Geotechnical Parameters for the Variegated Soils of São Paulo Formation by Means of In Situ Tests

The purpose of this paper is to present geotechnical parameters of the variegated soils from the São Paulo Formation, aiming in particular to establish correlations between stress history parameters, earth pressure coefficients at rest, deformability modules and resistance obtained through piezocone and dilatometer tests. The performance of in situ tests has, as the major advantage over laboratory tests for obtaining soil parameters of the project site, minimization of sample disturbance effects. Using data from geotechnical investigation carried out during the basic design of the Green Line expansion for the Metrô SP it was possible to obtain preconsolidation pressure ( ’p), over consolidation ratio (OCR), earth pressure coefficients at rest (K0), Young’s module (Ei), initial shear module (G0), constrained module (M) and undrained shear strength (su) for the variegated soils of the São Paulo Formation. These soils are characterized by interlayered levels of clays and sands, constituting a characteristic with different geotechnical parameters and perched water tables. These characteristics explain some of the knowledge gaps on these soils. Nevertheless, it was possible to validate the correlations between earth pressure coefficients at rest (K0), over consolidation ratio (OCR), Young’s module (Ei), undrained shear strength (su) and preconsolidation pressure ( ’p) through field tests and the results of laboratory testing conducted by other authors.


Introduction
Nowadays, the biggest cities in the world need more investment in infrastructure, particularly in large capacity public transportation, such as trains and subways.For such projects, and their execution, there is increasingly demand for safety and efficiency.For this reason, in situ tests have proved hugely advantageous, allowing information to be gathered on site, thus increasing overall safety and efficiency, whilst simultaneously decreasing uncertainty and minimizing costs.
Tunnel construction carried out by Companhia do Metropolitano de São Paulo (Metrô -SP), the state subway company of São Paulo State (Brazil), provided the opportunity to perform a comprehensive program of in situ tests, such as the piezocone (CPTu) and the dilatometer (DMT), both of which are still not frequently used in Brazil, especially in the soils from São Paulo Sedimentary Basin.Soils from São Paulo Sedimentary Basin comprise mostly the São Paulo and Resende Formations.Whilst many studies about the São Paulo Formation have been conducted, there is still a lack of information about some properties of the soils from São Paulo Sedimentary Basin.The most comprehensive and critical synthesis of knowledge on these soil properties to date can be found in Massad (2012).This paper presents the results of important geological and geotechnical investigations of variegated soils from the São Paulo Formation carried out during the basic design of the Green Line expansion for the Metrô -SP.These results cover soil parameters, such as earth pressure coefficients at rest (K 0 ) over consolidation ratio (OCR), preconsolidation pressure (s' p ),Young's module (E i ), initial shear module (G 0 ), constrained module (M) and undrained shear strength (s u ) obtained from the piezocones and the dilatometers carried out in twelve sites, numbered from 1 to 12.In order to validate the data obtained from this investigation, a comparison is made with the results of previous studies for the São Paulo Formation.

Studied Area
The geological and geotechnical investigations were carried out during the basic design of the Green Line expansion for the Metrô -SP, located in east of São Paulo, Brazil (Fig. 1).Soils from São Paulo Sedimentary Basin, which was formed by the Brazilian Southeast Continental Rift during the Paleogene Period, are comprised mostly of the Resende and São Paulo Formations.The Resende Formation is characterized by distinct packs of sand (known as basal sands) and stiff overconsolidated clays (locally known as "taguá").
The São Paulo Formation is a depositional environment associated with fluvial meanders, comprising the variegated soils, and generally overlaid by a porous red clay.According to Massad (2012), the variegated soils are highly weathered sediments deposited in alternating layers of sand and clay, and are very heterogeneous.Their engineering properties vary widely due to the occurrence of very different types of soils, such as sands, clayey fine sands and sandy clays with silts.The clay content ranges from 20 to 80%.The activity index averages 0.65.In general, these soils are overconsolidated, but the preconsolidation pressure is not correlated with the weight of current or past overburden pressure.Massad et al. (1992) speculated that successive sedimentation cycles, associated with the drying of the soil, have affected the preconsolidation pressures through capillary tensions, which are greater for the finer the soil particles, or that there was a chemical cementation of the soil particles as a result of pedological evolution.
The Green Line extension (Line 2) crosses the Tiete River and streams of east São Paulo.Thus, there are many different hydrogeological conditions with regard to the São Paulo Basin.An example of this would be the perched water tables present in the upper soil layers.These are independent water levels that cause changes in the value of pore pressure along depth in each region and, due to the interleaving of sands and clays, suction may arise in the unsaturated parts of the subsoil.Their presence would cause the porous stone of the piezocone to lose its saturation, made with glycerin, adversely affecting the pore pressure measurements and any analysis of soil parameters that depend on it.This will be evidenced in analyzes presented along this paper.
As an example, Figs. 2-c and 3-c show the direct parameters of the CPTu and the intermediate parameters of the DMT, and the geological profiles of sites 1 and 3.The latter show perched water tables.
In both figures, the letters signify: (a) q c and q t -measured and corrected cone tip resistances (from CPTu); (b) f s -sleeve friction (from CPTu); (c) u 2 -pore pressure measured at the base of the cone (from CPTu);  For other symbols, see the appended list of symbols.

