A numerical model based on fundamental engineering principles was developed and validated to establish a relationship between process parameters and the final properties of woodbased composite boards. The model simulates the mat formation, then compresses the reconstituted mat to its final thickness in a virtual press. The number of interacting variables during the hot-compression process is prohibitively large to assess a wide variety of data by experimental means. Therefore, the main advantage of the model based approach that the effect of the hot-compression parameters on the final properties of wood-based composite boards can be monitored without extensive experimentation. The mat formation part of the model is based on the Monte Carlo simulation technique to reproduce the spatial structure of the mat. The dimensions and the density of each flake are considered as random variables in the model, which follow certain probability density distributions. The parameters of these distributions are derived from data collected on industrial flakes by using an image analysis technique. The model can simulate the structure of a threelayer oriented strandboard (OSB) mat as well as the structure of random fiber networks. A grid is superimposed on the simulated mat and the number of flakes, the thickness, and the density of the mat at each grid point are computed. Additionally, the model predicts the change in several void volume fractions within the mat and the contact area between the flakes during consolidation. The void volume fractions are directly related to the physical properties of the mat, such as thermal conductivity, diffusivity, and permeability, and the contact area is an indicator of the effectively bonded area within the mat. The heat and mass transfer part of the model predicts the change of air content, moisture content, and temperature at designated mesh points in the cross section of the mat during the hotcompression. The water content is subdivided into vapor and bound water components. The free water component is not considered in the model due to the low (typically 6-7 %) initial moisture content of the flakes. The gas phase (air and vapor) moves by bulk flow and diffusion, while the bound water only moves by diffusion across the mat. The heat flow occurs by conduction and convection. The spatial derivatives of the resulting coupled partial differential equations are discretized by finite differences. The resulting ordinary differential equation in time is solved by a differential-algebraic system solver (DDASSL). The internal environment within the mat can be predicted among different initial and boundary conditions by this part of the hot-compression model. In the next phase of the research, the viscoelastic (time, temperature, and moisture dependent) response of the flakes was modeled using the time-temperature-moisture superposition principle of polymers. A master curve was created from data available in the literature, which describes the changing relaxation modulus of the flakes as a function of moisture and temperature at different locations in the mat. Then the flake mat was compressed in a virtual press. The stress-strain response is highly nonlinear due to the cellular structure of the mat. Hooke's Law was modified with a nonlinear strain function to account for the behavior of the flake mat in transverse compression. This part of the model gives insight into the vertical density profile formation through the thickness of the mat. Laboratory boards were produced to validate the model. A split-plot experimental design, with three different initial mat moisture contents (5, 8.5, 12 %), three final densities (609, 641, 673 kg êm3 or 38, 40, 42 lb ê ft3), two press platen temperatures (150, 200 °C), and three different press closing times (40, 60, 80 s) was applied to investigate the effect of production parameters on the internal mat conditions and the formation of the vertical density profile. The temperature and gas pressure at six locations in the mat, and the resultant density profiles of the laboratory boards, were measured. Adequate agreement was found between the model predicted and the experimentally measured temperature, pressure, and vertical density profiles. The complete model uses pressing parameters (press platen temperature, press schedule) and mat properties (flake dimensions and orientation, density distribution, initial moisture content and temperature) to predict the resulting internal conditions and vertical density profile formation within the compressed board. The density profile is related to all the relevant mechanical properties (bending strength, modulus of elasticity, internal bond strength) of the final board. The model can assist in the optimization of the parameters for hot-pressing woodbased composites and improve the performance of the final panel. / Ph. D.

