Made available in DSpace on 2015-03-26T13:50:19Z (GMT). No. of bitstreams: 1 texto completo.pdf: 5181773 bytes, checksum: 7144c056ee7a7f0c8dd5aaa582d3f5a7 (MD5) Previous issue date: 2007-06-18 / Conselho Nacional de Desenvolvimento Científico e Tecnológico / The goal of this study was to compare nine databases of precipitation and four databases of incident solar radiation of the South America, considering the meridional variation, and the variation across the main continental watersheds (Amazonas, Tocantins, São Francisco, Orinoco, Paraná/Prata, in addition to the main Patagônia basins) and across the main vegetation types (tropical evergreen forest, tropical deciduous forest, woodland, savanna and grassland/steppe). For the precipitation assessment, it was used three products derived from the interpolation of the observed data (CRU; Legates and Wilmott; Leemans and Cramer), three databases based on the composition of rain gauges with remote sensing (TRMM, CMAP and GPCP) and three reanalysis databases (NCEP/NCAR, ERA-40 and CPTEC). For the incident solar radiation assessment, in addition to the reanalysis products, the product generated by the algorithm GL1.2 was also included in the comparison. This product estimates the incident solar radiation from the GOES satellite images. The results show that the annual average precipitation fields of the different products present differentiated behavior between themselves. For example, the CPTEC reanalysis does not represent the main regimes of precipitation in the continent, overestimating the precipitation in the interior of northeast Brazil and underestimating the precipitation in the rest of the continent. There is a substantial agreement among annual mean rain gauge interpolation products, particularly between the products CRU and Leemans and Cramer, which tend to better represent the rainfall regime of Northeast Brazil, for example. The combined rain gauge remote sensing products obtained similar patterns, mainly CMAP and GPCP. Both represent the South Atlantic Convergence Zone (SACZ), a band of higher precipitation that spans from northwestern South America to the southeast of Brazil, extending itself over the Atlantic Ocean. The monthly average values of the incident solar radiation products of the NCEP/NCAR, ERA-40 reanalyses and the GOES estimates were similar for all vegetation types and watershed, while the CPTEC reanalysis presented higher values. / O objetivo deste trabalho é comparar nove bancos de dados de precipitação e quatro de radiação solar incidente para a América do Sul, considerando a variação meridional, a variação por diferentes bacias hidrográficas (Amazonas, Tocantins, São Francisco, Orinoco, Paraná/Prata, além das bacias dos rios da Patagônia) e pelos principais tipos de vegetação (floresta tropical, caatinga, cerrado e pampas). Para a precipitação foram utilizados três produtos provenientes da interpolação de dados observados (CRU; Legates e Willmott; Leemans e Cramer), três da composição de dados observados com sensoriamento remoto (TRMM, CMAP e GPCP) e três bancos de dados de reanálise (NCEP/NCAR, ERA-40 e CPTEC). Para a radiação solar incidente, além dos produtos de reanálise citados acima, também foi comparado o produto gerado pelo algoritmo GL1.2 que produz estimativas da radiação solar incidente a partir do conjunto de imagens do satélite GOES. Os resultados mostram que os campos de precipitação média anual dos diferentes produtos apresentam comportamento diferenciado entre si. Por exemplo, a reanálise do CPTEC não é capaz de representar os principais regimes de precipitação existentes no continente, demonstrando uma forte tendência em superestimar a precipitação no interior do nordeste brasileiro e subestimá-la na maioria das demais regiões. Nos produtos de interpolação de dados observados em superfície, verifica-se uma maior concordância na distribuição dos valores de precipitação média anual, principalmente entre os produtos CRU e Leemans e Cramer, que tendem a representar melhor os regimes de chuvas, como por exemplo, na região do bioma caatinga (nordeste do Brasil), onde o clima é semi-árido. Os produtos que combinam dados de estações pluviométricas com os de sensoriamento remoto mostram-se com valores bem próximos do esperado, principalmente o CMAP e GPCP. Ambos representam bem a Zona de Convergência do Atlântico Sul (ZCAS), que é a faixa de valores mais elevados de precipitação que abrange desde o noroeste do continente até o sudeste/sul do Brasil, estendendo-se até o oceano. Para a radiação solar incidente nota-se que os valores médios mensais dos produtos de reanálise do NCEP/NCAR, ERA-40 e as estimativas do satélite GOES se encontram bem próximos para todos os biomas e bacias estudados. Já a reanálise do CPTEC apresentou valores bem mais elevados para a radiação sobre a América do Sul, tanto para a média anual como nas médias mensais analisadas.

Thermal stresses can arise in glass used in architectonic realisations in buildings as a result of solar irradiation:  that is, sunlight.  The magnitude of these thermal stresses depends on the temperature conditions in the glass in terms of the temperature differences between warm and cold areas.  The work described here has verified the existence and distribution of thermal stresses in glass, both theoretically and experimentally, in a load-bearing element made of glass and wood, when exposed to incident solar radiation and various degrees of shading.  The finite element method was used in order to simulate the temperature distribution in the glass, together with the resulting associated thermal stresses.  Parts of these model results were then compared with results and data from experimental investigations in a solar simulator.  The experimental trials were carried out on a plain glass sheet and on a wood-framed glass sheet.  This frame assembly saw the glass fitted in a wooden frame such that its edge was flush with the edges of the frame and therefore freely exposed to the surroundings.  Two different grades of glass were analysed in the theoretical modelling and used in the experimental solar exposure:  a clear glass, with a low absorption coefficient, and a tinted glass having a relatively high absorption coefficient.  The experimental part of the work included two different finishes to the glass edge:  as-cut, with no further attention, and cut and smoothed.  The results from the finite element method calculations agree relatively well with the experimental results.  When simulating the temperature values and distributions, the value of the coefficient of heat transfer is an important factor in determining the results.  The coefficient can be given either a constant value, taken from a standard, or a calculated value that varies depending on the surface temperature and ambient temperature at every instant.  For the clear glass with a low coefficient of absorption, the calculated temperatures did not differ significantly depending on which method had been used to provide a value for the coefficient of heat transfer.  However, for the glass with a high coefficient of absorption, and when exposed to high solar intensity, a calculated value of coefficient of heat transfer should be used in order to arrive at relevant values of surface temperatures and stresses.  Thermal tensile stresses have more effect on the total stress level than have stresses arising from typical in-service vertical loads.