A New Approach for the Transformer Differential Protection Based on S-Transform and Fuzzy Expert System

This paper presents a novel method to discriminate between the magnetizing inrush and external and internal fault currents in power transformers. Fault type identification and faulted phase selection are also possible by the proposed algorithm. The proposed method has two main parts. First, by means of S-transform, which is the most accurate method in the field of signal processing, some useful features are extracted from the input signal. Then, the extracted features are converted to some numerical indices. In the second part, an effective decision maker is needed to classify the input signal. One of the best methods, which have been used for decision-making applications is fuzzy logic. So, the numerical indices are used as inputs for the fuzzy system. The output of the fuzzy system not only can reveal whether the input signal is the magnetizing inrush, external or internal fault, but it can also identify the fault type when there is an internal fault. Finally, the faulted phases can be identified with a supplemental algorithm. To generate the test signals, a three-phase transformer is modeled in PSCAD/EMTDC. Testing the proposed algorithm by different simulated data shows the robustness of the proposed method in the transformer differential protection.