Modern civilization relies on stable power systems to meet fluctuating energy demands. Isolated microgrids with conventional power stations can supply rural areas, but disruptions can propagate across connected networks. Load frequency control techniques are vital to maintain steady electricity flow across grids, especially in isolated microgrids, despite demand fluctuations. Modern isolated microgrid structures with inverter-based generation and renewable supplies reduce inertia and introduce significant frequency control challenges compared to traditional systems. In addition to reduced inertia, uncertainties in renewable sources also threaten frequency stability. Therefore, load frequency control is critical for sustaining frequency stability and operational performance in isolated microgrids. Enhancements in this domain fall into three primary categories: first, incorporating additional regulating sources such as high-power density or novel generation units; second, developing new single- or cascade-loop controller structures to effectively manage multiple sources; and third, proposing novel tuning strategies to optimize both conventional and advanced load frequency controllers.
This thesis introduces four LFC (Load Frequency Control) schemes addressing the second and third research directions. The first LFC scheme improves traditional proportional-integral (PI) and proportional-integral-derivative (PID) controllers by employing a novel multi-objective-based tuning method aimed at minimizing frequency deviations in low inertia, single-area isolated microgrids (IMGs). The second scheme proposes another advanced multi-objective tuning approach to optimize conventional PID, fractional-order PID (FPID), and cascade PID controllers. These tuning strategies outperform classical error-based methods—including integral absolute, integral time absolute, integral square, and integral time square errors—across various load and renewable energy scenarios. Additionally, anti-windup mechanisms are applied to ensure the optimized PID, FPID, and cascade PID controllers function effectively within the nonlinear operational boundaries and physical limitations of generation sources. During this tuning process, the schemes also evaluate the maximum practical generation capacities of the different sources involved.
To further improve LFC across multiple IMG sources, the third proposed scheme introduces a new single-loop controller combining a fractional-order proportional (FP) controller with a tilt-integral-derivative (TID) controller, termed the Fractional-Order Proportional-Tilted-Integral-Derivative (FPTID) controller. This FPTID controller demonstrates superior performance compared to traditional options like PID, FPID, and standalone TID controllers in multiple single-area IMG case studies. Lastly, the fourth LFC scheme integrates the fractional-order proportional derivative controller with an N-filter (FPDN) and the FPTID in a cascade-loop configuration, forming a novel FPDN-FPTID controller capable of enhancing frequency regulation in both single- and multi-area IMGs. Extensive MATLAB/Simulink simulations reveal that this cascade-loop controller outperforms numerous published solutions, including PID, FPID, PID with N-filter, cascade PID, PI-FPID, and PDN-PI controllers.
Furthermore, these LFC schemes leverage modern optimization algorithms to refine controller settings. Artificial rabbits optimization is utilized in the first and second schemes, geometric mean optimization in the third, and dandelion optimization in the fourth. The effectiveness of these algorithms has been validated through extensive comparative studies with other published methods in the LFC domain.
] 