Flotation stabilization and optimization

It is possible to obtain good performance from a flotation plant but it has proved difficult to maintain such performance. Recovery rates on a flotation plant can be around 90% and often lower, making flotation one of the least efficient processes in the concentration path. Hence, over the last few decades much research and development has gone into the stabilization and optimization of flotation circuits. Flotation is a process with many inputs and complex interactions. The modelling of flotation is reviewed by Mathe et al.1 Some current modelling techniques are explained by Manlapig and Franzidis2, who account for the effects of ‘true’ flotation, entrainment and froth recovery. Edwards and Flintoff3 were of the opinion that mathematical modelling and simulation of flotation systems have not evolved to the same extent as comminution. Technology for the control of flotation circuits has evolved somewhat independently from that of its modelling. McKee4 describes three ‘approaches’ required for flotation control: stabilizing, setpoint and optimizing. McKee5 explains that simple proportionalintegral controllers do not provide acceptable responses. Ding and Gustafsson6 describe the application of a multivariable control strategy for intended application on an industrial flotation plant. They used a linearized model and a linear quadratic gaussian controller. Their results of simulated control are good. Van Deventer et al.7 suggest that neural networks have provided good scope for the analyses of froth characteristics as functions of their dynamically measured images. Brown et al.8 describe advances in the use of froth imaging technology in process control, on an industrial plant, with improvements in metallurgical importance. It seems that imaging technology is capable of leading to successes in control, but it does also need a lot of work to relate the extra variables it introduces to underlying variables of importance, such as relative flotation rates, entrainment and optimum flow distributions. Mintek has been very active in the industrial implementation of advanced flotation control for over a decade. Stabilizing level control by its control system FloatStarTM (Schubert et al.9) has been applied to about thirty flotation circuits, some applications of which are described by Henning et al.10, Singh and Schubert11, and Muller et al.12. This paper includes Mintek’s recent industrial application of optimizing control, which is being Flotation stabilization and optimization