Experimental measurements are conducted on uninterrupted and interrupted aluminum microchannel heat sinks of 300, 500, 600 and 900 &mu / m channel widths. Two different versions of interrupted channels are tested / with single interruption and with 7 interruptions. Distilled water is used as the working fluid and tests are conducted at volumetric flow rates in a range of 0.5-1.1 lpm. Thermoelectric foils are used to supply uniformly distributed heat load to the heat sinks such that for all the tests the heat removed by water is kept constant at 40 W. Pressure drop and temperature increase are measured along the channels of different configurations for a number of different flow rates. For the interrupted channels thermal boundary layers re-initialize at the leading edge of each interrupted fin, which decreases the overall boundary layer thickness. Also the flow has been kept as developing, which results in better heat transfer performance. Due to the separation of the flow into branches, secondary flows appear which improves the mixing of the stream. Advanced mixing of the flow also enhances the thermal performance. In the experiments, it is observed that interruption of channels improved the thermal performance over the uninterrupted counterparts up to 20% in average Nusselt number, for 600 micron-wide channels. The improvement of average Nusselt number between the single interrupted and multi interrupted channels reached a maximum value of 56% for 500 micron-wide channels. This improvement did not cause a high pressure drop deviation between the uninterrupted and interrupted microchannels even for the maximum volumetric flow rate of 1.1 lpm. Highest pressure drop through the channels was measured as 0.07 bar, which did not require to change the pump. In the tests, maximum temperature difference between the inlet of the fluid and the base of the channel is observed as 32.8&deg / C, which is an acceptable value for electronic cooling applications.

Enhanced heat transfer in electronic hardware by direct, small-scale actuation is investigated experimentally in two test bed configurations. The first configuration exploits the unsteady motions induced by a vibrating reed embedded within a heated duct (in contact with hardware that needs cooling) to enhance forced convection transport heat from the duct surfaces. The flow within the duct is either exclusively driven by the reed or, for higher heat flux, is augmented by an induced core flow. The time harmonic motion of the reed results in the regular shedding of vortical structures that interact with the inner surfaces in the absence and presence of a core flow. The second configuration focuses on the effects of small scale motions induced by a synthetic jet on heat transfer within an advanced heat sink. The synthetic jets emanate directly through the base of the heat sink and induce a recirculating flow between the fins, resulting in a lower thermal resistance than what is typically achieved with traditional fans. The unsteady flow characteristics in both configurations are investigated using particle image velocimetry (PIV). Of particular interest are the effects of small-scale motions and enhanced mixing on heat transfer compared to conventional time-invariant flows at similar or higher Reynolds numbers.

A key challenge for three dimensional (3D) integrated circuits (ICs) is thermal management. There are two main thermal challenges in typical 3D ICs. First, in the homogeneous integration with multiple high-power tiers, an effective cooling solution that scales with the number of dice in the stack is needed. Second, in the heterogeneous integration, an effective thermal isolation solution is needed to ‘protect’ the low-power tier from the high-power tier. This research focuses to address these two thermal challenges through hybrid microfluidic cooling and thermal isolation technologies. Within-tier microfluidic cooling is proposed and demonstrated to cool a stack with multiple high-power tiers. Electrical thermal co-analysis is performed to understand the trade-offs between through silicon via (TSV) parasitics and heat sink performance. A TSV-compatible micropin-fin heat sink is designed, fabricated and thermally characterized in a single tier, and benchmarked with a conventional air-cooled heat sink. The designed heat sink has a thermal resistance of 0.269 K·cm2/W at a flow rate of 70 mL/min. High aspect ratios TSVs (18:1) are integrated in the micropin-fins. Within-tier microfluidic cooling is then implemented in 3D stacks to emulate different heating scenarios, such as memory-on-processor and processor-on-processor. Air gap and mechanically flexible interconnects (MFIs) are proposed for the first time to decrease the vertical thermal coupling between high-power (e.g. processor) and low-power tiers (e.g. memory or nanophotonics). A two-tier testbed with the proposed thermal isolation technology is designed, fabricated and tested. Compared with conventional 3D integration approach, thermal isolation technology helps reduce the temperature at a fixed location in the low-tier by 12.9 °C. The resistance of a single MFI is measured to be 46.49 mΩ.

