A new intelligent approach for frequency controller of autonomous hybrid power systems

• Published in: Neural computing & applications 2025-01
• Main Authors: Magdy, Gaber, Bakeer, Abualkasim, Bakeer, Mahmoud, Albalawi, Hani, Zaid, Sherif A.
• Format: Article
• Language: English
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-024-10635-y

An intelligent approach for load frequency control based on a proportional-integral-derivative (PID) controller, referred to as an intelligent PID (IPID) controller, is proposed to enhance the frequency stability of autonomous hybrid power systems (HPSs) considering the high penetration levels of renewables. The proposed IPID is based on the ultra-local mod4el, i.e., a model-free control, which operates independently of the power system parameters since the controller incorporates a term for unknown disturbances, enhancing its control action. Moreover, a new meta-heuristic optimization algorithm, i.e., Geometric Mean Optimizer (GMO), is used to fine-tune the proposed IPID controller parameters for frequency regulation of the studied HPS. Furthermore, the superior performance of the proposed IPID controller based on the GMO is validated by comparing it with other counterpart controllers used; such as the PID and fractional-order PID (FOPID) controllers under different load/renewables perturbations, system uncertainties, and high renewables penetration. The Matlab simulation results demonstrate that the proposed IPID controller using GMO offers greater stability in reducing response deviations compared to traditional PID and FOPID controller structures. Where the proposed IPID controller demonstrated a 80% improvement in performance compared to the traditional PID controller, and a 57% improvement over the FOPID controller in handling scenarios with random load demand and high renewables penetration. Finally, to combine the accuracy of physical simulation and the adaptability of numerical simulation, the proposed IPID controller based on the GMO is verified and implemented in a real-time environment based on the PLECS RT Box platform.