Piezocone and Dilatometer Tests -Theoretical Background
The in situ tests have the advantage of being easy to conduct and involve the soil at the construction site.These characteristics give the tests more reliability.
The piezocone (CPTu) consists of an instrumented cone provided with a porous stone that is pushed into the ground at a controlled rate (speed of 20 mm/s ± 5 mm/s), allowing us making it possible to obtain relevant soil parameters through the measurement of the tip resistance (q c ), the lateral friction (f s ) and the pore pressure (u 2 ) generated during the process.
The flat dilatometer, or DMT, is a device used to determine the soil in situ lateral pressure and soil lateral stiffness.The test involves driving into the ground a flat blade with a steel membrane that is then expanded.Measurements of the corresponding pressure and deformation are taken, permitting the determination of relevant soil parameters.

Piezocone
Empirical equations were taken from the literature in order to obtain design parameters such as earth pressure co-efficients at rest (K 0 ) , overconsolidation ratio (OCR), preconsolidation pressure (s' p ), Young's module (E i ), constrained module (M), initial shear module (G 0 ) and undrained shear strength (s u ).
Two criteria were used in selecting the empirical equations: a) to avoid those that depend on the pore pressure, due to the aforementioned effect of the perched water table; and b) to validate the results through comparisons with laboratory test findings.
The preconsolidation pressure (s' p ) was estimated through the Kulhawy & Mayne (1990) equation, given by: For cohesive soils, it was possible to correlate E i with the piezocone results, as in Eq. 2: and, as far as the initial shear module is concerned, application was made of the equation: from Watabe et al. (2004).
The undrained shear strength varies, as demonstrated in the following: The empirical parameters a and N kt were set below.The over consolidation ratio, defined through the relation between s' p and s' vo .

OCR
is relatively easy to obtain when the soil is submerged.However, for the variegated soils its determination is very difficult due to the aforementioned perched water tables and their effect on pore pressure.It therefore becomes necessary to make use of cones to measure, either directly or indirectly, the suction.This issue is currently being investigated by Giacheti (2015) in his research on the use of tip TDR (Time Domain Reflectometry) to evaluate the suction effect in the CPTu results.A method adapting this tip for use in conjunction with piezocone -measuring the suction in unsaturated soil -was presented by Esquivel & Vaz (2009).

Dilatometer
With regard to the DMT, the preconsolidation pressure values were obtained from OCR and the in situ vertical effective stress, as follows: Kamei & Iwasaki (1995) performed several tests in Japanese clay soils and proposed Eq. 7 to estimate the OCR. ) . . (7) The OCR can also be calculated from Eqs. 8 and 9, of Marchetti (1980) and Lunne et al. (1989), respectively.

OCR K D = ( .
) .05 1 56 (8) The choice of the equation that is best suited to variegated soils will be made later.
. for s (10) It also is possible to calculate K 0 from Marchetti (1980): For the initial Young's module (E i ) the following equation, from Robertson et al. (1989) and Campanella et al. (1985), was used: In this equation the conversion factor (F) was taken as equal to 5 for cohesive soils, and 2 for sandy soils.
For the initial shear modulus (G 0 ) Lunne et al. (1990) proposed the Eq. 13 for cohesive soils, that was adopted for the variegated soils: with a varying from 75 to 150.For the variegated soils a figure of 112.5 was adopted.
Finally, the determination of the undrained shear strength (s u ) was calculated by means of Eq. 14 from Marchetti (1980): ) .s (14)

