Les travaux présentés dans ce mémoire de HDR sont la synthèse de six années de recherche sur l'influence de la pesanteur lors des transferts de chaleur et de masse par ébullition et évaporation. Dans le premier chapitre de ce mémoire, je présente une synthèse de la littérature depuis 2003 sur mes activités de recherche. Je mets l'accent sur les techniques employées et les approches choisies pour comprendre les phénomènes physiques rencontrés que ce soit en évaporation ou en ébullition. A la fin de chaque section de ce chapitre, je fais le point sur notre positionnement par rapport à ces activités, je retrace le cheminement qui nous amenes à développer les méthodes inverses en conduction de la chaleur appliquées à l'ébullition convective, ou à développer l'imagerie infrarouge aux gouttes en évaporation. Dans le second chapitre de ce mémoire, je détaille les travaux effectués en présentant de manière synthétique les méthodes et outils ainsi que les résultats obtenus. Il ne s'agit pas dans ce chapitre de s'appesantir sur les détails techniques de construction des expériences qui ont déjà été publiés dans les articles en annexe. Dans le troisième chapitre de ce mémoire qui est divisé en deux sections distinctes, je présente dans un premier temps les projets acceptés qui vont démarrer dans les mois à venir. Dans un second temps, je présente des perspectives de recherche au travers de proposition de recherche qui déjà fait l'objet de demande de financement.

Although a tropical cyclone may conceptually be regarded as an axisymmetric vortex, there is substantial evidence that asymmetric dynamics play an important role. In this thesis, dynamical impacts of rotating convective asymmetries are examined in this thesis. Two types of rotating convective asymmetries are considered: rotating eyewall convective maximum which is located in the core region of the storm and spiral bands which are located outside the core. Both of them can be characterized as rotating asymmetric convective heat sources, and they are superimposed on a balanced, axisymmetric vortex to approximate the effect of rotating eyewall convective maximum and spiral bands on tropical cyclone by using a simple nonhydrostatic three-dimensional, but linear model that is based on vortex anelastic equations. The evolution of rotating convective asymmetric heat sources on a balanced, axisymmetric vortex, which is modeled after tropical cyclones, is investigated to examine angular momentum transport by gravity waves that radiate away from the core region. Results show that gravity waves can transport angular momentum away from a tropical cyclone, but a very small amount, which is several orders of magnitude smaller than the estimate by recent studies. The significantly large difference may largely be due to the difference between two-dimensional and three-dimensional adjustment processes. Assuming that the effects of spiral bands on tropical cyclone wind field are caused by the response to diabatic heating in their convection, rotating asymmetric heat sources are constructed to reflect observations of spiral bands. These heat sources are rotated around a realistic but idealized balanced axisymmetric vortex. Simulation results show that the response of tropical cyclone wind field to idealized spiral band heat sources can successfully capture a number of observed well-known features of spiral band circulation, such as overturning secondary circulation, descending mid-level inflow, and cyclonic tangential acceleration. Comparison to full-physics numerical simulations confirms the validity of this method which provides a simple dynamical framework to better understand the impact of spiral bands in tropical cyclone.

In this thesis the stability of generic boundary layer flows is studied from a global viewpoint using optimization methods. Global eigenmodes of the incompressible linearized Navier-Stokes equations are computed using the Krylov subspace Arnoldi method. These modes serve as a tool both to study asymptotic stability and as a reduced basis to study transient growth. Transient growth is also studied using adjoint iterations. The knowledge obtained from the stability analysis is used to device systematic feedback control in the Linear Quadratic Gaussian framework. The dynamics is assumed to be described by the linearized Navier-Stokes equations. Actuators and sensors are designed and a Kalman filtering technique is used to reconstruct the unknown flow state from noisy measurements. This reconstructed flow state is used to determine the control feedback which is applied to the Navier-Stokes equations through properly designed actuators. Since the control and estimation gains are obtained through an optimization process, and the Navier-Stokes equations typically forms a very high-dimensional system when discretized there is an interest in reducing the complexity of the equations. A standard method to construct a reduced order model is to perform a Galerkin projection of the full equations onto the subspace spanned by a suitable set of vectors, such as global eigenmodes and balanced truncation modes. / QC 20100924

