Silt mixture CPT characterization for a reliability evaluation of pile total stress bearing capacity

In the last decades variability in geotechnical characterization of soil deposits has been recognized to be another issue to manage by researchers and designers. Such perspective that treats soil strength properties as random variables can be considered a technique to quantify the spatial correlation of measures provided through experimental tests. Therefore sophisticated laboratory tests can reproduce almost all of the possible stress-path that soil samples can expertise but their responses represent the ones of little volumes respect to huge volume behaviour which geotechnical designers deal with. Conversely in situ test measures are strongly affected by measurement errors even though the measures relate to in situ soil conditions. Moreover such measures must be transformed in strength properties by correlation relations that introduce further sources of uncertainty in data set (model errors). So that statistical techniques and procedures can be quite useful to interpret data set and can be efficiently employed in geotechnical studies and designs. According to the previous issues the paper deals with the interpretation of CPT data from Brindisi urban area. Variability in CPT indexes as q1c is evident and correspond to the variability of undrained shear resistance su which is an important parameter of pile designing in clayey and silty soils. Consequently the study applies the most recent statistical tests aiming at: 1) reconstructing the spatial correlation structure of CPT data set related to the bearing resistance of this soil; 2) recognizing the outlier values due to measurement errors. Results from statistical techniques will be employed in Reliability-based design of pile foundation.