Data Mining Technology in Detection and Identification of Bad Data in Power System

Abstract

Traditional bad data detection methods are estimated algorithms that require repeated state estimations. A large number of calculations may also cause “residual flooding” or “residual pollution” phenomena, which is the ideal state. The bad data can be detected and identified before the estimation, and the bad data detection and identification method based on association rule mining studied in this paper can solve these problems to a certain extent. This paper first analyzes the traditional bad data detection and identification methods and then leads to data mining technology. Second, it delves into the classic algorithm Apriori and improvement in association rules and studies the basic algorithm and improvement of periodic association rule mining. Application of improved algorithm. The current, active, and reactive power data of a certain line collected in the SCADA system of a dispatching center from May to September and five months were selected as sample data to finally verify the feasibility and effectiveness of the method.