An improved model for simulation of flotation circuits

Abstract Batch testing of pulp samples from various streams in a flotation plant has made it possible to model plant behaviour more accurately. A new and very flexible model is introduced for characterizing the floatability of plant feed in terms of a continuous distribution of rate constants. This model, which is the sume of two normal distributions, provides a more accurate means of modelling a complete set of data from a flotation circuit, particularly the middlings. A new froth model is also introduced which relates the recycle of particles from the froth to the pulp under plant conditions. The models were fitted to published data from Cominco's Red Dog lead cleaning circuit and were compared to conventional discrete models. The recovery predicted by the models differed by up to 2%.