Parameter identification of an elastic-viscoplastic model for cohesive soils using self-boring pressuremeter tests
Abstract
Cohesive soils challenge accurate parameter determination due to time-dependent behavior and sampling sensitivity. The self-boring pressuremeter test (SBPT) minimizes disturbance and offers clear boundary conditions. This study develops an efficient optimization framework to identify parameters of an advanced elastic-viscoplastic model for cohesive soils using SBPT data. Comprehensive SBPTs on Kunming peaty soil included multiple loading, strain-holding, and pressure-holding tests. Results showed significant rate-dependency in cavity pressure and excess pore pressure, resulting from both partial drainage and intrinsic viscous soil behavior. A finite element model incorporating an overstress EVP constitutive law and partial drainage was established. Sensitivity analysis revealed cavity pressure is mainly governed by overconsolidation ratio, recompression index, and Poisson’s ratio, while excess pore pressure is most sensitive to permeability, recompression index, and Poisson’s ratio. An automated optimization framework integrating the numerical model with the multi-island genetic algorithm was implemented using Python scripting. Parameters optimized from two loading-rate tests deviated notably from laboratory values—especially in compressibility—underscoring SBPT’s advantage in capturing true in-situ behavior. Using these parameters, predictions for other tests agreed well with measurements. This approach reduces subjectivity, mitigates sampling bias, and utilizes excess pore pressure from loading tests as a practical alternative to strain- holding tests for permeability determination.