Mostrar registro simples

dc.contributor.advisorAlmeida, Ricardo Carvalho de, 1959-pt_BR
dc.contributor.otherUniversidade Federal do Paraná. Setor de Tecnologia. Programa de Pós-Graduação em Engenharia Ambientalpt_BR
dc.creatorCarvalho, Gleiciane Fernanda dept_BR
dc.date.accessioned2024-04-08T15:37:54Z
dc.date.available2024-04-08T15:37:54Z
dc.date.issued2015pt_BR
dc.identifier.urihttps://hdl.handle.net/1884/38179
dc.descriptionOrientador : Prof. Dr. Ricardo Carvalho de Almeidapt_BR
dc.descriptionDissertação (mestrado) - Universidade Federal do Paraná, Setor de Tecnologia, Programa de Pós-Graduação em Engenharia Ambiental. Defesa: Curitiba, 26/02/2015pt_BR
dc.descriptionInclui referências : fls. 99-103pt_BR
dc.description.abstractResumo: A utilização de modelos numéricos atmosféricos de mesoescala é de fundamental importância em projetos eólicos, visando obter conhecimento aprofundado do regime e padrão de vento em uma área de interesse. Sabe-se que a melhor configuração do modelo sempre irá depender do local, da aplicação e do período do ano. Para tanto, se faz necessária a execução de várias simulações na busca da melhor configuração, sendo que a maioria dos estudos foca nas resoluções espaciais e parametrizações físicas. Um dos recursos amplamente empregados em simulações de mesoescala é a regionalização dinâmica, que é o aninhamento de grades. A vantagem desse método é executar simulações em alta resolução em uma área específica sem que todo o domínio computacional tenha que ser processado com um espaçamento de grade muito pequeno. Apesar de muito utilizado não há uma análise sistemática sobre a influência da regionalização dinâmica sobre os resultados obtidos. Neste trabalho são apresentados resultados de simulações numéricas de alta resolução, com diferentes configurações de regionalização dinâmica, para analisar a sensibilidade do modelo numérico atmosférico Weather Research and Forescasting (WRF) na reprodução da velocidade do vento em duas estações meteorológicas no Estado do Paraná, Guaratuba, no litoral do estado, e Paranavaí, no interior. Foram realizadas simulações em um mês de verão, Janeiro, e de inverno, Julho, no ano de 2007 e comparados com dados de observação das estações meteorológicas de superfície. A velocidade do vento simulada a 10 metros de altura foi analisada por métodos estatísticos, avaliando-se o viés, a acurácia e a distribuição de frequência. Foi também utilizada a análise espectral de Fourier. Foram definidos oito experimentos com variação na quantidade de grades, fator de redução entre as grades e espaçamento de grade. Os resultados demonstraram que a utilização e a configuração do aninhamento interferem significativamente na forma do modelo representar as ondas meteorológicas de diferentes escalas, em função da dinâmica específica de cada região analisada. Guaratuba, região com terreno complexo e com circulação predominantemente local, obteve melhores resultados sem a utilização da regionalização dinâmica e com grades de domínio pequeno. Os melhores experimentos para essa região foram aqueles que conseguiram reproduzir as altas frequências do espectro corretamente. Paranavaí, região com terreno pouco acidentado e com circulação majoritariamente regida por eventos de mesoescala e de escala sinótica, foi melhor representada pelo experimento com o maior número de grades e de maior domínio. Esse experimento também conseguiu representar corretamente as baixas frequências do espectro, preponderante para esta estação. Portanto, a análise espectral prévia dos dados de observação do período de simulação desejado pode fornecer informações essenciais para a definição da melhor estratégia de regionalização dinâmica, otimizando o custo computacional e extraindo os melhores resultados do modelo.pt_BR
dc.description.abstractAbstract: The use of mesoscale atmospheric numerical models is fundamental in wind power projects, in order to obtain depth knowledge of the wind regime and patterns over an area of interest. It is known that the best configuration of the model always will depend on location, application and period of the year. For that it is necessary the execution of many simulations searching the best configuration, although the majority of studies focus in the spatial resolution and physical parameterization. One of the resources highly used in mesoscale simulation is the dynamic downscaling, which is the grid nesting. The advantage of this method is to execute high resolution simulations in a specific area without the entire computational domain had to be processed with very small grid spacing. Although used a lot there's not a systematic analysis of the influence of the dynamic downscaling over the results obtained. In this work are presented results of high resolution numerical simulations, with different configurations of dynamic downscaling, to analyze the sensitivity of the atmospheric numerical model Weather Research and Forecasting (WRF) in the reproduction of wind speed in two meteorological stations in Paraná State, Guaratuba, in the state coast, and Paranavaí, in the countryside. It was executed simulations in a summer month, January, and in winter, July, in the year 2007 and compared with observational data from the surface meteorological stations. The wind speed simulated at 10 meters of high was analyzed by statistical methods, evaluating the bias, accuracy and the frequency distribution. It was also used the Fourier spectral analysis. It was defined eight experiments with variation in quantity of grids, reduction factor between grids and grid spacing. The results showed that the use and the configuration of nesting interfere significantly in the way on the model represent the meteorological waves in different scales, in function of the specific dynamics of each analyzed region. Guaratuba, region with a complex terrain and with circulation predominantly local, got better results without the use of dynamic downscaling and with grid of small domain. The best experiments for this region were those which were able to reproduce the spectrum high frequencies correctly. Paranavaí, region with flat terrain and majority of circulation governed by mesoscale and synoptic scale events, was best represented by the experiment with major number of grids and with greatest domain. This experiment also manages to represent correctly the spectrum low frequencies, preponderant for this station. Therefore, the previous spectral analysis of the observational data from the wished simulation period can provide essential information for the definition of the best strategy of dynamic downscaling, optimizing the computational cost and extracting the best results of the model.pt_BR
dc.format.extent103 f. : il. algumas color., tabs.pt_BR
dc.format.mimetypeapplication/pdfpt_BR
dc.languagePortuguêspt_BR
dc.relationDisponível em formato digitalpt_BR
dc.subjectEngenharia Sanitáriapt_BR
dc.subjectVentos - Velocidadept_BR
dc.subjectFourier, Analise dept_BR
dc.subjectMeteorologia - Modelos matemáticospt_BR
dc.titleAnálise de sensibilidade do modelo WRF a diferentes configurações de regionalização dinâmica para modelagem da velocidade do ventopt_BR
dc.typeDissertaçãopt_BR


Arquivos deste item

Thumbnail

Este item aparece na(s) seguinte(s) coleção(s)

Mostrar registro simples