Introduction

Many methods and approaches are used to determine the bearing capacity and settlement properties of the foundation soils on which structures are placed. The most commonly used methods are the menard pressuremeter test (PMT), standard penetration test (SPT), cone penetrometer test (CPT) and plate loading test. Apart from these tests, some other laboratory tests are conducted. Due to possible degradation of the samples taken for the laboratory tests, and the inability of the sample to represent the general characteristics of the soil completely, the accuracy of parameters used in calculations is likely to decrease. The greatest advantage of in situ tests is that they reveal more realistic and reliable results as they are carried out without disturbing the soil. In our country (Turkey), the SPT and PMT tests are the most commonly used in situ tests performed in various soils. Although PMT is more complex to carry out compared to SPT, it is applicable to both soils and fissured rocks. In this test, various variables can be measured directly. However, the greatest drawback of this test is the likelihood that it can be misinterpreted. In recent years, along with widespread use of drilling works, PMT tests are increasingly being conducted in our country, too. In field studies, generally the Menard type pressuremeter is preferred. PMT can be applied to a wider range of soils, from very soft soils to rocks with a uniaxial compressive strength of 20 MPa, in which the SPT test cannot be conducted. PMT can be carried out far into the weak rock class identified in the ISRM (2007) rock characterization method using ground identifications and the uniaxial compressive strength of discontinuity surfaces (Kayabaşı 2012).

Only a limited number of studies between SPT and PMT have been conducted by individual researchers (Chiang and Ho 1980; Ohya et al. 1982; Gonin et al. 1992; Yagiz et al. 2008; Bozbey and Togrol 2010; Kayabaşı 2012; Kayabaşı and Gökceoğlu 2012; Aladağ et al. 2013; Cheshomi and Ghodrati 2015; Anwar 2016). The first studies were revealed by Chiang and Ho (1980) and Ohya et al. (1982). The linear relationship between SPT(N) and E PMT and P L was presented by Chiang and Ho in weathered granite in Hong Kong in 1980, whereas Ohya et al. (1982) presented the correlation between SPT(N) and E PMT for clay soils. Also, Yagiz et al. (2008) studied the relationship between N cor and E PMT and P L, and found that a linear relationship exists between the corrected N cor and E PMT and P L in Denizli (Turkey) for shallow sandy silty clays. Bozbey and Togrol (2010) presented the relationship between N 60 and E PMT and P L based on a study conducted in Istanbul (Turkey). Their results were based on 182 tests carried out on sand and clay soils, and present a distinctive linear correlation for each soil type. Gonin et al. (1992) correlated SPT results with E PMT and P L for nine different French soils.

In some of these studies, high determination coefficients (R 2) between the SPT N, and pure limit pressure (P L) and Menard pressuremeter (elasticity) modulus (E M) values were obtained for different soil classes. However, it has been particularly emphasized by the above-mentioned researchers that the empirical formula obtained may be misleading unless the pressuremeter test is applied precisely, the results for different soil classes are included, and different empirical equations for different soil groups by taking the geological characteristics of the study area into consideration are offered (Phoon and Kulhawy 1999; Yagiz et al. 2008; Kayabaşı 2012; Cheshomi and Ghodrati 2015; Bozbey and Togrol 2010).

In this study, the SPT and PMT tests were conducted at low and high plasticity (CL–CH) clay levels of soils with varying SPT N values from soft to firm clay, and at optimum levels in lithological units that can be classified as excessively stiffened or generally soft-supersoft rocks (e.g., marl). While measuring N blow counts via SPT tests, P L and E M values were determined via pressuremeter method. The results were then used to determine the relationship between the results of these two tests, and consequently compared with the results of previously conducted researches at similar soils.

Methods

Among a great number of test methods that determine in situ properties of soils, SPT and pressuremeter tests are primarily the most commonly used ones. The SPT test, which was developed in the 1920s with the intent of measuring penetration resistance of the soil, is known to be the most preferred field test of soil studies due to the ease of use it offers. After recording the number of blows that penetrate the tube into the ground as SPT-N 30 value, an N 60 value is calculated by revising it in accordance with the specifications in the literature (Bowles 1997; ASTM D1586 1999; Aggour and Radding 2001; British Standards Institution 2007; Sivrikaya and Togrol 2007).

The first study to determine the characteristics of lateral deformation in the soil was conducted in 1930s by Koegler. This was followed by contribution of Dr. Louis Menard in 1950s, who developed the device named Menard Pressuremeter (Menard 1957). Since the development of this device, the number of studies on this subject has increased, and the Menard Pressuremeter has led researchers to develop new approaches and the empirical formulas currently in use.

