Digital Ore Processing Simulation based on Image Characterization Techniques
Mineral system understanding
Mr. Javier Merrill
University of Tasmania
The current standard for metallurgical testing, whether it is for plant equipment sizing or operational control, is based on conventional practices that do not always lead to optimal setups. Typically this is due to the uniqueness of individual orebodies. Operational parameters such as comminution-targeted grain size, flotation cell pH, pulp and tailing solids concentration, are often selected by simplicity of their control/measurement, rather than because of its impact upon the desired behaviour. Recently, characterization techniques for images have become widely applied for research purposes, but their implementation in processing workflows is minimal.
In this study we assess the use of a mineral-oriented edge-detection on hyperspectral images to simulate rock breaking preferential pathways, generating a set of particles that can be fed to a flotation simulation. Using the simulation results, a targeted comminution grain size is calculated aimed to generate optimal liberation for the flotation process, at the same time estimating environmental impact through acid rock drainage potential of the tailings phase, generating the link between common plant parameters and phenomenologically-relevant rock properties obtained through this novel imaging characterization technique. Here, we present an example and a case study using imaging of rock from Chilean Cu-porphyry deposits.