FAM on grain scheduling control of highly nonlinear and disturbed processes

Abstract

This paper describes an adaptive computer-based controller by gain scheduling strategy on the basis of associative memories for the control of nonlinear disturbed processes. The proposed algorithm, search for a rule-base during a training phase for which is applied a pattern input/output signal by means of a set of well adjusted regulators distributed along the full operating range (from 0 to 100/% input reference). After such phase the obtained rule-base will be processed by classical fuzzy methods in on line feedback control.This algorithm is highly efficient for each disturbance under different operating conditions. Good results are obtained by computer simulation using the simulation language Simnon as well as in real time computer control on a training board.