Abstract

Microgrids are independent micro electric systems made up of locally controlled systems that can function both connected to the main grid (on-grid mode) or isolated from the main grid (off-grid or islanded mode). CIGRE defines Microgrids as “electricity distribution systems containing loads and distributed energy resources, (such as distributed generators, storage devices, or controllable loads) that can be operated in a controlled, coordinated way either while connected to the main power network or while islanded.”

As described by the IEEE Standards Coordinating Committee, microgrids have the ability to: (1) improve electrical reliability for customers; (2) relieve electric power system overload problems, in particular for highly congested power systems; and (3) resolve power quality issues. Most of the advantages offered by microgrids heavily rely on their predisposition to operate in isolated, independent mode. However, microgrids in islanded mode present technical operating challenges that need to be thoroughly investigated.

A microgrid must be able to independently meet the active and reactive power requirements of its assigned loads. In addition, it must also actively regulate voltage and frequency within a safe operating range in order to ensure system stability.

Investigating the technical challenges of islanded microgrids requires appropriate modeling tools. As is the case for high voltage (HV) power systems, the reliability of an isolated microgrid starts with a thorough investigation of its behaviour under various steady-state conditions and a derivation of the steady-state voltage profiles and transmission line loading levels throughout the system.
The present thesis investigates the steady-state analysis of islanded microgrid systems. To that end, an algorithm is developed using MATLAB to solve positive sequence (i.e., balanced) load-flow problems associated with isolated microgrids (IMGs). The proposed algorithm takes into account the specificities of IMGs and therefore yields more accurate results than those obtained with a conventional load flow algorithm. Compared to a conventional load flow algorithm, the algorithm that is most suitable for IMG has the following salient features: (1) no slack bus capable of supplying/absorbing the deficit/excess active and reactive power, (2) variable system frequency, and (3) part of, if not all, DG units operated in droop-control mode, which means that their active and reactive power outputs are not pre-specified, but are rather dependent on load flow variables (i.e., system frequency and bus voltages, respectively) at a given time.