Search

Renewable energy resources (RESs) are significantly integrated in distribution networks to promote green technologies in future power systems. The idea of microgrids (MGs) is developed for the efficient use of RESs through an appropriate control, monitoring and management system. Control and management of MGs are challenging tasks along with numerous economic and environmental benefits. The challenges of MGs operation include tie-line power fluctuations that have an adverse effect on the stability and quality of distribution networks. Tie-line power control in a residential MG is difficult due to dependency on RESs as a primary generation unit in MGs. Motivated by these, this thesis investigates the tie-line power control issues in grid-connected residential MGs and applies several controls and optimisation methods to achieve a smooth tie-line power satisfying system boundary conditions. First, a dynamic energy management system (EMS) is designed to reduce the tie-line fluctuation in a grid-connected MG through an indirect grid power control strategy. A fuzzy logic-based EMS is proposed to control the battery power due to the variations in generations and loads. The net power demand and battery state of charge (SoC) of an MG are considered inputs of the fuzzy controller to determine the battery power by keeping the battery SoC within limits. An offline optimisation method is used to optimise the membership functions and rules to shape the performance parameters. Thereafter, a golden section search-based non-linear programming method is applied to design a battery power management system to minimise the tie-line fluctuation in an MG counting the system constraints and disturbances. Two other rule-based methods are also demonstrated for comparative analysis of the proposed methods in terms of predefined performance parameters. Afterward, a dynamic grid power control method is presented to control the interlink inverters in grid-connected MGs. A grid power controller is designed based on a complete model of the MG systems to achieve a constant tie-line power on typical days of the year. The designed controller can effectively smooth tie-line fluctuation in a grid-connected residential MG. The charging/ discharging of the battery is controlled by a DC-DC converter which is also responsible to provide a stable DC bus to the input of an interlink inverter. The reference tie-line power is determined by a MG controller based on statistical power generations, load demand and battery SoC. Moreover, an eigenvalue-based stability analysis is performed to show the sensitivity of system parameters on system stability. Furthermore, the tie-line power control in a networked MG (NMG) is investigated to obtain a smooth tie-line power in an NMG connected to a common bus. A model predictive control-based distributed power flow controller is proposed to control the interlink inverters of the NMG in a distributed manner. Charging/ discharging of battery is controlled by a decentralised model predictive power controller to provide a stable DC voltage for MGs. Communication between MGs is performed for sharing the status of the tie-line power along with the scheduled tie-line reference. The information from the network is used to determine the instantaneous reference grid power of individual MGs for achieving a smooth tie-line power for the network. Inverter switching actions are performed to minimise the difference between predictions and references. In addition, a comparative study with a decentralised operation of MGs is conducted to show the benefits of networked operation. All the proposed methods are tested through rigorous case studies to validate the performance despite the variations in input and output system disturbances. Comparative analysis among different methods is also conducted to demonstrate the performance variations through adopting different methods. For the simulation experiment set up, MATLAB SIMULINK Simscape Electrical is used to develop a designed system model of MGs and experimental models of the proposed methods. Experiments are performed using real weather and residential load information in Queensland, Australia. The results demonstrate that the proposed methods have achieved the design objectives to solve the tie-line fluctuation problem of grid-connected residential MGs.