AN INTELLIGENT MULTI-LAYERED MACHINE LEARNING AND BIG DATA ANALYTICS FRAMEWORK FOR AUTONOMOUS CYBER THREAT DETECTION, ADAPTIVE DEFENSE MECHANISMS, AND SECURE CLOUD-NATIVE SMART ELECTRICAL POWER GRID CYBER-PHYSICAL INFRASTRUCTURE

Abstract

The rapid digitalization of smart electrical power grids has significantly improved operational efficiency, automation, and real-time decision-making. However, it has also increased the exposure of cyber-physical infrastructure to sophisticated and evolving cyber threats. As modern smart grids increasingly rely on cloud-native architectures, distributed sensors, intelligent control systems, and interconnected communication networks, traditional rule-based and static cybersecurity mechanisms are no longer sufficient to ensure resilient and adaptive protection. This paper proposes an intelligent multi-layered framework that integrates machine learning and big data analytics for autonomous cyber threat detection, adaptive defense mechanisms, and secure cloud-native management of smart electrical power grid cyber-physical infrastructure. The proposed framework is designed to address the growing complexity, scale, and heterogeneity of cyber threats targeting critical energy systems. The framework consists of multiple coordinated layers, including data acquisition, preprocessing, feature engineering, anomaly detection, threat classification, adaptive response, and secure cloud orchestration. At the core of the proposed architecture, machine learning models are employed to identify malicious activities, abnormal communication patterns, false data injection attacks, denial-of-service incidents, insider threats, and other advanced persistent threats affecting the cyber-physical components of the grid. Big data analytics techniques are incorporated to process high-volume, high-velocity, and high-variety operational data generated from smart meters, supervisory control and data acquisition systems, phasor measurement units, IoT devices, and distributed energy management platforms. This enables the framework to support real-time situational awareness, predictive threat intelligence, and continuous risk assessment. In addition, adaptive defense mechanisms are embedded within the framework to enable dynamic mitigation, automated policy updates, attack containment, and resilience enhancement under changing threat conditions. The cloud-native design further improves scalability, interoperability, and deployment flexibility while supporting secure resource management across distributed grid environments. The proposed framework aims not only to detect cyber threats with high accuracy but also to enhance response speed, reduce false alarms, and improve the overall security posture of critical smart grid infrastructure. This research contributes to the development of next-generation intelligent cybersecurity solutions for energy systems by combining autonomous learning, data-driven threat analytics, and resilient cloud-native defense strategies. The study offers a practical and scalable pathway toward securing future smart grid cyber-physical ecosystems against increasingly complex cyber attacks while supporting reliability, sustainability, and operational continuity in critical power networks.