The biogeochemical processes responsible for production and emission of ammonia from stored liquid dairy manure are governed by environmental factors (e.g. manure temperature, moisture) and manure characteristics (e.g. total ammoniacal nitrogen concentration, pH). These environmental factors and manure characteristics vary spatially as a result of spatially heterogeneous physical, chemical, and biological properties of manure. Existing process-based models used for estimating ammonia emission consider stored manure as a homogeneous system and do not consider these spatial variations leading to inaccurate estimations. In this study, a one-dimensional compartmental biogeochemical model was developed to (i) estimate spatial variation of temperature and substrate concentration (ii) estimate spatial variations and rates of biogeochemical processes, and (iii) estimate production and emission of ammonia from stored scraped liquid dairy manure. A one-dimension compartmentalized modeling approach was used whereby manure storage is partitioned into several sections in vertical domain assuming that the conditions are spatially uniform within the horizontal domain. Spatial variation of temperature and substrate concentration were estimated using established principles of heat and mass transfer. Pertinent biogeochemical processes were assigned to each compartment to estimate the production and emission of ammonia. Model performance was conducted using experimental data obtained from National Air Emissions Monitoring Study conducted by the United States Environmental Protection Agency. A sensitivity analysis was performed and air temperature, manure pH, wind speed, and manure total ammoniacal nitrogen concentration were identified as the most sensitive model inputs. The model was used to estimate ammonia emission from a liquid dairy manure storage of a dairy farm located in Rockingham and Franklin counties in Virginia. Ammonia emission was estimated under different management and weather scenarios: two different manure storage periods from November to April and May to October using historical weather data of the two counties. Results suggest greater ammonia emissions and manure nitrogen loss for the manure storage period in warm season from May to October compared to the storage period in cold season from November to April. / Ph. D. / Dairy manure is a byproduct of dairy farming that can be used as a fertilizer to provide essential plant nutrients such as nitrogen, phosphorus, and potassium. However, manure can only be applied to crop lands in a certain time of the year during growing seasons. Further, discharge of dairy manure into natural environment is prevented by the environmental regulations. Therefore, manure storage structures are used to store liquid dairy manure until time permits for land application or use for other purposes. During the storage, liquid dairy manure goes through biological, chemical, and physical processes and release manure gases that are linked to deteriorate human and animal health and contribute to environmental pollution. Ammonia is one of the manure gases released to atmosphere from stored liquid dairy manure. Furthermore, release of ammonia from stored manure reduce nitrogen content and reduce fertilizer value of stored manure. Implementing control measures to mitigate ammonia emission is necessary to prevent ammonia emission and reduce nitrogen loss from stored manure. Deciding and applying of appropriate control measures require knowledge of the rate at which ammonia emission occurs and when ammonia emission occurs. Use of process-based models is one of the less expensive and reliable method for estimating ammonia emission from stored liquid dairy manure. Process-based model is a mathematical model that simulates processes related to ammonia production and emission from stored manure. Even though, there are several process-based models available for estimating ammonia emission from stored liquid dairy manure, these models do not fully represent the actual processes and conditions relevant to production and emission of ammonia. For instance, spatial variation of temperature and total ammoniacal nitrogen concentration within stored manure is not considered in existing process-based models. Therefore, in this study a new compartmental process-based model was developed for estimating these spatial variations and production and emission of ammonia from stored liquid dairy manure. The model uses weather data and manure management information as inputs for estimating ammonia emission and nitrogen loss. The performance evaluation of the compartmental process-based model revealed that air temperature, manure pH, wind speed, manure total ammoniacal nitrogen concentration are important model inputs for estimating ammonia emission from stored liquid dairy manure. The model was used to estimate ammonia emission from a dairy farm located in Rockingham and Franklin counties in Virginia. Results suggest greater ammonia emissions and manure nitrogen loss for the manure storage period in warm season from May to October compared to the storage period in cold season from November to April.

