Benchmarking selection of parameter values for the Barcelona basic model

• Benchmarking exercise on selection of parameter values for the Barcelona Basic Model
• Some model parameters impact on several different aspects of soil behaviour.
• All these impacts should be considered before finalising selection of parameter values.
• Parameter optimizations based on inverse analysis should be used only with great caution.
• Extrapolating predictions far from the experimentally explored stress range in boundary value problems is risky.

Seven teams took part in a benchmarking exercise on selection of parameter values for the Barcelona Basic Model (BBM) from experimental data on an unsaturated soil. All teams were provided with experimental results from 9 tests performed on a compacted soil in order to determine values for the ten BBM soil constants and an initial value for the hardening parameter. The coordinating team then performed simulations (at stress point level) with the 7 different sets of parameter values, in order to explore the implications of the differences in parameter values and hence to investigate the robustness of existing BBM parameter value selection procedures. The major challenge was found to be selection of values for the constants λ(0), r, β, N(0) and pc   and an initial value for the hardening parameter p¯00, with the various teams proposing significantly different values for some of these key parameters. A key lesson emerging from the exercise is the importance of choosing a method for selecting values for the parameters β and pc which places the main emphasis on attempting to optimise the match to the experimental spacing of normal compression lines at different values of suction.