In this thesis, the moisture characteristics of spruce plywood are studied experimentally and numerically with special attention given to moisture storage and release as the indoor humidity changes diurnally. This is referred to as the moisture buffering capacity. Two test facilities (a glass jar facility and a transient moisture transfer facility) are used to measure the moisture accumulation and temperature and relative humidity profiles within spruce plywood. These measured data are used to determine the moisture buffering capacity of spruce plywood and validate a one-dimensional transient numerical model that can be used to calculate the transient heat and moisture transfer in spruce plywood. There is good agreement between the measured and simulated results over the range of test variables investigated. </p><p>This validated numerical model is used to investigate the effect of initial conditions, boundary conditions, thickness and humidity step change on the moisture buffering capacity. In addition, sensitivity studies are performed to investigate the effect of variations in material properties used in the numerical model. The properties that are considered in these sensitivity studies are the sorption isotherm, effective thermal conductivity, heat of sorption and effective diffusion coefficient. These studies show that the sorption isotherm has the greatest effect on the moisture buffering capacity, as well as the temperature and relative humidity profiles within spruce plywood. For example, a ± 10% change in sorption isotherm has a ± 7%, ± 6% and ± 10% effect on the moisture buffering capacity, and the relative temperature and relative humidity change, respectively. </p><p>This thesis also verifies the moisture diffusivity property for spruce plywood, which was developed by Olutimayin and Simonson (2005) to account for moisture storage in cellulose insulation for a single step change in humidity. It was found that for spruce plywood, the moisture penetration depth may be over predicted by an order of magnitude when moisture storage is neglected using a transient analytic solution which does not include moisture storage.

The Shared Mobile Atmospheric Research and Teaching Radar (SMART-R) is a truck-mounted C-band, Doppler radar that was deployed during the Dynamics of the Madden-Julian Oscillation (DYNAMO) / Cooperative Indian Ocean Experiment on interseasonal variability in the year 2011 (CINDY2011) campaign on Addu Atoll, Maldives. One of SMART-R’s objectives was to provide continuous volume scans of precipitating clouds during all phases of the Madden-Julian Oscillation (MJO) for the full duration of the campaign. Data from SMART-R is available for 2 October 2011 through 9 February 2012. Every 10 minutes a full volume scan was produced, which was subsequently run through quality control algorithms that, among other filters, performed a calibration correction, noise filtering, and an attenuation correction. It was observed that data from SMART-R appeared to be slanted towards the WNW, and after analysis, a 0.75◦ tilt correction was applied towards azimuth 285◦. The data was then converted into Cartesian coordinates and an additional noise filter was applied. NETCDF files with radial velocities and corrected reflectivity were produced. From the reflectivity observations, a suite of products including rain maps, echo- top heights and convective/stratiform separations were produced. A modified version of the convective/stratiform separation was developed in an attempt to classify shallow and weak convection more correctly. The modified algorithm utilizes an isolation parameter set to 10 km to the north, south, east, and west, a 10-dBz echo-top height threshold set to 9 km, and a 16-dBz reflectivity threshold at 3 km to ensure only isolated, shallow, and weak rain originally classified as stratiform, is reclassified as convection. Analyses of these products clearly suggest two MJO events occurring in October and November as indicated by the Wheeler and Hendon Multivariate MJO index. While stratiform rain almost always encompassed a larger area of the radar domain, convective rain was the larger producer of rain with the exception of active MJO periods. In addition, echo-top height counts are observed to increase in both vertical structure and frequency as the MJO initiates and becomes active over the radar domain. Possible connections are also made between echo-top height data and humidity retrievals from soundings launched on Addu Atoll. It appears that during MJO initiation, convective echo tops lead the moistening of the mid troposphere, while during suppressed phases, the convective echo tops lag behind the moistening of the mid troposphere. Wind shear also appears to be weaker during an active MJO event, and increase as the active MJO exits the region. From these observations, as well as other rain statistics including the diurnal cycle, indicators for a localized MJO index are proposed that are based on local radar and sounding data, rather than satellite and reanalysis observations of wind and outgoing long-wave radiation.

