Towards real-time power management of microgrids for power system integration : a decentralized multi-agent based approach
Colson, Christopher Michael.
MetadataShow full item record
The steadily increasing need for electrical power, rising costs of energy, market forces and industry deregulation, an aging infrastructure, tight constraints on new long distance transmission lines, global environmental concerns, and a public demand for greater electrical reliability and security are overwhelming our existing power system. One technology that offers solutions to many of these challenges and addresses smart grid objectives directly is: microgrids. A microgrid is a small (typically several MW or less in scale) power system incorporating distributed generators, load centers, potentially storage, and the ability to operate with or apart from the larger utility grid. Properly managed, assets connected within a microgrid can provide value to the utility power network, improve energy delivery to local customers, and facilitate a more stable electrical infrastructure, benefitting environmental emissions, energy utilization, and operational cost. While microgrids can achieve significant improvements for customers and utilities alike, microgrid research is in its infancy and, to date, a comprehensive means of managing microgrid operations has not been realized. In this work, two primary efforts are undertaken. First, given the lack of a comprehensive software test bed for microgrids, a simulation environment capable of incorporating microgrid operational concepts, electrical modeling, asset dynamics, and control conditions is developed. Second, using the simulation environment, an enhanced decentralized multi-agent power management and control system is designed and evaluated for the purpose of supervising multiobjective microgrid operations under normal and emergency conditions. Results presented demonstrate effective multi-agent methods that yield improved microgrid performance, as well as facilitate coordinated system decision-making without reliance on a centralized controller. These advancements represent innovation towards the autonomous operation of microgrids, as well as provide important insight into new tradeoff considerations associated with multi-objective design for power management. Microgrids are infrastructure elements that bridge the gap between emerging energy technologies and the existing power system. Simply put, smart grid objectives including higher penetration of renewables, integration of storage, delivery efficiency improvements, more responsive system elements, stronger resiliency, and improved flexibility will be difficult to achieve without microgrids. The simulation environment developed and the power management methodology presented are important steps towards enabling microgrids and realizing their benefits.