ANFIS based prediction of the aluminum extraction from boehmite bauxite in the Bayer process

Abstract

This paper presents the results of nonlinear statistical modeling of the bauxite leaching process, as part of Bayer technology for alumina production. Based on the data, collected during the year 2011 from the industrial production in the alumina factory Birač, Zvornik (Bosnia and Herzegovina), nonlinear statistical modeling of the industrial process was performed. The model was developed as an attempt to defi ne the dependence of the Al2O3 degree of recovery as a function of input parameters of the leaching process: content of Al2O3, SiO2 and Fe2O3 in the bauxite, as well as content of Na2Ocaustic and Al2O3 in the starting sodium aluminate solution. As the statistical modeling tool, Adaptive Network Based Fuzzy Inference System (ANFIS) was used. The model, defi ned by the ANFIS methodology, expressed a high fi tting level and accordingly can be used for the effi cient prediction of the Al2O3 degree of recovery, as a function of the process inputs under the industrial conditions.