Consistent sensing algorithm for distributed PV station area line loss anomaly fluctuation based on spectral clustering

The current conventional line loss anomaly sensing algorithm mainly realizes line loss anomaly sensing by calculating the station area line loss rate and comparing it with the set threshold value, which leads to poor sensing effect due to the lack of clustering analysis of line loss characteristics. In this regard, a consistent sensing algorithm based on spectral clustering of line loss abnormal fluctuations in distributed PV station areas is proposed. The line loss data is assimilated with time granularity and data cleaning, and combined with the spectral clustering algorithm to cluster the line loss abnormal data features and set the highest threshold of the Euclidean distance of the line loss rate to realize the line loss abnormal sensing. In the experiments, the proposed method is verified for the sensing effect. The experimental results show that when the proposed method is used to sense line loss data, the algorithm can accurately sense different types of line loss data and has a more desirable sensing performance.