An S-Transform based approach for fault detection and classification in power distribution systems

This paper presents a methodology to automate the process of fault detection and classification. The proposed method is based on a time-frequency analysis of current for fault detection, and for fault classification current and voltage are analyzed. Therefore, it utilizes the features extracted from the signals by the S-Transform and the concept of energy. A multilayer perceptron is used to classify the faults. Performance tests were done by simulations conducted on Alternative Transients Program (ATP) software using the Medium Voltage European benchmark system from CIGRE, adapted with the inclusion of a distributed generator. The scenarios analyzed accomplished the distributed generation impact on the classification process under different situations varying fault resistance, load profile, distributed generation output power, and also the presence of noise in the measurements. Overall, the proposed method presented favorable results for both detection and classification.