<p>Co-axial dripping is one of the many ways to make drops with a core-shell structure for encapsulated materials. However, in systems where the capsule components are not density matched or surfactants are not used, the shell will eventually thin and break if not solidified in time. If the shell fails before solidifying, the core will leak out and result in a non-functional capsule. This study assumes that all capsules will fail once the core has reached 80% eccentricity, meaning a shell region has thinned to 20% of its original thickness (~70 µm). In reality, rupture of the shell depends more on stochastic defects and disturbances, but locally decreasing the shell thickness will increase the probability of capsule rupture. With this assumption, the survival time of a core-shell drop is inversely proportional to the relative velocity of the inner drop, where the greater this relative velocity, the faster the shell phase will thin. Stoke's law is generally used to approximate the speed of a sphere in a fluid. However, this study demonstrates that Stoke's law is insufficient for predicting the inner drop's motion for a compound drop. This is due to internal flows that develop within all fluid drops because of shear forces on the drop’s external face during freefall. For core-shell drops, prior studies report how the inner drop velocity can change in magnitude and direction as a function of its eccentricity, meaning its position within the outer drop. Since previous studies did not analyze this core-shell drop relationship with a 50 vol% core and a high viscosity shell, a model was built in COMSOL Multiphysics to understand how the claims from literature would apply to a previous encapsulation study (Betancourt, 2021). The model was also put through a series of validation tests that confirmed the model’s ability to accurately represent the speed and direction of inner drop motion. The final model configuration was then used to identify the transition point between buoyancy-driven and internal flow-driven failure modes observed during the production of core-shell drops in a previous encapsulation study for phase change materials (Betancourt, 2021). The model results showed how the estimated inner drop velocity was significantly reduced once accounting for the internal flows within the shell phase of a compound drop. While this study does help characterize the motion of an inner drop and could be used to find a material system with a favorable velocity profile, it is still recommended to use an in-air curing system to produce concentric capsules. Achieving a concentric capsule would still require this co-axial dripping setup to be modified significantly. </p> <p>Betancourt-Jimenez, D., Wells, B., Youngblood, J. P., & Martinez, C. J. (2021). Encapsulation of biobased fatty acid amides for phase change material applications. <em>Journal of Renewable and Sustainable Energy</em>, <em>13</em>(6), 064101. https://doi.org/10.1063/5.0072105</p>

<p>  </p> <p>Heat pipes and vapor chambers are passive two-phase heat transport devices that are used for thermal management in electronics. The passive operation of a heat pipe is facilitated by capillary wicking of the working fluid through a porous wick, and thus is subject to an operational limit in terms of the maximum pressure head that the wick can provide. This operational limit, often termed as the capillary limit, is the maximum heat input at which the pressure drop in the wick is balanced by the maximum capillary pressure head; operating a heat pipe or a vapor chamber above the capillary limit at steady-state leads to dryout. It thus becomes important to predict the performance of heat pipes and vapor chambers and explore the parametric design space to provide guidelines for minimized thermal resistance while satisfying this capillary limit. An increasingly critical aspect is to predict the transient thermal response of vapor chambers. Moreover, heat pipes and vapor chambers are extensively being used in electronic systems where the power input is dictated by the end-user activity and is expected to even exceed the capillary limit for brief time intervals. Thus, it is imperative to understand the behavior of heat pipes and vapor chambers when operated at steady and transient heat loads above the capillary limit as dryout occurs. However, review of the literature on heat pipe performance characterization reveals that the regime of dryout operation has been virtually unexplored, and thus this thesis aims to fill this critical gap in understanding.