Paper Details

Name: Azhar Ashraf, Shishir Singh, Devesh Kumar Upadhyay, Shruti Sharma, & Nimmanagoti Anil
Country: India

Year: 2026
Volume: Volume No: 13, January, Year: 2026 (Special Issue)
Page Number: 160-167
DOI: https://doi.org/10.5281/zenodo.18975445

ABSTRACT—
For smart grids dealing with changing cyber threats, guaranteeing a cyber resilient network is important. This project uses artificial intelligence, specifically Convolutional Neural Networks (CNN), to improve grid security by detecting and mitigating phishing, malware, and DDoS threats. The system correctly detects phishing 33.1% of the time, malware 31.4% of the time, and DDoS attacks 32.4% of the time, with low false positive rates of 0.7% for phishing, 1.0% for malware, and 1.4% for DDoS. To precisely spot all unusual login behaviours, the system thoroughly integrates real-time threat intelligence, in-depth behavioural analysis, and immediate proactive alerts, ultimately achieving a 90% accuracy rate. User feedback confirms its effectiveness and usability through a satisfaction rate of 92% and an adoption rate of 78%. The system secures all communications through encryption, two-factor authentication, and identity verification, fully complying with GDPR and NIST SP 800-63B. This AI framework both strengthens how secure the grid is and keeps operations strong by always
adjusting to new dangers.

Keywords: cybersecurity automation in power systems, secure grid communication, AI-powered security, smart grid security compliance, and machine learning for cybersecurity; Adaptive security architecture; proactive cyber defence; intrusion detection and prevention; threat intelligence in energy networks; and cyber- resilient smart grids.