Menard type pressuremeter equipment consists of two main parts: the read-out unit, which rests on the ground surface, and the probe that is inserted into the borehole. The pressuremeter was lowered into the pre-formed hole and a uniform pressure is applied to the borehole walls by means of an inflatable flexible membrane. The pressure applied to the borehole walls is increased every 60 s in order to deform the borehole walls.

The pressure is held constant for 30 s and 60 s, and the increase in volume required to maintain constant pressure is recorded. A load-deformation diagram was determined to calculate P L and E M values (Shields and Bauer 1975; Baguelin et al. 1978; Mair and Wood 1987; Clarke 1995; Coduto 1999; ASTM D4719 2000).

The data used in this study were obtained from 20 exploration boreholes drilled as a part of a soil investigation program. The study area consists of two main layers. The upper levels, between 1.0 and 5.0 m, consist of silty clays and clay. The deeper levels consist of marl/claystone. The depths of the borings range between 1.5 m and 4.5 m. Groundwater was encountered at depths ranging from 2 m to 14 m. SPT and PMT tests were performed in those boreholes at clay levels having different consistencies. The pressuremeter test was performed in accordance with ASTM D4719-00. Soil characteristics were also determined by conducting laboratory tests on disturbed samples taken from these levels where SPT and PMT tests were performed.

Laboratory tests (e.g., soil gradation, Atterberg limits) were carried out to determine the index properties of soil, which were used in soil classifications. In the present study, the results of in situ tests and laboratory test results were evaluated together to determine the relationship between the SPT N, and E M and P L values at low and high plasticity (CL–CH) clayey soils in the Kurubaş district of Van, Turkey, and, finally, empirical formulas showing this relationship were developed through regression analyses.

Results

Geology

Two different geological units are exposed in the study area (Fig. 1). The basement rocks are Oligocene-Lower Miocene Van formation and consist of alternation of marl/clay and sandstones. Loose sandy-silty clay units cover the Van formation with an approximate maximum thickness of 15 m (Fig. 2).

Fig. 1
figure 1

Site location and geology map of the study area (modified from MTA 2007)

Full size image
Fig. 2
figure 2

NW-SE cross section of the study area

Full size image

Van formation

Van formation (Tv) is the most widespread sedimentary unit in the studied area, and is characterized by alternation of sandstones and marls. Sandstones are relatively thin and firmer than the marl layers (Fig. 3).

Fig. 3
figure 3

Marl and sandstone layers from the eastern parts of the study area

Full size image

Clay and fill

Van formation is overlain by a loose sandy-silty clay unit (Qk—Qd) (Fig. 4) which is possibly a result of the weathering of the Van formation. The approximate thickness of the unit is about 3–5 m, but, it sometimes reaches a maximum of 15 m. Plasticity is dominantly low within this unit; however, higher levels also exist.

Fig. 4
figure 4

A view from the cores indicating lower and upper levels of a borehole

Full size image

Index properties of clay levels

Based on the grain size distribution tests, the particle size of soil differs from very fine to fine. Soil characteristics of the upper levels within the study area are classified as low (CL) and high (CH) plasticity clay-silty clay (Fig. 5). The water content of test samples was determined at between 11% and 33%. Depending on the water content, the physical state of a clayey soil may change.

Fig. 5
figure 5

Soil classification of fine-grained units in the investigated area in accordance with TS 1500 (2000)

Full size image

Statistical evaluation

SPT and PMT tests were performed at 34 various levels of clay soils with different consistencies at varying depths between 1.50 and 4.50 m. E M and P L values were calculated from the pressuremeter curve determined after pressuremeter tests whereas SPT (N 60) values were calculated using SPT (N 30) values (Table 1). A statistical study was performed to determine the relationship between SPT (N 60) values and E M and P L values in CL and CH type of clayey soils.

Table 1 The measured soil parameters
Full size table

The coefficient of correlation determines the linear dependency between two variables. When the coefficient of correlation ρ xy is equal to zero, there is no linear dependency between x and y; however, when the absolute value of ρ xy approaches 1, the dependency between variables becomes increasingly stronger, and turns into a deterministic relation. But, this relation does not always indicate a cause and effect relationship between two variables. It may lead to a greater correlation coefficient where both variables have a relationship with another variable as well. On the other hand, it is important to assess whether the distribution is suited for the data set obtained in this study. Otherwise, this may result in a great difference between measured and calculated values using proposed empirical formulas. The t test and the F test were used to assess the validity of the proposed equations together with the coefficient of regression (the r value). Since the t test could not be applied for a limited number of data, the data in this study were evaluated initially in terms of concordance with the commonly used F test analysis. The computed F test value was found to be greater than the tabulated F value (Table 2), supporting a reliable correlation between measured and predicted values.

