Evaluating Power System Stability in PV-Integrated Grid with AI-Based Solar Forecasting Model

This research focuses on forecasting solar irradiance for the year 2025 by applying Artificial Neural Network (ANN) based on historical data collected over three years (2022–2024) for the humidity, earth temperature, and wind speed. The primary aim is to assess the potential output power of a photovoltaic (PV) system and its reliability in covering electrical loads. The developed forecasting model demonstrated promising accuracy, achieving accepted values for the Mean Squared Error (MSE). To evaluate the system's performance under adverse conditions, two fault scenarios were simulated by MATLAB-Simulink software: one where the load was supplied entirely by the grid (no PV system), and another where both the grid and the PV system supplied the load. Results indicated that the inclusion of the PV system introduced challenges in system recovery due to its lack of inertia, exacerbating instability during faults. To address this limitation, a battery energy storage system was integrated into the PV setup. This addition significantly improved the system's resilience and ensured smooth recovery once the fault is removed. The study underscores the importance of incorporating energy storage solutions to enhance the reliability and stability of PV-based power systems, offering critical insights for designing sustainable and robust energy infrastructures.