Self-learning Diagnosis of Transmission Line Fault Type Based on Deep Forest and SMOTE

Accurately identifying the fault type of transmission lines is of great significance to the safe operation of transmission lines. In this paper, the distributed detection method is adopted to collect and classify transmission line fault data, while 15 time-domain features and 12 frequency-domain features are extracted from each fault waveform to build a transmission line fault database. Then, based on this database, study on classification and identification of faults by Deep Forest t algorithm, and eliminate the imbalance of data by SMOTE, finally achieve intelligent diagnosis of fault types of transmission lines. The correctness rate of the proposed method is 92% in the binary diagnosis of lightning and non-lightning strike, and 79% in the full classification diagnosis. Compared with the traditional classification algorithm, the proposed method has higher accuracy and better generalization performance. The research has reference significance for the digital construction of transmission lines and the intelligent upgrading of power grid.