D.Ing. / According to literature 6% of South Africa’s primary energy consumption could be saved if heat pumps were used to their full technical potential. Although there is world-wide interest in the use of heat pumps and considerable effort has been expended on heat-pump research, heat pumps are not commonly used in South Africa. The objective of this thesis is to determine the possibility of a combined evaporator or condenser with a normal pump. This will reduce cost and space of a normal heat pump and make heat pumps economically more competitive against resistance element geysers. In order to investigate this combination research is done on semicircular heat exchangers, since this is the primary geometry of the heating channels in the condenser pump. Analyses is done experimentally on a standard 28.58 mm hard drawn copper tube, cut trough the middle, with a 1.6 mm copper plate in between to obtain a semicircular heat exchanger. Turbulent flow is investigated with the flat side of the semicircular heat exchanger being horizontal or vertical, a spiralled and a s-shape semicircular heat exchanger. In each case the heat transfer coefficient is determined with the use of the Wilson plot technique. It is found that there is a significant increase in Nusselt number for semicircular heat exchangers above a normal tube-in-tube heat exchanger but the pressure loss coefficient increase with an equal amount.

This thesis reviews the development and current methods of numerical convection heat transfer from available literature, encompassing an analysis of the various finite element formulations available for investigating convection. It further describes the finite element formulation for the primitive variable convection heat transfer equations via a Galerkin weighted residual scheme and using mixed interpolation, and it demonstrates the capability of this method by means of five practical examples, namely natural convection in a thermally driven square cavity, a thermally driven vertical slot, a thermally driven triangular cavity, and a liquid convective diode, and forced convection in a cooling pond. This study also provides the background and framework for the problem of transient convection heat transfer, and for further steady-state studies using parameters outside those considered herein.

Within a gas turbine engine, the turbine nozzle guide vanes are subjected to very harsh conditions from the highly turbulent and hot gases exiting the combustor. The temperature and pressure fields exiting combustors are highly nonuniform and dictate the heat transfer and aero losses that occur in the turbine passages. To better understand these effects, the goal of this work was to develop an adjustable combustor exit profile simulator for the Turbine Research Facility (TRF) at the Air Force Research Laboratory. The TRF is a high temperature, high pressure, short duration blow-down test facility that is capable of matching several aerodynamic and thermal nondimensional engine parameters including Reynolds number, Mach number, pressure ratio, corrected mass flow, gas to metal temperature ratio, and corrected speed. The primary research objective was to design, install, and verify a non-reacting simulator device that can provide representative combustor exit total pressure and temperature profiles to the inlet of the TRF turbine test section. This required the upstream section of the facility to be redesigned into multiple concentric annuli that serve the purpose of injecting high momentum dilution jets and low momentum film cooling jets into a central annular chamber, similar to a turbine engine combustor. The design of the section allows for variations in injection levels to generate different pressure profiles with elevated turbulence. The dilution and film cooling temperatures can also be varied to create a variety of exit temperature profiles similar to real combustors. The impact of the generated temperature and pressure profiles on turbine heat transfer and secondary flow development was ultimately investigated. Proposed optimal inlet conditions for the turbine tested in this research effort were determined based on the measured data corresponding to the combustor simulator exit profiles that minimized vane heat transfer and total pressure loss. / Ph. D.

The study of two phase flow, where a finely divided solid is suspended in a fluid medium, has received considerable attention in the last five years due to the increased number of systems in which this phenomena occurs. The two most important areas involving the gas- particle mechanism are nuclear reactor heat exchangers and liquid and solid rockets.

Thesis (Ph.D.)--Massachusetts Institute of Technology, Dept. of Nuclear Engineering, 1981. / MICROFICHE COPY AVAILABLE IN ARCHIVES AND SCIENCE. / Includes bibliographies. / by Stanley R. Shanfield. / Ph.D.

Thesis. 1976. M.S.--Massachusetts Institute of Technology. Dept. of Nuclear Engineering. / Microfiche copy available in Archives and Science. / Includes bibliographical references. / M.S.

No description available.

The main purpose of this work is to find a physical and numerical description related to the reaction of the multilayer nano energetic material (nEM) in dense film. Energy density of nEM is much higher than the conventional energetic material; therefore, nEM finds more applications in propulsions, thermal batteries, material synthesis, nano igniters, waste disposals, and power generations. The reaction model of a multilayer nEM in a dense film of aluminum and copper oxide deposits on a composite substrate of silica/silicon is studied and solved in different stages. The two main interests in this study are propagation speed and maximum temperature of the reaction. In order to relate speed of reaction and maximum flame temperature a number of other variables such as heat loss, the product porosity, and the reaction length should be estimated. The main aim of this study is to introduce a numerical model which estimates and relates these values in multilayer nEM in a dense film. The following is a summary of the execution steps to achieve this goal. In Part I of this thesis, flame front speed and the reaction heat loss were the main targets. The time-of-flight technique has been developed to measure the speed of flame front with an accuracy of 0.1 m/s. This measurement technique was used to measure the speed of propagation on multilayer nEM over different substrate material up to 65 m/s. A controllable environment (composite silicon\silica) was created for a multilayer standard thin film of aluminum and copper oxide to control the reaction heat loss through the substrate. A number of experimental results show that as the thickness of silica decreases, the reaction is completely quenched. Reaction is not in self-sustaining mode if the silica thickness is less than 200 nm. It is also observed that by ii increasing silica’s thickness in substrate, the quenching effect is progressively diminished. The speed of reaction seems to be constant at slightly more than 40 m/s for a silica layer with thickness greater than 500 nm. This would be the maximum heat penetration depth within the silica substrate, so the flame length was calculated based on the measured speed. In Part II, a numerical analysis of the thermal transport of the reacting film deposited on the substrate was combined with a hybrid approach in which a traditional two-dimensional black box theory was used, in conjunction with the sandwich model, to estimate the maximum flame temperature. The appropriate heat flux of the heat sources is responsible for the heat loss to the surroundings. A procedure to estimate this heat flux using stoichiometric calculations is based on the previous author’s work. This work highlights two important findings. One, there is very little difference in the temperature profiles between a single substrate of silica and a composite substrate of silicon\silica. Secondly, by increasing the substrate thickness, the quenching effect is progressively diminished at given speed. These results also show that the average speed and quenching of flames depend on the thickness of the silica substrate and can be controlled by a careful choice of the substrate. In Part III, a numerical model was developed based on the moving heat source for multilayer thin film of aluminum and copper oxide over composite substrate of silicon\silica. The maximum combustion flame temperature corresponding to the speed of flame front is the main target of this model. Composite substrate was used as a mechanism to control the heat loss during the reaction. Thickness of the substrate, the length of flame front, and the density of the product were utilized for the standard multilayer thin film with 43 m/s flame front speed. The calculated heat penetration depth in this case was compared to the experimental result for the same flame front iii speed. Numerical model was also used to estimate three major variables for a range of 30-60 m/s. In fact, the maximum combustion flame temperature that corresponds to flame speed along with the length of the flame, density of the product behind the flame, and maximum penetration depth in steady reaction, were calculated. These studies will aid in the design of nEM multilayer thin film. As further numerical and experimental results are obtained for different nEM thicknesses, a unified model involving various parameters can be developed.