During 1951-2009, 47% of all tropical systems (TSs) within the Atlantic Basin transitioned to post-tropical (PTC) extratropical classification. These systems have shown the capability of producing hurricane-force winds and gusts for portions of the eastern United States. This study provides a climatological foundation for high-wind observations that were contributed from PTCs. In this study, 76 PTC systems were identified and tracked using six hourly observations from the National Hurricane Center’s HURDAT dataset. Mean wind radii buffers were calculated and used to determine the high-wind observations attributed by PTCs. High-wind climatology was developed by using hourly surface wind data from the National Climatic Data Center (NCDC) and deploying the current NWS high-wind criteria. For this study, the geography and climatology of PTCs and resultant high winds were analyzed using geographic information systems (GIS). Findings show that < 1% (270) of all high-winds events that occur within the U.S. were contributed from PTCs. The highest frequencies were found in three regions: Midwest, Mid-Atlantic, and New England. Due to the low number of high-wind events produced from PTCs, an adjusted wind scale was created by using standard deviations of sustained and gust observations. The goal of this study is determine the contribution of high winds from PTCs, with the aim of improving our understanding of the hazardous outcomes of such events.

In this thesis, the moisture characteristics of spruce plywood are studied experimentally and numerically with special attention given to moisture storage and release as the indoor humidity changes diurnally. This is referred to as the moisture buffering capacity. Two test facilities (a glass jar facility and a transient moisture transfer facility) are used to measure the moisture accumulation and temperature and relative humidity profiles within spruce plywood. These measured data are used to determine the moisture buffering capacity of spruce plywood and validate a one-dimensional transient numerical model that can be used to calculate the transient heat and moisture transfer in spruce plywood. There is good agreement between the measured and simulated results over the range of test variables investigated. </p><p>This validated numerical model is used to investigate the effect of initial conditions, boundary conditions, thickness and humidity step change on the moisture buffering capacity. In addition, sensitivity studies are performed to investigate the effect of variations in material properties used in the numerical model. The properties that are considered in these sensitivity studies are the sorption isotherm, effective thermal conductivity, heat of sorption and effective diffusion coefficient. These studies show that the sorption isotherm has the greatest effect on the moisture buffering capacity, as well as the temperature and relative humidity profiles within spruce plywood. For example, a ± 10% change in sorption isotherm has a ± 7%, ± 6% and ± 10% effect on the moisture buffering capacity, and the relative temperature and relative humidity change, respectively. </p><p>This thesis also verifies the moisture diffusivity property for spruce plywood, which was developed by Olutimayin and Simonson (2005) to account for moisture storage in cellulose insulation for a single step change in humidity. It was found that for spruce plywood, the moisture penetration depth may be over predicted by an order of magnitude when moisture storage is neglected using a transient analytic solution which does not include moisture storage.

