A MILP model for design of flotation circuits with bank/column and regrind/no regrind selection

Abstract A method for the design of flotation circuits is presented. The design problem is represented by several superstructures. The first superstructure represents separation tasks (STS), which include: feed processing superstructure (FPS), concentrate processing superstructure (CPS), and tail processing superstructure (TPS). The FPS commonly uses a single stage, i.e., rougher. The CPS represents the circuit needed to carry out the cleaner task, and the TPS represents the circuit needed to carry out the scavenger task. These superstructures are flow networks between several separation stages. In each separation stage two kinds of cells are allowed, bank and column. In several streams in the CPS and TPS, the incorporation of regrind mills is also included. The optimal selection of the circuit is made with an appropriate objective function, upon which the values of the operational and structural variables may be determined. The problem is formulated using disjunctive programming, which is converted to a Mixed Integer Linear Programming (MILP) problem. The model includes mass balance, equipment models, operational conditions, and logic relationship. The approach is illustrated for a copper concentration plant.