silva (1970); s pðBÞ , Becker et al. (1987); s pðPÞ , proposed.

Published by NRC Research Press


Boone 289

Fig. 12. Comparison of s p0 interpreted using proposed method and methods of (a) Casagrande (1936), (b) Becker et al. (1987), and
(c) Pacheco Silva (1970). r2, square of correlation coefficient.
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIV WINDSOR on 11/11/14

0 0 0 0 0 0
Fig. 13. Relationship between ðs pmax  s pmin Þ=s pavg and natural water content of 169 samples. s pmax , s pmin , and s pavg are the maximum,
minimum, and average values of s p0 , respectively.
For personal use only.

0 0
Pacheco Silva (1970) (s pðPSÞ ), and proposed method (s pðPÞ ) by Mesri (1975, 1989) be considered applicable and rela-
are identified. In each case where the authors provided val- tively constant for these soil types.
ues for s p0 based on any of the interpretation methods, these Figure 12 compares the proposed interpretation method
values were used. Because none of the natural samples had a with those of Casagrande (1936), Pacheco Silva (1970), and
‘‘known’’ preconsolidation pressure to which all interpreta- Becker et al. (1987) for 169 tests. This figure illustrates that,
tion methods could be compared, the data were compared among those methods examined, the s p0 values determined
with measurements of undrained shear strength, su, using using the proposed approach are more closely correlated
the principles outlined by Mesri (1975, 1989). The measure- with those determined via the strain energy density method,
ments of undrained shear strength were obtained using a va- which is a more rigorous theoretical approach than either of
riety of methods including the Nilcon push-in field vane the two other graphical methods.
device, field vane shear tests conducted within boreholes, Figure 13 illustrates a comparison of the range between
and cone penetration tests calibrated to field vane shear the maximum and minimum values of s p0 derived for each
tests. The results of this comparison are summarized in Ta- test using the four methods of interpretation identified in Ta-
ble 3. This table illustrates that the Casagrande and Pacheco ble 1, normalized by the average s p0 , and the natural water
Silva approaches produce average and standard deviation content of the samples. A frequency diagram illustrating the
su =s p0 values significantly greater than when using either the number of samples and their respective water contents is
strain energy density or proposed methods, indicating a also provided in Fig. 13. This figure generally suggests that
greater degree of variability should the principles suggested the interpreted s p0 values from all methods tend to converge

Published by NRC Research Press


290 Can. Geotech. J. Vol. 47, 2010

for the higher water content (or voids ratio) samples be- Fig. 14. Comparison of s p0 interpreted using proposed method and
cause, certainly for the test results used as part of this study, methods of Casagrande (1936), Pacheco Silva (1970), and Becker
these soils also exhibited clearly defined ‘‘breaks’’ between et al. (1987) (shown as strain energy density) for samples with wn >
the recompression and virgin compression portions of the 45%.
test curves. This trend, as well as consideration of scale ef-
fects (as in Figs. 2, 3, and 5), and the observed reverse cur-
vature exhibited by tests of sensitive soils (see curves a to d
in Fig. 2) at high stress levels after ‘‘collapse’’ of an open
fabric suggest that the shape of oedometer test curves for
soils of lower initial voids ratio (generally less than 1) and
potentially higher OCR may be the result of the mechanical
characteristics of particle-to-particle contact within the soil
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIV WINDSOR on 11/11/14

fabric rather than disturbance. As this study utilized a


greater number of tests for soils at natural water contents
ranging between about 20% and 35%, a separate comparison
was also made between the various methods of interpreta-
tion only for those tests at higher natural water contents that
exhibited ‘‘clear breaks’’ in the plotted test curves. This
comparison of 43 test results is illustrated in Fig. 14. Least
squares regressions for each of the methods in comparison
with the proposed method all indicate very close correlation,
thus supporting the conclusion that s p0 interpreted from all
methods converge for such soils. the values of s p0 estimated using the three different approaches
As another measure of the validity of this approach, the (Casagrande (1936), strain energy density (Becker et al.
preconsolidation pressures defined by this method have 1987), and proposed method) differed by less than about 12%
been used in back-analysis comparison with four cases of (see Fig. 16 and Table 4), indicating that using any of these
For personal use only.

