Geotechnical properties of Brazilian marine clays

Propriétés géotechniques des argiles marines brésiliennes

J. M. C. Barros & R. M. S. Silveira

Institute for Technological Research of the State of São Paulo - IPT, São Paulo, Brazil
C. S. Amaral
Petrobras – Petróleo Brasileiro S.A., Rio de Janeiro, Brazil

ABSTRACT
This paper presents the main results obtained from an extensive laboratory investigation on Brazilian marine clays. Samples were
collected from different depths in eight petroleum fields (Marlim, Marlim Leste, Marlim Sul, Espadarte, Jabuti, Jubarte, Cachalote
and Roncador) located in Campos basin and one petroleum field (Piranema) located in Sergipe - Alagoas basin, offshore Brazil. Grain
size analyses, Atterberg limits, oedometer tests, isotropically consolidated undrained compression (CIUC) and K0-consolidated
undrained compression (CK0UC) triaxial tests, direct simple shear (DSS) and resonant column tests were carried out in samples from
56 cores. Moreover, in some of CIUC triaxial tests, maximum shear moduli (Gmax) were determined by using bender elements inserted
into the specimens. The SHANSEP technique was used in the investigation to study the undrained behavior of the marine clays. A
summary of the main characteristics of compressibility, deformability and undrained strength (su) of the soils investigated is
presented. Finally, a correlation between Gmax values obtained with bender elements and su from CIUC tests is shown in the paper.
RÉSUMÉ
Cet article présente des principaux résultats obtenus d’une large investigation de laboratoire sur des argiles marines brésiliennes. Des
échantillons ont été extraits de différentes profondeurs dans huit champs de pétrole Marlim Leste, Marlim Sul, Espadarte, Jabuti,
Jubarte, Cachalote, Roncador et Marlim) localisés dans le Bassin Campos et un champ de pétrole (Piranema) localisé à Sergipe –
Bassin Alagoas (Brésil). L’analyse granulométrique, limites d’Atterberg, essais de consolidation, essais de compression triaxiale
isotropiquement consolidés et non drainés (CIUC), essais de compression triaxiale K0 (CK0UC) et essais de cisaillement simple direct
(DSS), essais à la colonne résonnante ont été effectués sur 56 différents échantillons. En plus, dans quelques essais triaxiaux CIUC, le
module de cisaillement maximum (Gmax) a été déterminé par l’usage d’éléments fléchissants insérés dans les éprouvettes. La
technique SHANSEP a été utilisée dans l’investigation pour l’étude du comportement non drainé des argiles marines. Un résumé des
caractéristiques de compressibilité, déformabilité et résistance non drainée (su) des sols examinés est présenté. Finalement, une
corrélation entre valeurs de Gmax obtenus par des éléments fléchissants et su obtenus des essais CIUC est présentée dans cet article.
Keywords : marine Brazilian clays, undrained shear strength, soil compressibility, maximum shear modulus

1 INTRODUCTION Moreover, in some CIUC triaxial tests, maximum shear moduli

(Gmax) were determined by using bender elements inserted into
The oil industry in Brazil is in large expansion, boosted by its the specimens.
recently discovered deepwater fields. As a result, every effort
has been made to provide the necessary geotechnical knowledge
of the marine subsoil, so that the offshore structures can be
constructed and the fields can start producing. With this aim, an 2 SOIL CHARACTERIZATION
extensive laboratory investigation on the geotechnical properties
of Brazilian marine clays has been developed and part of the The range of the index parameters of the soils from the nine
results is presented in this paper. fields investigated is presented in Table 2. As it can be seen,
Samples were collected in a total of 56 cores from eight there are several types of soils with void ratios varying from
petroleum fields (Roncador, Marlim, Marlim Sul, Marlim Leste, 0.90 to 4.2, plasticity index from 5 to 70% and clay fraction
Jubarte, Cachalote, Espadarte and Jabuti) located in Campos from 5 to 80%.
basin and one petroleum field (Piranema) located in Sergipe -
Alagoas basin, offshore Brazil. The locations of the petroleum
fields are presented in Figure 1. As shown in Table 1, the water
depth in these fields varies from 600 to 2000 m. Table 1 also Table 1. Identification of the petroleum fields.
shows the number of cores and the depth sampled in each field. Field
number depth water depth
In Roncador and Marlim fields, the samples were collected with of cores sampled (m) (m)
Kullemberg corers, while in the remaining fields Jumbo Piston Marlim Leste 8 4 to12 940 to 1350
Corers were used. It is impossible not to recognize the Marlim Sul 6 12 to 19 1590 to 1970
enormous difficulty in collecting undisturbed samples in so Espadarte 3 8 to 16 1260 to 1610
Jabuti 9 2 to 5 1100 to 1250
deep seabed and that an inevitable disturbance should be Jubarte 6 18 to 20 1170 to 1440
occurred. Cachalote 2 17 to 19 1370 to 1500
Grain size analyses, Atterberg limits, oedometer tests, Piranema 7 3 to 21 1230 to 1670
isotropically consolidated undrained compression (CIUC) and Roncador 10 3 to 5 1500 to 2000
K0-consolidated undrained compression (CK0UC) triaxial tests, Marlim 5 3 to 5 600 to 1000
direct simple shear and resonant column tests were carried out.