</p> <p>The design for minimized thermal resistance of a vapor chamber or a heat pipe is guided by the relative contribution of thermal resistance due to conduction across the evaporator wick and the saturation temperature gradient in the vapor core. In the limit of very thin form factors, the contribution from the vapor core thermal resistance dominates the overall thermal resistance of the vapor chamber; recent work has focused on working fluid selection to minimize overall thermal resistance in this limit. However, the wick thermal resistance becomes increasingly significant as its thickness increases to support higher heat inputs while avoiding the capillary limit. A thermal resistance network model is thus utilized to investigate the importance of simultaneously considering the contributions of the wick and vapor core thermal resistances. A generalized approach is proposed for vapor chamber design which allows <em>simultaneous</em> selection of the working fluid and wick that provides minimum overall thermal resistance for a given geometry and operating condition. While the thermal resistance network model provides a convenient method for exploring the design space, it cannot be used to predict 3-D temperature fields in the vapor chamber. Moreover, such thermal resistance network models cannot predict transient performance and temperature evolution for a vapor chamber. Therefore, an easy-to-use approach is proposed for mapping of vapor chamber transport to the heat diffusion equation using a set of appropriately defined effective anisotropic thermophysical properties, thus allowing simulation of vapor chamber as a sold conduction block. This effective anisotropic properties approach is validated against a time-stepping analytical model and is shown to have good match for both spatial and transient temperature predictions.</p> <p>Moving the focus from steady-state and transient operation of vapor chambers, a comprehensive characterization of heat pipe operation above capillary limit is performed. Different user needs and device workloads can lead to highly transient heat loads which could exceed the notional capillary limit for brief time intervals. Experiments are performed to characterize the transient thermal response of a heat pipe subjected to heat input pulses of varying duration that exceed the capillary limit. Transient dryout events due to a wick pressure drop exceeding the maximum available capillary pressure can be detected from an analysis of the measured temperature signatures. It is discovered that under such transient heating conditions, a heat pipe can sustain heat loads higher than the steady-state capillary limit for brief periods of time without experiencing dryout. If the heating pulse is sufficiently long as to induce transient dryout, the heat pipe may experience an elevated steady-state temperature even after the heat load is reduced back to a level lower than the capillary limit. The steady-state heat load must then be reduced to a level much below the capillary limit to fully recover the original thermal resistance of the heat pipe. The recovery process of heat pipes is further investigated, and a mechanism is proposed for the thermal hysteresis observed in heat pipe performance after dryout. A model for <em>steady-state</em> heat pipe transport is developed based on the proposed mechanism to predict the parametric trends of thermal resistance following recovery from dryout-induced thermal hysteresis, and the model is mechanistically validated against experiments. The experimental characterization of the recovery process demonstrates the existence of a maximum hysteresis curve, which serves as the worst-case scenario for thermal hysteresis in heat pipe after dryout. Based on the learnings from the experimental characterization, a new procedure is introduced to experimentally characterize the steady-state dryout performance of a heat pipe.</p> <p>To recover the heat pipe performance under steady-state, it has been shown that the heat input needs to be lowered down or <em>throttled</em> significantly below the capillary limit. However, due to the highly transient nature of power dissipation from electronic devices, it becomes imperative to characterize heat pipe recovery from dryout under transient operations. Hence, power-throttling assisted recovery of heat pipe from dryout has been characterized under transient conditions. A minimum throttling time interval, defined as time-to-rewet, is identified to eliminate dryout induced thermal hysteresis using power throttling. Dependence of time-to-rewet on throttling power is explored, and guidelines are presented to advise the throttling need and choice of throttling power under transient conditions. </p> <p>The experimental characterization of heat pipe operation at pulse loads above the capillary limit and power throttling following the pulse load helped define the dryout and recovery performance of a heat pipe. Next, a physics-based model is developed to predict the heat pipe <em>transient</em> thermal response under dryout-inducing pulse load and power throttling assisted recovery. This novel model considers wick as a partially saturated media with spatially and temporally varying liquid saturation, and accounts for the effect of wick partial saturation in heat pipe transport. The model prediction are validated against experiments with commercial heat pipe samples, and it is shown that the model can accurately predict dryout and recovery characteristics, namely time-to-dryout, time-to-rewet, and dryout-induced thermal hysteresis, for heat pipes with a range of wick types, heat pipe lengths and pulse loads above