Abstract
The increasing intelligence of systems leads to enhanced flexibility and control, enabling the development of more efficient systems capable of adapting to environmental changes. However, the design of dynamic control mechanisms for such systems presents challenges due to their inherent nonlinearity and stochastic nature. This thesis tackles the challenges associated with the efficient operation of systems that employ dynamic control mechanisms, specifically microgrids, and 5G wireless communication networks. Microgrids are electricity distribution networks that can serve in both grid-connected and islanded modes. In the islanded mode, microgrids are solely dependent on the electricity sources (i.e., dispatchable, and non-dispatchable) to maintain stable operation. In this case, the droop control mechanism is often embedded into the network to adjust the active/reactive power output of the dispatchable electricity sources in the microgrid. Due to the nonlinearity of the system and stochasticity associated with renewable energy sources and load demands, the problem of optimal selection of droop coefficients is further complicated. We present a novel solution approach to this problem by integrating the Newton-Raphson (NR) algorithm into sequential sampling-based particle swarm optimization (PSO). The scalability and fast quadratic convergence make the NR algorithm ideal for solving the power flow problem of islanded microgrids. Subsequently, PSO provides a versatile option to carry out optimization since it is not dependent on assumptions about the structure of objectives or constraints. The proposed approach improves the voltage and frequency stability of microgrids while minimizing active/reactive distribution losses. The results also demonstrate significant improvements in the voltage profiles of the buses while maintaining a stable frequency during the islanded mode of operation compared to their conventional counterparts.