The influence of pore pressure gradients in soil classification during piezocone penetration test
Highlights
► A new parameter “i” proposed to classify soils with cone penetration test data. ► A new soil type behavior index “Ic” proposed. ► A new soil classification chart with seven zones has been developed.
Introduction
Cone penetration test can be done easily then most of the other in-situ tests and its results are reliable and repeatable. It can be said that a major advantage of this test is that CPT provides a continuous profile. The scope of this paper is to estimate soil class by using in-situ cone penetration test results. Several investigators have attempted to classify soils by using the test data. The early methods have employed qc and fs to prepare classification charts without attempting to correct these for overburden and porewater pressure (Begemann, 1965). Sanglerat et al. (1974) have asserted that the type of soil is a function of the tip resistance and the friction ratio Rf, whereand sand, silt and clayey soils were represented in separate closed polygons in their chart.
Schmertmann (1978) represented cone tip resistance (qc) on a log and Rf on arithmetic axis to define the different zones. His chart differed from that of Begemann (1965) because sands are classified according to relative density and clays with their consistency. However, it is seen that fine grained soils are represented in limited bands of consistency that do not cover the whole spectrum. He emphasised that results from different regions may influence the shape of the chart due to factors such as sensitivity of the soils and their creep behaviour, roughness of the sleeve and the groundwater regime suggesting that it would be expedient to develop charts for local use.
Douglas and Olsen (1981) are the first investigators who attempted to include some of the USCS symbols in the qc–Rf (log) chart. In addition, they incorporated properties such as liquidity index, sensitivity, earth pressure coefficient and void ratio. Their chart is the predecessor of the currently existing charts and its striking difference from that of Schmertmann is the concave upwards shapes of the lines separating soil zones.
Jones and Rust (1982) have subsequently initiated the use of a piezometer in the cone (CPTU), where the change of porewater pressures during penetration was measured. The chart they developed is based on readings of net cone tip resistance (qc − σv0) versus excess pore pressure (Δu = umax − u0). This chart is unique because it comprises relative density and consistency values. Vermeulen and Rust (1995) have used this chart with minor changes to illustrate its use with a lot of data.
Robertson and Campanella (1983) modified the Douglas and Olsen (1981) chart and reported that mean grain size can be estimated by using the concentric circles. They also argued that measuring excess porewater pressures will improve the soil classification process.
Senneset and Janbu (1985) developed a classification system where a pore pressure coefficient Bq was defined. In addition to the use of qt, tip resistance corrected for pore pressure u2 was henceforth adopted.a is the ratio of the cone base cross section and total cross section. Bq is thus defined asu0 represents the hydrostatic pressure, u2 the dynamic pore pressure measured immediately behind the cone, σv total stress at specified depth and qt net cone tip resistance.
Robertson et al. (1986) used the expression for Bq to develop another classification chart where 12 zones were defined using the axes qt–Rf (%) and qt–Bq. Senneset et al. (1989) proposed a similar chart where Bq which is a function of corrected tip resistance qt and u2 with the difference that qt axis was arithmetic. Additionally, the maximum tip resistance is limited to below 16 MPa.
Robertson (1990) made a critical appraisal of their 1986 charts and changed the labels of the axes to normalised sleeve friction (F)–normalised tip resistance (Q). The accompanying chart uses Qt and Bq. The soil zones were reduced to 9 in this study. The F–Q chart is currently the most referred to where
Jefferies and Davies (1991) contested the Robertson (1990) charts claiming that two charts showing the relationship among Q, F and Bq is not essential. The chart was then modified by changing the Bq axis to Q(1 − Bq) to show all parameters in a single chart. It was then possible to express the influence of porewater pressure in the same chart. They claimed that such a grouping duly enlarged the zone for fine grained soils whilst no significant change emerged for sands.
Schneider et al. (2008) proposed using the ratio Δu2/u0 instead of Bq which may be more suitable for identifying clays, silts and sands. He claimed that soil behaviour is governed by dissipation of pore pressures that emerge during loading.
It can be deducted from above discussion that each parameter involved plays an important role to classify the soil. Generally, coarse grained soils give higher cone resistances (qc) than the fine grained. On the other hand, friction ratio (Rf) is bigger for high plasticity soils. Robertson et al. (1986) are of the opinion that Rf gives more reliable results than qc in general.
