The hidrogeradores are keys-parts in the Brazilian circuit of energy, its non-availability and bad functioning can causes highest fines for the concessionaires applied for the ANEEL for not taking care of the demands and for improper to accomplish maintenances, beyond aggravating trustworthiness in the guarantee of the supply to the final consumers. To guarantee that this does not come to happen, the preditive maintenance supplies techniques that can point the imperfections analyzing the “vital signals” originated by the proper equipment. Thus the mechanical and operational conditions periodically are monitored and when the trends are detected unhealthy, the bothering parts in the machine are identified and programmed for maintenance. So that this trend is found, logic fuzzy is used of the shape the behavior of the hidropower, being more specifies: “mancais”,“estator”, “anel coletor”, inferring probable conclusions of imperfections. In this work, she shows the process of construction of the system that assists diagnosiss in it of the preditive maintenance, since its methodology of development for “macro-activities”, architectural definition, conformity of the requirements and analysis of the inserted knowledge the intelligence of the system. The system was developed in platform labview to serve as support tool. All the inserted knowledge in the system was gotten with the body specialist of Eletronorte and another part in literature, was necessary to apply the concept of rules to it schemes of inference fuzzy, for a linguistic form of easy understanding, so that the proper specialists extend and evolve software.
|
