The convective mass transfer coefficient is determined for evaporation in a horizontal rectangular duct, which forms the test section of the transient moisture transfer (TMT) facility. In the test facility, a short pan is situated in the lower panel of the duct where a hydrodynamically fully developed laminar or turbulent airflow passes over the surface of the water. The measured convective mass transfer coefficients have uncertainties that are typically less than ±10% and are presented for Reynolds numbers (ReD) between 560 and 8,100, Rayleigh numbers (RaD) between 6,100 and 82,500, inverse Graetz numbers (Gz) between 0.003 and 0.037, and operating conditions factors (H*) between -3.6 and -1.4. The measured convective mass transfer coefficients are found to increase as ReD, RaD, Gz and H* increase and these effects are included in the Sherwood number (ShD) correlations presented in this thesis, which summarize the experimental data.<p> An analogy between heat and mass transfer is developed to determine the convective heat transfer coefficients from the experimentally determined ShD correlations. The convective heat transfer coefficient is found to be a function of ShD and the ratio between heat and moisture transfer potentials (S*) between the surface of the water and the airflow in the experiment. The analogy is used in the development of a new method that converts a pure heat transfer NuD (i.e., heat transfer with no mass transfer) and a pure mass transfer ShD (i.e., mass transfer with no heat transfer) into NuD and ShD that are for simultaneous heat and mass transfer. The method is used to convert a pure heat transfer NuD from the literature into the NuD and ShD numbers measured in this thesis. The results of the new method agree within experimental uncertainty bounds, while the results of the traditional method do not, indicating that the new method is more applicable than the traditional analogy between heat and mass transfer during simultaneous heat and mass transfer.<p>A numerical model is developed that simulates convective heat and mass transfer for a vapor permeable Tyvek® membrane placed between an airflow and liquid water. The boundary conditions imposed on the surfaces of the membrane within the model are typical of the conditions that are present within the TMT facility. The convective heat and mass transfer coefficients measured in this thesis are applied in the model to determine the heat and moisture transfer through the membrane. The numerical results show that the membrane responds very quickly to a step change in temperature and relative humidity of the air stream. Since the transients occur over a short period of time (less than 1 minute), it is feasible to use a steady-state model to determine the heat and mass transfer rates through the material for HVAC applications.<p>The TMT facility is also used to measure the heat and moisture transfer through a vapor permeable Tyvek® membrane. The membrane is in contact with a water surface on its underside and air is passed over its top surface with convective boundary conditions. The experimental data are used to verify the numerically determined moisture transfer rate through the Tyvek® membrane. The numerical model is able to determine the mass transfer rates for a range of testing conditions within ±26% of the experimental data. The differences between the experiment and the model could be due to a slightly different mass transfer coefficient for flow over Tyvek® than for flow over a free water surface.

