Reagent feed intelligent setting based on belief rule base for antimony roughing flotation

The reagents' feeds are the most important operating variable in the process of antimony flotation. It is mainly regulated by the experienced operators. Due to the complicated nonlinear relationships between the antimony roughing flotation process and reagents' feeds, as well as the uncertainty in the industrial data and on-site conditions, this paper proposes an intelligent setting method for antimony roughing flotation reagent feed based on belief rule base (BRB). This method, by utilizing a large number of historical data in the flotation process, firstly the relationships between the working conditions, froth image features and reagents' feeds are summarized into expert rules in the form of BRB structure. Then through the parameter learning model, the parameters of BRB are optimized. Finally, by using evidential reasoning algorithm (ER) the feed values of the reagents are obtained. The experimental results show that the BRB system can describe the uncertain information well and improve the precision of the reagents' feeds.