Study on Power State Data Anomaly Detection Under Grid Dynamic Behavioral Constraints
At present,conventional methods for detecting abnormal power state data have the problem of poor detec-tion accuracy.To address this issue,researchers have proposed a power state data anomaly detection method based on dynamic behavior constraints of the power grid.This method first analyzes the dynamic behavior constraints of the power grid,and takes source load data and operational data as extraction targets.Then,interpolation method is used to fill in the missing values of the obtained raw data to ensure the integrity of the data.Next,the grey correlation analysis method is used to analyze the degree of correlation between data to extract key parameter data.At the same time,select the cluster center and identify abnormal data by determining whether the distance from the new input da-ta to the cluster center is higher than the anomaly threshold.The experimental results show that when using this method for anomaly detection in power data,the AUC value of the algorithm is low,and it has relatively ideal detec-tion accuracy.
power systemsoperational dataanomaly detectionclustering algorithms
万方数据
维普
