Este trabalho tem como objetivo implementar e testar modelos de gás cinza atualizados na rotina de radiação térmica do software Fire Dynamics Simulator (FDS), além da utilização do próprio modelo de gás cinza disponível no software, para a predição do fluxo de calor radiativo. Os modelos de gás cinza estudados foram o modelo padrão do software FDS (aqui denominado como GC1), e os modelos de gás cinza mais atuais: o GC2, no qual o coeficiente de absorção do meio participante é dado por relações polinomiais, e o GC3, sendo este um modelo de gás cinza que baseia o cálculo do coeficiente de absorção no modelo WSGG. Os novos modelos de gás cinza foram implementados no código fonte do software FDS, o qual é um código aberto, e a verificação da implementação foi realizada através da solução numérica do equacionamento utilizando os valores reportados pelo software. Com os novos modelos de gás cinza já corretamente implementados, passou-se então para a simulação computacional dos casos previamente selecionados. Para todos os modelos de gás cinza, foram simulados incêndios em poças, para diferentes combustíveis (etanol, n-heptano e metanol) em diferentes cenários de incêndio, considerando ou não a presença de fuligem no sistema. Os cenários de incêndio eram: (i) totalmente fechado, (ii) totalmente aberto e (iii) com uma condição intermediária, fechado, porém com uma abertura para o meio externo. Um estudo de análise de malha e de diferentes parâmetros, como o estudo da quantidade necessária de ângulos sólidos discretos, foram realizados para correta padronização dos parâmetros. As simulações computacionais foram validadas para o modelo de gás cinza padrão do FDS através da comparação de resultados com aqueles reportados na literatura específica de cada caso. Com os modelos já validados simulou-se novamente cada cenário de incêndio com os diferentes modelos de gás cinza anteriormente implementados. A partir da análise dos resultados obtiveram-se boas concordâncias para os campos de temperatura, frações molares tanto de CO2 quanto de H2O e para as frações volumétricas de fuligem. Os fluxos de calor radiativos foram corretamente preditos para todos os modelos de gás cinza implementados. O modelo GC2 apresentou resultados com desvios médios na faixa de 15%, o modelo de gás cinza baseado no WSGG (GC3) apresentou os melhores resultados, com erros médios inferiores a 10%, enquanto que o modelo padrão do software, GC1, apresentou resultados intermediários. / This work aims to implement and test updated gray gas models in the thermal radiation routine of the Fire Dynamics Simulator (FDS) software, as well as the use of the gray gas model available in the software to the prediction of radiative heat flux. The gray gas models studied were the default model of the FDS software (determined GC1), and the most current gray gas models: the GC2, in which the absorption coefficient of the participant medium is given by a polynomial relations, and the GC3, which is a gray gas model that was based on the calculation of the absorption coefficient in the WSGG model. The most recently gray gas models were implemented in the source code, which is an open source, and the verification of the implementation was performed by the numerical solution of the equations from the reported values of the software. With the new gray gas models already implemented, the next step was the computational simulation of the previously selected cases. For all the gray gas models, pool fires were simulated different scenarios of fire for different fuels (ethanol, nheptane and methanol), with and without considering soot presence in the system. The fire scenarios were: (i) fully closed, (ii) fully open and (iii) with an intermediate condition, closed but with an opening to the external environment. A study of a mesh analysis and different parameters, such as the study of the required amount of discrete solid angles, were performed to correct the standard parameters. The computational simulations were verified for the default gray gas model of the FDS by comparing the simulations results with those reported in the specific literature of each case. With the models already verified, each fire scenario was simulated with the different gray gas models previously implemented. From the analysis of the results, good agreements were obtained for the fields of temperature, molar fraction of CO2 and H2O and soot volume fraction. The radiative heat fluxes were correctly predicted for all gray gas models early implemented. The GC2 model present results with average deviation in the range of 15%, the gray gas model based on WSGG (GC3) presented the best results, with average deviation lower than 10%, while the default software model (GC1) presented intermediate results.

The Meyers et al. (2007) Tampa Bay Model produces water level and three-dimensional current and salinity fields for Tampa Bay. It is capable of computing temperature but is presently run without active thermodynamics. Variations in water temperature are driven by heat exchange at the water-atmosphere boundary and advective heat flux at the mouth of the bay. The net heat exchange surface boundary condition is required for computations of three-dimensional temperature fields. Components of the surface heat budget were measured or derived at an observational tower in Middle Tampa Bay. Net heat exchange at the surface of Tampa Bay was computed from June 2002 to May 2005. Total heat energy gained or lost at the bay-atmosphere interface includes turbulent and radiative heat fluxes. An initial examination of turbulent heat exchange, the portion of total surface heat flux driven by atmospheric turbulence, demonstrated the skill of a bulk flux algorithm (TOGA COARE v. 3.0) in predicting measured sensible heat flux over Tampa Bay (R² = 0.80 and RMSE of 11.02 W/m² from June through November of 2002). Insolation was measured directly at the observational tower. Solar radiation is reflected in proportion to sea surface albedo, computed following Payne (1972). Based upon Secchi depth readings, Tampa Bay was classified as a water body type 7. The amount of penetrating insolation reflected from the bottom was computed for this type 7 estuary. Upwelling longwave radiation is emitted in proportion to the water temperature according to the Stefan-Boltzmann law. Eleven bulk formulas for computing downwelling longwave radiation were assessed for skill in reproducing observations made at buoys moored on the West Florida Shelf. Berliand and Berliand (1952) best represented downwelling longwave heat flux measurements at the buoys and is appropriate for application over Tampa Bay. Surface heat flux dominates cooling in fall and warming in spring while advective heat exchange becomes important during the summer. Extreme events, including tropical cyclones and extratropical fronts, dramatically impact surface heat exchange, driving rapid cooling. The methods applied in computation of heat flux components are amenable to real-time modeling exercises.

A new method of power production, called pressurized oxy-fuel combustion, burns coal with CO2 and oxygen, rather than air, bringing us closer to the end goal of developing zero emission coal-fired utility boilers. However, high-pressure, high-temperature systems such as these are under-studied, and their behavior is difficult to measure. An accurate model for previously untested conditions requires data for validation. The heat release profile of flames and their radiative intensity is one of the key data sets required for model validation of an oxy-coal combustion system. A radiometer can be used to obtain the necessary radiative heat flux data. However, several studies show significant measurement errors of past radiometer designs. This work focuses on developing a narrow angle radiometer that can be used to describe radiative heat transfer from a pressurized oxy-coal flame. The sensitivity of the instrument to outside environmental influences is thoroughly examined, making it possible to obtain the axial radiative heat flux profile of the flame in a 100kW pressurized facility by accurately converting the measured quantities into radiative heat flux. Design aspects of the radiometer are chosen to improve the accuracy of radiative heat flux measurements as well as conform to the physical constraints of the 100kW pressurized facility. The radiometer is built with a 0.079-inch aperture, an 8.63-inch probe internally coated with high emissivity coating, four baffles spaced evenly down the length of the probe, no optic lens, a thermopile as the sensor, argon purge gas, and a water-cooled jacket. The radiometer has a viewing angle of 1.33 degrees. The instrument is calibrated with a black body radiator, and these calibration data are used in combination with radiation models to convert the radiometer signal in mV to radiative heat flux in kW/m2. Environmental factors affecting accuracy are studied. The results of the calibration data show that the radiometer measurements will produce a calculated heat flux that is accurate to within 5.98E-04 kW/m2.