Improving mixing model calibration using scenario simulation in block caving mines

Authors: Villa, D; Soto, N Download Paper Open access courtesy of:

DOI https://doi.org/10.36487/ACG_repo/2645_49

Cite As:
Villa, D & Soto, N 2026, 'Improving mixing model calibration using scenario simulation in block caving mines', in A van As, D Cumming-Potvin & J Wesseloo (eds), Caving 2026: Proceedings of the Sixth International Conference on Block and Sublevel Caving, Australian Centre for Geomechanics, Perth, pp. 1-16, https://doi.org/10.36487/ACG_repo/2645_49



Abstract:
Predicting ore grades and dilution in block caving mines is inherently complex due to the nature of the extraction method, where material flow and cave performance strongly influence outcomes. Numerous factors affect flow behaviour, including cave propagation, preferential flow, differential migration of fine and coarse fragments, draw cone geometry, lateral movement caused by airgap formation (rilling), and surface material displacement from toppling or large-scale failures. Although extensive literature exists on grade prediction and flow modelling, discrepancies between forecasted and actual grades are still common. As a result, mixing models require frequent calibration to reflect new data, inputs and assumptions. This calibration is typically manual and iterative process, often requiring weeks of effort without necessarily improving understanding of the underlying cave behaviour. This paper presents the application of a scenario‑simulation approach in which hundreds of model realisations are generated by systematically varying key inputs and assumptions. Statistical analysis of the results in 2D and 3D outputs enables identification of the parameter combinations that best improve grade forecasts and provides insight into the influence of different mixing mechanisms across the footprint and at individual drawpoints.

References: