DIAGNOSTIC SYSTEM OF FAILURES IN BEARINGS OF ROTARY ELECTRICAL MACHINES BY MEANS OF SPECTRAL CURTOSE AND QUADRATIC ENVELOPE OF STATORIC CURRENT

摘要

"Failure diagnosis system for rolling bearings of rotating electric machines through spectral kurtosis and quadratic envelope of the stator current". The present invention, pertaining to electronics, describes a system capable of predictively detecting mechanical failures of a localized and incipient nature in bearings of alternating current three-phase electric motors (1), in order to avoid unscheduled production shutdowns. The method used for the detection of mechanical failures in electric motor bearings (1) is based on the analysis of electrical quantities, especially on the electric motor statistic current (1), the time domain and the frequency domain, considering that disturbances in torque and air gap affect the electric motor's statistic current (1). In this way, the motor is used as a transducer to evaluate the bearing condition. Once the motor electrical signals have changed, specific fault patterns can be identified and predictive bearing condition monitoring can be performed. Are the techniques used to analyze the currents that feed the electric motor (1) based on the analysis of the motor current signature? mcsa, spectral kurtosis and quadratic envelope analysis.