Proceedings of the 17th International Conference on Soil Mechanics and Geotechnical Engineering 143
M. Hamza et al. (Eds.)
© 2009 IOS Press.
doi:10.3233/978-1-60750-031-5-143
144 J.M.C. Barros et al. / Geotechnical Properties of Brazilian Marine Clays

Figure 2. Position of Brazilian marine soils in the Plasticity Chart.

Figure 1. Location of petroleum fields.

Table 2. Index parameters.

Clay
γ w LL PI
Field e fraction
(kN/m3) (%) (%) (%)
(%)
14.29 30.8 0.90 35 15 24
Marlim Leste
18.65 96.1 2.81 94 62 63
13.64 54.6 1.57 58 28 43
Marlim Sul
16.50 122 3.56 107 70 70
13.01 50.2 1.37 48 19 45
Espadarte
16.86 151 4.20 97 65 67 Figure 3. Plasticity index as a function of clay fraction – all fields with
13.50 43.6 1.21 28 5 5 exception of Piranema
Jabuti
17.38 90.2 2.61 72 43 50
13.68 39.1 1,25 55 28 9
Jubarte
16.91 121 3.41 108 66 68
13.98 71.6 2.03 65 32 6
Cachalote
15.40 112 3.10 96 58 69
13.34 31.7 0.88 28 10 15
Piranema
18.37 120 3.46 91 58 80
14.21 77.7 2.09 46 23 32
Roncador
15.46 105 2.86 83 51 55
15.48 64.4 1.81 54 23 37
Marlim
18.04 71.7 1.97 68 37 43
Note: The minimum and maximum values of each parameter
are indicated in the table.

The position of the soils in the plasticity chart is shown in

Figure 2. It is well-known that the “A” line equation is PI = Figure 4. Plasticity index as a function of clay fraction – Piranema
Field.
0.73(LL-20). Nagaraj and Jayadeva (1983) obtained the
correlation PI = 0.74(LL-8) for organic clays. Almeida et al.
(2008), in a study on Rio de Janeiro soft clays, obtained for the
3 COMPRESSIBILITY AND STRESS HISTORY
majority of the soils analyzed results that were well fitted by
“A” line. For the set of 463 samples analyzed in this
A total of 165 oedometer tests were performed on the soils
investigation, the correlation obtained was:
analyzed. Table 3 shows values of compression index (Cc),
swelling index (Ce), compression ratio (CR = Cc / (1+e0)) and
PI = 0.74(LL − 17 ) (1) overconsolidation ratio (OCR).
As shown in Table 3, CR values vary from 0.16 to 0.39. It
with a coefficient of determination (R2) of 0.91. was observed that CR has the tendency to increase with the
With regard to the activity index, the soil from Piranema liquid limit (CR = 0.0035LL). The OCR values are in general
Field presents a different behavior in comparison with the data smaller than 2 but in three fields higher values are also
set. For this reason, the results are presented separately (Figures observed: Marlim Leste, Marlim Sul and Espadarte.
3 and 4). The activity index for Piranema soil is about 0.65 The obtained average value of Ce/Cc is 0.14. Marlim Leste,
(inactive) and for the remaining fields is about 0.82 (normal Piranema and Jabuti fields present the highest average values
activity). (0.16 to 0.18) whereas Roncador and Marlim fields present the
smallest ones (0.10 to 0.12).
 J.M.C. Barros et al. / Geotechnical Properties of Brazilian Marine Clays 145

Table 3. Parameters of compressibility.

Field Cc Ce CR OCR Gmax = 14.2
(7.32 − e)2 pa0.4 σ 00.6 (4)
Marlim Leste
0.34 0.06 0.16
<5 1+ e
1.2 0.15 0.37
0.53 0.08 0.20 where pa is the atmospheric pressure. Expression (4) was
Marlim Sul <5
1.20 0.20 0.36 obtained for normally consolidated alluvial clays with void
0.7 0.09 0.20 ratios varying from 1.4 to 4. As can be noticed, the expression
Espadarte <3
1.06 0.16 0.33 obtained in the present investigation leads to greater values of
Jabuti 0.66 0.12 0.23 ~1 Gmax (approximately 30%).
0.56 0.08 0.19
Jubarte <2
1.17 0.20 0.37
0.6 0.08 0.20
Cachalote ~1
1.07 0.14 0.30
0.41 0.06 0.18
Piranema <2
1.31 0.20 0.39
0.67 0.07 0.21
Roncador <2
1.26 0.17 0.34
0.62 0.05 0.22
Marlim <2
0.7 0.08 0.24
Note: The minimum and maximum values of each parameter
are indicated in the table.

