Abstract. The digitalization of energy infrastructure has become an integral component of modern operation, concurrently increasing exposure to cyber threats—particularly distributed denial-of-service (DDoS) attacks. These attacks disrupt the normal functioning of SCADA systems, IoT devices, and intelligent power grids, thereby posing significant risks to critical infrastructure. This study investigates contemporary approaches to detecting and mitigating DDoS attacks targeting energy systems through the application of machine learning techniques. A range of models is examined, including classical algorithms (Random Forest, Decision Tree, Gradient Boosting, SVM), deep learning architectures (CNN, LSTM), and hybrid models (LSTM-CNN). Model performance was evaluated using benchmark datasets (CICDDoS2019, KDD-CUP) and validated in a simulated SCADA environment. Emphasis is placed on developing an adaptive and intelligent protection framework capable of real-time anomaly detection within energy network traffic. The findings indicate that hybrid models can achieve detection accuracies of up to 99% under certain scenarios. Furthermore, the study explores the potential of integrating blockchain and cloud-based technologies to enhance the robustness and scalability of cybersecurity solutions. These outcomes provide practical guidance for designing comprehensive defense mechanisms in digitalized energy systems.

Keywords: Cybersecurity, DDoS attacks, energy facilities, SCADA, IoT, machine learning, deep learning, hybrid models, blockchain, intelligent networks.