embankment settlement. One case is located in the area of methods for interpretation of s p0 could have produced similar
Sault St. Marie, Ontario, and the other three are located in settlement estimates. Settlements estimated using the different
southwestern Ontario. Case 4, of the Tilbury Creek Bridge methods of estimating s p0 are summarized in Table 5.
in Tilbury, Ontario, was reported in some detail by Soder-
man and Kim (1970). In all back-analysis comparisons, ap- A study was recently carried out in southwestern Ontario
plied stresses were computed based on two primary to calibrate foundation settlement estimates, field and labo-
assumptions, common in such work in practice. First, stress ratory data, and measured long-term settlements for bridges
distributions from embankment and foundation loads were and embankments constructed over soft to stiff cohesive
estimated using conventional Boussinesq (1885) methods. soils (MTO 2006). Historically, settlement estimates have
Second, primary consolidation settlements were calculated not compared well with field results (e.g., Soderman and
using conventional one-dimensional consolidation eqs. [1a] Kim 1970) for projects in the region and, with the expansion
to [1c]. For each of the sites, correlations between natural of highways, calibration of settlement estimates with field
water content and the compression indices for similar soils performance was seen to potentially improve decisions re-
were used to supplement oedometer test data to provide a lated to foundation choices and the magnitudes of differen-
more detailed interpreted profile of compressibility. tial settlement between new and older structures that could
Case 1 consisted of a 6.5 m high embankment (including be expected and tolerated. Case 2 consists of a multi-span
surcharge loads) with 228 side slopes built in Sault Ste. highway bridge over a railway corridor for which the em-
Marie, Ontario, placed over approximately 8 m of silt and bankments (26.58 side slopes) and abutment foundations,
sand, and 25 m of soft and sensitive silty clay in which pre- each measuring about 27 m by 3 m, were built over low-
fabricated vertical drains were installed to accelerate consol- plasticity cohesive soils, commonly considered to be fine-
idation. In this case, settlement magnitudes were initially grained glacial till-like mud deposited in a lacustrine envi-
estimated during design using a s p0 profile based on oedom- ronment (e.g., Soderman and Kim 1970). Initially, a 3 m
eter test results, su =s p0 ¼ 0:22 based on correlating cone pen- high embankment was constructed leading to the edge of
etration test (CPT) results with push-in field vane shear tests the bridge foundations, with the subsequent height of the
and oedometer results interpreted using the Casagrande 9.8 m embankment constructed after the integrated bridge
(1936) and Becker et al. (1987) methods, and Cc and Cr val- abutment and retaining wall construction was finished. Case
ues based on oedometer tests and site-specific correlations 3 consists of a single rigid-frame highway bridge over a lo-
between Cc, Cr, and natural water content (as shown in cal road, supported by shallow spread abutment foundations,
Fig. 15). In this case, the maximum centreline settlement at measuring about 36 m by 1.8 m, with 7.3 and 7.8 m high
about 800 days after completion of staged embankment con- approach embankments (26.58 side slopes). As with case 2,
struction (end of primary consolidation) estimated using a 3 m high embankment was built to the edge of the existing
conventional settlement calculation methods was about roadway before building the bridge and remaining embank-
460 mm. Measured settlement at this time was about ment. This bridge was constructed at the same general site
425 mm. The oedometer tests completed for case 1 exhib- as case 2. A profile of the subsurface conditions for these
ited clear ‘‘breaks’’ in the data plots shown in Fig. 16, and two cases, along with site-specific correlations of consolida-

Published by NRC Research Press


Boone 291

Fig. 15. Profile of subsurface conditions, settlement estimation parameters, and stress conditions for case 1.
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIV WINDSOR on 11/11/14
For personal use only.

Fig. 16. Four example oedometer test curves for case 1 with un- Table 4. Summary of estimated preconsolidation pres-
loading curves shown by dashed lines. sure values, case 1.
0 0 0 0
s pðCÞ s pðPSÞ s pðBÞ s pðPÞ
Test (kPa) (kPa) (kPa) (kPa)
BH03-02-10 182 174 188 167
BH03-02-18 224 230 225 224
E-2-10 153 113 149 127
E-2-15 227 221 223 216

grande (1936) method and measured settlements are summar-


ized in Table 5. As part of a follow-up study, a CPT was
completed in a ‘‘green field’’ area adjacent to the embank-
ment areas, but outside of the area that may have been influ-
enced by the original construction in the mid-1960s. Based
on a calibration of cone penetration testing with field vane
shear testing in other nearby areas, an estimated profile of su
and subsequently, by assuming su =s p0 ¼ 0:22, a profile of s p0
was developed for this location. The subsequent settlement
estimates using this profile are also summarized in Table 5
and s p0 values derived from the oedometer tests using differ-
ent methods for these cases are summarized in Table 6.
Case 4, originally reported by the Department of High-
ways Ontario (1959, 1960) and later re-examined by Soder-
man and Kim (1970), consists of two 15 m long simply
tion parameters, is illustrated in Fig. 17 and example oed- supported single-span highway bridges spanning over a
ometer tests are illustrated in Fig. 18. Estimated long-term creek. The abutments are supported on shallow spread foun-
settlements of the embankments and overpass structures dations, each measuring about 11 m by 4.9 m, with a 4 m
based on interpretation of oedometer tests using the Casa- approach fill behind each abutment. A profile of the subsur-