Ammonia is a nuisance gas in many swine barns. The overall objective of this research project was to model ammonia formation and transmission processes in a grower-finisher swine barn, by first modelling the ammonia production and emission from urine puddles on the floor surface and the ammonia emission from the slurry pit, and then incorporating these emission rates in a dynamic model that separates the room and slurry pit headspace as two separate, but linked, control volumes. A series of studies were conducted to gather more information about the processes affecting the ammonia emission rate from the floor surface and the slurry that were later included in the overall room model developed. The model was then used to investigate ammonia reducing techniques and technologies based on the understanding of ammonia production and transmission incorporated in the model. The first step in modelling the ammonia emission rate from the floor surface was to determine the frequency of urinations by grower-finisher pigs. Male and female pigs were observed three times during their finishing phase to determine their urination frequency over the course of a day. The average measured urination frequency was 0.62 ± 0.11 urinations pig-1 h-1. A sinusoidal dromedary model was developed to describe the daily variation in urination frequency for male and female pigs between 51 and 78 kg.<p>In order for the deposited urinations on the floor surface to emit ammonia, the urea in the urine must first be converted to ammonia and the urease enzyme catalyzes this reaction. Two methods, a fixed-time-point method using the indophenol assay for ammonium-nitrogen analysis and a continuous method using the coupled enzyme assay, were used to measure enzyme activity at the floor surface of a swine barn and were compared to reported urease activity levels in the literature. Using both methods, there appeared to be an ammonia-producing site on the floor surface or within the collected samples that made accurate measurements of urease activity impossible. A review of urease activity levels in the literature from dairy-cow houses suggest that urease activity will be lowest following floor-cleaning and increase quickly following fouling of the floor surface. Based on the literature review, a urease activity value of 5 g NH¬3 m-2 h-1 was suggested for use in ammonia emission modelling of fouled floor surfaces in swine barns until better measurements become available. <p>The ammonia emissions from 36 simulated urine puddles under a variety of temperature, air velocity and initial urea concentration conditions were measured in a bench-scale experimental set-up. The measurements were used to calibrate and validate a dynamic, mechanistic, urine puddle emission model that considered the processes of evaporation, urea conversion, change in liquid concentration and puddle pH in order to simulate the amount of ammonia emitted from a puddle. Based on the correlation coefficients (R) between measured and simulated values for water volume (R=0.99), total ammoniacal nitrogen concentration (R=0.90), and total emission (R=1.00), along with five other statistical tests for each simulated variable, the model was deemed accurate. The measurements and simulations in this experiment showed the impact of puddle pH, urease activity and changing environmental conditions on the average puddle emission rate. Puddle emission continued to occur as long as there was still water.<p> The impact of different slurry compositions on the ammonia emission rate from slurry pits was tested in another bench-scale experimental set-up with emission chambers. The emission chamber concentration data collected was used to calibrate and validate a developed slurry emission model. The collected slurry samples were concentrated mixtures of urine and feces from individually-housed animals fed different diets. An empirical equation was developed to express the amount of total ammoniacal nitrogen in the slurry that was in the form of ammonia (f) and thus volatile to the surroundings. Based on the empirical equation, the simulated value of f was between 0.03 and 0.08 and did not show the sensitivity to slurry pH that has been reported by other authors. The slurry emission model with the empirical equation for f was validated with ammonia emission measurements from eight different slurry samples and simulated hourly concentration measurements within 17% and five-day average concentration measurements within 3%. Further testing was recommended to ensure the model developed for concentrated manure in this study was applicable to the more dilute slurry found in swine barns. <p>Using the information gained in the previous experiments, a mechanistic model describing the dynamic ammonia concentration in the room and in the slurry channel headspace of grower-finisher swine barns, as well as the ammonia emitted to the surrounding environment was developed. Data was collected from two grower-finisher rooms to use as input data to the model and for calibration and validation purposes. The model calibration procedure determined that the amount of emissions originating from the slurry for the simulated room conditions was generally less than 5% of the total room emissions, the air exchange rate through the slatted floor was approximately 4% of the room ventilation rate, and that in the first two weeks of animal activity in a room the urease activity at the floor surface will increase. The model was validated using separate data from that used in the calibration process. The model simulated hourly room concentration levels within 2.2 ppm and 3-day average concentration levels within 1.6 ppm. The model simulations were more accurate for one room that was fed a typical grower-finisher diet compared to another room fed an experimental diet with lower protein content and sugar-beet pulp inclusion. <p>The dynamic model was tested for its sensitivity to various input factors in terms of the floor emission rate, slurry emission rate and total emission rate. An interesting aspect of the simulations was that increases in either floor or surface emission rate were compensated to a small extent by decreases in the other emission rate as a result of a reduced concentration gradient for mass transfer. The ammonia emission rate from the floor was most sensitive to changes in urease activity, fouled floor area and puddle area. The ammonia emission rate from slurry was most sensitive to changes in slurry pH. The impact of input variables on the total emission rate was dependant on the simulated proportion of the total ammonia emission coming from either the floor surface or slurry channel. Three ammonia reduction techniques were tested and evaluated on their impact to the total ammonia emission rate from a room compared to a given set of control conditions.<p>The work in this thesis highlighted the importance of ammonia emission from the floor surface. The proportion of ammonia originating from the slurry and from the floor surface respectively will vary on the specific conditions within the barn, and will impact the effect of any ammonia mitigation technique that is investigated or used.

