Enhanced DE-TMCMC and its application in identifying parameters of advanced soil model

Abstract

The parameter identification using Bayesian approach with Markov chain Monte Carlo (MCMC) has been verified only for certain conventional simple constitutive models up to now. An enhanced version of the differential evolution transitional Markov chain Monte Carlo (DE-TMCMC) method and a competitive Bayesian parameter identification approach for use in advanced soil models are presented. The DE-TMCMC, enhanced through implementing a differential evolution into TMCMC to replace the process of proposing a new sample, is proposed. To verify its robustness and effectiveness, the triaxial tests on Toyoura sand are selected as objectives to identify the parameters of the critical state-based sand model SIMSAND. The original TMCMC is also used as a reference to compare the results of DE-TMCMC, which indicates that the DE-TMCMC is highly robust and efficient in identifying the parameters of advanced soil models. All the results demonstrate the excellent ability of the enhanced Bayesian parameter identification approach in identifying the parameters of advanced soil models from both laboratory and in situ tests. 目前基于贝叶斯结合马尔可夫链蒙特卡罗（MCMC）的参数识别方法仅在某些传统的简单本构模型的参数识别上得到了验证。鉴于此，提出了一种效率更高的基于差分进化算法的过渡马尔可夫链蒙特卡罗方法（DE-TMCMC），并基于此提出了一种高效的贝叶斯参数识别方法，应用于高级土体本构模型的参数识别。为了验证其稳健性和有效性，选取丰浦砂的常规室内试验结果作为目标试验来识别考虑临界状态的砂土本构模型的参数。通过对比原始 TMCMC 方法在参数识别上的表现，突显了 DE-TMCMC 在识别砂土高级本构模型参数方面的能力。