Nombre: LUCAS CASTRO DE REZENDE

Fecha de publicación: 17/08/2022
Supervisor:

Nombre	Papel
AUGUSTO CÉSAR RUEDA MEDINA	Advisor *

Junta de examinadores:

Nombre	Papel
AUGUSTO CÉSAR RUEDA MEDINA	Advisor *
CLAINER BRAVIN DONADEL	External Examiner *
HÉLIO MARCOS ANDRÉ ANTUNES	Internal Examiner *

Sumario: In recent years, the concern regarding the use of energy from clean sources has increased considerably, resulting in several new technologies and system models in order to alleviate the greenhouse effect related to the burning of fossil fuels. Electric vehicles and distributed generators from renewable sources are pointed out as solutions to the problem; however, its implementation in the system can cause adversity in the power system, if done in an uncoordinated way. In this study, the objective was to develop a solution using the Genetic Algorithms Method, with the aid of the A Mathematical Programming
Language (AMPL) software to obtain an algorithm that optimally coordinates both the loading of the electric vehicles, as the power supplied by each distributed generator in each period of time, in order to respect the operational and safety limits of the electric energy distribution system, of the electric vehicles and of the distributed generators. The method showed satisfactory results in its application and reduced the system cost by approximately 12.19% compared to the system without distributed generators and electric
vehicles. In the second stage of the research, hardware parallelization methods were presented, which consists of applying specific tasks to a dedicated hardware, increasing the computation speed. From the demonstrations, it was applied in a GA and observed that, in fact, there are considerable gains in processing speed, being a viable alternative to implement together with the GA.
Keywords: Genetic Algorithm; Coordinated Charge; Distributed Generation; Real-time Simulation; Electric Vehicles, Hardware-in-the-loop; High-Level Synthesis; Hardware parallelization methods.

Acceso al documento