The heat transfer characteristics of airflows in serpentine cooling channels in stator vanes of gas turbines and the novel QuTech® Heat Pipe (QTHP) for electronic cooling applications were studied. The cooling channels are modeled as smooth and roughened two-pass trapezoidal channels with a 180° turn over a range of Reynolds numbers between about 10,000 and 60,000. The naphthalene sublimation technique and the heat and mass transfer analogy were applied. The results showed that there was a very large variation of the local heat (mass) transfer distribution in the turn and downstream of the turn. The local heat (mass) transfer was high near the end wall and the downstream outer wall in the turn and was relatively low in two regions near the upstream outer wall and the downstream edge at the tip of the divider wall in the turn. The variation of the local heat (mass) transfer was larger with ribs on two opposite walls than with smooth walls. The regional average heat (mass) transfer was lower in the turn and higher in the entire channel with the flow entering the channel through the larger straight section than when the flow was reversed. The pressure drop across the turn was higher with the flow entering the channel through the larger channel than when the flow was reversed. Thermal performance of the QuTech® Heat Pipe was identified over a range of inclination angles between 90° and -90° and thermal mechanism of the QTHP was studied with GC-MS, ICP-OES, XRD, XPS, and DSC. This study resulted in the following findings: the performance of the QTHP was severely dependent on gravity; the QTHP utilizes water as working fluid; there were inorganic components such as Na, K, P, S, and Cr, etc.; and the vaporization temperature of the working fluid (mostly water) was lower than the boiling temperature of pure water. This was due to the presence of inorganic salt hydrates in the QTHP. It may be concluded that thermal performance of heat pipes increases with additional latent heat of fusion energy and energy required to release water molecules from salt hydrates.

Recent emphasis on the prediction of temperature and concentration fields in the atmosphere has led to the investigation of accurate solution methods of the time-dependent conservation equations for mass, momentum, energy and species. Atmospheric radiation is the key component of this system. Therefore, atmospheric radiation models were developed in isolation from the climate models. The time-dependent multi-dimensional governing equations of atmospheric models must be solved in conjunction with the radiative transfer equation for accurate modeling of the atmosphere. In order to achieve this objective, a 1-D Radiative-Convective Model for Earth-Atmosphere System (RCM4EAS) was developed for clear and cloudy sky atmospheres. The radiative component of the code is Santa Barbara DISORT (Discrete Ordinate Radiative Transfer) Atmospheric Radiative Transfer (SBDART) integrated with exponential sum-fitting method as the radiative property estimation technique. The accuracy of SBDART was tested by comparing its predictions of radiative fluxes with those of Line-by-Line Radiative Transfer Model (LBLRTM) for 1-D longwave (3.33-100 &micro / m) clear sky atmosphere and a good agreement was obtained. A parametric study aiming at finding the optimum parameters to be used as input in SBDART regarding the wavelength increment and order of approximation was also carried out. Variable wavelength and eight streams were selected as optimum parameters for the accuracy and computational efficiency. The code was then coupled with a 1-D Radiative-Convective Model (RCM) to obtain the time dependent code, (RCM4EAS), which was applied to the investigation of the sensitivity of climate to changes in the CO2 concentration for clear and cloudy sky conditions. CO2 sensitivity analyses revealed that doubling the CO2 concentration in the earth&rsquo / s atmosphere from its present value (387 ppm) results in an increase in equilibrium surface temperature of 4.2 K in the clear sky atmosphere as opposed to 2.1 K in cloudy sky atmosphere with typical cloud physical parameters. It is worth noting that times required to reach equilibrium surface temperatures are approximately 2000 and 6000 days for clear and cloudy sky atmospheres, respectively and these temperature increases are calculated assuming that all the other parameters except CO2 concentration remain unchanged within these time periods. Therefore, it should be noted that these temperature increases reflect only the effect of CO2 doubling and excludes the effect of other forcings which might positively or negatively affect these temperature increases. Overall evaluation of the performance of the code developed in this thesis study indicates that it can be used with confidence in 1-D radiative-convective modeling of the earth-atmosphere systems.