Inrush current in the transformer may assume as fault current for the protective equipment. Thus, it is important to distinguish the inrush current from the internal fault current in the transformer. In the present paper, along with the transformer nonlinear modeling, the extended Kalman filter (EKF) is studied to estimate the transformer inrush current. Three phase transformers inrush currents are simulated using the nonlinear model, as well as the initial inrush currents are estimated using the extended Kalman filter formulation. Then, by comparing this current and the current calculated from the transformer model, we decide on the type of current (whether it is a fault or an inrush current). Now, if the inrush current estimation error exceeds a certain limit compared to the transformer actual inrush current, it is considered as a fault current. In this paper, the transformer winding fault location has been also presented which has been less investigated in previous studies. The relation of the fault location to the estimation conditions and the estimation error standard deviation is also studied . In addition, an index is defined so that if the value of this index is greater than a specific value in any phase of the transformer, a fault condition is recognized.
Taghikhani et al. (Sun,) studied this question.