You are in the accessibility menu

Please use this identifier to cite or link to this item: http://acervodigital.unesp.br/handle/11449/90698
Title: 
Estimativa da radiação solar ultravioleta em Botucatu/SP Brasil utilizando técnicas de aprendizado de máquina
Author(s): 
Almeida, Thiago do Nascimento Santana de
Institution: 
Universidade Estadual Paulista (UNESP)
Abstract: 
  • O presente trabalho tem como objetivo avaliar a estimação da radiação solar ultravioleta diária (UV) utilizando técnicas de aprendizado de máquina em Botucatu/SP/Brasil. Para a geração dos modelos foram utilizadas as redes neurais artificiais com função linear (RNA), a máquina de vetores de suporte com função linear (SVM-Linear) e a máquina de vetores de suporte com função RBF (SVM-RBF). Como entrada, para cada uma das técnicas, foram testados cinco grupos contendo diferentes variáveis meteorológicas medidas como rotina na estação de radiometria solar de Botucatu. A maior precisão na estimação da UV foi obtida utilizando a SVM-RBF e, como variáveis de entrada, foram utilizados dados de insolação, umidade relativa, temperatura do ar, precipitação, fotoperíodo, radiação solar no topo da atmosfera, radiação ultravioleta no topo da atmosfera, radiação solar global, transmissividade atmosférica e massa ótica relativa. Este modelo apresentou MBE = 0,321%; RMSE = 5,712%; d = 0.991 e R² = 0.969, porém o conjunto de entradas utilizando apenas a radiação global, radiação solar no topo da atmosfera e radiação ultravioleta no topo da atmosfera merece atenção por apresentar resultados próximos a partir de apenas uma variável medida (radiação global) e duas calculadas. Este grupo apresentou MBE = 1,614%; RMSE = 6,749%; d= 0,989 e R² = 0,959. Tais resultados foram comparados com modelos encontrados na literatura e se mostraram equivalentes aos que utilizaram técnicas de aprendizado de máquina para estimar a radiação UV em outras cidades e superiores aos modelos estatísticos gerados com o intuito de estimar a UV para a cidade de Botucatu
  • In this papper is evaluated the estimation of daily solar ultraviolet radiation (UV) using machine learning techniques in Botucatu / SP / Brazil. To develop the model was utilized the artificial neural networks with linear function, the support vector machine with linear function and with RBF function. As input to each of the techniques, were tested five groups containing different weather variables measured as routine in Botucatu radiometry solar station. A more precise estimation of UV was obtained using the SVM-RBF and as input variables, we used data of insolation, relative humidity, air temperature, precipitation, photoperiod, solar radiation at the top of the atmosphere, ultraviolet radiation at the top of atmosphere, solar radiation, atmospheric transmissivity and relative optical mass. This model presented MBE = 0.321%; RMSE = 5.712%; d = 0991 and R²=0969, but the set of entries using only the global radiation, solar radiation at the top of the atmosphere and ultraviolet radiation at the top of the atmosphere deserves attention for presenting similar results from only one measured variable (global solar radiation) and two calculated. This group presented MBE = 1.614%; RMSE = 6.749%; d = 0.989 and R²=0.959. These results were compared to models found in the literature and proved equivalent to that used machine learning techniques to estimate the UV radiation in other cities and higher than the statistical models generated in order to estimate the UV to the Botucatu city
Issue Date: 
13-Feb-2013
Citation: 
ALMEIDA, Thiago do Nascimento Santana de. Estimativa da radiação solar ultravioleta em Botucatu/SP Brasil utilizando técnicas de aprendizado de máquina. 2013. viii, 57 f. Dissertação (mestrado) - Universidade Estadual Paulista, Faculdade de Ciências Agronômicas de Botucatu, 2013.
Time Duration: 
viii, 57 f. : il. color. , gráfs, tabs.
Publisher: 
Universidade Estadual Paulista (UNESP)
Keywords: 
  • Radiação ultravioleta
  • Radiação solar
  • Modelos matematicos
  • Redes neurais (Computação)
  • Radiation
URI: 
Access Rights: 
Acesso aberto
Type: 
outro
Source:
http://repositorio.unesp.br/handle/11449/90698
Appears in Collections:Artigos, TCCs, Teses e Dissertações da Unesp

There are no files associated with this item.
 

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.