This thesis numerically studied mass transfer during fully confined liquid jet impingement on a rotating target disk of finite thickness and radius. The study involved laminar flow with jet Reynolds numbers from 650 to 1500. The nozzle to plate distance ratio was in the range of 0.5 to 2.0, the Schmidt number ranged from 1720 to 2513, and rotational speed was up to 325 rpm. In addition, the jet impingement to a stationary disk was also simulated for the purpose of comparison. The electrochemical fluid used was an electrolyte containing 0.005moles per liter potassium ferricyanide (K3(Fe(CN6)), 0.02moles per liter ferrocyanide (FeCN6?4), and 0.5moles per liter potassium carbonate (K2CO3). The rate of mass transfer of this electrolyte was compared to Sodium Hydroxide (NaOH) and Hydrochloric acid (HCl) electrochemical solutions. The material of the rotating disk was made of 99.98% nickel and 0.02% of chromium, cobalt and aluminum. The rate of mass transfer was also examined for different geometrical shapes of conical, convex, and concave confinement plates over a spinning disk. The results obtained are found to be in agreement with previous experimental and numerical studies. The study of heat transfer involved a microchannel for a composite channel of trapezoidal cross-section fabricated by etching a silicon wafer and bonding it with a slab of gadolinium. Gadolinium is a magnetic material that exhibits high temperature rise during adiabatic magnetization around its transition temperature of 295K. Heat was generated in the substrate by the application of magnetic field. Water, ammonia, and FC-77 were studied as the possible working fluids. Thorough investigation for velocity and temperature distribution was performed by varying channel aspect ratio, Reynolds number, and the magnetic field. The thickness of gadolinium slab, spacing between channels in the heat exchanger, and fluid flow rate were varied. To check the validity of simulation, the results were compared with existing results for single material channels. Results showed that Nusselt number is larger near the inlet and decreases downstream. Also, an increase in Reynolds number increases the total Nusselt number of the system.

The objective of this research was to design aluminum powder metallurgy (PM) alloys and processing strategies that yielded sintered products with thermal properties that rivaled those of the cast and wrought aluminum alloys traditionally employed in heat sink manufacture. Research has emphasized PM alloys within the Al-Mg-Sn system. In one sub-theme of research the general processing response of each PM alloy was investigated through a combination of sintering trials, sintered density measurements, and microstructural assessments. In a second, the thermal properties of sintered products were studied. Thermal conductivity was first determined using a calculated approach through discrete measurements of specific heat capacity, thermal diffusivity and density and subsequently verified using a transient plane source technique on larger specimens. Experimental PM alloys achieved >99% theoretical density and exhibited thermal conductivity that ranged from 179 Wm-1K-1 to 225 Wm-1K-1. Thermal performance was largely dominated by the amount of magnesium present within the aluminum grains and in turn, bulk alloy chemistry. Data confirmed that the novel PM alloys were highly competitive with even the most advanced heat sink materials such as wrought 6063 and 6061. Two methods of thermal analysis were employed in order to determine the thermal conductivity of each alloy. This first consisted of individual analysis of the specific heat capacity (Cp), thermal diffusivity (?) and density (?) as a function of temperature for each alloy. The thermal conductivity (K) was subsequently determined through the relationship: K=C_p ??. The second means of thermal analysis was a direct thermal conductivity measure using a transient plane source (TPS). The thermal diffusivity and density of samples were both found to decrease with temperature in a linear fashion. Conversely, the specific heat capacity was found to increase with temperature. The only measured thermal property that appeared to be influenced by the alloy chemistry was the thermal diffusivity (and subsequently the calculated thermal conductivity). Both means of thermal analysis showed high thermal conductivity in alloys with low concentrations of magnesium, demonstrating the significance of having alloying elements in solid solution with aluminum. Overall, several alloys were developed using a press and sinter approach that produced higher levels of thermal conductivity than conventional aluminum heat sink materials. The highest thermal conductivity was achieved by alloy Al-0.6Mg-1.5Sn with a calculated value of 225.4 Wm-1K-1. This novel aluminum PM alloy was found to exceed both wrought 6061 and 6063 (195 and 217 Wm-1K-1 respectively). Furthermore, PM alloy Al-0.6Mg-1.5Sn was found to have a significant advantage over die-cast A390 (142 Wm-1K-1).

Exhausting the thermal power from a fusion tokamak is a critical engineering challenge. The life of components designed for these conditions has a strong influence on the availability of the machine. For a fusion power plant this dependence becomes increasingly important, as it will influence the cost of electricity. The most extreme thermal loading for a fusion power plant will occur in the divertor region, where components will be expected to survive heat fluxes in excess of 10 MW/m² over a number of years. This research focussed on the development of a heat sink module for operation under such conditions, drawing on advanced cooling strategies from the aerospace industry. A reference concept was developed using conjugate Computational Fluid Dynamics. The results were experimentally validated by matching Reynolds numbers on a scaled model. Heat transfer data was captured using a transient thermochromic liquid crystal technique. The results showed excellent agreement with the corresponding numerical simulations. To facilitate comparison against other divertor heat sink proposals, a nondimensional figure of merit for cooling performance was developed. When plotted against a non-dimensional mass flow rate, the reference heat sink was shown to have superior cooling performance to all other divertor proposals to date. Results from Finite Element Analysis were used in conjunction with the ITER structural design criteria to life the heat sink. The sensitivity of life to both boundary conditions, and local geometric features, were explored. The reference design was shown to be capable of exceeding the life requirements for heat fluxes in excess of 15 MW/m². A number of heat sinks, based on the reference design, were fabricated. These underwent non-destructive testing, before experimentation in a high-heat flux facility developed by the author. The heat transfer performance of the tested modules was found to exceed that predicted by numerical modelling, which was concluded to be caused by the fabrication processes used.

