Abstract In underground mines, stability problems can cause unplanned dilution, increasing the costs of mining operations, such as loading, transport, crushing and grinding. This is usually dealt with by simply increasing investments with support. However, the design geometry of stopes and drifts also has a great effect on a stope’s stability. The use of empirical stability methods is a very common practice, however because of their few input parameters, it cannot predict the actual performance related to each stope and drift design. This study has as main goals a back analysis of the actual dilution, using the Equivalent Linear Overbreak Slough (ELOS) method and analysis in the drift developments at the hanging wall contact, using numerical models. The case study is an underground hard rock gold mine, with a pillar-less Transversal Stope method. The hanging wall failure was quantified using a database from 19 stopes measured by the Cavity Monitoring System (CMS). Numerical finite elements models (FEM) were used, and 5 primary stopes were selected to calibrate models. The volumes of actual dilution showed a good correlation with the volumes of the relaxation zones produced by the models. Then, new optimized drift geometry layouts at the hanging wall contact were proposed, showing a decrease in 2.35 times in the relaxation zone of the numerical models, therefore justifying the use of these models as a dilution optimization tool. This methodology also proved to be useful in the definition of cable bolt length.

Resumen Las excavaciones subterrâneas se encuentran entre las obras de ingeniería más complejas que existen, ya que involucran muchas variables, desde el entorno de trabajo hasta los métodos y equipos adoptados para la excavación. Históricamente, los proyectos preliminares de excavación se han desarrollado con base en métodos empíricos y clasificaciones cualitativas o semicuantitativas de macizos rocosos. Dada la información insuficiente sobre las propiedades del macizo rocoso, debido a las limitaciones técnicas relacionadas con los sondeos y sus interpretaciones, existe, desde los estudios conceptuales y la ejecución de un proyecto, una gran variabilidad en las decisiones a tomar. Las decisiones incorrectas con respecto al método de excavación, el tipo de soporte y los estimados de avances pueden resultar muy costosos para la empresa, lo que genera gastos y/o riesgos para las vidas humanas no planificados o innecesarios. Por lo tanto, este estudio propone el uso de Análisis de Riesgo cuantitativo por Simulaciones de Monte Carlo para determinar la clase de soporte más probable que se aplicará en un proyecto de excavación subterránea.

Abstract Underground excavations are among the most complex engineering works in existence, as they have many variables involved, from the working environment to the methods and equipment adopted for excavation. Historically, preliminary excavation projects have been developed based on empirical methods and qualitative or semi-quantitative classifications of rock mass. Given insufficient information regarding rock mass properties, due to technical limitations related to soundings and their interpretations, there is-from conceptual studies and project executions-great variability in the decisions to be made. Wrong decisions regarding the excavation method, support type, and projections of advances can be highly costly to the enterprise, leading to unplanned or unnecessary expenses and/or risks to human lives. Thus, this study proposes the use of quantitative Risk Analysis by Monte Carlo Simulations to determine the most likely support class to be applied in an underground excavation project.