Cyber Attack Detection in Distribution Networks with Topological Data Analytics aided Learning

The integration of smart grid technologies has brought significant advancements to power systems, yet it has also increased its vulnerability to cyber threats. False data injection attacks (FDIAs) pose a substantial risk to grid data integrity, particularly in critical areas like voltage control and state estimation. This study centers on leveraging the latest advancement in topological data analysis (TDA), specifically multi-parameter persistent homology, which has shown remarkable effectiveness in graph representation learning in recent years. Our objective is to utilize this approach to bolster the detection of FDIA using data collected from voltage sensors. By integrating topological methods, our approach aims to fortify the resilience of power systems against cyber threats, thereby ensuring the reliability and security of smart grid operations.