%0 Conference Paper %A Ghazvinian, E. %A Fuenzalida, M. %A Orrego, C. %A Pierce, M. %D 2020 %T Back analysis of cave propagation and subsidence at Cadia East Mine %P 535-550 %E R. Castro, F. Báez & K. Suzuki %C Online %8 9-11 December %B MassMin 2020: Proceedings of the Eighth International Conference & Exhibition on Mass Mining %X Cadia East panel cave mine is one of the three mines comprising the Cadia Valley Operations, one of Australia’s largest gold mining operations, located in New South Wales. Production at Cadia East commenced in 2013 from Panel Cave 1 (PC1) with its extraction level set at a depth of 1,200 m below the surface. Currently, ore is extracted from PC1 and its neighboring panel (PC2) with an extraction level at 1,400 m depth. A critical component of effective production for Cadia East is successful cave propagation to the surface considering significant cave heights. To ensure cave performance, the rock mass immediately above the footprint of PC1 and PC2 was pre-conditioned by means of hydraulic fracturing (combined with blasting for PC1). Furthermore, a surface hydraulic fracturing program had to be instigated in the hard and competent near surface rock mass above PC1 to assist cave propagation through the final 450 m and its breakthrough to the surface. This paper discusses the significance of using a strain-softening model with the ability to capture the correct mechanics of rock mass frictional strength mobilization for a precise back-analysis of cave performance. IMASS (Itasca Model for Advanced Strain Softening) was used for this study in FLAC3D. IMASS uses two-mode softening yield surfaces. The first residual envelope represents the post-peak strength, and the second residual envelope represents the ultimate rock mass residual strength. The two-mode softening allows for mobilization of high apparent friction angles at low confinement when the blocks are formed in the rock mass. This, in combination with implementation of pre- and post-conditioning of the rock mass (hydraulic fracturing program) in the simulation, were critical in the successful calibration of the model. %1 Santiago %I University of Chile %U https://papers.acg.uwa.edu.au/p/2063_36_Ghazvinian/ %R 10.36487/ACG_repo/2063_36