Development of multi-agent system based energy management system for micro grids

Abstract

In this thesis, our objective is to introduce multi agent concept for the energy management of micro grids. This objective is very significant as micro grids are emerging as a method to integrate the operation of distributed energy sources (DES) in modern power systems. They introduced advanced communication technologies, optimizing techniques, sensing and monitoring features for the power distribution network. However, optimal energy management is still a challenge for microgrids when controlling renewable DES’s with intermittent generating patterns. In this research, a Multi Agent System (MAS) based architecture is used for the operation of energy management system (EMS) of a microgrid. The objective of this research is to implement an agent based control architecture for DES’s, loads and energy storage systems (ESS) of a microgrid to achieve optimal energy management. This document discusses the modeling, simulation and hardware implementation of agent based energy management system. Initially, JADE (Java Agent Development Environment) is used to implement the agent based control architecture. A microgrid is modeled in Matlab/Simulink and interlinked with agents developed in JADE through a middle layer. The decision making authority is shifted to ground hierarchy, enabling local agents to take control over DES’s to optimize the renewable power consumption. The outcome of this research shows that MAS based control architecture can be used to optimize the energy management in a microgrid. This thesis presents a complete literature review about the background of this research and similar projects, and the results obtained through the research in both simulation and hardware implementation in developing a multi agent based energy management system for a micro grid.