Aplicação de análises de vibração para detecção de defeitos em rolamentos do motor de indução

Abstract

The induction motor is widely used in the industry and monitoring it is essential to avoid emergency stops caused by the aggravation of defects that arise in its components, of which the bearings are the most susceptible. This work consists in the study of methods of vibration analysis to detect bearing defects and that will be applied to the vibration signals collected by two accelerometers, one on each side of an induction motor that has been tested several times with normal and defective bearings. The methods applied are spectral, envelope and cepstral analysis, the latter two being used to identify defects in rolling bearings in predictive maintenance in order to take advantage of the component's useful life and plan the intervention to correct the anomaly, avoiding thus, losses due to an unexpected stop of the machine and high maintenance costs. The analyzes were performed using the MATLAB® software and compared, verifying which methods are applicable for the identification of defects in the induction motor bearing under different operating conditions. It was found that the spectral and envelope analysis were not adequate to identify the defect in the bearing ball. The cepstral analysis, on the other hand, proved to be applicable in both load conditions in the signals collected by the accelerometer on the same side of the defective bearing, and in the accelerometer signals on the other side, only in the condition close to the rated load of the motor.