Frontiers in Energy Research (Jan 2023)

A novel fault diagnosis method for power grid based on graph Fourier transform

  • Xiaojun Liao,
  • Xiaojun Liao,
  • Senlin Yu,
  • Li Zhang,
  • Xianzheng Feng,
  • Xiaoru Wang

DOI
https://doi.org/10.3389/fenrg.2022.1020687
Journal volume & issue
Vol. 10

Abstract

Read online

The fast and robust identification of fault elements is essential for the security and continuous operation of the power grid. The existing methods might be maloperation for bad data disturbance and require strict and exact synchronization. To address the challenge, this paper uses the conflict graph to propose a new sensitivity graph signal model for the power grid fault diagnosis. Next, a novel graph Fourier transform (GFT)-based method is proposed to diagnose the fault branch. Firstly, the measurement sensitivity graph signals are constructed by the conflict graph model, where the data is from activated recorders and protection devices. Next, the eigenvalue and GFT coefficient are used to extract the frequency characteristics of the signals. The fault branches provide the maximum contribution rate to the high-frequency coefficient of GFT. Then, for each node, the importance degree of the measurement sensitivity conflict graph signal is defined. The high-frequency importance degree-based method is proposed to discriminate the fault branch. Finally, simulations and practical cases verify the correctness and effectiveness of the proposed method. The proposed method owns fast faults diagnosis and good practicability. Additionally, the identification accuracy is high and the method is robust to bad data interference, due to considering measured data from whole activated fault recorders and protection devices.

Keywords