Abstract

An L-150 pilot Jameson flotation cell was instrumented and a distributed control system was developed. The parameters of a metallurgic phenomenological model were estimated from industrial data. A steady state simulator was built based on this nonlinear model. This hybrid system combines on-line measured operating variables with virtual variables, characterizing the feed. All these variables are fed on-line to a simulator to predict the characteristics of the concentrate and tailings. The expert system modifies the set points of the distributed control system, including two routines: expert feedback and feed forward control. Several cases for different feed conditions are discussed.