The convective mass transfer coefficient is determined for evaporation in a horizontal rectangular duct, which forms the test section of the transient moisture transfer (TMT) facility. In the test facility, a short pan is situated in the lower panel of the duct where a hydrodynamically fully developed laminar or turbulent airflow passes over the surface of the water. The measured convective mass transfer coefficients have uncertainties that are typically less than ±10% and are presented for Reynolds numbers (ReD) between 560 and 8,100, Rayleigh numbers (RaD) between 6,100 and 82,500, inverse Graetz numbers (Gz) between 0.003 and 0.037, and operating conditions factors (H*) between -3.6 and -1.4. The measured convective mass transfer coefficients are found to increase as ReD, RaD, Gz and H* increase and these effects are included in the Sherwood number (ShD) correlations presented in this thesis, which summarize the experimental data.<p> An analogy between heat and mass transfer is developed to determine the convective heat transfer coefficients from the experimentally determined ShD correlations. The convective heat transfer coefficient is found to be a function of ShD and the ratio between heat and moisture transfer potentials (S*) between the surface of the water and the airflow in the experiment. The analogy is used in the development of a new method that converts a pure heat transfer NuD (i.e., heat transfer with no mass transfer) and a pure mass transfer ShD (i.e., mass transfer with no heat transfer) into NuD and ShD that are for simultaneous heat and mass transfer. The method is used to convert a pure heat transfer NuD from the literature into the NuD and ShD numbers measured in this thesis. The results of the new method agree within experimental uncertainty bounds, while the results of the traditional method do not, indicating that the new method is more applicable than the traditional analogy between heat and mass transfer during simultaneous heat and mass transfer.<p>A numerical model is developed that simulates convective heat and mass transfer for a vapor permeable Tyvek® membrane placed between an airflow and liquid water. The boundary conditions imposed on the surfaces of the membrane within the model are typical of the conditions that are present within the TMT facility. The convective heat and mass transfer coefficients measured in this thesis are applied in the model to determine the heat and moisture transfer through the membrane. The numerical results show that the membrane responds very quickly to a step change in temperature and relative humidity of the air stream. Since the transients occur over a short period of time (less than 1 minute), it is feasible to use a steady-state model to determine the heat and mass transfer rates through the material for HVAC applications.<p>The TMT facility is also used to measure the heat and moisture transfer through a vapor permeable Tyvek® membrane. The membrane is in contact with a water surface on its underside and air is passed over its top surface with convective boundary conditions. The experimental data are used to verify the numerically determined moisture transfer rate through the Tyvek® membrane. The numerical model is able to determine the mass transfer rates for a range of testing conditions within ±26% of the experimental data. The differences between the experiment and the model could be due to a slightly different mass transfer coefficient for flow over Tyvek® than for flow over a free water surface.

In a gas turbine, a small portion of air is bled from the compressor to provide cooling to keep the turbine at a safe operating temperature. The air flows through several passages in between where the components of the turbine are assembled. In this study, the heat transfer and pressure loss characteristics of two of these passages are investigated experimentally. The first of the two passages investigated is the passage in between the turbine blade root and disc. This passage has a unique geometry resembling an S-shape. The heat transfer and pressure loss characteristic of this passages in not well documented. For this study, a model of the realistic S-shaped passage has been made. In addition, a simplified rectangular duct with hydraulic diameter similar to that of the realistic S-shaped passage was constructed along with three other rectangular passages at aspect ratios, 17.33, 8.81, 3.93, and 2.02. This study aims to determine if rectangular duct correlations are valid for the realistic S-shaped model. Specifically, flow in low Reynolds number ranges of less than 3000 are of interest. With the effect or rotation and aspect ratio being of primary concern in the study, an experimental rig was constructed to measure the heat transfer and pressure loss in these models. The experiments were conducted with both clockwise and counterclockwise rotation to account for the passage on the pressure side and suction side of the passage. The centerline Nusselt number distribution measured to check if the flow was fully developed. The effect of rotation caused swirling, increasing the entrance length in the duct and also enhanced heat transfer. The rotation also enhanced the heat transfer in the fully developed region. The fully developed experimental data for the simplified rectangular ducts showed good correlation with that of literature. However, the realistic S-shaped duct showed lower heat transfer values than the simplified rectangular ducts. Still, the effect of rotation is seen