Centralised vs decentralised PI control for an islanded multi-inverter microgrid

Abstract

Energy supply security is arguably one of the most pressing challenges for both emerging and established countries around the world, as technology grows increasingly reliant on reliable energy. With the increase of Renewable Energy Sources (RES), low maintenance costs in comparison with utility grids, and low energy losses, microgrids are a preferred solution for addressing energy security. Microgrids are not without their challenges, therefore, a lot of research is focused explicitly on the voltage and frequency stability of a microgrid. This study aims to compare the performance of a centralised and decentralised controller for an island multi-inverter microgrid. Microgrid control is frequently complex, financially and computationally expensive, and difficult to scale as the microgrid grows, therefore, a simple control technique (PI controller) will be used in both configurations to determine if it can be used to maintain stability within a microgrid after being subjected to various types of disturbances. The current challenges in microgrid control are firstly identified from literature. A simulation model from the literature is retrieved, however, due to model constraints, a mathematical model must be derived that can be used to approximate the simulation model before any sort of control can be implemented. The mathematical model is used to design the control systems, which are then applied to the simulation model in the Simulink environment. To compare the centralised and decentralised control strategies, an experimental design is created in which scenarios are built to assess the performance of the control systems under extreme conditions derived from the literature. The results of each scenario will be analysed to determine the best performing approach for each scenario/application, and the techniques will be rated based on their capacity to minimise disruptions. This work underlines the importance of simplicity; the fact that a basic PI controller can sustain stability in a multi-inverter microgrid demonstrates the effectiveness of resilient control techniques. Future research may include the integration of Neural Networks into the control system to allow for approaches such as load forecasting, although this will significantly increase the control system’s complexity.