Detection of mixed eccentricity fault in doubly-fed induction generator based on reactive power spectrum

Wide application of wind energy in the world towards reduction of global warming, have made the low cost maintenance of these turbines very important. Basically, mechanical faults such as gear-box faults in these turbines lead to a long down time and consequently heavy financial loss. Generally, these faults cause eccentricity in the generator air gap, so by its early detection the fault can be largely prevented. In this paper, eccentricity fault detection is studied in doubly-fed induction generator (DFIG) which is the most established generator in wind turbines and also is more prone to mechanical faults compared to other technologies. One of frequency components of reactive power is introduced as eccentricity fault index. As the reactive power is always evaluated in the control system, there is no need to include a new hardware for fault detection. Finite elements method is used to model the wound rotor induction machine (WRIM) which is coupled with closed-loop control system. Following the validation of the model in the healthy case experimentally, eccentricity under different operating conditions is included in the modeling. Finally, a simple procedure based on simple signal processing methods is introduced for eccentricity fault diagnosis considering the operating conditions of the wind turbine.