Early detection of impending slope failure in open pit mines using spatial and temporal analysis of real aperture radar measurements

Abstract

Slope monitoring in open pit mines is an essential component of day-to-day operations and plays a key role in assisting geotechnical engineers and mine operators in maintaining mine safety and production schedules. Pit slope monitoring techniques have advanced significantly within the past decade, most notably in ground-based radar technology. Ground-based radar allows real-time monitoring of slope deformation across a broad coverage area, alerting mine staff to wall movements exceeding established thresholds. Lineof-sight measurements derived from the radar can be presented as 3D point clouds for the scan area, allowing mine staff to view the distribution of slope movements across the pit wall with each progressive scan. This paper presents a new methodology for spatial and temporal analysis of deformation point clouds captured by ground-based radar. The methodology builds on two existing early warning methods, the Fukuzono inverse-velocity method and the SLOpe gradient (SLO) method, which are based on the analysis of point measurement data derived from traditional geodetic prism monitoring. However, similar methodologies that fully utilise the spatial and temporal characteristics of ground-based radar data are yet to be developed. Radar data from historical slope failures captured by GroundProbe Slope Stability Radar (SSR) at a number of hard rock mines was utilised in the development of the new spatial and temporal analysis methodology. A slope failure that occurred at an open pit copper mine is presented throughout as a case example. The spatial analysis component of the methodology uses a benchmark point (or pixel), based on an alarm threshold specific to each failure case, and averaged deformation increments based on a percentage of the deformation measured by the benchmark pixel at the time of alarm. The temporal analysis component of the methodology examines deformation and velocity trends for all spatial analysis cases. The results of the spatial and temporal analysis were then used to evaluate the inverse-velocity and SLO time of failure prediction methods. Overall, the proposed methodology will improve, and provide a more systematic means of interpreting spatial and temporal ground-based radar data, aiding geotechnical engineers in managing slope movement alarms and alarm responses to provide a safer working environment for mine employees.