Modeling 3D soil lithotypes variability through geostatistical data fusion of CPT parameters

Abstract

The Cone Penetration Test (CPT) measures enable to recognize vertical lithological sequence at each investigated point. From the tip resistance q_c and sleeve resistance f_s profiles, the Soil Behavior Type index I_SBT has been calculated, in order to identify the lithotypes alongside depth. The present study focuses on the combination of different variables to provide a lithological and mechanical subsoil characterization. The main objectives of the paper are: (1) to model the 3D spatial variability structure of the soil lithotypes and mechanical properties using q_c and f_s profiles; 2) to evaluate the uncertainties of the estimates for designing purposes. 182 CPTs were collected in a 900 km² area (corresponding to a subsoil volume of about 12 km³) located in the study site in the Bologna province (Italy). The study area is made up of fine-grained soils, silt and clay mixtures that are intercalated at different depths by sandy and gravelly soils. These variations of each soil fraction affect the engineering properties of these alluvial deposits. For 3D modeling, two geostatistical methods, Ordinary Kriging (stationary method), and Intrinsic Random Function theory (non-stationary method) have been used. The results show that the non-stationary method allows to obtain more reliable q_c and f_s predicted values. The final stochastic mechanical and lithological model enables engineers and geologists to detect the emerging of fan and paleochannels bodies where mean resistance values can abruptly change in terms of bearing capacity, liquefaction potential and static and dynamic settlement occurrence.