An application of Artificial Neural Network for prediction of densities and particle size distributions in mineral processing industry

This paper demonstrates an application of Artificial Neural Network {ANN) for determination of undedlow and ovedlo w densities of Hydrocyclone separators. The discussions are extended and further results are presented for the prediction of particle size distributions in the underilow and ovedlow streams. The fit the experimental results against the predicted results are illustrated and statistical analysis are made. Jt is shown that once the history of the operations are known, ANN proves to be a useful tool for predicting future separation eHciencies. This approach has a potential to eliminate the need for installation of expensive on-line instruments for density measurements and particle size analyses. This approach can be applied in similar situations in mineral processing industry.