Incorporating Liberation Information in Flotation Models

Flotation of a mineral bearing particle is the result of successful completion of a chain of events, such as, producing particles with sufficient exposure of the valuable mineral on the surface, reagent adsorption, suspension of the particle in the pulp, collision of particle with air bubbles and subsequent attachment, ascending to the froth column and transport to the concentrate stream, among others. These sub-processes in turn are affected by the feed properties such as size and density, and process variables such as size of air bubbles, degree of turbulence in the cell, particle/particle interactions etc.. Thus, inefficiencies occurring in any one or more of the above processes will result in inefficient flotation performance. In spite of its dependency on a wide variety of sub-processes, flotation performance is commonly modelled based on first order kinetics using two or three parameter models. As the model parameters may be affected by several of the sub-processes mentioned above, interpretation of their variations and identifying the specific causes is extremely difficult. Of the many factors that influence the flotation performance, liberation information of the particles such as the grade distribution and surface exposure are considered vitally important. In this paper, it has been demonstrated a) how a predictive liberation model may be derived and used in quantifying the liberation information of the particles produced at various grind sizes and b) how the flotation process may be modelled by explicitly incorporating the liberation information, thereby improving the accuracy of the model. Comparisons of the observed data with the proposed model and those of published liberation/flotation models have also been made.