Published by NRC Research Press


292 Can. Geotech. J. Vol. 47, 2010

Table 5. Summary of back-analysis comparisons of settlement estimates.

Settlement (mm)
Embankment Casagrande
Case Structure type height (m) (1936) Proposed Measured
1 Embankment 6.5 453 465 425
2 (east) Rigid frame 10.0 202 90 88
2 (west) Rigid frame 10.2 228 99 125
3 (east) Simple span 7.3 282 120 121
3 (west) Simple span 8.5 283 120 136
4 Simple span 4.0 356* 63 total{ 12 (abutment)
*As reported by Soderman and Kim (1970).
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIV WINDSOR on 11/11/14

{
46 first stage, 17 abutment only

Fig. 17. Profile of subsurface conditions, settlement estimation parameters, and stress conditions for cases 2 and 3.
For personal use only.

face conditions is provided in Fig. 19 along with the corre-  In situ preconsolidation pressures were likely higher than
sponding oedometer test results in Fig. 20. Consistent with indicated by conventional interpretation of the oedometer
construction practice at the time, the embankment was parti- tests and, geologically, were the result of lower past
ally constructed through the bridge area to provide access, groundwater levels.
the embankment at and between the bridge abutments was  The Casagrande construction was ‘‘very difficult to use’’
removed, the bridge and remaining embankment was con- for these soils.
structed, and the creek was diverted into a new channel be-  The parameters cv (coefficient of consolidation) and Cc
tween the abutments. Settlement of the bridge foundations can be ‘‘grossly underestimated’’ due to sample distur-
was measured once the abutments were constructed and, bance as it was interpreted that the lack of a ‘‘well-
therefore, settlements of the embankments constructed prior defined break’’ in the consolidation test plots was indica-
to bridge construction remain unknown. tive of disturbance (note that the original estimate applied
Comparing the large differences between the original set- the Schmertmann (1955) approach to deriving an esti-
tlement predictions and field measurements, Soderman and mated ‘‘field’’ compression index, see Figs. 20 and 21).
Kim (1970) derived a number of conclusions from their study  The most probable factor in the overestimation of settle-

Published by NRC Research Press


Boone 293

Fig. 18. Two oedometer test curves for cases 2 and 3 with unload- consistent with the other cases showing reasonably close
ing curves shown by dashed lines. agreement between the estimated and observed settlements.

Conclusions
The ‘‘preconsolidation pressure’’ has been proven to be a
useful characteristic yield stress for defining a mathemati-
cally convenient bilinear approximation to the nonlinear
stress–strain behaviour of soils undergoing one-dimensional
consolidation within the standard oedometer test. The advan-
tages of the proposed method of interpretation of this char-
acteristic stress, when compared with others, are that it
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIV WINDSOR on 11/11/14

 is based on simple slope–intercept mathematics


 can be directly adapted to manual or spreadsheet calcula-
tions, and thus produces a single value result
 does not require subjective and graphical interpretations
or ‘‘best-fit’’ approaches to nonlinear data
 is independent of plotting scale (maxima, minima)
 correlates well within the range of results from other ap-
proaches when applied to very soft and sensitive soils
 is directly linked to the in situ stress and voids ratio state
 correlates well with su =s p0 concepts
 correlates well with estimates and measurement of settle-
Table 6. Summary of estimated preconsolidation pres-
ment for field cases
sure values, cases 2 and 3. This method, as with all others, remains subject to the ef-
fects of poor quality sampling and testing and consequent
For personal use only.

