Open Cut » General
Systems to detect, monitor and analyse rock falls in open pit mining operations have the potential to improve operational safety, improve calibration of rockfall simulators (restitution coefficients), and importantly provide quantitative data to justify current standoff designs. Although several technologies have potential to assist in the monitoring and analysis of rockfalls and their trajectories, currently no effective systems exist.
A monitoring system that can accumulate a large database of rockfall events across the full strike length of highwalls is required. The use of existing monitoring systems (radar, vision, lidar) as well as knowledge of the wall geometry, geology (rock mass types) and structural characteristics (defect orientations and intensities) has potential to support detection of events and detect rockfall movements accurately enough for determination of trajectories (bounce kinematics), impact locations and final resting positions.
To achieve this goal, CSIRO undertook a scoping study to define the technical requirements, specifications and system design for a fused sensor system to support the analysis of rockfalls on highwalls. The technologies required included computer vision in combination with new rockfall specific radar systems, and acoustic (seismic) monitoring. CSIRO also undertook benchtop analyses and initial field-based feasibility studies to progress this development.
The technical requirements analysis was conducted which identified that the system must be capable of detecting small (~10cm) rocks with full 3D trajectory estimation. This analysis then informed the development of the system design, which incorporated the fusion of the aforementioned sensors and databases. Radar companies currently deploying or developing rockfall monitoring radars were identified as a potential vehicle to deploy this system.
The benchtop assessment of vision, radar and microseismic sensing modalities determined that these technologies were well suited to meeting the previously defined requirements, when combined in a formal sensor fusion framework.
Previously recorded video of a highwall failure was used to determine the feasibility of computer vision algorithms to detect small rockfall events. It has been demonstrated that even moderate quality standard video footage can be processed to precisely detect rockfalls.
CSIRO based field trials to demonstrate feasibility of a microseismic monitoring network's sensitivity to rockfall events were performed. The results were very encouraging with small (10cm) size rockfall impacts being detectable by geophones around 30m away.
Several potential issues with implementation have been discussed, including radar sample rates, microseismic sensor ground-coupling and impairment to video analysis from rockfall related dust generation. Mitigation of these issues has been discussed and the fused sensor approach outlined as a way of addressing the limitations of any particular sensing modality.