An Operator Oriented Flotation Model

"A kinetic model was developed for the conditioning and flotation processes of an industrial lead concentrate dezincing circuit. The models were incorporated into a flowsheet simulator. Simulation results demonstrate the advantages of split conditioning of coarse and fine particles and the potential of this approach for the production of the heuristics required for the development of intelligent control strategies.The conditioning model relates the proportions of activated and hydrophobic surfaces and residual reagent concentrations to activator and collector reagent flowrates. The model accounts for the solids and water flowrates, and for the solids grade and size distribution. The flotation model uses the predicted proportion of hydrophobic mineral surface, the aeration rate and the froth depth to estimate cell, bank and circuit performances. IntroductionFlotation has been the object of modelling and simulation studies for more than two decades. The process is usually modelled as a kinetic one in which the particle flotation rate constants are key parameters. Rate constants are related to the operating variables (pH, reagent flowrates, aeration rate, etc), however these relationships are seldom established in the literature.In spite of the large amount of literature on flotation modelling and simulation, few papers deal with the interaction between flotation rate constants and operating variables. This important gap, which was identified in the early seventies by Woodburn (l970), still exists. One of the first attempts to include operating conditions in a flotation simulator was made by King (l973), who explicitly introduced the aeration rate into an air bubble loading model. Lynch et al. (1981) used an empirical model to relate the proportion of slow floating particles to the ratio of collector per ton of ore. Such empirical relationships are difficult to calibrate and cumbersome to use when an activator and a collector are added to slurries in which particle size distribution and surface characteristics are changing. Eascur and Herbst (l982) relate in a fundamental way the mineral flotation rate constants to the aeration rate, froth depth, agitation and frother concentration, however the model does not account for reagent flowrates. Niemi et al. (1982) have proposed a model for the conditioning of minerals and included it into a conventional flotation simulator."