Table 2 Analysis of variance for the significance of the regressions and r values
Full size table

Statistical compatibility analysis was performed for the measured E M, P L and SPT (N 30) values, and finally for the SPT (N 60) values calculated using SPT (N 30) values (Table 1). The results of this analysis are given in Tables 2 and 3. After determining the compatibility of the data set that will be used in statistical analysis, regression analysis using a commercial software package (SPSS v23; SPSS 2002) was performed.

Table 3 Results from F tests of proposed equations and the significances of the r values
Full size table

The regression coefficients (R 2) between E M and SPT (N 60) and P L and SPT (N 60) were calculated as 0.75 and 0.76, respectively when the correlation was performed for 34 different levels having different types of clayey soils (Fig. 6a, b). The sub foundation soils at three points where SPT and PMT had been conducted were found to be marl and saturated clay. Since, the scope of this study was to determine the relationship between SPT and PMT test results in CL–CH class soils; the values obtained for marl and saturated clay were not taken into consideration in the analyses. Therefore, regression analysis was repeated using the values gathered from 31 different levels of CL and CH soils, and the determination coefficients (R 2) between E M and SPT (N 60) as well as P L and SPT (N 60) were found to be 0.91 and 0.89, respectively (Fig. 6c, d; Table 4).

Fig. 6
figure 6

Regression and confidence intervals between E M–SPT (N 60) and P L–SPT (N 60) of all samples (34 samples) (a, b) and (c, d) CL and CH soils samples (31 samples) in the study area

Full size image
Table 4 Summary of the developed model
Full size table

The values of E M and P L obtained from PMT tests in this study were compared with the values calculated using the proposed equations, in Bozbey and Togrol (2010), Kayabaşı (2012) and Cheshomi and Ghodrati (2015). The results are given in Tables 5 and 6. A similar linear relationship exists between E M and SPT (N 60) and P L and SPT (N 60) in all proposed equations (Fig. 7a, b). It is important to note that the empirical equations developed for calculating E M and P L values in this study reveals more accurate results than those developed by other researchers.

Table 5 Comparison between measured E M and P L with estimated E M and P L
Full size table
Table 6 Empirical relationships among E M, P L and SPT (N 60) by several studies
Full size table
Fig. 7
figure 7

Comparison of measured and calculated E M and P L value for clayey soils

Full size image

In all proposed equations, it was determined that the coefficient of determination between SPT (N 60) and E M was lower than SPT (N 60) and P L (Fig. 8). This may be attributed to the fact that P L values are read through pressuremeter curve directly, while E M is calculated from the below equation proposed by Baguelin et al. (1978), which is dependent on the Poisson ratio of the soil, which, in many cases, is hard to determine.

$$E_{\text{M}} = k\frac{P}{V}$$

where; \(k = 2(1 + \upsilon )(V_{\text{c}} + V_{\text{m}} )\).

Fig. 8
figure 8

Regression and confidence interval curves between a SPT (N 60) and E M and b SPT (N 60) and P L generated from the data regarding four studies

Full size image

In this equation, V c and V m values are the volume of the probe used in the pressuremeter, and the mean value of volume of the probe determined from the linear side of the pressuremeter curve, respectively.

Conclusion

In this study, the results of in situ tests (SPT and PMT) and laboratory tests were evaluated together in order to determine the relationship between SPT and PMT tests, which are commonly being used to determine the characteristics of clayey soils. High determination coefficients (R 2) showing the relationship between SPT and PMT test results suggest the use of SPT (N 60) values higher than ten for calculating E M and P L values.

The values of E M and P L obtained from PMT tests were compared with the values calculated using the developed formulas in this study and previously developed formulas by other researchers. Although all equations yield similar linear relationship between E M and SPT (N 60) and P L and SPT (N 60), the variation in the results suggests an increased number of the data set used. Since, the calculation of E M depends on the Poisson ratio of soils, the determination coefficient calculated for the relationship between SPT (N 60) and E M values was found to be lower than that calculated for SPT (N 60) and P L values. Therefore, it is recommended to determine the Poisson ratio for each level separately when conducting PMT tests, for more accurate results. It is also recommended to establish multiple regression analysis among the SPT, PMT and the results of laboratory tests for determining the physical and mechanical characteristics of clayey soils.