The evolution of technologies, the progress of power electronics and economic stakes, the use of rotating machines is becoming increasingly important in all fields, including electrical drives and energy production. Their robustness, reliability and low cost are particularly appreciated. These generator-mode machines are the basis for the wind turbines of today. However, despite all the research and improvements that have been made, these machines still represent potential sites for stator and rotor failure. These faults need to be detected and located in good time, as they can cause serious damage to the system. In recent years, the diagnosis of faults affecting wind energy systems has been extensively studied. A large number of methods are available to improve system design, increase power quality and output, and reduce costs. Despite the results of this research, these multi-complex systems are still the subject of numerous laboratory research, both industrial and academic. The aim of this paper is to use knowledge-based methods, these methods do not use mathematical models to describe cause-and -effect relationships. The only knowledge is based on human experience through feedback. The aim is to generate healthy and faulty operating data for state variables and to use an advanced approach, based on deep learning, to detect and locate faults through stator current analysis.