A real time neuro-fuzzy production system for ITAIPU Binacional

The paper presents the development of an expert system modeled by using connectionist and fuzzy paradigms for fault detection in the Itaipu Hydroelectric company. For this purpose, a rule based knowledge system already in use by Itaipu operation has been taken as reference. The main aim was the development of the NEUFI Production System (NEUro-Fuzzy for Itaipu), whose knowledge base contains rules modeled by the neuro-fuzzy architecture. To make the implementation of the NEUFI Production System easier, a simulator, entitled Neuro-Fuzzy Simulator, was also developed, which allows training and writing of neuro-fuzzy networks