Faults Detection and Localization in a Connected Wind Turbine

Like any other industrial process, a wind system can be subject to various failures. Defects can occur at the level of electrical and mechanical systems or at the sensor level, which reduces the performance of the wind system. An accurate and consistent evaluation of wind system performance is crucial for sustainable development as it is considered an essential criterion for energy quality. In this paper, we utilized the statistical method PCA (Principal Component Analysis method (PCA) without a priori model for the detection and localization of faults in a wind system connected to the grid using a permanent magnet synchronous generator (PMSG). The proposed technique is a method that utilizes data from the implicit model of the machine. The choice of the number of components to retain is crucial for the PCA model. Different approaches for determining this number are proposed. Thus, to validate this method, graphs with real data are presented, followed by the diagnosis, detection, and localization of faults using the residuals of the state variables. The detection index Di is proposed for the detection and localization of machine faults.