Deteção de falhas em motores elétricos através da classificação de padrões de vibração utilizando uma rede neural elm

Abstract

The condition monitoring of industrial electric motors provides information to help planning maintenance interventions before the occurrence of failures. This paper proposes a new approach for monitoring the operational condition of three-phase induction motors based on the extraction of characteristics of a signal obtained with MEMS accelerometers. The data extracted from the decomposition of the vibration signal using Haar Transform and the fractal dimension are used to train a ELM neural network. The results of our experiments demonstrated the feasibility of the proposed methodology in detection and identification of mechanical and electrical failures