ISRCKF‐Based Dynamic State Estimation for Cyber‐Physical Wind Energy Systems With Bad Data Identification

ABSTRACT This article proposes an improved square root cubature Kalman filter algorithm for dynamic state estimation in cyber‐physical wind energy systems (CPWESs). The integration of wind power generation equipment into the power grid may amplify fluctuations in system measurement data. Bad data can easily be disguised as fluctuation data, invading CPWESs and damaging their accuracy in state estimation. It is difficult to distinguish between measurement data with significant fluctuations and bad data in wind energy access networks using traditional methods. The proposed algorithm can effectively handle the identification and correction of bad data while maintaining high accuracy in CPWESs with strong data fluctuation. First, it aims to identify the mutation data of the new information sequence. By calculating the innovation covariance based on the innovation sequence and comparing the measured variance with the total variance, it constructs a dynamic high‐dimensional observation noise scaling factor to adjust the measurement noise matrix to obtain the first layer of suspicious data set. Then, based on the measurement sequence, it calculates the correlation of the data as the second identifier to distinguish between fluctuation and bad data, so that the bad data set is found and corrected. Simulation examples on the IEEE 57‐bus test system illustrate the effectiveness of the proposed method.