Simplified architecture of a neurofuzzy controller applied on industrial systems for fluid flow control

Abstract

This work presents a simplified architecture of a neurofuzzy controller for general purpose applications that tries to minimize the processing used in the several stages of hazy modeling of systems. The basic procedures of fuzzification and defuzzification are simplified to the maximum while the inference procedures are computed in a private way. The simplified architecture allows a fast and easy configuration of the neurofuzzy controller and the structuring rules that define the control actions is automatic. Th controller's Limits and performance are standardized and the control actions are previously calculated. For application, the industrial systems of fluid flow control will be considered.