Cause Identification Method of Power Quality Problems in Low-voltage Distribution Network Based on Multi-label Scene Division

With the popularization of "carbon peak and carbon neutralization", large-scale new power supply and new load are connected to the low-voltage distribution network, and the power quality problem in low-voltage distribution station are becoming more and more serious, and the causes of power quality problems are becoming more and more complicated. Moreover, the boundary of traditional gene pool scene division is not clear enough, and it is not suitable for developing new distribution system. For this reason, a method for identifying the causes of power quality problems in multi-label scenarios of low-voltage distribution network based on mechanism data joint driving is proposed. Firstly, according to the designed multi-tags, the low-voltage distribution scenarios are divided to form a corresponding power quality problem cause library; Secondly, according to the localization data, the localization neural network model is trained in two layers; Then, through the trained neural network agent, the threshold value of localized gene pool is revised; Finally, the proposed multi-label new rural gene database is taken as an example to verify the method.