Read the full-text of this article for free

Title: An accident chain-based risk assessment method for power system faults

Authors: Xuming Liu; Xiaokun He; Yonglin Li

Addresses: Jinling Institute of Technology, Nanjing, 211100, China ' Jinling Institute of Technology, Nanjing, 211100, China ' Jinling Institute of Technology, Nanjing, 211100, China

Abstract: A chain fault risk assessment method based on transformer-federation migration learning algorithm is proposed. Firstly, this paper describes the chain fault evolution path of AC-DC hybrid grid, constructs fault simulation model, and clarifies the selection principle of the initial fault set of the accident chain. Secondly, the chain fault probability assessment model based on the accident chain is established according to the key indicators of the accident chain. Then, the weighted fuzzy C-mean clustering algorithm is used to cluster and analyse the correlation indicator values, and the fault set feature extraction module is obtained according to the transformer model. Finally, a multimodal transformer architecture based on federated learning cooperative work is designed to realise the accurate estimation. The experimental results show that the proposed method for AC-DC hybrid grids has a good generalisation capability and can quickly and accurately determine the fault set feature extraction module.

Keywords: AC-DC hybrid grid; accident chain; transformer model; federated migration learning; fault analysis; risk assessment.

DOI: 10.1504/IJICT.2025.148823

International Journal of Information and Communication Technology, 2025 Vol.26 No.34, pp.60 - 81

Received: 20 May 2025
Accepted: 06 Aug 2025
Published online: 26 Sep 2025 *