A comparative study of various metamodeling approaches in tunnel reliability analysis

Thapa, A., Roy, A. and Chakraborty, S. (2024) A comparative study of various metamodeling approaches in tunnel reliability analysis. Probabilistic Engineering Mechanics, 75, 103553. (doi: 10.1016/j.probengmech.2023.103553)

Abstract

Various metamodeling approaches are applied in conjunction with Monte Carlo simulation and or the second moment-based method for reliability analyses of underground tunnels. However, there is no study regarding the suitability of such metamodels for reliability analyses of tunnels. An attempt is made here to make a comparative assessment of different metamodeling approaches for tunnel reliability analysis to comprehend the performances of various metamodels from the subset of machine learning methods. In doing so, the least square method based polynomial response surface method (RSM), mostly used in tunnel reliability analyses, and its improved version i.e., moving least square method-based RSM, are taken up for comparison. Further, the most successful empirical risk minimization-based Kriging model and the structural risk minimization principle-based support vector regression model are considered for comparison. Also, the sparse Bayesian regression found to be useful in solving various structural reliability analysis problems, is taken up for the present comparative study. Two numerical examples demonstrate the effectiveness of the selected metamodels in tunnel reliability analysis. It has been generally noted that the Kriging and SVR-based metamodels outperform in reliability estimates of underground tunnels.

• Item Type: Articles
• Status: Published
• Refereed: Yes
• Glasgow Author(s) Enlighten ID: Roy, Dr Atin
• Authors: Thapa, A., Roy, A., and Chakraborty, S.
• College/School: College of Science and Engineering > School of Engineering > Infrastructure and Environment
• Journal Name: Probabilistic Engineering Mechanics
• Publisher: Elsevier
• ISSN: 0266-8920
• ISSN (Online): 1878-4275
• Published Online: 13 November 2023
• Copyright Holders: Copyright: © 2023 Elsevier Ltd.
• First Published: First published in Probabilistic Engineering Mechanics 75: 103553
• Publisher Policy: Reproduced in accordance with the publisher copyright policy