Scaling effect in cone penetration testing in sand

The Cone Penetration Test (CPT) was developed originally in Holland in the 1930’s as a device which provides a small scale model of a pile foundation. Early versions were simple cone points for which the only measurement was the thrust required to push the point through the ground. Over the past 20 years, the cone was standardized to a tip area of 10 cm², and an electrical version was produced, which allows for continuous measurement of the cone tip resistance and sleeve friction along with a computer-based data acquisition system. The electrical cone represents a significant step forward for the CPT, since it provides a continuous profile of information that can be used to identify soil type and define important engineering parameters. More recently, the CPT has shown considerable potential for calculation of settlements of footings on sand, determination of pile capacity, assessment of ground pressures, and evaluation of liquefaction potential for cohesionless soils.

Along with the widening application of the CPT, new varieties of cone penetrometers have appeared, with different sizes than the standard. Smaller cones are used for instances where relatively small depths of soil need to be penetrated, and larger cones have been developed for penetrating dense and gravelly soils. With the introduction of the new cones, there has been a tendency to assume that the methods for reducing CPT data for the standard sized cone can be extrapolated to the other sizes of cones. That is, it is assumed that there are no scale effects in cones of different sizes. While this may be true, to date, little direct evidence has been produced to support this view, and the issue is an important one from two points of view:

1. The present data analysis technology is based on that primarily from testing with a standard cone. lt is important to know if any changes are needed in this approach, or if the existing methods can be used with confidence for any size cone.

2. If it can be shown that no scale factor exists, then this will allow the use of new, smaller cones in experimentation in modem calibration chambers with the knowledge that the test results are applicable for the cones that a.re more widely used in practice. The smaller cones offer several advantages in this type of work in that they facilitate the research considerably by reducing the effort involved in sample preparation, and they are less likely to produce results influenced by boundary conditions in the chamber.

One of the major objectives of this research is to develop an insight into the issue of the scale factor caused by the use of different sizes of cones. This is accomplished through an experimental program conducted in a new large scale calibration chamber recently constructed at Virginia Tech.

Many of the latest developments in cone penetration testing have been forthcoming from testing done in calibration chambers where a soil mass can be placed to a controlled density under known stress conditions. To conduct the experimentation of this work, it was necessary to design, fabricate, and bring to an operational stage a calibration chamber. The Virginia Tech chamber is one of the largest in the world. A significant portion of the effort involved in this thesis research was devoted to this task. In particular, attention was devoted to the development of a system for placement of a homogeneous soil mass in the chamber, and the implementation of a microcomputer-based data acquisition unit to record and process the test results.

The scale effects investigation was performed using three different sizes of cone penetrometers in a test program conducted in the calibration chamber. Of the three cones, one is smaller than the standard with a tip area of 4.23 cm², one was a standard cone with a tip area of 10 cm², and one was larger than the standard cone with a tip area of 15 cm². A total of 47 tests were carried in the chamber using two different levels of confining stress and two different sand densities. The test results show that while a scale factor might exist, the degree of its influence on interpreted soil parameters for a practical problem does not appear significant.