enhancing the rotation inf the realistic S-shaped duct. Additionally, the friction factor which was measured using the cross-sectional average static pressure showed similar results for the realistic S-shaped duct and the simplified rectangular duct. The passage between turbine disc bore and shaft is modeled as an annular duct with inner surface rotation. Flow in the turbulent region is studied and the test sections are made to have short length to hydraulic dimeter ratios. Along the centerline, the onset of Taylor vortices can be seen enhancing the Nusselt number in regions where the flow should be fully developed. This effect can also be seen enhancing the heat transfer in the fully developed region. The presence of Taylor vortices also adds resistance increasing the pressure loss across the duct. / Master of Science / Industrial gas turbines are designed to have an optimum overall pressure ratio for target temperatures rise. The demand for higher efficiency and power continues to push the operating pressure and temperature. Air systems is the flow network to provide necessary cooling to keep the machinery at a safe operating temperature. In this study, two passages of the air system in the turbine are of interest. The passage between turbine blade root and disc, and the passage between the turbine disc and shaft. The effect of rotation on the flow through the two passages are of primary interest with focus on heat transfer and pressure loss characteristics. This experimental study presents unique results as a realistic model of the passage which resembles an S-shape was constructed and tested. The passage in between the turbine disc and shaft forms a rotating annular passage. There is limited data available representing the realistic geometrical shape of the annular passage under rotation. Therefore, the present study aims to present data for more realistic geometry and operating conditions. In addition, simplified rectangular ducts and annular ducts are also tested for comparison purpose. The results of the study showed that the rotation does provide a significant increase in heat transfer and pressure loss in experiment modeling the passage between the turbine blade root and disc. Comparing the realistic S-shape passage and the rectangular passage with similar aspect ratio, the realistic S-shape passage showed less heat transfer and less sensitivity to the effect of rotation. The pressure loss characteristics on the other hand proved to be very similar. For the experiments modeling the passage between turbine disc and shaft, the effect of rotation once again showed to increase the heat transfer and pressure loss. The effect is more prominent when there is less axial flow.

Les gaines rectangulaires utilisées dans les réseaux de transport de gaz, notamment dans l'industrie de l'aluminium, sont soumises à des excitations provenant du flux s'écoulant à l'intérieur. Ce travail de thèse vise à comprendre quels sont les phénomènes impliqués dans la vibration des gaines rectangulaires de transport de gaz. Dans un premier temps nous présentons un modèle semi-analytique de gaine rectangulaire homogène basé sur le couplage de 4 plaques. Cette modélisation nous permet ainsi de définir les caractéristiques modales de la gaine. Ce modèle a été validé expérimentalement et numériquement par un code élément-finis. Dans un second temps, nous appliquons diverses excitations aérodynamiques et aéroacoustiques à notre gaine afin de déterminer quelles sont les contributions prépondérantes. Les comparaisons entre les résultats numériques et expérimentaux dans le cas d'un écoulement en gaine droite montrent que les contributions aéroacoustiques sont prépondérantes. Les mêmes tendances ressortent lors d'essais dans des configurations en coude, sauf à proximité de celui-ci où les sources aérodynamiques sont également importantes. Pour finir, nous appliquons ces recherches à une gaine rectangulaire utilisées dans l'industrie de l'aluminium. Nous constatons que le modèle que nous avons développé permet d'obtenir des tendances sur la réponse vibratoire de la gaine et met de nouveau en avant l'importance des contributions de type aéroacoustiques. Des pistes de réduction des niveaux vibratoires sont explorées et notamment celles de l'impact des paramètres géométriques de la structure. / Rectangular duct used for gas transport ductwork, especially in the aluminium industry, are excited by the internal flow. This thesis seeks to understand what are the phenomena involved in the vibration of the gaz transport ductwork. Firstly, we present a semi-analytical model of a homogeneous rectangular duct based on the coupling of 4 plates. This modeling allow us to define the duct modal characteristics and is validated by using both experimental and numerical (FEM) ways. Secondly, we applied aeroacoustic and aerodynamic excitations to our duct in order to determine which are the main contributions. Comparisons between numerical and experimental results, in the case of a straight duct highlight that aeroacoustic sources are predominant. The same trends are found with a bend configuration with few changes close to the band where aerodynamic sources seems to be predominant also. Finally, we apply our model to a large rectangular duct used in the aluminium industry. We note that the model gives good trends and highlights again the importance of the aeroacoustic contributions. Some reflexions about how to reduce the vibration levels by changing geometrical parameters are given at the end.