Prognóstico de Radiação Solar Através Modelos que Combinam as Técnicas Wavelet e Redes Neurais (Forecast of Solar Radiation Through Models that Combine the Wavelet and Neural Networks Techniques)
DOI:
https://doi.org/10.26848/rbgf.v7.5.p808-817Keywords:
Wavelet - Rede Neural - PrognósticoAbstract
O prognostico de variáveis meteorológicas, como radiação solar, sempre foi de grande importância para a tomada de decisão em ocasião de ocorrências de eventos incomuns. Nesse contexto é justificável a busca por modelos matemáticos e estatísticos que produzam melhores prognósticos para tais variáveis. Desta forma investiga-se a estratégia de conjunção que se compõe de duas técnicas muito utilizadas no tratamento de série temporal; a transformada Wavelet que mostra analiticamente o sinal no domínio do tempo e da frequência; e as RNA’s a quais são modelos de inteligência artificial. A combinação dessas duas técnicas, o que se denomina modelo híbrido, tem se mostrado eficaz no prognóstico de variáveis meteorológicas. Os dados diários de radiação solar são do Instituto Agronômico do Paraná/PR coletados no período de 1990 até 1995. Neste trabalho são estudadas conjunções de modelos híbridos com Redes Neurais e técnicas Wavelets, apresentando o resumo de alguns dos modelos sintetizados na literatura, para o prognóstico de radiação solar. Tais modelos híbridos estudados se mostraram satisfatórios no prognóstico dessa variável, pois apresentaram um melhor desempenho em relação aos modelos que não são híbridos, sendo o modelo que se mostrou mais eficiente no prognóstico foi o que utiliza as sub-séries da decomposição como entrada da Rede Neural, pois apresenta regressão com valores significativos (R próximo a 1).
A B S T R A C T
The prognosis of meteorological variables such as solar radiation has always been of great importance for decision making in time of occurrence of unusual events. In this context it is justifiable to search for mathematical and statistical models that produce better prognosis for these variables. Thus investigates the combination strategy which uses two techniques widely used in the treatment time series; Wavelet transform analytically shows that the signal in the time domain and frequency; and the RNA's which models of artificial intelligence. The combination of these two techniques, which is called the hybrid models, has proven effective in predicting meteorological variables. Daily data of solar radiation are the Agronomic Institute of Parana / PR collected from 1990 to 1995. In this work conjunctions of hybrid models with neural networks and wavelets techniques are studied, presenting a summary of some of the synthesized models in the literature for the prediction of solar radiation. Such hybrid models studied were satisfactory prognosis with this variable because it showed better performance compared to models that are not hybrids in the model that is more efficient prognosis was that uses the sub-series decomposition as input the Network neural, it presents significant regression values (R close to 1).
Keywords: Wavelet, Neural Network, Forecast.
Downloads
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2015 Samuel Vitor Saraiva, Ricardo Ferreira Amorim, Frede Oliveira Carvalho, Leonardo Domingues

This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with Revista Brasileira de Geografia Física agree to the following terms:
Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under the Creative Commons Attribution 4.0 International (CC BY 4.0) license that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
Authors are permitted to make their work available online before or during the editorial process, on academic social networks, digital repositories, or preprint servers. After publication in Revista Brasileira de Geografia Física, authors are expected to update the preprint or postprint versions on the platforms where they were originally made available, providing a link to the final published version and any other relevant information, with proper recognition of authorship and the initial publication in this journal.
You are free to:
Share — copy and redistribute the material in any medium or format for any purpose, even commercially.
Adapt — remix, transform, and build upon the material for any purpose, even commercially.
The licensor cannot revoke these freedoms as long as you follow the license terms.
Under the following terms:
Attribution — You must give appropriate credit , provide a link to the license, and indicate if changes were made . You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.
No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.