COBEM99

SR21 Redes Neurais, Algoritmos Genéticos e Lógicas Nebulosa/Neural

Title:

INTELLIGENT DIAGNOSIS OF MALFUNCTION ENGINES - DIAM-

Summary :

ABSTRACT: IN RECENT YEARS THE NEW TECHNOLOGY OF GENETIC ALGORITHMS (GA) AND ARTIFICIAL NEURAL NETWORKS (NN) HAVE BEEN SUCCESSFULLY APPLIED TO COMPLEX PROBLEMS OF OPTIMISATION. IN THIS PAPER WE PRESENT AN EXPERT SYSTEM CALLED DIAM TO AUTOMATICALLY DIAGNOSE POSSIBLE CAUSES FOR MOTOR MALFUNCTIONING. THE SYSTEM WAS DEVELOPED WITH THE CONCERN OF BEING A USEFUL TOOL IN THE SPECIALISED AUTOMOBILE WORKSHOPS. THE SYSTEM WAS DESIGNED AND DEVELOPED IN INSTITUTO SUPERIOR DE ENGENHARIA DO PORTO (ISEP) IN COOPERATION WITH THE ENTERPRISE SALVADOR CAETANO (TOYOTA). EXPERT KNOWLEDGE IS NECESSARY IN THIS DOMAIN DUE TO THE COMPLEXITY AND DIVERSITY OF POSSIBLE PROBLEMS OCCURRING IN AUTOMOBILE ENGINES. THE GOAL OF THE GAS IS TO TUNE THE NN USED IN THE EXPERT SYSTEM. THE JOINT USE OF GAS AND NNS, CALLED GANN, ENABLES TO PROFIT FROM ADVANTAGES OF BOTH GAS AND NNS. THE COMBINATION OF GAS AND NNS IS USED, WITH ADVANTAGE, IN THE DETECTION AND DIAGNOSING OF AUTOMOBILE MOTOR FAULTS. THE SYSTEM IS CURRENTLY IN THE TESTING PHASE (IN ISEP) AND WE ESTIMATE THAT IT WILL SOON BE TESTED IN A REAL AUTOMOBILE WORKSHOP. KEYWORDS: GENETIC ALGORITHMS, ARTIFICIAL NEURAL NETWORKS, TIME SERIES FORECASTING

Author :

da Silva, Antônio J. F.
Ferreira, Fernando J.

Paper View :