Results and Analysis
Analyzing the twelve sites, it was possible to conclude that the subsoil is very heterogeneous, confirming the already mentioned extreme heterogeneity of the variegated soils, comprising alternating layers of sandy clays, and clayey sand with silt fractions.However, the horizontal stress index (K D ) revealed values higher than two, confirming the overconsolidation of the variegated soils, already mentioned in the literature.
Site 5 showed different results, potentially indicating a geological anomaly or problems in carrying out the tests.In order to avoid distortion of the data; it was disregarded.
For illustration purposes, the graphical representations with respect to depth of soil parameters obtained from the CPTu and the DMT are shown in Figs. 4 and 5, for sites 1 and 3 respectively.
The results obtained from the analysis of the sites are presented in Tables 1 and 2. The most reliable test was chosen to give each parameter, so as not to depending on the value of poropressure.In some cases, DMT and CPTu values were considered together.
From the analysis of the data presented in Table 1 it can be concluded that: • The values of the preconsolidation pressure ranged between 100 and 5000 kPa; • The OCR values ranged between 1 and about 40.The values below 1, associated with K D lower than 2, were disregarded; and • The values of the earth pressure coefficients at rest varied between 1 and 3.9.
In the same way, from the analysis of the data in Table 2 it can be concluded that: • The range of undrained shear strength values (s u ) was 15-600 kPa; • The Young's module (E i ) from CPTu data showed variation between 6 and 300 MPa and, from the DMT data, between 1 and 300 MPa;  vealed by the figures given in Table 3 is a result of the previously mentioned heterogeneity of variegated soils.
As a result, it can be concluded that for the DMT, Eq. 16 fits fairly well the variegated soils: it is close to the point of Camkometer (Abramento & Pinto, 1998).It refers to the equations of Lunne et al. (1990) for K 0 and Kamei & Iwasaki (1995) for OCR.A similar result was obtained by Massad (2012) for the "taguá" of the Resende Formation.

Young's modulus (E i ) and the undrained shear strength (s u )
To define which equations best fit with the variegated soil in relation to initial Young's modulus (E i ) and the undrained shear strength (s u ), the equations from CPTu were compared with triaxial test data (consolidated undrained) carried out on samples of variegated soil of Clovis Bevilacqua Square, during construction of the Metro Blue Line-SP and presented by Massad (1980) and Massad et al. (1992).The comparison between the results is presented in Fig. 7.
As regards the results of laboratory tests (triaxial CU) conducted on variegated soil samples, these authors reported a range of values for the ratio E i /s u between 300 and 600, with an average equal to 400 or 0.40 if one takes into consideration E i in MPa and s u in kPa.
As far as the parameters E i and s u are concerned, the following procedure was adopted: • s u values were estimated using Eq. 4, with N kt = 20, to be justified further on; and • E i values were estimated from Eq. 2, with a determined by comparison with results of laboratory tests (Clovis Bevilacqua Square).In fact, combining Eq. 2 with Eq. 4 results in: As the laboratory average of E i /s u is equal to 400, it was possible to obtain: Figure 7 shows the ratio E i /s u from the CPTu of the twelve test sites together with the laboratory range of 300 to 600, for comparison.
For dilatometer tests the parameter s u depends on s' vo (Eq.14) which, in turn, is affected by the aforementioned uncertainty in the value of the pore pressure, due to the perched water tables.This uncertainty has caused a great

Undrained shear strength (s u ) and preconsolidation pressure (s' p )
To define which equations best fit with the variegated soil in relation to the undrained shear strength (s u ) and preconsolidation pressure (s' p ), the equations from CPTu and DMT were compared with laboratory tests (triaxial CU) presented by Massad (1980 and2012).
For dilatometer tests the following equations were used: a) Equation 14from Marchetti (1980) to determine the undrained shear strength (s u ); and b) Equation 6 from Kamei & Iwasaki (1995) to determine preconsolidation pressure (s' p ).
Extracting the K D from Eq. 6 and replace it in Eq. 14 one can get the following expression: The OCR depends obviously on s' vo and is therefore affected by perched water table problems.But fortunately the OCR exponent in Eq. 20 is small, which makes the relationship s u /s' p relatively constant and independent of s' vo .This is demonstrated by the graphs shown in Fig. 8.It therefore follows that the ratio s u /s' p ranges between 0.16 and 0.20 for DMT data.
For the piezocone tests the Eqs. 4 (N kt = 20) and 1 (K 1 = 0.333) were used for determining s u and s' p , respectively.Based on these equations it is easy to establish the following relationship: also included in the graphs of Fig. 8. Massad (1980 and2012) showed that for the variegated soil of the São Paulo Formation the ratio c'/s' p @ 0.10 is valid, where c' is the effective cohesion intercept from CU triaxial tests in the overconsolidated range.As it is well known, in this range, the strength envelopes in terms of effective and total stresses practically coincide (see, for example, Pinto, 2000).Then, s u /s' p @ 0.10, which was included in Fig. 8 under the heading "Laboratory" and can be taken as a lower limit of this ratio, due to soil disturbance in laboratory tests.
As a result of this analysis, it is possible to conclude that the ratio s u /s' p varies between 0.16 and 0.20 for the DMT data, averaging 0.18 with a standard deviation of 0.013.
Reporting referring again to Fig. 8, the ratio s u /s' p from CPTu data ranges from 0.10 to 0.20, with the value of 0.15 in an intermediate position.This result can be considered a validation of the value of N kt = 20 of Eq. 18, adopted in the analysis.To correlate the initial shear module (G 0 ) with preconsolidation pressure (s' p ) Fig. 9 was prepared.
For CPTu data Eq. 3 from Watabe et al. (2004) was adopted to estimate the G 0 and Eq. 1 to estimate s' p with K 1 equal to 0.33.It is easy to see that the ratio G 0 /s' p = 50/0.333= 150 or 0.15 if the G 0 are in units of MPa and s' p in kPa.To determine G 0 and s' p from DMT data, Eqs. 6, 7 and 13 were used.As previously mentioned, the G 0 and s' p for dilatometer tests depend on s' vo with all aforementioned perched water table problems.Manipulating these three equations, it was possible to arrive at the following expression: Similar to Eq. 20, there is a marked influence of s' vo but still relatively small in relation to G 0 /s' p .The graphs in Fig. 9 show that this ratio varies from 0.10 to 0.20, with the figure of 0.15 from CPTu data in an intermediate position.OCR and K 0 ), deformability modules (E i , M and G 0 ) and undrained shear strength (s u ).It became clear that the parameters K 0 and OCR from DMT data may be estimated using, respectively, the equations of Lunne et al. (1990) and Kamei & Iwasaki (1995).It was then possible to determine the correlation between these two parameters, resulting in K 0 = 0.92OCR 0.38 .This correlation was validated with a K 0 value given by a test made with the pressuremeter (Camkometer) at the base of a stiff variegated clay layer in Ibirapuera, São Paulo.

