A robust identification method for transmission line parameters based on BP neural network and modified SCADA data

Accurate transmission line (TL) parameters are the basis of power system calculations. In recent years, artificial intelligence (AI) develops rapidly, which has been applied widely in power systems. However, AI is rarely applied to TL parameter identification. Thus, combining the TL model and AI, this paper proposes a robust identification method for TL parameters combined with BP (back propagation) neural network and median robust estimation, with the modified SCADA measurements based on TL model. Specifically, first, the robust identification method for TL parameter combined with BP neutral network and median estimation is proposed. And then, the training set that considers various working conditions and different line parameters is constructed based on the π-equivalent model. Furthermore, the input data of BP neural network is construed by modifying the SCADA data based on TL model. In addition, the median estimation is used to obtain the final result, which could reduce the interference of noise. Finally, the results with simulated data and measured SCADA measurements data show the effectiveness and practicality of the proposed method, respectively.