Review of flyrock incidents in surface mining and the limitations of current predictive models

This paper will review several flyrock incidents that occurred recently in the South African mining industry to highlight that potential difficulties of the historical approaches to the problem of flyrock based on correlation studies and regression analysis, including artificial neural networks and similar techniques. These process are not capable of addressing two core issues – root causes of flyrock and projection velocity. A further shortcoming of correlation techniques is that they give no information on the influence of rock size and shape on the flight distance. The scaled depth of burial model for crater blasting in the collar zone and bench face does not specifically address the question of flyrock velocity. A third approach, based on flight trajectory calculations, often neglects the very significant effects of air resistance on the trajectory. Some trajectory models incorporate air resistance but use an implausible fragment velocity model that cannot propel sizeable rocks to distances much beyond 150 m. (Approximately 500 ft.) Nonetheless, trajectory calculation incorporating the effects of air drag affords the most promising approach to the prediction of flyrock range. A unique and insightful feature of the proposed realistic flight modelling is that it collapses all suspected causes of flyrock, many of which are not well understood, to just a single parameter – the launch velocity. This indicates that the root causes of flyrock lie in the mechanisms of momentum transfer to broken rock and suggests new avenues of study for the explosives industry. In this paper we use the trajectory model to identify certain conditions on the bench whereby the impulse of gas pressure acting on fractured material on or close to a free face is enhanced, resulting in a greater transfer of momentum to potential flyrocks.