Probabilistic methods for flotation circuit mass balance estimation

Abstract:

Varying feed conditions, operational challenges, changing mineralogy and suboptimal reagent dosing can cause abnormal and undesirable events that destabilize and reduce flotation performance. An industrial case study (copper concentrator) on the use of real-time pulp measurements to detect abnormal events and aggregate their impact on flotation performance to support early intervention by metallurgists is presented. Since mechanisms of flotation in the pulp and froth phases are complex, multivariate, and sensitive to disturbances, the availability of real-time pulp bubble measurements can provide valuable information on flotation performance to allow timely corrective actions. Online pulp sensors in a copper concentrator plant were validated through step tests, and historical data analysis shows the potential of flotation performance monitoring with these sensors. A pulp performance monitoring dashboard has been deployed to the copper concentrator. Events detected include sub-optimal frother dosing and abnormal mineralogy (e.g., carbonaceous material).