Fault Detection Method for Permanent Magnet Synchronous Generator Wind Energy Converters Using Correlation Features Among Three-phase Currents

In a permanent magnet synchronous generator (PMSG) system, conversion systems are major points of failure that create expensive and time-consuming problems. Fault detection is usually used to achieve a steady system. This paper presents a full analysis of a PMSG system for wind turbines (WT) and proposes a fault detection method using correlation features. The proposed method is motivated by the balance among the three-phase currents both before and after an open-circuit fault occurs in a converter of the PMSG system. It is unnecessary to analyze the output waveforms of a converter during fault detection. In this study, two correlation features of stator currents, the mean and covariation, are extracted to train an artificial neural network (ANN), thereby enhancing the performance of the proposed method under different wind speed conditions. Moreover, additional sensors and the collection of a massive amount of data are not required. Model simulations of an ideal inverter and a PMSG system are conducted using PSCAD software. The simulation results show that the proposed method can detect the locations of faulty switches with a diagnostic rate greater than 99.4% for the ideal inverter, and the PMSG drives settings at different wind speeds.