Banca de DEFESA: ERIC BERNARDES CHAGAS BARROS

Uma banca de DEFESA de DOUTORADO foi cadastrada pelo programa.
DISCENTE : ERIC BERNARDES CHAGAS BARROS
DATA : 23/09/2021
HORA: 09:00
LOCAL: Sala Virtual (https://meet.google.com/hdh-qyrg-ais)
TÍTULO:

Jemadar-AI: A Fog-Based Framework for Microgrid Power Management

PALAVRAS-CHAVES:

smart-grids, microgrids, fog computing, cloud computing, scheduling

PÁGINAS: 123
GRANDE ÁREA: Ciências Exatas e da Terra
ÁREA: Ciência da Computação
RESUMO:

The constant increase in demand for electricity and the reduction of conventional means of production point to a future energy crisis. One of the factors that impact the growing demand is uncontrolled consumption, generating an increasing energy dispatch to meet the consumer's needs. Thus, production needs to use a greater amount of energy sources, especially during peak consumption hours. To control demand, demand management techniques are being used by consumers to reduce peak periods and energy bills. Likewise, energy production can combine the use of renewable and non-renewable energy sources with a focus on reducing non-renewable energy production. In this scenario, intelligent energy production requires automated control to avoid energy losses. Currently, several researchers propose the use of optimization algorithms for electrical equipment scheduling as a demand management solution, so the use of equipment be displaced to avoid peak energy periods. Research has also been proposed on the use of the Proportional-Integral-Derivative (PID) control to balance renewable and non-renewable production. However, few researchers consider fog computing with the help of the cloud to perform the processing of these algorithms. Thus, this work proposes the JEMADAR-AI framework to deal with the data traffic of electrical networks. JEMADAR-AI trains two neural networks based on reinforcement learning: (i) the first one to performs equipment scheduling; (ii) the second adjusts the PID controller dynamically. For this, JEMADAR-AI uses an architecture that combines fog computing with cloud computing to reduce request response times for equipment scheduling and power adjustment requests. The results obtained through the performance evaluation showed that the use of JEMADAR-AI reduces up to 19% the electrical power at peak hours and 21.6% the values of the energy bills when compared with scenarios that do not perform any optimization.

MEMBROS DA BANCA:
Presidente - 1850683 - MAYCON LEONE MACIEL PEIXOTO
Interno - 1814505 - GUSTAVO BITTENCOURT FIGUEIREDO
Interno - 1764465 - LEOBINO NASCIMENTO SAMPAIO
Externo à Instituição - LUIZ FERNANDO BITTENCOURT
Externo à Instituição - EDWARD DAVID MORENO ORDONEZ