Fault detection for stator inter-turn short circuit in doubly fed induction generators using adaptive observers

This paper presents an observer-based fault detection approach for inter-turn stator short circuit fault in doubly fed induction generators (DFIG). A state-space model is developed, which can represent both the healthy and faulty conditions. Based on this model, two different adaptive observers: asymptotic adaptive observer and exponential adaptive observer are constructed, which are both able to estimate the fraction of short circuited winding online. The case when the rotor speed is immeasurable is also considered in this paper. The proposed exponential adaptive observer is reformulated into a new observer, which is capable of estimating the fault level and the rotor speed simultaneously. The simulation results demonstrate that this approach has effectiveness on fault detection under both steady state operation and transient state operation. It has strong robustness against the speed and load variations. This advantage has significance on fault detection of wind turbine DFIGs, since the DFIGs are predominantly under varying speed operation.