An identification method for vulnerable lines based on combination weighting method and GraphSAGE algorithm. (July 2023)

• Record Type: Journal Article
• Authors: Sun, Ji; Liu, Jiajun; Li, Chenjing; Zhi, Na
• Is Part Of: International journal of electrical power & energy systems. Volume 149(2023)
• Publication Date: 2023-07
• Subjects: Cascading outages -- Combination weighting Method -- GraphSAGE algorithm -- Vulnerable lines identification; Electrical engineering -- Periodicals; Electric power systems -- Periodicals; Électrotechnique -- Périodiques; Réseaux électriques (Énergie) -- Périodiques; Electric power systems; Electrical engineering; Periodicals; 621.3
• DOI: 10.1016/j.ijepes.2023.109035
• Languages: English
• ISSNs: 0142-0615

Abstract

Highlights: We establish the overload-dominated cascading outages model, through which the transmission lines dataset is obtained. We propose an assessment method of line vulnerability index based on the combination weighting method. We propose a method based on GraphSAGE algorithm that can quickly and correctly identify vulnerable lines in the power grid. Abstract: The cascading outages (COs) of power grid is the necessary stage of blackout. It is of great significance to accurately identify the vulnerable lines (VLs) in the grid at this stage for the assessment of the overall state of power grid and the suppression of blackout. For this reason, we propose an identification method for VLs based on the combination weighting method (CWM) and the Graph Sample and Aggregate (GraphSAGE) algorithm. We analyze the development process and transmission mode of COs, establish the COs model of power grid and obtain the relevant data set. Based on the data set, we define the line vulnerability index (LVI) and apply the CWM to calculate it. Through sorting and screening the index, we can get the VLs. Then, the mapping relationship between the line characteristics and the vulnerability is clarified, and the VLs identification model based on GraphSAGE algorithm is established and an online VLs identification scheme is proposed. Finally, the COs data set under different initial failures are constructed in the 39-bus system and the 118-bus system, and the identification model is trained, verified and tested. The simulation results show that the proposed method can quickly and accurately identify the VLs in the COs, and provide a theoretical basis for online identification of VLs with COs.