A cyber attack detection strategy for plug-in electric vehicles during charging based on CEEMDAN and Broad Learning System

Abstract

Cyber attacks targeting the charging process of plug-in electric vehicles (PEVs) have become more frequent with the popularity of PEVs, posing a significant threat to the battery life of PEVs and the stability of the power distribution network system. To mitigate this threat, a cyber attack detection strategy based on Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) and Broad Learning System (BLS) is proposed in this paper. The strategy exploits the coupling between PEVs and the power distribution network system during the charging process to achieve the detection of cyber attacks. CEEMDAN is implemented to extract features of system signals, and BLS uses the features extracted by CEEMDAN for feature mapping and enhancement to capture the presence of cyber attacks. Simulation analysis demonstrates that the proposed strategy can achieve more than 85% detection accuracy, with CEEMDAN bringing about 7% improvement in accuracy.