Other investigators (Zhang and Tumay, 1999, Cetin and Ozan, 2009) followed a different path to tackle the problem. They used probabilistic methods for soil characterisation and classification. Zhang and Tumay (1999) proposed a classification method to classify soil from CPT data by using statistical and fuzzy subset approaches. A continuous profile of the difference of having each soil type (silty, clayey, and sandy) can be obtained with this method. Cetin and Ozan (2009) proposed a simplified soil classification scheme based on probabilistic method. Cai et al. (2011) compared the CPT soil classification charts by using CPTU data obtained from clay deposits in Jiangsu Province, China. Researchers concluded that using only cone resistance and sleeve friction parameters to classify the soils with CPT gives less reliable results than using pore pressure ratio and net cone resistance.
Section snippets
Soil behaviour type index (ic)
Efforts for understanding the response of soil to penetration have recently been directed to the study of soil behaviour type index Ic, a value that represents the dimensionless radii of the concentric circles in several publications.
Jefferies and Davies (1993) have demonstrated that the curves in the Robertson chart (1990) are indeed concentric circles. They developed a chart where the axes were labelled as F − Q(1 − Bq) and soil type behaviour index was formulised as
Database
The first step in site investigation is to describe the soil profile by classifying the layers using two or four letter symbols. ASTM D 2487-93 (1994), BS 5930 (1981) and TS1500 (2000), a modified version of the British standard, are used for the purpose. No similar approach by the use of CPT data apart from Douglas and Olsen (1981) has been attempted so far.
This study has been conducted using the rich database obtained from Adapazari, Turkey, the site of the catastrophic earthquake in 1999.
Dynamic porewater pressure gradient (i)
One gets the impression, upon reviewing the existing work on classification by CPTU that several attempts have been made but the final solution still evades the researcher. The developments in the derivation of the formula for Ic indicates that some further refinements are still needed. The author believes that the next step should be to minimise the incursions of neighbouring zones in the classification chart in order that each group is identified in the USCS in a single zone rather than the
Multiple linear regression analysis, MLR
Multiple linear regression analysis (MLR) was implemented to see the contribution of i to the classification process done with variables Q, F, qc1n, and Bq derived from the CPTU data. This would enable the investigator to establish the relationship by evaluating the worthiness of dependent variables wL, Ip, D50, %Clay and SCN in order to see which parameter is to be included in the final solution. Accordingly, every dependent variable (wL, Ip, D50, %Clay, and SCN) was incorporated in the MLR.
Derivation of the formula for soil type behaviour index
This part of the paper gives an account of the derivation of the formula for the determination of the soil class where parameter i was also employed. Since apart from Li et al. (2007) all investigators used a circular form to define the soil type behaviour index Ic as (a2 = b2 + c2), its use was preferred in this study. The equation used to identify standard soil groups in this study was therefore adopted as
Fig. 5 shows the relationships between the liquid
Discussion and conclusions
This paper is about the prediction of soil classes assigned by laboratory through the use of data collected during a cone penetration test. A comprehensive evaluation of the existing knowledge available in the literature has been carried out and a new parameter to perform the analysis is proposed. The parameter i representing the porewater pressure gradients along the soil profile during cone penetration is a dimensionless number describing the changes in pore water pressures during flight. The
Acknowledgement
This research has been funded by TUBITAK [Projects no: 104M387 (Önalp et al., 2007) and 106M042 (Önalp et al., 2010)], the Turkish Foundation for Scientific and Technical Research. Their generous support is gratefully acknowledged.
References (31)
- et al.
Comparison of CPT charts for soil classification using PCPT data: example from clay deposits in Jiangsu Province, China
Engineering Geology
(2011) Standard classification of soils for engineering purposes (Unified Soil Classification System)
(1994)- et al.
Towards systematic CPT interpretation
The friction jacket cone as an aid in determining the soil profile
Determination of the relationship between soil properties and earthquake damage with the aid of neural networks: a case study in Adapazari, Turkey
Natural Hazards and Earth System Sciences
(2012)- et al.
Liquefaction of silts: the Adapazari criteria
Bulletin of Earthquake Engineering
(2010) - et al.
Documenting incidents of ground failure resulting from the August 17, 1999 Kocaeli, Turkey earthquake
(2001) Anon. British soil classification system for engineering purposes
BS 5930 Code of Practice for Site Investigations
(1981)- et al.
CPT-based probabilistic soil characterization and classification
Journal of Geotechnical and Geoenvironmental Engineering
(2009) - et al.
Soil classification using electric cone penetrometer