The accuracy of the state parameter inferred from miniature CPTu probing in the geotechnical centrifuge

Abstract

The state parameter (ψ) is a key indicator of soil behaviour and liquefaction susceptibility, yet the determination thereof in-situ remains challenging. Current practice relies on empirical correlations developed from piezocone penetration tests (CPTu), most of which were calibrated using clean, dilative sands. Their applicability to silty, contractive tailings materials therefore remains uncertain. This study evaluates the accuracy of five commonly used CPTu-based state parameter inference methods developed by Been et al. (1987), Been and Jefferies (1992), Plewes et al. (1992), Robertson (2010) and Robertson (2022). The assessment was carried out using geotechnical centrifuge modelling and miniature CPTu testing, which enables realistic stress conditions around the cone, conditions that are difficult to replicate in calibration chambers. Four saturated soil models were prepared to different densities and tested at different stress levels. These included two silica sand models and two gold tailings models, each representing different soil states. The known in-situ state parameter profiles were calculated using critical state soil properties and compared to the inferred profiles obtained directly from CPTu data using each empirical method. Quantitative performance was assessed using the root-mean-square error (RMSE). Results show strong method-dependent biases. Both Been methods consistently underpredicted ψ across all models, yielding strongly dilative estimates. The Plewes method demonstrated improved accuracy but high variability. The Robertson methods provided the best performance in loose, contractive tailings, but weakened in dense states due to stress-normalisation limitations. The findings highlight limitations in the generalised use of current state parameter-inference procedures and provide guidance for selecting the most reliable correlations for sands and gold tailings under different states and drainage conditions.