Ensemble Calibration of Tunnel Deformation in Weak Rock Masses

Abstract

This study evaluates the mechanical behavior of weak and heterogeneous rock masses along the Honaz Tunnel (SW Türkiye) through 2D finite-element simulations. To address the extreme prediction scatter of classical empirical equations in weak formations, an ensemble-based calibration framework was developed by averaging the rock mass deformation modulus (Emass\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${E}_{mass}$$\end{document}) derived from 15 different empirical formulations. To ensure rigorous methodological validation and prevent circularity, the in-situ dataset comprising nine geological units was strictly divided into a six-unit calibration set for parameter back-analysis and a three-unit independent validation set. Three constitutive criteria—Mohr–Coulomb (MC), Hoek–Brown (HB), and Generalized Hoek–Brown (GHB)—were compared under the implemented NATM support scenario. Using the validated MC parameters, alternative support layouts derived from comprehensive rock mass classification systems (e.g., Q, RMR, RMQR, RMi) were further evaluated. All numerical predictions were subsequently benchmarked against in-situ convergence measurements. Results indicate that, under the specific calibration and stress conditions of this shallow tunnel, the MC model provided the most accurate fit to the field data. Conversely, the HB and GHB models tended to safely overestimate deformations in highly weathered, soil-like units. The findings highlight the effectiveness of integrating an unweighted ensemble empirical framework with a stratified validation strategy for robust tunnel support design in complex geomechanical environments.