Effect of soil characteristics on relative values of piezocone, T-bar and ball penetration resistances

It has been reported in the literature that net cone penetration resistance is normally higher than the penetration resistance measured using full-flow cylindrical (T-bar) or spherical (ball) penetrometers, and also that the ball penetration resistance is generally very close to the T-bar penetration resistance. In this paper, a worldwide high-quality database (three onshore sites and seven offshore sites) was used to identify soil characteristics that contribute to differences in the relative magnitudes of penetration resistance measured with cone, T-bar and ball penetrometers. For soils covered in this database, the ratio of T-bar to net cone penetration resistance (q T-bar /q net ) lay between 0·75 and 1, decreasing with depth from near unity at shallow depths. For the four sites where ball penetrometer data were available, the ratio of ball to T-bar penetration resistance (q ball /q T-bar ) was found to fluctuate around unity, with an average just greater than 1, but no distinguishable trend with depth. The measured q T-bar /q net and q ball /q net followed theoretical trends for the cone resistance to increase with rigidity index, G/s u , with the best quantitative agreement obtained using a rigidity index defined in terms of the small strain stiffness, G 0 , as measured by in situ seismic cone tests. The resistance ratio also followed the theoretical trend predicted with certain assumptions for the effects of strength anisotropy. However, while q T-bar /q net and possibly q ball /q net appear to depend slightly on the liquidity index, liquid limit and plasticity index, the resistance ratios appeared independent of other soil parameters, including normalised in situ shear stress, sensitivity and yield stress ratio, for the soils covered in this study.