AI-Based Forecasting and MPC-Inspired Strategy with Battery Dispatch for PID-Enhanced Frequency Regulation in Hybrid Power Systems

The work presents a unified control framework for frequency regulation in a two-area hybrid power system comprising solar photovoltaic (PV) and thermal generation. The proposed strategy consists of three major components: Support Vector Regression (SVR) for short-term forecasting, Model Predictive Control (MPC) based strategy for dispatching battery storage energy, and Differential Evolution (DE) for tuning gains of Proportional–Integral–Derivative (PID) controller. To test the framework, actual solar generation data from the Mirzapur Solar Power Plant in India is used. In the absence of real-time demand and thermal data, synthetic load and thermal generation profiles are created to simulate realistic operating conditions. The SVR model forecasts near-term forecasts of generation and load behavior, which then guides battery dispatch actions that maintain system balance. Concurrently, the DE-optimized PID controller actively reduces frequency deviations in the solar region by processing Area Control Error (ACE) values. Simulation results demonstrate that this integrated SVR–MPC–DE-PID system improves system stability by damping ACE oscillations, regulating battery state-of-charge (SOC), and reducing frequency deviations. Although traditional performance metrics improve moderately, the system exhibits stronger overall responsiveness and robustness—highlighting its potential for practical deployment in renewable-integrated smart grids.