Prediction of the Dynamic Behaviour of Iron and Sulphur During Nickel-Copper Matte Conversion Using Semi-Empirical Models

Abstract The Ausmelt ® converter is being introduced into South Africa by Anglo Platinum Ltd. as an alternative to the traditional Pierce Smith converter process to convert a nickel copper matte containing platinum group metals (PGM's). The control of iron and sulphur endpoints during the matte conversion process is of paramount importance in the recovery of the PGM's. Linear systems models were used to model the dynamic behaviour of sulphur and iron compositions in the matte. The models characterised the dynamic behaviour of sulphur and iron well and can be used for model predictive control. A backpropagation feedforward neural network was also trained on equilibrium data generated using the FACT database. Comparison of actual and equilibrium values could then be performed continuously.