Diagnóstico de Curto-Circuito entre Espiras de Estator de Motores de Indução Usando Fluxo de Dispersão e Dados Estatísticos

Abstract

This work proposes the investigation of stator short circuit fault (SSCF) in Three Phase Induction Motors (TIM) related to stray flux. The level of defect is analyzed by means of an exploratory coil inserted in the motor housing, whose function is to sample the stray flux. Obtaining the voltage signals induced in the coil, it is possible to extract characteristics that can help in the automated identification of SSCF in TIM’s. In this way, considering the need to identify these faults in an incipient, automatic way and considering the importance of finding relevant attributes for the adequate characterization of these faults, this article presents an attribute engineering method aimed at fault identification, presenting the process of data acquisition, extraction, selection, reduction and evaluation of statistical attributes to obtain an accurate diagnosis. The objective is to choose only potentially relevant attributes based on statistical information for the identification of SSCF, allowing the implementation in embedded systems with small computational capacity. The results show a reduction of 57 attributes initially extracted to 8, guaranteeing an accuracy of 86% in the identification of failures.