A series of large-scale laboratory experimental work was performed to demonstrate the failure characteristics of rock cores confined with backfills. The 136 mm diameter core specimens, representing model pillars, were confined with "soft backfills", tailings of varying cement content, in a 10 mm thick and 30 cm diameter aluminum cylinder. The complete load-deformation responses were recovered from simultaneous compression on composite pillar-backfill specimens. This paper presents the preliminary results obtained on failure characteristics of model pillars and the interaction of pillar-backfill under confinement and evaluates the post-failure behaviour of pillars. The relationship between the residual strength of pillars as a function of stiffness of confining fill material indicates that the residual strength progressively increases with increasing cement content in backfill. The pillars confined with stiffer fills have an average residual strength of about 70% of peak strength, while the pillars confined with leaner fills and uncemented tailings show lower residual strengths, ranging from 24% to 55% of peak strength. The experimental data also show that changes in the peak strength of rock pillars may only be partially related to the degree of confinement provided by fill material. The failure mode of pillars confined by fills with low cement content or uncemented fill is semi-violent and violent, while the pillars confined with stiffer fills fail non-violently.
Experimental studies have demonstrated that "soft backfills" (cemented tailings) used as confinement of rock pillars act as passive lateral support in room-and-pillar-mines. The backfill confinement provides significant ground stabilization limiting the advance of the failure zones and, by controlling violent pillar failure in brittle rock masses, alleviates seismic activities in rockburst-prone underground excavations. Small-scale laboratory tests conducted on core specimens surrounded by various soft fills have shown the effects of backfills on post-failure behaviour of pillars (Blight and Clarke 1983). Similar tests were repeated, to evaluate the failure characteristics of confined and unconfined core specimens. Tests were performed on "tight fill", and "fill with gap" (Swan and Board 1989). In the present laboratory tests larger diameter rock core specimens were used as "model pillars" confined with various types of cemented soft fills. An aluminum cylinder provided confinement for the composite rock-fill test specimen. This paper represent the data obtained from these large-scale laboratory tests and evaluates the experimental results.
The laboratory experiments performed comprised the following testing material, preparation and procedures:
Core specimens representing the pillars were composed of andésite rocks with quartz stringers and chlorite filled fractures obtained from the testholes drilled for stress measurement purpose at an underground mine (Arjang 1987). The average determined rock properties were: Elastic modulus 85.0 ±1.3 GPa, Poisson's ratio 0.20 and uniaxial compressive strength of 193.0 ±42.0 MPa.
The tailings material was obtained from the belt filters of a uranium mine after the leaching stage. The unclassified pyritic tailings consisted of a mixture of sand and silt (50% less than 150 µm) with a bulk density of about 65% solid at a moisture content of approximately 40%.