A presente tese envolve um estudo experimental da ebulição convectiva no interior de um dissipador de calor baseado em multi-microcanais. Resultados experimentais para perda de pressão e coeficiente de transferência de calor foram levantados para os hidrocarbonetos R600a (isobutano), R290 (propano) e R1270 (propileno), fluidos com reduzido GWP (Global Warming Potential) e ODP (Ozone Depletion Potential) nulo. O desempenho termo-hidráulico destes fluidos foi avaliado em um dissipador de calor de cobre, contendo cinquenta canais paralelos com seção transversal retangular de 123x494 µm2 , 15 mm de comprimento e área de base de 15x15 mm2. Os experimentos foram realizados para fluxos de calor de até 400 kW/m2, velocidade mássica variando entre 165 e 823 kg/m2s, graus de sub-resfriamento do líquido na entrada da seção de testes de 5, 10 e 15°C e temperaturas de saturação de 21 e 25°C. Os dados experimentais foram amplamente analisados e discutidos, focando o efeito do fluido refrigerante. Oscilações dos sinais de temperatura e pressão foram analisadas parametricamente visando caracterizar efeitos de instabilidades térmicas. Adicionalmente, realizou-se análise comparativa de desempenho dos refrigerantes baseada na 2ª Lei da Termodinâmica. Os dados para hidrocarbonetos foram comparados com resultados de trabalhos prévios para o refrigerante R134a levantados na mesma seção de testes e utilizando a mesma bancada experimental. A partir destes dados, conclui-se que os hidrocarbonetos proporcionam coeficientes de transferência de calor superiores ao R134a. Em geral, o coeficiente de transferência de calor apresenta a seguinte ordem decrescente: R290, R1270, R600a e R134a. No entanto, o R290 necessitou superaquecimentos da parede superiores ao R1270 para iniciar o processo de ebulição. O refrigerante R1270 proporcionou perdas de pressão totais inferiores aos demais fluidos segundo a seguinte ordem decrescente: R600a, R134a, R290 e R1270. O refrigerante R1270 apresentou frequências de oscilação inferiores na temperatura da câmara de saída. Baseado na análise de desempenho da 2ª Lei da Termodinâmica, conclui-se que, as irreversibilidades devido ao processo de transferência de calor foram predominantes quando comparadas àquelas devido à perda de pressão. Através desta análise também constatou-se o melhor desempenho para o refrigerante R290. / The present thesis concerns an experimental study on flow boiling inside a microchannel array. Experimental results for two-phase pressure drop and heat transfer coefficient were acquired for the hydrocarbons R600a (isobutane), R290 (propane) and R1270 (propylene). These fluids present low Global Warming Potential (GWP) and null Ozone Depletion Potential (ODP). The cooling performance of these hydrocarbons were evaluated for a copper heat sink containing fifty parallel microchannels. The microchannels are rectangular with cross section of 123x494 µm2, 15 mm length and a footprint area of 15x15 mm2. The experimental evaluation was performed in a test facility located at the Laboratory of Thermal and Fluid Engineering of School of Engineering of São Carlos, University of Sao Paulo. The experiments were performed for heat fluxes up to 400 kW/m2, mass velocities from 165 to 823 kg/m2s, degrees of liquid subcooling at the test section inlet of 5, 10 and 15°C and saturation temperatures of 21 and 25°C. The experimental data were carefully analyzed and discussed focusing on the effects of the fluid on the heat sink thermal hydraulic performance. Fluctuations in the temperature and pressure were analyzed parametrically in order to evaluate thermal instability effects. Additionally, an exergy analysis was performed to evaluate the refrigerant efficiency during convective evaporation. Subsequently, the parametric effects and performance of hydrocarbons were compared with previous results for refrigerant R134a obtained in the same test facility and under the same experimental conditions. The refrigerant R290 provided heat transfer coefficients higher than R600a and R1270. However, R290 needed a degree of wall superheating for the onset of nucleate boiling higher than R1270. Based on the exergy analysis it was concluded that, the irreversibility associated to the heat transfer process are predominant compared with the irreversibility due to the pressure drop. According to the Second Law analyses it was also concluded R290 as the fluid providing the best performance.

