Calibration of BBM Parameters using the Modified State Surface Approach

¹ Missouri University of Science and Technology, Department of Civil, Architectural, and Environmental Engineering, 1401 N Pine St., Rolla, MO 65409 USA
² Technical University of Catalunya, Department of Geotechnical Engineering and Geosciences, Jordi Girona 1-3, 08034 Barcelona, Spain

Abstract

The Barcelona Basic Model (BBM) developed by Alonso et al. [1] is the first and the mostwidely used elasto-plastic model for unsaturated soils. The BBM successfully explained many key features of unsaturated soils and received extensive acceptance. However, there is lack of a well-establishedmethod for selecting parameter values for the BBM from laboratory tests, although a variety of methods have been recently developed for calibrating model parameters for the BBM. Concerns still exist on the correctness and robustness of such parameter value selection procedures. The above statements were evidenced by a recent benchmark exercise on selection of parameter values for the BBM organized within a "Marie Curie" Research Training Network on "Mechanics of Unsaturated Soils for Engineering"(MUSE)[2]. Experienced constitutive modelers in unsaturated soils from 7 prestigious teams were provided with the same experimental results on an unsaturated soil to calibrate the parameter values in the BBM. Theoretically, the calibrated parameters from different teams are expected to be the same or at leastvery close. However, the selected parameter values by the 7 teams are surprisingly widely different. This paper first discussed the limitations in the existing methods to calibrate the parameter values in the BBM. A novel approach was then proposed to calibrate the parameter values for the BBM. The approach takes advantage of the close-form solution of the BBM, which is derived based upon a newly proposed Modified State Surface Approach to study the unsaturated soils [3-6]. The same experimental data, used in the MUSE benchmark exercise, were reanalysed using the proposed approach to calibrate parameters for the BBM. The results were compared with those in the MUSE benchmark exercise from which thesimplicity, effectiveness, and robustness of the proposed method were evaluated.