A Transformer Neural Network-Based Cyberattack Detection Technique in Hybrid Power System

Due to the dynamic evolution of power systems by incorporating renewable sources, it is now more difficult for power networks to regulate frequency and power deviation in the tie-line. Hybrid power systems (HPSs) use different independent generation systems such as wind turbines, solar photovoltaics (PV), diesel engines, fuel cells (FCs), aqua electrolyzers (AE), and energy storage devices such as flywheels and batteries. To create a smart grid, recent advancements in power systems rely heavily on communication networks, offering numerous benefits like complete system accessibility, tracking, control, and protection. However, it makes the system more complicated and susceptible to cyberattacks, which endangers the smooth functioning of the system. It must be successfully discovered to mitigate the attack's impact and return to normalcy rapidly. This work adopts the transformer neural network to sense the presence of false data injection (FDI) attacks in the HPSs. Different parameter metrics like recall, accuracy, precision, and F1-metric are calculated for each classifier technique for the comparison.