Fault discrimination scheme for power transformer using random forest technique

• Published in: IET generation, transmission & distribution Vol. 10; no. 6; pp. 1431 - 1439
• Main Authors: Shah, Ashesh M, Bhalja, Bhavesh R
• Format: Journal Article
• Language: English
• Published: The Institution of Engineering and Technology 21.04.2016
• Subjects: CT saturation condition; Discrimination; Disturbances; EMTDC software package; fault diagnosis; Faults; feature extraction; Indian power transmission network; Joints; learning (artificial intelligence); power system analysis computing; power transformer; power transformers; power transmission faults; primary‐to‐secondary winding faults; PSCAD software package; random forest‐based fault discrimination technique; Ratings; Regular Papers; Simulation; Software packages; Transformers; turn‐to‐turn winding faults; Indian power transmission network; fault diagnosis; power transformers; turn-to-turn winding faults; primary-to-secondary winding faults; CT saturation condition; power system analysis computing; power transformer; feature extraction; random forest-based fault discrimination technique; power transmission faults; EMTDC software package; PSCAD software package; learning (artificial intelligence)
• ISSN: 1751-8687; 1751-8695; 1751-8695
• DOI: 10.1049/iet-gtd.2015.0955

This study presents random forest-based fault discrimination technique for power transformer. The proposed scheme relies on extracting features from the measured data of differential current signals of a power transformer. Various simulation cases consisting of internal faults including special types of turn-to-turn and primary-to-secondary winding faults and other disturbances (over-excitation and different types magnetising inrush such as initial, residual, recovery and sympathetic) have been generated with varying fault and system parameters for an existing power transformer of an Indian power transmission network using PSCAD/EMTDC software package. The performance of the proposed scheme has been evaluated over a simulation dataset of 5442 cases and the overall fault discrimination accuracy of more than 98% is achieved. The proposed scheme gives promising results for different connections and various ratings of the transformer, even though it is trained only once for a single rating and connection of a transformer. Comparative evaluation of the proposed scheme with the existing scheme clearly indicates the superiority of the proposed scheme as it remains stable during CT saturation condition and gives better stability during disturbances compared with conventional/existing schemes.