Keywords: system identification, Bayesian inverse problem, nonlinear optimization, global sensitivity analysis, surrogate modeling, mechanized tunnel simulation.

This paper presents deterministic and probabilistic approaches for the identification of soil layer changes ahead of a mechanized tunnel face. In the deterministic approach an optimization algorithm, namely differential evolution, has been utilized for minimizing the discrepancy between numerical model responses and in-situ recorded data. For considering the uncertainties associated in the identification process and the prior knowledge about the unknown parameters that is obtained from bore holes, the Bayesian updating technique has been used in the probabilistic approach. Moreover, a variance-based global sensitivity analysis for detecting key unknown model parameters is implemented. In addition to and as a result of the computational effort needed to perform the global sensitivity analysis and the Bayesian updating, a surrogate modeling based on quadratic polynomial regression has been considered as a subsitute for the computationally expensive finite element numerical simulation of the mechanized tunneling.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £65 +P&P)