Assessment of Site Variability from Analysis of Cone Penetration Test Data

Abstract

Soil property values for use in geotechnical design are often estimated from a limited number of in situ or laboratory tests. The uncertainty involved in estimating soil properties from a limited number of tests can be addressed by quantifying the variability within individual soundings and of the collection of soundings at a site. It has been proposed that factors of safety or resistance factors used in design be linked to site variability. Site variability can be assessed by studying the correlation structure of in situ test data. The cone penetration test (CPT), which is a reliable and widely‐accepted in situ test, can be used for this purpose. Soil behavior type (SBT) charts are often used to obtain the subsurface soil profile from CPT parameters such as the cone resistance and the sleeve friction. A soil profile generation algorithm was developed in this research to generate a soil profile from an individual CPT sounding using two modified SBT charts. Soils are variable in both the vertical and horizontal directions. A vertical variability index (VVI) was defined to quantify variability in a CPT sounding. The average of the VVIs for all CPT soundings performed at a site is the site VVI. A site horizontal variability index (site HVI) was also developed, based on cross‐correlation between cone resistances, the cone resistance trend differences and the spacing between every pair of CPTs considered, to quantify the soil variability of a site in the horizontal direction. A site variability rating (SVR) system, integrating the vertical and horizontal site variability, was developed to assess the overall site variability. Depending on the SBT chart selected, the soil profile generated using the soil profile generation algorithm may be slightly different; however, the SBT chart effect on the variability indices that compose the SVR index is small. Close agreement was found between the SVRs obtained using the two SBT charts selected for this research. In order to illustrate the use of the algorithms for VVI and HVI calculations and SVR of sites, CPTs from across the state of Indiana were analyzed. CPT data were obtained from Purdue's own database, Indiana Department of Transportation's (INDOT’s) data repository and the U.S. Geological Survey (USGS) website. Site variability is calculated for specific depths of interest. For example, that depth of interest will be shallower for shallow foundations than for deep foundations. Site variability rating maps (SVR maps) for various depths of interest were constructed for the state of Indiana, illustrating the potential use of the site variability assessment methodology. An optimal sounding spacing calculation methodology was also developed to make the site investigation process more efficient, cost‐effective and reliable.