Comparative study of the S-shaped and Hoek-Brown failure criterion by finite-difference modelling, Northparkes, NSW, Australia

Abstract

Cave mining operations produce bulked ore from self-sustainable propagating caves in high stress regimes. Undercut development creates even higher abutment stresses and a stress-shadow which have significant impact on extraction level stability. Geotechnical assessment of the infrastructure is vital for the safety performance and continuity of production. Northparkes Mines (NPM) is a block caving operation located in New South Wales, Australia. They deployed a post-undercutting strategy to develop the E48 deposit, exposing all tunnels and draw bells at the extraction level to high abutment stresses. Convergence of the extraction drifts and relative displacements in the brittle rock mass have been monitored during this process. The current industry standard for geotechnical analyses of hard rock in stressed environments is the non-linear Hoek-Brown failure criterion. It estimates the resistance to shear failure of jointed to heavily jointed rock masses by rock mass strength degradation. However, brittle rock masses fail in a tensile mode in low confinement zones around excavations. The most recently developed hypothesis in this field of work is the S-shaped failure criterion. It assumes spalling at low confinement and an inhibition of tensile failure processes at relatively high confinements, where shear bands are the dominant structures causing failure. In other words, it suggests that the Hoek-Brown failure criterion underestimates the rock mass strength in highly confined zones of brittle rock masses. Validated principles to establish the rate of degradation in high confinement zones are currently not available and the S-shaped approach is therefore not yet ready for engineering purposes. The aim of this project is to indicate differences in results derived from numerical modelling when either the Hoek-Brown or the S-shaped failure criterion is applied and relate those differences to geotechnical data obtained at NPM. A three-dimensional, parameterized environment has been created by applying the mesh generator KUBRIX Geo on a Computer-Aided Design of the extraction level. Undercut development was simulated by an adequate stress-path using the Finite Difference Method in FLAC3D. An evaluation of the response of extensometer stations to an advancing undercut front at NPM has revealed immediate and abrupt dilative behaviour of the rock mass once the cave line has passed. Scenarios using the S-shaped failure criterion were able to simulate this behaviour if rock mass degradation was sufficiently contained at high confinement, while the scenario using the Hoek-Brown failure criterion was unable to do so. The extent of yield, stresses in pillar cores and stress-strain responses of the rock mass indicate stronger pillars in scenarios using the S-shaped failure criterion. The S-shaped failure criterion is a state-of-the-art approach and although the basic concept is finding more support, the definition of its parameters will require more research in the coming years. This project improves the understanding of the application of the S-shaped failure criterion in continuum modelling and highlights its strengths and shortcomings.