196 complemented with field observations and empirical knowledge, the study confirms its potential as a cost-effective and practical tool in long-term mine planning. KEYWORDS 3D numerical modelling; underground mine design; deep mining geomechanics; stress redistribution; pillar geometry optimisation; excavation stability; 1. INTRODUCTION Three-dimensional numerical simulations are increasingly used to support decisionmaking in deep underground mining, where rock mass response is controlled by high stress levels, complex excavation geometries, geological heterogeneity, and discontinuities. In such settings, key design choices—mining layout, pillar dimensions, development sequence, and support strategy—are tightly coupled with geomechanical risk, including excessive convergence, brittle failure, and mining-induced seismicity that may escalate into damaging rockbursts (Gibowicz & Kijko, 1994; Brady & Brown, 2006). The literature consistently indicates that these hazards cannot be evaluated reliably using single simplified indicators alone; rather, they require an integrated understanding of stress redistribution, stress-path effects, and damage development around excavations (Martin et al., 1999; Cai et al., 2004; Diederichs, 2004). State-of-the-art mine design commonly combines empirical rules and classificationbased guidance with simplified analytical calculations for scoping, complemented by targeted numerical analyses at selected critical locations. While this tiered approach remains practical, its reliability decreases when outcomes are driven by inherently three-dimensional interactions—such as stress variation around advancing faces, pillar–abutment coupling near panel boundaries, and the cumulative influence of multiple excavations developed over time. Moreover, studies of brittle damage around underground openings demonstrate that local failure often depends on excavation-induced stress concentration and confinement loss, which can be strongly geometry- and sequence-dependent and therefore difficult to capture using simplified methods (Martin et al., 1999; Cai et al., 2004; Diederichs, 2004; Kaiser & Cai, 2012). In parallel, research and practice in mining seismology and seismic risk management emphasize that effective control of seismic hazard requires linking mine geometry and extraction sequence to expected seismic response and damage potential, supported by both monitoring and designbased mitigation (Mendecki, 1997; Hudyma & Potvin, 2010; Simser, 2019). Despite progress in algorithms and computing power, the broader integration of 3D simulations into routine planning is still limited by practical barriers. Model construction and calibration can be time consuming, particularly when it involves realistic geometry, appropriate boundary conditions, rock mass parameterisation, and uncertainty checks. Moreover, many operations still lack a repeatable workflow that translates 3D model outputs into concise decision metrics that mine planners can use to compare competing layouts in a consistent way
RkJQdWJsaXNoZXIy MTM0Mzk2