A DBSCAN algorithm-based transformer poor leakage reactance parameter identification method

The invention provides a DBSCAN algorithm-based transformer poor leakage reactance parameter identification method in the field of electric power system planning. The method comprises the steps of selecting a to-be-identified data set; performing normalization processing on the to-be-identified data set to form a relevant coefficient matrix; setting relevant clustering parameters and performing DBSCAN algorithm clustering on the to-be-identified data set to form clustering clusters and noise clusters thereof; calculating typical feature vectors of the clustering clusters to form a typical feature vector library; dividing and distinguishing noise vectors according to different dubieties and marking corresponding levels; outputting noise vectors of different dubieties to allow related personnel to perform corresponding treatment. The method can rapidly and efficiently achieve automatic error calculation of large-scale power grids and automatic identification of unreasonable transformer leakage reactance paramet