Bayesian model selection for sand with generalization ability evaluation

Abstract

Summary Current studies have focused on selecting constitutive models using optimization methods or selecting simple formulas or models using Bayesian methods. In contrast, this paper deals with the challenge to propose an effective Bayesian‐based selection method for advanced soil models accounting for the soil uncertainty. Four representative critical state‐based advanced sand models are chosen as database of constitutive model. Triaxial tests on Hostun sand are selected as training and testing data. The Bayesian method is enhanced based on transitional Markov chain Monte Carlo method, whereby the generalization ability for each model is simultaneously evaluated, for the model selection. The most plausible/suitable model in terms of predictive ability, generalization ability, and model complexity is selected using training data. The performance of the method is then validated by testing data. Finally, a series of drained triaxial tests on Karlsruhe sand is used for further evaluating the performance.

Three‐dimensional hydromechanical modeling of internal erosion in dike‐on‐foundation

“…At the present time, well‐documented erosion tests and triaxial compression tests with different fines content are still needed. The erosion law and model parameters can be either identified by optimization methods or Bayesian‐based methods …”

Soil‐structure interface modeling with the nonlinear incremental approach

“…Moreover, due to the discrepancy of the tests and lack of experimental data (eg, GSD before and after shearing), the parameters are not easy to be quantified accurately by the standard way. Another efficient procedure for parameter identification is to apply the recently developed optimization methods 54–57 or Bayesian‐based methods 58,59 . Thus, the newly developed optimization method 57 was adopted to identify three selected sets of sand‐steel interface tests.…”

Enhancement of backtracking search algorithm for identifying soil parameters

“…To generate the synthetic tests, the SIMSAND model proposed by Jin et al and a set of typical parameters of were adopted, as summarised in Table . Figure shows the illustration of synthetic test data generated by SIMSAND.…”

“…Yin et al distinguished three approaches: analytical methods, empirical correlations, and optimisation methods to determine soil parameters based on experimental data. Among these techniques, the inverse analysis by optimisation has been successfully used in the geotechnical area because it produces a relatively objective determination of the parameters for an adopted soil model, even for those that have no direct physical meaning. The existing optimisation techniques can be divided into two categories: (a) deterministic optimisation techniques; and (b) stochastic optimisation techniques.…”