Using dynamic state estimation to detect loss of excitation in synchronous generators

Abstract

AbstractOne of the common methods to detect Loss of Excitation (LOE) is the impedance‐based technique. But it is too slow, and may show incorrect performance when faced with power system disturbances like stable power swings (SPS). To improve the aforementioned problems, a new index is introduced based on the change of transient states of the synchronous generator which are estimated by Extended Kalman filter (EKF). The principal idea of the proposed method is that the angle between the transient voltage and q‐axis internal voltage phasor will be increasing during LOE events. Therefore, this angle and its derivative are used. Estimation of the internal variables of the generator is used to achieve the angle between transient voltage and internal voltage phasor. To evaluate the efficiency of this algorithm, numerous simulations are performed on various networks and in different operating conditions. The simulation results of the method demonstrate that this approach both reduces the detection time in most cases and improves its security in the encounter of the other disturbances. Experimentally, the performance of the proposed method is evaluated by using a 5 kVA laboratory generator. Experimental studies show that the presented algorithm performed correctly detects LOE and SPS conditions.