Power Systems Observability Analysis Based on Parallel Gaussian Belief Propagation

• Published in: 2022 IEEE 10th International Conference on Computer Science and Network Technology (ICCSNT) pp. 19 - 23
• Main Authors: Ren, Chunhui, Wang, Xuan, Cao, Chunna, Chen, Jinming
• Format: Conference Proceeding
• Language: English
• Published: IEEE 22.10.2022
• Subjects: Computational modeling; Computer science; Factor Graphs; Gaussian Confidence Propagation; Mathematical models; Observability; Observability Analysis; Parallel Computing; Parallel processing; Power measurement; Power systems
• ISSN: 2690-5892
• DOI: 10.1109/ICCSNT56096.2022.9972970

Power system state estimation is fundamental for a range of online applications of power systems in energy management systems. The premise of state estimation calculation is that we need to make sure the system of equations should have a feasible solution. As a preprocessing step of the power state estimation algorithm, observability analysis plays a crucial role in power state estimate, which needs to determine whether the equation of state estimation formed by the measurements has a distinct solution. In this paper, we propose a parallel observability analysis based on Gaussian belief propagation to address the low efficiency of observability analysis in large-scale power systems. The Gaussian belief propagation process is implemented with multi-core and multithread parallel computing to achieve efficient observability analysis. Simulation calculations were performed using multiple power systems with 10, 000 to 70, 000 buses. The experimental results show that the proposed parallel observability analysis algorithm is more efficient than the conventional algorithm.