Recognition of Cyclostationary Signal Features Using Nonlinear Oscillators

The important engineering problem is monitoring the working state of complex systems such as electric machines. One of tasks is to detect the inductive motor bearing faults by tracking windings currents change. The aim of this study is to analyze the response of nonlinear systems to cyclostationary signals recorded from the three-phase motor. The Duffing oscillator system is studied. The numerical simulation of nonlinear systems with the recorded signals as input is performed. The technique for the analysis of the observed dynamics is proposed. We show that such an approach allows detecting the bearing faults with a high level of confidence.