The demand for smaller but more powerful electronic components is ever increasing. This demand puts a strain on engineers to produce optimal cooling designs for these electronic components. One method for cooling these electronic components is with heat sinks which effectively increase the surface area available for extracting the heat from the electronic components. Computational Fluid Dynamics (CFD) software is sometimes used to aid in the design process, but CFD simulations are computationally expensive and take long to complete. This causes the design engineer to test only a few proposed designs based on his/her experience and select the design that performs the best out of the tested designs, which might not be the optimum. The temperature distribution inside the heat sink can be solved relatively quickly with the diffusion equation, but the flow around the heat sink complicates the CFD simulation and increases the solving time significantly. Therefore, applications have been developed where the interaction between the heat sink and the flow around the heat sink is replaced by heat transfer coefficients. These coefficients are calculated from correlated equations which contain the flow properties. The flow properties are extracted from a flow network solver, which solves the flow around the heat sink. This procedure results in less expensive simulations, which can be used together with an optimisation procedure to develop an optimum cooling design. In this dissertation, a correlation for the contraction heat transfer coefficients of rectangular pin fin heat sinks was developed. A methodology was developed where consecutive regression lines were fitted to a large set of data extracted from numerous CFD simulations. The combination of these regression lines formed the basis of the correlation, which was divided into two correlations; one for laminar flow and another for turbulent flow. The correlations were tested against CFD simulations as well as experimental data. The results indicate that these correlations can be effectively used to calculate the contraction heat transfer coefficients on pin fin heat sinks. / Dissertation (MEng)--University of Pretoria, 2011. / Mechanical and Aeronautical Engineering / unrestricted

<p>Heat sinks have the capability of increasing operating heat flux limits for improved thermal management in the immersion cooling of electronics using dielectric fluids. However, even for arrays of simple, straight fins, the generation of vapor between and along fins during pool boiling lead to performance effects that are not well understood. Further investigation of the heat-flux-dependent variation of boiling modes that can manifest along the fin height is required. Although methods for the prediction of fin boiling heat transfer exist that incorporate a variable heat transfer coefficient determined from a flat surface, they have been developed and assessed for single, isolated fins under the assumption that the sides of the fin at any location behave like that of a flat surface. As a result, when applied to fin arrays, these methods may not always be accurate for the full range of heat flux operation along boiling curve up to the critical heat flux, due to the fins interfering with each other when arranged in arrays of differing spacing and height. To establish when the fins in an array can be described as isolated and having the flat surface boiling behavior, pool boiling experiments are performed using copper heat sinks in two fluids with vastly different properties: HFE-7100 and water. The spacing and height of the longitudinal fins are varied across a range from much larger to less than half of the scale of the capillary length scale of both fluids, <em>L</em>_<em>b</em>. High-speed visualizations enable the identification of different boiling regimes to identify correspondence between flow observations and the boiling performance, such as when there is bubble confinement from fin interference. Trends in the pool boiling data are also compared, noting changes in superheat at various heat fluxes to establish when fin height or spacing affects boiling behavior. The experimental boiling performance is compared to predictions developed assuming isolated fins so as to identify the spacings and heights for which the fin arrays follow this behavior. Overall, the data from both fluids strongly support a hypothesis that <em>L</em>_<em>b</em> is the key length scale. Heat transfer from fin array heat sinks with heights and spacings above <em>L</em>_<em>b</em> are shown to be accurately predicted in both fluids. However, spacings smaller than <em>L</em>_<em>b</em> lead to bubble confinement which affects the superheat, particularly at low heat fluxes, while heights shorter than <em>L</em>_<em>b</em> are unable to support multiple boiling regimes along the fin sidewall. This work identifies the capillary length as the key length scale at which confinement and height effects need to be considered for accurate predictions of immersion cooling applications.</p>

Analytical, closed-form solutions governing thermoelectric behavior are derived. An analytical model utilizing a thermal circuit is presented involving heat transfer into, through, out of, and around a thermoelectric device. A nondimensionalization of the model is presented. Linear heat transfer theory is applied to the model to obtain a series of closed form equations predicting net power output for the thermoelectric device. Fluid streams flowing through shrouded heat sinks with square pin fins are considered for the thermal pathways to and from the device. Heat transfer and pressure drop are characterized in a manner conducive to an analytical model using previously published experimental results. Experimental data is presented which validates and demonstrates the usefulness of the model in predicting power output for commercially available thermoelectric generators. A specific design for a thermoelectric power harvester is suggested consisting of a pattern of thermoelectric generators. An economic model for calculating payback time is developed. An optimization process is demonstrated that allows for the payback time of such a system to be minimized through optimization of the physical design of the system. It is shown that optimization of the thermal pathways dramatically reduces payback time. Optimized design of a system is discussed in light of theoretical cases with feasible payback times.