0
s pðCÞ 0
s pðPSÞ 0
s pðBÞ 0
s pðPÞ disturbance, as well as issues associated with the rate of
Test (kPa) (kPa) (kPa) (kPa) strain and choice of using either end of increment time or
BH101-7 250 279 333 316 end of primary consolidation for the selection of stress and
BH101-13 250 254 387 298 voids ratio data points. When examining oedometer test re-
sults and completing settlement calculations using oedome-
ter data, it is further concluded that
ments was an underestimation of the preconsolidation
pressure.  the shape of more ‘‘rounded’’ test data curves, when
plotted on semi-logarithmic scales, may be representative
Their study considered that sample disturbance was a sig- of the character of the soils, rather than solely the result
nificant factor in the derivation of preconsolidation pressures of disturbance
for settlement estimates. While the study provided some  for soils of a particular mineralogical composition,
useful explanations for the differences between estimated particle-size distribution, and depositional environment,
and observed settlements, the use of higher Cc values based overconsolidation can produce more ‘‘rounded’’ test
on the Schmertmann (1955) construction coupled with curves than these same soils when normally consolidated
higher preconsolidation pressures may have had compensat-  soils with in situ voids ratios less than about 1 will sel-
ing effects, potentially obscuring the underlying factors in dom produce oedometer test curves that exhibit clearly
the settlement estimate discrepancies. Tables 5 and 7 sum- defined ‘‘breaks,’’ as the soil fabric has less potential for
marize settlement estimates and estimated s p0 values for this significant particle rearrangement
case, respectively.  e0, derived from oedometer tests, should not be used in
Based on a review of over 100 oedometer tests completed conventional consolidation settlement equations as this
on soils from the southwestern Ontario region, it was con- represents only the greatest value of e following unload-
sidered that the ‘‘rounded’’ curves were representative of ing and swelling, and a more appropriate value on which
the soil’s composition and mechanical behaviour, rather to base such settlement calculations is ev0 calculated di-
than directly the result of sample disturbance. Given that rectly from water content determinations for saturated
the soils are plastic and of a relatively higher density than soils
sensitive or quick Leda clays from other regions of Ontario,  the recompression index, Cr, should be based on an
these soils may be less subject to the disturbing effects of unload–reload cycle that bounds the in situ vertical effec-
sampling and testing than quick clays. Therefore, the param- 0
tive stress, s v0 , to provide a reasonable approximation of
eters illustrated in Figs. 19 and 20 were based directly on the soil behaviour when subject to additional loads be-
0
the proposed approach for preconsolidation pressure deter- tween s v0 and s p0 , and the use of unloading–reloading cy-
0
mination and the Cc values derived from the oedometer tests cles above s v0 (or swelling index, Cs) may result in
and correlated Cr values. Based on these profiles, settle- overestimating s p0 and any settlement calculated within
ments were estimated for the bridge abutments and these re- the recompression range.
sults are also provided in Table 5. These results are Judgement of the character of oedometer test results and

Published by NRC Research Press


294 Can. Geotech. J. Vol. 47, 2010

Fig. 19. Profile of subsurface conditions, settlement estimation parameters, and stress conditions for case 4 showing s p0 profiles estimated at
the time of design and by Soderman and Kim (1970) with consideration of alternative geologic characterization of stress history.
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIV WINDSOR on 11/11/14
For personal use only.

Fig. 20. Twelve oedometer test curves for case 4 with unloading Fig. 21. Two oedometer test curves for case 4 showing unloading
curves shown by dashed lines. See Fig. 21 for interpretation of two curves and estimated field consolidation curve (after Soderman and
example curves. Kim 1970).

subsequent interpretation is critical when using the proposed influenced by anomalies in the test data, e.g., significant
method of s p0 estimation, as with all other methods. Indis- changes in Cc resulting from data permutations between
criminate application of this mathematical method may pro- very small increments in a controlled-rate-of-strain test.
duce inappropriate results if derivation of Ccmax or Cr are Comparison of the mathematical results to a graphical

Published by NRC Research Press


Boone 295

Table 7. Summary of estimated preconsolidation Brumund, W.F., Jonas, E., and Ladd, C.C. 1976. Estimating in situ
pressure values, case 4. maximum past preconsolidation pressure of saturated clays from
0 0 0 0
results of laboratory consolidometer tests. In Special Report 163.
s pðCÞ s pðPSÞ s pðBÞ s pðPÞ Transportation Research Board, Washington, D.C. pp. 4–12.
Test (kPa) (kPa) (kPa) (kPa) Burland, J.B. 1990. On the compressibility and shear strength of
BH2B-569 640 375 640 302 natural clays. Géotechnique, 40(3): 329–378.
BH2B-563 250 337 480 269 Butterfield, R. 1979. A natural compression law for soils (an ad-
BH2B-559 250 217 390 213 vance on e-logp’). Géotechnique, 24(4): 469–479.
BH2B-551 113 165 270 183 Casagrande, A. 1936. The determination of the preconsolidation
BH3B-550 158 200 240 179 load and its practical significance. In Proceedings of the First In-
BH2B-543 137 128 150 163 ternational Conference on Soil Mechanics and Foundation Engi-
BH2B-533 162 270 230 263 neering, Cambridge, Mass., 22–26 June 1936. Harvard Printing
BH2B-515 223 183 240 272 Office, Cambridge, Mass. Vol. 3, pp. 60–64.
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIV WINDSOR on 11/11/14

