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

Area of Science:

• Geotechnical Engineering
• Applied Geology
• Spatial Statistics

Background:

• Cone Penetration Tests (CPTs) provide essential data on subsurface soil layers.
• Soil Behavior Type index (I_SBT) aids in lithological identification from CPT data.
• Understanding 3D soil variability is crucial for accurate geotechnical analysis.

Purpose of the Study:

• To model the 3D spatial variability of soil lithotypes and mechanical properties using CPT data (tip resistance q_c and sleeve resistance f_s).
• To evaluate the uncertainties associated with these estimates for engineering design purposes.
• To develop a stochastic mechanical and lithological model for subsoil characterization.

Main Methods:

• Collection and analysis of 182 CPT soundings across a 900 km² area.
• Application of two geostatistical methods: Ordinary Kriging (stationary) and Intrinsic Random Function theory (non-stationary).
• Integration of CPT tip resistance (q_c) and sleeve friction (f_s) for 3D soil modeling.

Main Results:

• The non-stationary geostatistical method (Intrinsic Random Function theory) yielded more reliable predictions for q_c and f_s.
• The developed 3D stochastic model effectively captures the spatial heterogeneity of soil properties.
• Identified abrupt changes in bearing capacity, liquefaction potential, and settlement due to fan and paleochannel bodies.

Conclusions:

• Non-stationary geostatistical modeling offers superior accuracy for CPT data interpretation in heterogeneous subsoils.
• The 3D lithological and mechanical model enhances the understanding of complex alluvial deposit structures.
• This approach provides critical insights for engineers and geologists in assessing subsoil risks and design parameters.