Diagnóstico de falhas em máquinas elétricas rotativas por meio de técnicas de Análise Modal Operacional

Abstract

The interest in systems for monitoring and detecting faults in rotating machines has become the scope of several studies due to the wide applicability of this type of machinery in industry and the quest to promote safer and more efficient operating conditions. The improvement for predictive analysis of failures of this equipment impacts on financial and operational bonuses so that the useful life of the machinery is optimized. Research in the field of maintenance and reliability addresses the use of engine vibration analysis to predict mechanical defects such as imbalance, misalignment, component failure, among others. The methodology addressed by this work proposes an alternative of predictive analysis based on models identified through vibration data in order to predict such mechanical failures. Thus, Modal Operational Analysis techniques such as Ibrahim’s Time Domain Method (ITDM) associated with the Random Decrement (RD) algorithm are used to describe different operating conditions of the machines. From the identification of the models, it is possible to classify the state of the machine, providing a predictive action to possible failures.