C Conferentia Proceedings
COB-2015-0750 Energy and Thermal Sciences

MULTIPLE LINEAR REGRESSION MODEL APPLIED TO A FRAMEWORK FOR INDICATORS ESTIMATE IN MAINTENANCE MANAGEMENT

José Arnóbio de Araújo Júnior1; Robson da Silva Magalhães2; Emerson Felipe Araújo Magalhães3

1 PEI/UFBA/IFAL; 2 Universidade Federal do Sul da Bahia; 3 Instituto Federal de Alagoas

doi:10.20906/CPS/COB-2015-0750

Resumo

The distribution systems of electric utilities in the Northeast are antique and have several types of problems. In an indirect way, these problems can be observed through the collective continuity indicators, provided by ANEEL (National Electric Energy Agency). These indicators are DEC (Equivalent Duration of Interruption per Consumer Unit) and FEC (Equivalent Frequency of Interruption per Consumer Unit). Observing these indicators, we found that most of the energy distribution feeders (Northeast Region), have values above the permitted levels, and as a result, these distributors are required to pay monthly fines. Interruptions of these feeders cause several problems of technical and financial order to their consumers. The objective of this work is to provide for businesses utilities, a support tool for decision making regarding the efficient use of predictive techniques such as thermography. A structure of model using Multiple Linear Regression was developed. The entry of the model is the data of all modes of failure that caused the interruption of the electric power distribution system. The outputs in each model are the estimates of the main quality indexes of a power distribution substation (DEC and FEC). The model obtained presented appropriate to perform the DEC and FEC's prediction of a given month, allowing the concessionaire to act directly on the feeders and the failure modes that contribute to the increase of these indices. Througt the simulation obtained by application of model, it is possible to identify what are the failure modes that can be detected by thermography, which are responsible for the variation of collective indicators of continuity.

Palavras-chave: Modeling, Thermography, Concessionaire of Electrical and Maintenance.

Como citar

José Arnóbio de Araújo Júnior; Robson da Silva Magalhães; Emerson Felipe Araújo Magalhães. “MULTIPLE LINEAR REGRESSION MODEL APPLIED TO A FRAMEWORK FOR INDICATORS ESTIMATE IN MAINTENANCE MANAGEMENT”. 23rd ABCM International Congress of Mechanical Engineering. COBEM2015. 2015. DOI: 10.20906/CPS/COB-2015-0750