Conclusions
The correlation between undrained shear strength (s u ) and preconsolidation pressure (s' p ), obtained through the DMT data, was set using the Marchetti (1980) and Kamei & Iwasaki (1995) equations, respectively.This led to values between 0.16 and 0.20 for the ratio s u /s' p , practically the same as that indicated by Massad (2012) for "taguá" from the Resende Formation.In addition, Massad (1980 and2012) found that s u / s' p @ 0.10 for the variegated soils of São Paulo using triaxial laboratory tests.As far as the CPTu data are concerned, a ratio s u /s' p = 0.15 was achieved adopting N kt equal to 20 in the equation s u = (q t -s vo )/N kt , i.e., the formula of the cylindrical cavity expansion.These findings were taken as a validation of the empirical factor N kt .
The correlation of initial shear modulus (G 0 ) with preconsolidation pressure (s' p ) from DMT data was obtained applying the Lunne et al. (1990) equation (for G 0 ) and the Kamei & Iwasaki equation (1995) (for OCR and then s' p ).Although both parameters depend on s' vo with all the already mentioned perched water table problems, it was shown that the ratio G 0 /s' p varied in a relatively narrow range, 0.10 to 0.20.On the other hand, for CPTu data, using the Watabe et al. equation (2004) for G 0 , and the Kulhawy & Mayne equation (1990) for s' p , it was possible to obtain G 0 /s' p = 0.15, an intermediate value between the former range of 0.10 to 0.20.
For the correlation between the Young's modulus (E i ) and undrained shear strength (s u ) the Kulhawy & Mayne equation (1990) (for E i ) and the formula of the cylindrical cavity expansion (for s u ), with N kt = 20 were applied to the data of CPTu.Imposing E i /s u = 400, the average value obtained in laboratory by Massad (1980) and Massad et al. (1992), it was possible to conclude that a = 20 in the Kulhawy & Mayne equation (1990) E i = a(q t -s vo ).
Finally, as a general conclusion, despite the extreme heterogeneity of the variegated soils, comprising alternating layers of sandy clays, and clayey sand with silt fractions, well defined correlations were obtained, consistent with results presented in the technical literature, confirming the potentiality of the piezocone and the dilatometer tests.However, future research should provide accurate measurements of the pore pressures, particularly the suction in layers of partially saturated soils, which can occur at different depths due to the occurrence of perched water tables, particularly in the upper parts of the São Paulo Formation.

Figure 1 -
Figure 1 -Location of the studied site.

Figure 4 -
Figure 4 -Typical results of parameter from CPTu and DMT from Site 1.

Figure 5 -
Figure 5 -Typical results of parameters from CPTu and DMT from Site 3.
Soils and Rocks, São Paulo, 39(2): 189-200, May-August, 2016.195 Geotechnical Parameters for the Variegated Soils of São Paulo Formation by Means of In Situ Tests

Figure 7 -
Figure 7 -Correlation between E i e s u from CPTu and DMT tests.

Figure 7 (
Figure 7 (cont.)-Correlation between E i e s u from CPTu and DMT tests.

Figure 8 -
Figure 8 -Correlation between s' vo and s u .

Table 1 -
Values from s' p, OCR and K 0 from the investigated sites.

Table 2 -
Values from s u , E i, G 0 and M from the investigated sites.

Table 3 -
Values of relevant parameters from the investigated sites.