Resumo:
This thesis presents a methodology, based on techniques of artificial
intelligence, especially of specialist and united systems approximate, to aid in the process of
electric outlet of decision. The central idea was to develop a group of rules that you/they can
aid the operators during the analysis of occurrences and in the system recompose of the net
of ELETRONORTE. For this, it was developed a methodology that uses besides an own
knowledge base, the integration with the systems INFO_OPR, SAP/R3 and SAGE, already
available in the concessionary. It was also become pregnant and implemented a system
extractor of knowledge of great bases of data that works together with a specialist system.
This hierarchical intelligent system works routines external computations of support, as well
as bases of existent data in the company.
To reach this I aim at the work it was divided in stages:
• Integration of systems of operation information and maintenance, where she looked
for to centralize and to optimize the necessary historical information to the formation
of the base of data.
• Extraction and storage of the existent knowledge of the integrated base, through an
algorithm based on the Theory of the Approximate Groups.
• Monitoring of the new events through the execution in time-Real of the rules of the
knowledge bank.
• Analysis of the new events under the focus of the rules of the knowledge bank.
• Consolidation of the results through the structuring and the calculation of the
indicators of acting of ELETRONORTE. These acting indicators are governed by the
manuals and rules of consolidation avaliable by the National Operator of Systems -
ONS.
• This system is in operation in ELETRONORTE.