Research Of Bad Data Identification And Correction Based On Time Series Method

With the rapid development of smart grid, the automation degree of power system significantly improved, the running of power grid has increasingly depended on the data information, and in order to ensure the security and stability of power system operation, it is necessary to improve the quality of the data. The main researches of the bad data identification and correction methods are as follows.Firstly, the research status of bad data identification and correction in power system is reviewed in this paper. Throug’ (?)e analysis of research methods at home and abroad, and based on the time series characteristics of the data in power system, the method of bad data identification and correction from the perspective of time series is proposed in this paper. Then, the basic concepts of time series analysis, the basic model of time series and the method of time series analysis modeling theory is introduced in this paper, which is the theoretical foundation for bad data identification and correction in power system.The data of power system can be seen as time series, which contains regularity and randomness, so the method for bad data identification in power system based on time series analysis is proposed in this paper. First of all, the Lagrange interpolation algorithm is used to fill the missing data to realize the data preprocessing; then the ARIMA model is used to fit the data after pretreatment, so as to realize the description of the statistical law of power data; at last, because bad data generally has the characteristics of the larger fitting residual, the identification of the bad data in power system is realized by setting the identification interval.According to the different types and numbers of bad data, the methods for bad data correction in power system based on power balance and neural network is respectively proposed in this paper. The bad data correction method based on power balance according to the law of KCL and KVL, which uses the power balance principle to correct a single bad data; the bad data correction method based on RBF neural network according to the strong learning ability of neural network, which uses the trained neural network to correct various types of bad data.The study of this paper is based on the real data of power system, and the simulation results shows that the proposed method for bad data identification and correction in power system is feasible and effective.