Diagnosis of Broken Bars in Wind Turbine Squirrel Cage Induction Generator: Approach Based on Current Signal and Generative Adversarial Networks

To ensure the profitability of wind industry, one most important objectives is to minimize maintenance costs. For this reason, components turbines are continuously monitored detect any type failure by analyzing signals measured sensors included in condition monitoring system. Most proposals for detection and diagnosis faults based on signal processing artificial intelligence models use a fault-free acquired system which fault has been provoked; however, when failures incipient, frequency associated with very close fundamental component there incomplete data, difficult. Therefore, purpose research diagnose electric generator operation, using current applying generative adversarial networks obtain synthetic data that allow counteracting problem an unbalanced dataset. The proposal useful broken bars squirrel cage induction generators, which, according control system, were healthy state.