BH2B-507 251 202 300 299 Clementino, R.V. 2005. Discussion: An oedometer test study on the
BH2B-498 197 192 240 354 preconsolidation stress of glaciomarine clays. Canadian Geo-
technical Journal, 42(3): 972–974. doi:10.1139/t05-010.
Department of Highways Ontario. 1959. Report on foundation con-
check, s p0 values interpreted through other testing ap- ditions at the Tilbury Creek Crossing, Highway 401. Prepared
proaches, an understanding of the geologic history, and con- by H.G. Acres Co. Ltd. Department of Highways Ontario, Tor-
sideration of other mineralogical or mechanical (sampling, onto, Ont.
testing, etc.) factors should be part of any s p0 interpretation Department of Highways Ontario 1960. Report on settlement study
of the Tilbury Creek Bridge. Prepared by H.G. Acres Co. Ltd.
process. Ontario Ministry of Transportation, Toronto, Ont.
As noted by Becker et al. (1988) ‘‘. . . it is not a question of Grozic, J.L.H., Lunne, T., and Pande, S. 2003. An oedometer test
which technique is correct; rather the issue is which techni- study on the preconsolidation stress of glaciomarine clays. Cana-
que provides the most repeatable result and is least ambigu- dian Geotechnical Journal, 40(5): 857–872. doi:10.1139/t03-043.
ous.’’ It is anticipated that the proposed method may prove a Grozic, J.L.H., Lunne, T., and Pande, S. 2005. Reply to the discus-
useful tool for obtaining less subjective estimates of s p0 and,
For personal use only.

sion by Clementino on ‘‘An oedometer test study on the precon-


consequently, better future assessments of the relationships solidation stress of glaciomarine clays’’. Canadian Geotechnical
between s p0 and su (as well as other geotechnical engineering Journal, 42(3): 975–976. doi:10.1139/t05-011.
Holtz, R.D., and Kovacs, W.M. 1981. An introduction to geotech-
parameters), and better calibration of settlement studies.
nical engineering. Prentice Hall, N.J. pp. 323–342.
Acknowledgements Jacobsen, H.M. 1992. Bestemmelse af forbelastningstryk i labora-
toriet. In Proceedings of Nordiske Geotechnikermonde NGM-
The author thanks P.R. Bedell, Dr. D.E. Becker, and 92, May 1992. Aalborg, Denmark. Danish Geotechnical Society
M. Devata for their constructive comments and discussions Bulletin No. 9. Vol. 2, pp. 455–460. [In Danish.]
during preparation of this paper; Dr. R. Finno of Northwest- Janbu, N. 1969. The resistance concept applied to deformation of
ern University for providing high quality data from the Chi- soils. In Proceedings of the 7th International Conference on Soil
cago area; staff at the Ontario Ministry of Transportation for Mechanics and Foundation Engineering, Mexico City, 25–
providing the historical studies; and to the Ontario Ministry 29 August 1969. A.A. Balkema, Rotterdam, the Netherlands.
of Transportation for permission to use the results of studies Vol. 1, pp. 191–196.
carried out on their behalf. Kaufman, R.I., and Sherman, W.C., Jr. 1964. Engineering measure-
ments for the Port Allen lock. Journal of the Soil Mechanics and
References Foundations Division, ASCE, 90(5): 221–247.
Ladd, C.C., and Foott, R. 1974. New design procedure for stability
Andresen, A., Berre, T., Kleven, A., and Lunne, T. 1979. Proce- of soft clays. Journal of the Geotechnical Engineering Division,
dures used to obtain soil parameters for foundation engineering ASCE, 100(7): 763–786.
in the North Sea. Marine Geotechnology, 3(3): 201–266. doi:10. Leonards, G.A. 1976. Estimating consolidation settlements of shal-
1080/10641197909379804. low foundations on overconsolidated clays. In Special Report
ASTM. 2004. Standard test methods for one-dimensional consoli- 163. Transportation Research Board, Washington, D.C. pp. 13–
dation properties of soils using incremental loading. ASTM stan- 16.
dard D2435-04. American Society for Testing and Materials Mayne, P.W. 1980. Cam-clay predictions of undrained strength.
(ASTM), West Conshohocken, Pa. Journal of the Geotechnical Engineering Division, ASCE,
Becker, D.B., Crooks, J.H.A., Been, K., and Jefferies, M.G. 1987. 106(11): 1219–1242.
Work as a criterion for determining in situ and yield stresses in Mesri, G. 1975. Discussion of New design procedure for stability
clays. Canadian Geotechnical Journal, 24(4): 549–564. doi:10. of soft clays. Journal of the Geotechnical Engineering Division,
1139/t87-070. ASCE, 101(4): 409–412.
Becker, D.B., Crooks, J.H.A., Been, K., and Jefferies, M.G. 1988. Mesri, G. 1989. A reevaluation of su(mob) = 0.22sp’ using laboratory
Work as a criterion for determining in situ and yield stresses in shear tests. Canadian Geotechnical Journal, 26(1): 162–164.
clays: Reply. Canadian Geotechnical Journal, 25(4): 848–850. doi:10.1139/t89-017.
doi:10.1139/t88-097. Morin, P. 1988. Work as a criterion for determining in situ and
Boussinesq, J. 1885. Application des potentiels à l’étude de l’équi- yield stresses in clays: Discussion. Canadian Geotechnical Jour-
libre et du mouvement des solides élastiques. Gauthier-Villars, nal, 25(4): 845–847. doi:10.1139/t88-096.
Paris. MTO 2006. Structure settlement study, Highway 401 Reconstruc-

