International Journal of Wireless and Ad Hoc Communication IJWAC 2692-4056 10.54216/IJWAC https://www.americaspg.com/journals/show/2872 2019 2019 Gwinnett Technical College (GTC), Lawrenceville , GA, 30043, USA Manal Manal Libyan Iron and Steel Company (LISCO), Misrata, Libya Salim M. Hebrisha Smart grids (SGs) offer can ensure that users with a continuous power supply, decreased line losses, improved renewable output and storing, user participation in current electricity, and demand-side responsiveness. The development of cyberphysical SG (CPSG) systems has transformed the standard power grid by allowing bi-directional energy flow among utilities and users. But, because of increased data change among consumers, it is presented a major problem to the firewall systems for the transmission networks at either cyber or physical planes. Intrusion Detection Systems (IDSs) can role an essential play in maintaining SGs systems against cyber threats by generating a second wall of defense, complementing conventional preventive security procedures (for instance, authorization, encryption, and authentication). Therefore, this article concentrates on the design and development of Hybrid Metaheuristics with Deep Learning Assisted Intrusion Detection in a Cyber-Physical Smart Grid (HMDL-IDCPSG) infrastructure. The major objective of the HMDL-IDCPSG system provides the effectual recognition of the intrusions using feature selection and classification processes in the CPSG infrastructure. In the presented HMDL-IDCPSG method, a binary dragonfly algorithm with the hybrid directed differential operator (BDA‐DDO) algorithm could be implemented for the feature selection (FS) method. Besides, attention-based bi-directional long short-term memory (ABiLSTM) algorithm could be carried out for the recognition and classification of the intrusions. At last, the sparrow search algorithm (SSA) can be exploited for highest chosen the hyperparameter values of the ABiLSTM algorithm which supports in achieving a better solution. For demonstrating the greater outcome of the HMDL-IDCPSG technique, a comprehensive simulation value can be executed. The obtained results reported the supremacy of the HMDL-IDCPSG methodology with other existing approaches 2024 2024 53 66 10.54216/IJWAC.080206 https://www.americaspg.com/articleinfo/20/show/2872