Figure 5 shows the relationship between compression index and

water content. This correlation presented a better fit than the
usual correlation between Cc and LL. The correlation found is:

Cc = 0.0106 w (2) Figure 6. Comparison between expressions (3) and (4) for three
different confining stresses.
which is very similar to those presented by Koppula (1981) and
Bowles (1979). Almeida et al. (2008) and Coutinho et al. (2000)
presented correlations between Cc and w for Rio de Janeiro and 5 SHEAR STRENGTH
Recife soft clays, respectively, that result in a little higher Cc
values.
5.1 The use of SHANSEP technique

Three cores, two from Roncador field and one from Marlim
field, were selected to verify the applicability of the SHANSEP
technique (Ladd and Foott, 1974) for Brazilian marine clays.
The results are presented in Table 4. The n value in the
expression su/σ´vo=A(OCR)n varied from 0.69 to 1.02. In Table
4, the expressions are presented setting n=0.8, which is the
typical value found in the literature.

Table 4. Correlations obtained for su/σ´vo (Barros et al., 2006).

Field Cores Test Correlation
CK0UC su/σ´vo=0.356(OCR)0.8
Roncador K04A
DSS su/σ´vo=0.304(OCR)0.8
CK0UC su/σ´vo=0.320(OCR)0.8
Roncador K05
DSS su/σ´vo=0.278(OCR)0.8
CK0UC su/σ´vo=0.381(OCR)0.8
Figure 5. Compression index and water content relationship. Marlim 03
DSS su/σ´vo=0.325(OCR)0.8

4 MAXIMUM SHEAR MODULUS

5.2 Undrained shear strength for normally consolidated
A correlation was determined between Gmax, void ratio at the condition
end of primary consolidation (e) and confining stress (σ0), valid
for the normally consolidated condition. For this purpose 80 For Marlim Sul, Marlim Leste, Jubarte, Cachalote, Espadarte
tests on soils from eleven cores of Marlim Leste, Marlim Sul e and Piranema fields, CIUC, CK0UC and direct simple shear
Jubarte were considered (40 Gmax values obtained from resonant (DSS) tests were carried out with effective vertical stresses
column tests and 40 Gmax values obtained from bender higher than the preconsolidation stress, in order to evaluate the
elements). The following expression was determined: ratio su/σ’vo for the normally consolidated condition. Table 5
shows the mean values of su/σ’vo obtained in each type of test.
With regard to the su/σ’vo values from CK0UC tests, it can be
Gmax = 1.83σ 00.62 e −0.83 (3) seen that the average value, 0.311, is close to the value 0.32
indicated by Ladd (1991). However, for the DSS tests the values
where σ0 is in kPa and Gmax in MPa (see also Barros et al., obtained in this investigation (average of 0.29) are considerably
2008). The R2 coefficient of expression (2) is equal to 0.87. higher than that suggested by Ladd (0.22 to 0.27).
Figure 6 shows a comparison between expression (3) and the It is important to mention that while in the DSS test the
well-known expression of Kokusho et al. (1982): deformation rate was that recommended by Ladd and Foott
146 J.M.C. Barros et al. / Geotechnical Properties of Brazilian Marine Clays

