A detection strategy based on deep learning against sequential outages induced by false data injection attacks

LR attacks pose a significant menace to the safety of smart grids as they involve the manipulation of accurate measurements, resulting in system line overload and potential sequential outages. While previous research has focused on detecting LR attacks that cause line overloads, this paper introduces a mechanism capable of identifying both line overload and sequential output-generating by attacks. The proposed approach employs a deep learning network to analyze cyber load data estimated by the energy management system. To evaluate the effectiveness of the identification process, the IEEE standard 118 bus system is subjected to various attack scenarios and parameters. Results demonstrate that the proposed mechanism can effectively differentiate among LR attacks that aim to overload the lines, those that have cascading output potential, and a secure system state with a high degree of precision. In contrast to previous studies, the authors consider loads as input features to the network, improving accuracy and reducing delicacy to factors like measurement noise. The proposed detection mechanism offers an efficient, fast, and practical approach to identifying LR attacks in smart grids.