Research And Application Of Abnormal Data Recognition And Verification Technology For Power Quality

As the carrier of information technology and information resources,the quality of data is very important.How to ensure the quality of data has become an important research topic in recent years.However,in practical applications,due to various unavoidable causes in the process of transmission and operation,abnormal data often occur in the database system.The abnormal data that appears in the data set reduces the quality of the data to a large extent.Therefore,how to test the abnormal data in massive data sets and how to process abnormal data is a research topic with research significance and application value.With the rapid development of power quality in China's power system in recent years,the scale of power quality data accumulated by various power enterprises is expanding rapidly.The abnormal data of power quality greatly reduces the quality of power quality data,and cause variation in data calculation,analysis,and prediction.The basic data of power quality,as an important data source of subsequent advanced application module and prediction module,will seriously affect the correctness of power quality analysis results if it can not guarantee the correctness of the basic data of power quality.This paper analyzes the current development of abnormal data detection and power quality data detection,and analyzes the characteristics of power quality basic data from the perspective of power quality related indicators,and discusses the necessity of detecting the data quality of active power distribution network.Several existing methods of identifying and verifying abnormal data are studied.According to the characteristics of power quality basic data,three abnormal data detection methods are selected,including setting threshold discriminant method,data transverse comparison method and improved K-MEANS algorithm.The power quality anomaly data recognition and verification system is designed and implemented in this paper.The above three methods are used to identify and verify the abnormal data of power quality,and generate detailed PDF detection report stored in the client local system.We also verify the detection performance of the improved K-MEANS clustering algorithm.By comparing with the traditional K-MEANS clustering algorithm,it is proved that the improved K-MEANS algorithm is suitable for the recognition and verification of power quality abnormal data.