Underground construction : case studies and analysis

• Main Author: Boon, Chia Weng.
• Other Authors: Low Bak Kong
• Format: Final Year Project
• Language: English
• Published: 2009
• Subjects: DRNTU::Engineering::Civil engineering::Geotechnical
• Online Access: http://hdl.handle.net/10356/15971
• Institution: Nanyang Technological University

This report addresses two most common stress-controlled stability problems in underground excavation, namely squeezing and spalling. An extensive review of case studies have been included in the Appendix, from which two have been selected for discussion in the main text: the Kaligandaki ‘A’ hydroelectric project in Nepal and the Mine-by Experiment in Manitoba, Canada. Chapter 3 establishes the limitations of using the equivalent Mohr-Coulomb criterion to approximate the radial convergence for a circular opening surrounded by an elastic-perfectly plastic rock mass that obeys the Hoek-Brown failure criterion. In turn, guidelines against potential pitfalls are recommended. This is achieved by comparing the analytical solutions proposed by Duncan Fama (1993) and by Carranza-Torres (2004). The equivalent Mohr-Coulomb parameters are obtained by transforming the Hoek-Brown parameters using methods proposed by Hoek et al. (2002). The accuracy of the semi-analytical solution proposed by Hoek and Marinos (2000) is also investigated. Chapter 3 also presents how the strain distribution for a circular excavation can be approximated through Monte Carlo simulations based on the rigorous analytical model proposed by Carranza-Torres (2004). Monte Carlo simulations and first-order reliablity method (FORM) are also carried out on the solution proposed by Duncan Fama (1993). Chapter 4 illustrates how the depth of brittle spalling for an underground circular opening can be approximated. Comparisons are made between the Hoek-Brown brittle parameters proposed by Martin et al. (1999) and Diederichs (2007), based on two-dimensional elastic plane-strain analysis using Examine2D. Chapter 4 also illustrates how the distribution of depth of failure can be approximated based on two approaches. The first approach is by performing probabilistic analysis on trendlines that have been best-fitted to the results of boundary element analysis, plotted in terms of normalized depth of failure Rf/r0 against normalized maximum stress with respect to damage-initiation strength σmax/σdi. The second approach is by approximating the probability of brittle failure at varying distance from the excavation boundary. The Probability Density Functions (PDFs) generated using parameters proposed by Diederichs et al. (2007) compare well with the measured distribution. On the other hand, parameters proposed by Martin et al. (1999) turned out to be conservative. This report also illustrates how the first-order reliablity method (FORM) using the computational approach proposed by Low and Tang (2007) can be implemented for squeezing and spalling analyses. This report only deals with aspects such as approximating the probability density functions (PDF) and cumulative distribution functions (CDF), and not reliability in its widest sense. In conclusion, this report summarizes some (but not all) of the techniques and considerations required for making preliminary design estimates for squeezing or spalling rock excavations.