Published by NRC Research Press


296 Can. Geotech. J. Vol. 47, 2010

tion, Southwestern Region, Geocres 40J2–79. Prepared by oedometer consolidation of a soft varved clay. Canadian Geo-
Golder Associates Ltd. Ontario Ministry of Transportation technical Journal, 9(2): 165–175. doi:10.1139/t72-017.
(MTO), Toronto, Ont. Report GWP 64–00–00. Quigley, R.M., Haynes, J.E., Bohdanowicz, A., and Gwyn, Q.H.J.
Oikawa, H. 1987. Compression curve of soft soils. Journal of the 1981. Geology, geotechnique, mineralogy and geochemistry,
Japanese Geotechnical Society, Soils and Foundations, 27(3): Leda clay from deep boreholes, Hawkesbury, Ontario. Ontario
99–104. Geological Survey, Toronto, Ont. Open File Report 5357.
Onitsuka, K., Hong, Z., Hara, Y., and Shigeki, Y. 1995. Interpreta- Schmertmann, J.M. 1955. The undisturbed consolidation of clay.
tion of oedometer test data for natural clays. Journal of the Japa- Transactions of the American Society of Civil Engineers, 120:
nese Geotechnical Society, Soils and Foundations, 35(3): 61–70. 1201–1233.
Pacheco Silva, F. 1970. A new graphical construction for determi- Soderman, L.G., and Kim, Y.D. 1970. Effect of groundwater levels
nation of the preconsolidation stress of a soil sample. In Pro- on stress history of the St. Clair till deposit. Canadian Geotech-
ceedings of the 4th Brazilian Conference on Soil Mechanics nical Journal, 7(2): 173–193. doi:10.1139/t70-022.
and Foundation Engineering, Rio de Janeiro, August 1970. Vipulanandan, C., Ahossin Guezo, Y.J., Bilgin, Ö., Yin, S., and
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by UNIV WINDSOR on 11/11/14

Vol. 2, No. 1, pp. 225–232 [in Clementino 2005]. Khan, M. 2008. Recompression index (Cr) for overconsolidated
Pinto, C.S. 1992. First Pacheco Silva Conference — Part I: Distin- soft clay soils. In Proceedings of GeoCongress 2008, New Or-
guished aspect of Pacheco Silva’s activities. Brazilian Geotech- leans, La., 9–12 March 2008. Geotechnical Special Publication
nical Journal (Solos e Rochas), ABMS-ABGE, 15(2): 49–87 [in 179. American Society of Civil Engineers (ASCE), Reston, Va.
Clementino 2005]. pp. 68–75.
Quigley, R.M., and Ogunbadejo, T.A. 1980. Clay layer fabric and
For personal use only.

Published by NRC Research Press

You might also like