the capillary limit. </p> <p>The work discussed in this thesis opens certain questions that are expected to guide further research in this area. First, the thermal hysteresis mechanism proposed could be further validated with direct visualization of the liquid in a vapor chamber. To achieve this, X-ray microscopy is proposed as a viable option for the imaging <em>in situ</em> wetting dynamics in a vapor chamber. Second, the model developed to predict the dryout and recovery characteristics of the heat pipe can be used to design heat pipe with improved performance under pulse loads and power throttling. Third, novel wick designs can be explored that utilize the understanding developed of governing mechanisms for recovery from dryout, and can eliminate thermal hysteresis at powers closer to capillary limit. Fourth, the modeling approach can be extended to predict dryout and recovery trends in vapor chamber since the heat transfer pathways in a vapor chamber are different than those of a heat pipe. Fifth, and lastly it was observed several times during experiments that some of the heat pipe samples would exhibit complete dryout (sudden catastrophic rise in temperature and thermal resistance at the point of dryout) whereas other samples would exhibit partial dryout (noticeable but small increase in thermal resistance at dryout) at operating powers just above the capillary limit. Exploring and explaining the cause of complete dryout, in particular, would be an extremely valuable contribution to the heat pipe research. </p> <p>The work discussed in this thesis has led to the comprehensive development of a functional and mechanistic understanding of heat pipe operation above the notional capillary limit. The experimental procedures developed in this work are utilized to characterize a heat pipe performance under dryout and recovery. The models based on the mechanistic understanding developed from experimental characterization of dryout and recovery operation of a heat pipe have been experimentally validated and are useful for predicting heat pipe performance under dryout-inducing pulse loads and power-throttling.   </p>

<p>  </p> <p>Cavitation is one of the most common causes of failures in axial piston machines. Due to the detrimental effects that cavitation has on unit performance, it is of important consideration both in the design of new units and in defining the operational limits of existing market products. The work in this thesis aimed to contribute to the current knowledge in both areas, with a focus on design considerations with respect to cavitation scalability, and on operating conditions by measuring cavitation severity under separate and combined inciting parameters. Though the application of unit scaling is common in industry for the design of pump families, there have been no comprehensive attempts to quantify whether cavitation in fluid power units may be adequately accounted for in published scaling laws. In this thesis, the scalability of cavitation phenomena was examined through a CFD scaling study performed using a modified version of the Full Cavitation Model.  Results indicate that linear scaling is consistent in maintaining volumetric efficiency performance within 1% across scaled units up to eight times larger or smaller than the baseline. However, the gas and vapor volume distributions vary significantly between scaled units, due largely to the linear non-scalability of fluid inertia and turbulent factors. Physical exchange between phases within a working fluid was shown to be time-dependent, such that the scaled-down unit exhibits bubble collapse rates up to 30% and 150% greater than the baseline and scaled-up units, respectfully. Considering these effects, the presented work demonstrates a potential for increased cavitation damage area when downscaling a unit and reduced risk in upscaling, despite the scaling law being a reliable indicator for volumetric efficiency. </p> <p>To define a more complete study of cavitation under a variety of operating conditions and inciting parameters, this a new experimental procedure and testing circuit was proposed with focus on repeatability by controlled pressure drops and preliminary quantification of inlet fluid quality. By measuring cavitation conditions under pressure starvation, incomplete filling, and combinations thereof, the direct effect of different inception methods on unit performance was shown to be readily identifiable. Through visualization of the inlet flow, reduction in inlet pressure levels was correlated to fluid cloudiness levels and bubble size, with transparency loss at 0.0 bar_g and transition from bubbly to plug flow at -0.4 bar_g. Incomplete filling-induced cavitation was also shown to be detectable by inlet flow conditions, with a distinct change in bubble coalescence rate when operating under shaft speeds greater than or equal to fill speed for a given inlet pressure. </p>

