ABSTRACT:

The process of modelling and design of a yielding pillar is a common necessity in rock mechanics as applied to deep underground mines. There are numerous reasons to design a pillar to yield. A large-scale example would be the convergence of two mining fronts, with the goal of managing the yielding of the merge zone and minimizing worker exposure to the yielding stress front. A stope-scale example would be in the typical primary-secondary stope mining sequence, where secondary stopes are intentionally left behind to support an ever-increasing stress-load. As the leading primary stopes are mined, the secondary stopes eventually yield, which are then mined under relatively low-stress, aseismic conditions.

In this analysis, a small-scale example of a yielding pillar will be modelled and compared to results from three case studies. When preparing a stope for production blasting, it is common to take long-wall slashes in the bottom sill. This is generally intended to open the bottom of the stope to its full floor area, increasing the void available for the initial production blast, and allowing ITH drill holes to break-through into the back, which can then be surveyed. These surveyed breakthrough holes allow for greater control and quality assurance during production blasting. Finally, the long-wall slashes increase the recovery rate for the stope block since the blasted walls represent ore that would otherwise be left behind as a pillar.

All three case studies presented are taken from Vale's Copper Cliff Mine, at a depth of 5,130 ft. The rockmass is quartz diabase of approximately 138 MPa uniaxial compressive strength and with blocky structure, but good quality (Approximately 90 rock quality designation). The long-wall slashes range in their cut depth between 10-14 ft and are blasted against pre-existing 20-22 ft stable rib pillars that separate the parallel orebody crosscuts. All three case studies are from stopes in the same area of the mine and therefore share a similar in-situ primary stress of 70 MPa. However, the overall geometries are different in each case. A Map3D model based on a linear-elastic constitutive model was used to characterize pillar average stress. The pillar dimensions were then compared using the Lunder & Pakalnis (1997) empirical method. All three cases have been discussed with reference to these two design methodologies and the actual pillar behavior in each case is compared to those predicted results.

The results show that the employed design methodologies are adequate for the purpose of this application and accord with site experience regarding pillar performance in this small-scale scenario.

This content is only available via PDF.
You can access this article if you purchase or spend a download.