Multilayer insulation (MLI) has been shown to be the best performing cryogenic insulation system at high vacuum (less than 10-3 torr), and is widely used on spaceflight vehicles. Over the past 50 years, many numerous investigations of MLI have yielded a general understanding of the many variables associated with MLI. MLI has been shown to be a function of variables such as warm boundary temperature, the number of reflector layers, and the spacer material in between reflectors, the interstitial gas pressure and the interstitial gas. Because conduction between reflectors increases with the thickness of the spacer material, and yet the radiation heat transfer is inversely proportional to the number of layers, it stands to reason that the thermal performance of MLI is a function of the number of layers per thickness, or layer density. Empirical equations that were derived based on some of the early tests showed that the conduction term was proportional to the layer density to a power. This power depended on the material combination and was determined by empirical test data. Many authors have graphically shown such optimal layer density, but none have provided any data at such low densities, or any method of determining this density. Keller, Cunnington, and Glassford showed MLI thermal performance as a function of layer density of high layer densities, but they didn’t show a minimal layer density or any data below the supposed optimal layer density. However, it was recently discovered by the author that by manipulating the derived empirical equations and taking a derivative with respect to layer density, a solution for on optimal layer density may be obtained. iv Several manufacturers have begun manufacturing MLI at densities below the analytical optimal density. This trend is apparently based on the theory that increased distance between layers lowers the conductive heat transfer and that there are no limitations on volume. By modifying the circumference of these blankets, the layer density can easily be varied. The most direct method of determining the thermal performance of MLI at cryogenic temperature is by evaporation (or “boil-off”) calorimetry. Several blankets were procured and tested at various layer densities by the Cryogenics Test Laboratory at NASA Kennedy Space Center. The blankets were tested over a wide range of layer densities including the analytical minimum. Several of the blankets were tested at the same insulation thickness while changing the layer density (thus a different number of reflector layers). Heat transfer optimization of the layer density of multilayer insulation systems would remove the variable of layer density from the complex method of designing such insulation systems. Since the layer density is one of the variables that in those complex equations that require more experience to understand fully grasp, this significantly simplifies the blanket design process. Additional testing was performed at various warm boundary temperatures and pressures. The testing and analysis was performed to determine thermal performance data and to simplify the analysis of cryogenic thermal insulation systems.

Uncertainty in accurately knowing applied internal heat transfer coefficients inside of a cooling passage can lead to variability in predicting low cycle fatigue life of a turbine vane or blade. Under-predicting a life value for a turbine part can have disastrous effects on the engine as a whole, and can negate efforts in innovative design for advanced cooling techniques for turbine components. Quantification of this fatigue life uncertainty utilizing a computational framework is the primary objective of this thesis. Through the use of probabilistic design methodologies a process is developed to simulate uncertainties of internal heat transfer coefficient, which are then applied to the aft section of a non-rotating turbine blade component, internally cooled through a multi-pass serpentine channel. While keeping all other parameters constant internal heat transfer coefficients are varied according to a prescribed uncertainty range throughout the passages. The effect on the low cycle fatigue life of the airfoil is then evaluated at three discrete locations: near the base of the airfoil, towards the tip, and at mid-span. A generic low cycle fatigue life prediction model is used for these evaluations. Even though the probabilistic design process uses independent random numbers to simulate the variation, in reality, heat transfer coefficients at points located closely together should be correlated. For this reason, an autocorrelation function is implemented. By changing the value of this function the strength of the correlation of iv neighboring internal heat transfer coefficients to each other over a certain distance can be controlled. In order to test the effect that this correlation strength has on the low cycle fatigue life calculation, low and high values are chosen and analyzed. The magnitude of the prescribed uncertainty range of the internal heat transfer coefficient variation is varied to further study the effects on life. Investigated values include 5%, 10% and 20% for the straight ribbed passages and 10%, 20%, and 40% for both the tip and hub turns. As expected there is a significant dependence of low cycle fatigue life to the variation in internal heat transfer coefficients. For the 20/40% case, variations in life as high as 50-60% are recorded, furthermore a trend is observed showing that as the magnitude of the uncertainty range of internal heat transfer coefficients narrows so does the range of the low cycle fatigue life uncertainty.