Abstract: Deep learning (DL) models have been proven to be highly effective in detecting false data injection attacks (FDIAs) in smart grids relying only on the measurements of the secured component.