A Novel Approach to Dynamic State Estimation Through DMD and Adjacency Matrices in Power Converter Dominated Power Grids

Authors

Baltas G.N., Singh S., Cao J., Chamarthi P., Rodriguez P.

Reference

2024 IEEE Energy Conversion Congress and Exposition, ECCE 2024 - Proceedings, pp. 1854-1858, 2024

Description

The integration of renewable energy sources and reliance on power electronics converters challenge grid stability and reliability, necessitating comprehensive system observability. Static state estimation methods fail to capture dynamic changes while Dynamic state estimation (DSE) offers continuous updates. However, DSE methods face limitations in real-time accuracy, computational efficiency, and reliance on high-quality training data, highlighting the need for efficient algorithms adaptable to diverse power system configurations.To overcome the above limitations, this paper presents a novel method for estimating the dynamics of unmeasured or 'missing' buses by using Dynamic Mode Decomposition (DMD) and adjacency matrices, addressing the challenges for power converters control to enhance grid stability and reliability. By reconstructing the state variables of unmeasured buses of the grid without direct measurements, the method offers a scalable and efficient solution, demonstrating remarkable accuracy in simulated scenarios across various network scales. The method is tested in two benchmark systems, a small scale for proof of concept and a large scale system for demonstrating the effectiveness of the proposed apporach. This method significantly improves grid observability and resilience, advancing grid management tools for enhanced power system stability and reliability. Most importantly its sipmicity and flexibility renders this approach a practical and scalable solution.

Link

doi:10.1109/ECCE55643.2024.10860955