A hybrid approach to modeling the flotation process from the 'Veliki Krivelj' plant

Abstract

The purpose of the flotation process control is optimization the concentrate grade and recovery, while maximizing profits. Consequently, the research into modeling and control of this process has always been an important area in control engineering practice. This paper presents the results of development and validation the predictive models, based on the ANFIS hybrid system. Models predict the values of copper concentrate and tailings grade as well as copper recovery in the flotation plant 'Veliki Krivelj'. The copper content in the feed ore, collector consumption in the rough flotation stage and consumption of frother, were selected as the independent variables. Other technical and technological parameters, relevant for the process of flotation concentration were considered constant. The results of the models validation have showed that the models provide the good predictions of changes in the copper concentrate grade, while the predictions of changes in the copper recovery and tailings grade are somewhat poorer.

Keywords

• flotation
• model
• anfis
• copper
• concentrate
• tailings
• recovery