Real-time adaptive control strategies for frequency regulation in isolated microgrids

The integration of renewable energy sources into power systems with distributed and variable characteristics has introduced challenges in maintaining stability and reliability within conventional energy infrastructures. In this context, islanded microgrids have emerged as a strategic solution, providing both energy security and support for sustainability targets. However, uncertainties in generation and rapid system dynamics in such configurations limit the effectiveness of fundamental control processes such as load frequency control. This necessitates the development of control strategies that are more flexible, adaptive, and resilient. Within the scope of this dissertation, four original control strategies with cascade structures have been developed for islanded microgrid systems by combining interval Type-2 fuzzy logic with classical and fractional order control techniques. For the optimization of controller parameters, powerful metaheuristic algorithms namely, the CIOA and the ISSA have been employed. The proposed structures have been tested under various microgrid configurations, including renewable energy sources, micro-generators, and hydrogen-based energy storage components such as fuel cells and electrolyzers. These tests were conducted under dynamic conditions involving generation uncertainties and load fluctuations, enabling a comprehensive performance assessment of the proposed control strategies. Additionally, improvements in load frequency control performance were evaluated by integrating demand-side flexibility into the control schemes for selected scenarios. The robustness of the proposed controllers was also analyzed under cybersecurity threats, specifically FDI and DoS attacks. Another critical aspect of this study is the discrete-time stability analysis conducted to assess the reliable digital implementation of the controllers. The instability risks that may arise from operating continuously modeled systems in discretized simulations were investigated. Furthermore, the representational accuracy of software-based simulations in reflecting physical systems and the necessity of real-time simulators were examined. The developed control schemes were tested in the MATLAB/Simulink environment and validated in real-time using the OPAL-RT platform for selected scenarios, with consistency between platforms analyzed through comparative evaluation.