부산대학교 도서관

Cyber-physical attacks are increasing in frequency across a variety of industries, including manufacturing, as a result of recent advancements in networking and Internet technologies. Cyber-physical attacks harm the system’s physical components in addition to stealing intellectual property, as the term implies. This research proposes the novel technique in security-based manufacturing of industrial process in smart grid with data classification using deep learning techniques. Here, the manufacturing process security has been enhanced using secure heterogeneous stochastic-based edge networks. Then the smart grid-based data has been classified using Boltzmann Markov model with ResNet+ architecture. The experimental analysis has been carried out in terms of training accuracy, validation accuracy, MAP, MSE, QoS, latency, and energy efficiency. The simulation findings demonstrate that the PHSA has superior defence capability and system dependability than the conventional smart grid security architecture, greatly enhancing the security of edge-enabled smart grid systems. The proposed technique attained validation accuracy of 86%, MAP of 55%, MSE of 51%, QoS of 55%, latency of 62%, and energy efficiency of 93% based on the number of users.