(1974) (5% / hour), in the CIUC and CK0UC triaxial tests, the 6 CONCLUSIONS
deformation rate was 8%/hour, about ten times that indicated by
Ladd and Foott (0.5 a 1.0 % / hour). This paper presented the results of an extensive laboratory
It can be estimated that the su/σ’vo values from CIUC and investigation on the behavior of Brazilian marine clays. The
CK0UC tests obtained in this investigation should be reduced in main conclusions are:
4 to 8% (Chandler, 1988 & Bjerrum, 1973) in order to be - these soils show a large variability of void ratios (0.9 to
compared with data from literature. 4.2), plasticity indexes (5 to 70%) and clay fractions (5 to 80%).
If this correction is done, the su/σ’vo values from CK0UC and - the mean line that indicates the position of these soils in the
DSS tests become very close. These results suggest that the plasticity chart lies slightly above and is parallel to the “A” line.
investigated Brazilian marine clays present a very small degree - the activity index of these soils is about 0.82 (normal
of anisotropy. activity) with exception of Piranema field, which is
approximately 0.65 (inactive).
- the OCR values are in general smaller than 2, but in three
Table 5 - su/σ’vo values for normally consolidated condition fields there are higher values.
Field CIUC CK0UC DSS - the relationship between Cc and water content obtained is
Marlim Leste -- 0.33 (10) 0.33 (9) very similar to those found in the literature.
Marlim Sul 0.41 (14) 0.31 (9) 0.30 (16) - it was possible to establish a correlation of Gmax with
Espadarte 0.39 (12) 0.33 (5) 0.29 (4) confining pressure and void ratio for the normally consolidated
Jubarte 0.37 (41) 0.31 (16) 0.28 (15) condition.
Cachalote 0.40 (10) 0.32 (4) 0.28 (4) - the very similar su/σ’vo values from CK0UC and DSS tests
Piranema 0.34 (109) 0.30 (24) 0.29 (96)
obtained ins this investigation suggest that Brazilian marine
Note: the number in parentheses indicates the totality of tests. clays present a very small degree of anisotropy.
- it was possible to establish a correlation of Gmax with the
undrained strength obtained in CIUC test.
5.3 Relationship between Gmax and su
ACKNOWLEDGMENTS
In an attempt to determine a correlation between Gmax and su,
Gmax values were obtained from bender elements in 60 CIUC The authors wish to thank Petrobras for financial support.
tests performed in normally consolidated conditions on soils
from 19 cores of Marlim Leste, Marlim Sul, Jubarte, Cachalote
and Espadarte fields. REFERENCES
Figure 7 shows the relationship between Gmax and su. The
correlation found is: Almeida, M.S.S & Futai, M.M. 2008. Laboratory behaviour of Rio de
Janeiro soft clays – Part 1: Index and compression properties. Soils
Gmax = 573su (5) and Rocks 31, no. 2: 69-75.
Arango, I.; Moriwaki, Y. & Brown, F. 1978. In situ and laboratory shear
with a R2 coefficient equal to 0.83. velocity and modulus. Proceedings of Specialty Conference on
Several correlations between Gmax and su are presented in the Earthquake Engineering and Soil Dynamics: 198-212.
technical literature. However, it is difficult to compare the Barros, J.M.C. & Ozório, D.C.D.S. 2006. The use of Shansep technique
results here obtained with others, since different ways of in marine Brazilian clays. Proceedings of Brazilian Conference of
determining both su and Gmax are usually employed. Soil Mechanics and Geotechnical Engineering, 13. Curitiba, Brazil.
Expression (5) is similar to the one obtained by Arango et vol.1: 489-494. (In Portuguese).
al. (1978) in tests on different types of soils, varying from low Barros, J.M.C., Silveira, R.M.S. & Amaral, C.S. 2008. Comparison
compressibility silts to high compressibility clays: Gmax= between Gmáx values obtained from resonant column tests and with
813su. They determined the undrained strength, su, from bender elements. Proceedings of Brazilian Conf. of Soil Mechanics
triaxial CIUC tests, but obtained Gmax from in situ tests. The and Geotechnical Engineering, 14. Búzios, Brazil. (In Portuguese).
greater value of the ratio Gmax/su obtained by them can be Bjerrum, L. 1973. Problems of Soil Mechanics and Construction on Soft
explained by the fact that they determined Gmax in situ, what Clays. State-of-the-Art Report, Session 4, Proceedings of VIII
usually results in values greater than the ones determined in International Conference on Soil Mechanics and Foundation
laboratory. Engineering, Moscow, vol. 3.
Bowles, J.E. 1979. Physical and geotechnical properties of soils. New
York: McGraw-Hill, 410 pp.
Chandler 1988. The in situ measurement of the undrained shear strength
of clays using the field vane. STP 1014, ASTM : 13-44.
Coutinho, R.Q.; Oliveira, J.T.R. & Oliveira, A.T.J. 2000. Geotechnical
properties of Recife soft clays. Soils and Rocks 23, no. 3: 177-203.
Kokusho, T. Yoshida, Y. & Yasuyuki, E. 1982. Dynamic properties of
soft clay for wide strain range. Soils and Foundations. JSSMFE,
vol. 22, no.4:1-18.
Koppula, S.D. 1981. Statistical estimation of compression index.
Geotechnical Testing Journal. vol.4, no.2:68-73.
Ladd, C. C. & Foott, R. 1974. New design procedure for stability of soft
clays. Journal of Geotechnical Engineering Division, ASCE, no.
GT7 : 763-786.
Ladd, C. C. 1991. Stability evaluation during staged construction.
Journal of Geotechnical Engineering Division, ASCE, vol. 117, no
4 : 540-615.
Nagaraj, T.S. & Jayadeva, M.S. 1983. Critical reappraisal of plasticity
index of soil. Journal of Geotechnical Engineering Division,
Figure 7. Maximum shear modulus and undrained strength relationship. ASCE, vol. 109, no 7: 994-1000.