<p>The present study investigates the capability of computational fluid dynamics (CFD) extensively to predict hydrodynamics and heat transfer characteristics of FC-72 flow boiling in a 2.5-mm ´ 5.0-mm rectangular channel and experimentally explores system instabilities: <em>density wave oscillation</em> (DWO), <em>pressure drop oscillation</em> (PDO) and <em>parallel channel instability</em> (PCI) in a micro-channel heat sink containing 38 parallel channels having a hydraulic diameter of 316-μm. </p> <p>The computational method performs transient analysis to model the entire flow field and bubble behavior for subcooled flow boiling in a rectangular channel heated on two opposite walls at high heat flux conditions of about 40% – 80% of <em>critical heat flux</em> (CHF).  The 3D CFD solver is constructed in ANSYS Fluent in which the <em>volume of fluid</em> (VOF) model is combined with a <em>shear stress transport</em> (SST) <em>k</em>-<em>ω</em> turbulent model, a surface tension model, and interfacial phase change model, along with a model for effects of shear-lift and bubble collision dispersion to overcome a fundamental weakness in modeling multiphase flows.  Detailed information about bubble distribution in the vicinity of the heated surface, thermal conduction inside the heating wall, local heat fluxes passing through the solid-fluid interface, and velocity and temperature profiles, which are not easily observed or measured by experiments, is carefully evaluated.  The simulation results are compared to experimental data to validate the solver’s ability to predict the flow configuration with single/double-side heating.  The added momentum by shear-lift is shown to govern primarily the dynamic behavior of tiny bubbles stuck on the heated bottom wall and therefore has a more significant impact on both heat transfer and heated wall temperature.  By including bubble collision dispersion force, coalescence of densely packed bubbles in the bulk region is significantly inhibited, with more giant bubbles even incurring additional breakup into smaller bubbles and culminating in far less vapor accumulation along the top wall.  Including these momentums is shown to yield better agreement with local interfacial behavior along the channel, overall flow pattern, and heat transfer parameters (wall temperature and heat transfer coefficient) observed and measured in experiments.  The computational approach is also shown to be highly effective at predicting local phenomena (velocity and temperature profiles) not easily determined through experiments.  Different flow regimes predicted along the heated length exhibit a number of dominant mechanisms, including bubble nucleation, bubble growth, coalescence, vapor blankets, interfacial waviness, and residual liquid sub-layer, all of which agree well with the experiment.  Vapor velocity is shown to increase appreciably along the heated length because of increased void fraction, while liquid velocity experiences large fluctuations.  Non-equilibrium effects are accentuated with increasing mass velocity, contributing minor deviations of fluid temperature from simulations compared to those predicted by the analytical method.  Predicted wall temperature is reasonably uniform in the middle of the heated length but increases in the entrance region due to sensible heat transfer in the subcooled liquid and decreases toward the exit, primarily because of flow acceleration resulting from increased void fraction.  When it comes to analyzing heat transfer mechanisms at extremely high heat flux via CFD, predicted flow pattern, bubble behavior, and heat transfer parameters (such as wall temperature excursion and thermal energy concentration) clearly represent phenomena of premature CHF, which take place slightly earlier than actual operating conditions.  But, despite these slight differences, the present computational work does demonstrate the ability to effectively predict the severe degradation in heat transfer performance commonly encountered at heat fluxes nearing CHF.  </p> <p>Much of the published literature addressing flow instabilities in thermal management systems employing micro-channel modules are focused on instability characteristics of the module alone, and far fewer studies have aimed at understanding the relationship between these characteristics and compressive volume in the flow loop external to the module.  From a practical point of view, developers of micro-channel thermal management systems for many modern applications are in pursuit of practical remedies that would significantly mitigate instabilities and their impact on cooling performance.  Experiments are executed using FC-72 as a working fluid with a wide range of mass velocities and a reasonably constant inlet subcooling of ~15°C.  The flow instabilities are reflected in pressure fluctuations detected mainly in the heat sink’s upstream plenum.  Both inlet pressure and pressure drop signals are analyzed in pursuit of amplitude and frequency characteristics for different mass velocities and over a range of heat fluxes.  The current experimental study also examines the effects of compressible volume location in a closed pump-driven flow loop designed to deliver FC-72 to a micro-channel test module having 38 channels with 315-μm hydraulic diameter.  