---

Today, differential relays are used in order to protect power transformers against all kinds of faults and events. Despite advances in relay fabrication technology, the detection and discrimination of different events is still one of the most important challenges for the protection engineers in this field. In this paper, an intelligent hybrid method has been proposed to detect and classify internal electrical faults, external faults while saturating Current Transformers (CTs) and inrush current in transformers. First, the internal and external fault currents and the inrush currents of power transformers are simulated by the Real-Time Digital Simulator (RTDS) and its software package (RSCAD). Then, the sampled signals in different events are transmitted to MATLAB software for detection and discrimination. At this stage, using the Bayesian Classifier method, which directly evaluates the training data information, external faults are separated from the other operating conditions of the transformer. Then, other events such as inrush current and internal electrical faults will be distinguished from each other by Decision Tree (DT) and Support Vector Machine (SVM) methods. The results show that the proposed intelligent hybrid protection method has the ability to detect and classify different disturbances in transformers in real time state with appropriate accuracy, which is one of the main innovations of this study compared to other published research.

---

This paper investigates the performance of differential protection of power transformers in the presence of internal faults, external faults, and cross-country faults in the presence of current transformers saturation, which is one of the main innovations of this study. Today, detection and discrimination of cross-country faults from other disturbances are one of the most important challenges facing protection engineers. Therefore, in this study, maximum overlap discrete wavelet transform has been used in order to accurately detect and classify these disturbances based on the extraction of energy coefficient indices of superior features. First, the cross-country faults, internal faults and external electrical faults, and inrush current phenomenon on the system under study in the EMTP software are simulated and differential current is sampled in different disturbances. Then, the mean indices of the sum of energy coefficient each level are calculated by MODWT by MATLAB software, and based on the values of indices, discrimination and classification of events are done. The results obtained from the simulations confirm that the proposed protection algorithm can detect and classify cross-country faults from other disturbances. Also, this method will improve the differential protection performance in different operating conditions and increase the reliability of power systems.