Date of Award


Document Type

Open Access Dissertation


Electrical Engineering


College of Engineering and Computing

First Advisor

Herbert L. Ginn III


A microgrid may have numerous multi-functional power electronic converters connecting sources, loads, and storage to the system. Systems where converters are the interface between many of the main sources of energy and load centers have the ability to direct the flow of energy if the control of the converters is coordinated. The influence of energy flow in a microgrid by coordinated action of converters is referred to here as ‘energy routing’. Energy routing allows for reduction of systems losses by optimizing source operating points and reducing transmission and distribution path losses. Energy ramp rates at various points in the system can also be manipulated by coordinated control of energy flow through the converters.

Converters can be coordinated centrally or in a distributed fashion. A distributed coordination system approach can enable system level converter control while avoiding single points of failure that are inherent in a centralized hierarchical control system, and that is robust and expandable. Research performed in the area of distributed control indicates that a control based on a multi-agent system (MAS) has the potential to satisfy the distributed converter control requirements. Here an optimization technique is developed that can be distributed for parallel computation by MAS type control systems.

An optimization algorithm will be presented that dynamically determines global optimal values of discretized command variables to the converters in a distributed fashion in order to ensure most economic fuel usage of the sources and minimization of distribution loss simultaneously. Converter command variables are discretized in order to formulate the optimization problem as a Mixed Integer Quadratic Programming (MIQP) problem. The MIQP framework allows decomposition of the optimization algorithm as well as pruning of the search span by a factor of hundreds. Thus, it provides very fast convergence to the optimal solution and ensures that the communication requirements are feasible for real-time system level coordination.

order to validate the distributed optimization and control method developed in this research, a simplified shipboard DC power distribution system and CERTS (Consortium for Electric Reliability Technology Solutions) microgrid are used for case studies. These are isolated microgrids with converters between all sources of energy and the main buses as well as between all load centers and the main buses. Energy routing through the branches is directly maintained by controlling the command variables input to the converters. Sources as well as storage are indirectly manipulated to their optimal set points by these discretized command variables. Simulation based validation is performed for both test systems.