Adam Pidlisecky

In general, my research focuses on characterization of the unconsolidated near-surface using cone-based geophysics. Cone-based geophysics is a new concept that involves making geophysical tomography measurements on specially modified cone penetrometers. Cone penetration testing (CPT) has long been used by geotechnical and environmental engineers as a method for obtaining 1D profiles of the physical properties of the top 100m of the earth. However, cone penetrometers suffer from the problem that measurements are made over a small volume immediately adjacent to the testing device. This leads to the need for geophysical techniques that can characterize the spatial heterogeneity between the 1D cone profiles. Geophysical data can be acquired rapidly using cone penetrometers; however, due to the perturbation caused by the cone, interpreting these data offers significant challenges.

A Little About Cones  

CPT involves pushing an instrumented 36mm diameter steel rod with a cone-shaped tip into the ground while making measurements close to the tip. The technology is classified as minimally invasive (the small holes are grouted after measurement), and is an efficient, inexpensive alternative to borehole installation. Of particular interest for characterizing zones of electrically conductive contamination is resistivity cone penetrometer testing (RCPT) which involves a module, immediately behind the standard cone penetrometer, equipped with several ring electrodes (figure. 1).

The module operates by injecting current into the ground using two of the electrodes, and measuring the resulting potential drop across the same two electrodes, or an adjacent pair of electrodes (Lunne et al., 1997). The shortcoming of the cone is that, as with direct sampling, measurements are made over small volumes of the subsurface.

My Part In This Idea  

Specifically, I am researching the use of cone penetrometers to obtain resistivity tomography data. Figure 2 illustrates the basics of a cone based tomography experiment; in this case the target is a salt water wedge. At various locations near the suspected salt water wedge, we emplace permanent current electrodes. Once the current electrodes have been emplaced, we begin pushing the resistivity cone into the ground using a cone truck. Measurements are made in the “standard” RCPT mode, thus providing a high resolution 1D profile of resistivity as a function of depth at the truck location. At regular intervals (e.g. every 0.5m) resistivity tomography measurements are made. To do this, the cone is and current injected into a pair of the current electrodes. The potential drop between the cone–mounted electrode and a surface reference electrode is then measured. This pattern is repeated for all independent current pairs and until the cone is at the maximum depth of interest. The truck is then moved, and the experiment repeated at a different location.

Figure 2 : Schematic of a cone-based resistivity imaging system

The advantage of obtaining data using the cone penetrometer is that it allows us to conduct a more complete experiment, this is to say that for a given current electrode configuration, we obtain many sub-surface potential measurements. This allows us to use fewer current configurations which results in fewer forward solves in our inversion and a better determined problem. However, the cone is highly conductive and represents a large perturbation to the field near the measurement location. As the cone is only 3.5cm in diameter, explicitly modeling it in 3D would result in a prohibitively expense inversion. At present I am working on ways to correct for this “cone-effect” without specifically modeling it. In addition to this I am interested in techniques for rapid inversion of these data. These cone measurements are made in a successive fashion, one profile followed by another followed by another.. etc. This type of acquisition would greatly benefit from the ability to invert these data in real time, as the results could be used to determine the optimum location of the next profile. Currently, I am exploring rapid Quasi-Newton inversion techniques that may be used for this as well as working with parallel computing to facilitate rapid inversions.