Enhancing Cybersecurity and Frequency Control in Smart Grids Through AI-Driven Short-Term Load Forecasting

Efficient management of energy systems is essential in today's world, where balancing electricity supply and demand ensures reliability and cost-effectiveness. This article emphasizes the critical role of short-term load forecasting (STLF) in stabilizing power operations and optimizing resource utilization while ensuring cybersecurity in smart grids. The threat of false data injection (FDI) attacks, which compromise data integrity within cyber/communication networks, poses significant challenges for smart microgrids, impacting state estimation, frequency control, and protection systems. Accurate forecasting helps schedule maintenance, adjust electricity prices, and inform strategic decisions, leading to significant savings. FDI attacks are well detected due to short-term prediction method and the system operator stick to the predicted load therefore, balancing powers occur and FDI attacks are eliminated. The study focuses on predicting electricity demand in Belgium for July 2020, using data collected from 2016 to mid-2020. Techniques like Support Vector Regression (SVR) and Multi-Layer Perceptron (MLP) are employed to create forecasts, which are then compared with actual demand to measure accuracy. To enhance SVR's performance, Particle Swarm Optimization (PSO) is applied to fine-tune the model's parameters, with various kernel functions tested for better results. The findings provide valuable insights for electricity providers, helping them plan production more effectively and manage peak energy demand. This research also supports smart grid efforts aimed at incorporating decentralized energy sources. Overall, it highlights the increasing need for advanced forecasting methods as energy systems evolve and electricity markets become more flexible, while also addressing the importance of cybersecurity against emerging threats.