Despacho Econômico Ótimo de Microrrede Usando Algoritmos de Controle Distribuído no Esquema Matricial e no Sistema Multiagente

Abstract

The growth of distributed generation has been accountable for changes in characteristics of the power distribution systems (SDEs), from passive to active networks, from unidirectional to bidirectional power flow and with consumers and prosumers. The multiple energy resources in the SDEs are best managed in the form of microgrids, which are small power systems made up of sources, energy storage devices and controllable loads. For the management of a microgrid, optimal economic dispatch (DEO) methods are used, which aim to obtain the lowest total operating cost of the microgrid, taking into account technical and regulatory restrictions. The objective of this work is to develop and simulate an optimal dispatch of a microgrid that comprises dispatchable sources, non-dispatchable sources, batteries, and controllable loads, using distributed optimization algorithms such as Consensus, Diffusion, Exact Diffusion, and Multi-agent systems (MAS). The Python Agent Development (PADE) platform was used to create and run the agents that participate in the optimal dispatch of the microgrid. The performance of the methods is compared in terms of computational cost and results obtained from the DEO. Among the optimization methods with matrix modeling, exact diffusion proved to be the fastest method with the lowest total computational cost when compared to its counterparts. Additionally, the Exact Diffusion method is the method that most respects the privacy of the data of the agents, with the diffusion parameter the only variable interchanged between agents, respecting the confidentiality of other data. As for the application of a Multiagent system on the PADE platform, an algorithm was developed to use the mosaik API enabling synchronized and successful communication between the microgrid agents. The Consensus algorithm was applied in the SMA, successfully performing the DEO of the microgrid.Finally, an extra SMA exclusive case was made to show the Plug and Play tool.