Banca de DEFESA: PAULO GUILHERME SILVA DOS SANTOS

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
STUDENT : PAULO GUILHERME SILVA DOS SANTOS
DATE: 25/08/2021
TIME: 14:00
LOCAL: ttps://meet.google.com/wup-gwpz-zyk
TITLE:

FORECAST OF ENVIRONMENTAL VARIABLES IN THE AMAZON USING LONG SHORT-TERM  ARTIFICIAL NEURAL NETWORKS

KEY WORDS:

Temperature forecasting, Solar irradiance forecasting, Deep Learning, Long short-term Memory

PAGES: 42
BIG AREA: Outra
AREA: Ciências Ambientais
SUMMARY:

With the expansion of the photovoltaic solar energy industry came the search to improve the efficiency of its systems. The Prediction of environmental variables such as temperature and solar irradiance help in decision making on the use of these systems. In this work we use a long short-term memory (LSTM)-based neural network to generate different models in different architectures and evaluate them. Data were obtained for two cities located in the Amazon region, where we obtained two temperature datasets and one solar irradiance dataset. During the treatment of datasets, characteristics of the data series such as autocorrelation and stationarity were verified, and the data were divided into training-test and validation. The architectures used have differences in their number of layers, to analyze the influence of the complexity of the architectures on their performance. In this study, the results demonstrated that a less complex LSTM architecture obtains superior performance, which decays as the number of layers increases. In data validation, the 3-layers architecture presented a statistically significant difference from the 7-layers architectures. For the LABIC temperature dataset the RMSE averages of the two architectures were 0.9393°F and 1.4531°F, for 3 and 7 layers, respectively; for the GEDAE temperature dataset the mean RMSE was 1.6499°F and 1.9767°F, for 3 and 7 layers, respectively; and in the solar irradiance dataset we obtained an average RMSE of 170.6649 W/m² and 204.7825 W/m², for 3 and 7 layers, respectively. The 3-layer architecture proved to be the best architecture among those used

BANKING MEMBERS:
Externo à Instituição - ROBERTO SCHIRRU - UFRJ
Presidente - 1963026 - ANDERSON ALVARENGA DE MOURA MENESES
Externa ao Programa - 2143011 - HELAINE CRISTINA MORAES FURTADO
Interno - 1794276 - JOSE MAURO SOUSA DE MOURA
Externa à Instituição - MARLA TERESINHA BARBOSA GELLER - ULBRA