Three accumulator locations are investigated: upstream of the test module, downstream of the test module, and between the condenser and pump.  Both high-frequency temporal parameter data and high-speed video records are analyzed for ranges of mass velocity and heat flux, with inlet subcooling held constant at ~15°C.  PDO is shown to dominate when the accumulator is situated upstream, whereas PCI is dominant for the other two locations.  Appreciable confinement of bubbles in individual channels is shown to promote rapid axial bubble growth.  The study shows significant variations in the amount of vapor generated and dominant flow patterns among channels, a clear manifestation of PCI, especially for low mass velocities and high heat fluxes.  It is also shown effects of the heat sink’s instabilities are felt in other components of the flow loop.  The parametric trends for PCI are investigated with the aid of three different types of stability maps which show different abilities at demarcating stable and unstable operations.  PDO shows severe pressure oscillations across the micro-channel heat sink, with rapid bubble growth and confinement, elongated bubble expansion in both directions, flow stagnation, and flow reversal (including vapor backflow to the inlet plenum) constituting the principal sequence of events characterizing the instability.  Spectral analysis of pressure signals is performed using Fast Fourier Transform, which shows PDO extending the inlet pressure fluctuations with the same dominant frequency to other upstream flow loop components, with higher amplitudes closer to the pump exit.  From a practical system operation point of view, throttling the flow upstream of the heat sink eliminates PDO but renders PCI dominant, and placing the accumulator in the liquid flow segment of the loop between the condenser and pump ensures the most stable operation.</p>

<p dir="ltr">As aircraft engine manufacturers continue to embark on their pursuit of higher-efficiency, lower-emissions gas turbines, a prevailing theme in the industry has been the increase of the engine bypass ratio. As the optimization space for engine bypass ratios trends towards smaller and smaller engine core sizes, the feasibility of centrifugal compressors as the final stage in an axial-centrifugal compressor becomes apparent due to their performance advantages at smaller scales.</p><p dir="ltr">This study performed an investigation into the aerodynamics of a stationary diffusion system intended for use with a final stage aeroengine centrifugal compressor using experimental and numerical techniques. Experimental work was performed at the Purdue Compressor Research Lab at Purdue University’s Maurice J. Zucrow Laboratories. Data were collected from several iterations of rapidly prototyped, additively manufactured diffuser and deswirl parts with internal instrumentation features. Furthermore, computational work on the stage was conducted using the Ansys Turbosystem.</p><p dir="ltr">The goal of this research is to identify trends in stationary diffusion system designs and the geometric features that cause them. Furthermore, the ability of steady computational fluid dynamics methods to predict these changes was evaluated using two turbulence models to produce a simulation of the compressor flow field. When used in conjunction with one another, the efficient use of computational methods to identify an optimal design and rapid prototyping to validate it leads to the determination of the best diffusion system design at a lower cost and time requirement than what is otherwise currently possible.</p><p dir="ltr">The different geometries which were tested identified the negative effects of spanwise vane contouring on the diffuser performance and the ability of endwall divergence to augment the pressure recovery performance of a design at the expense of increased losses. A full understanding of the effect of each design parameter is enabled by iterative inclusion or exclusion of certain design parameters. Furthermore, the use of computational fluid dynamics showed that the BSLEARSM turbulence model performs reasonably well in predicting the build-to-build performance trends. However, neither the BSLEARSM nor the SST turbulence model were able to accurately identify performance trends for the deswirl. For this reason, additional work is warranted to identify an optimal set of parameters to characterize the high axial and meridional turning present in this component.</p>

<p dir="ltr">With increasing power densities in electronic devices, effective thermal management has become an indispensable aspect of electronic systems design. Although phase change materials (PCMs) have been studied as a potential solution, their integration into microelectronic and high-power devices presents a significant challenge due to low thermal conductivity and lack of effective thermal pathways from the heat source to the heat sink. While much work has focused on integrating thermal storage into heat sinks, this dissertation instead investigates integrating PCMs between the heat source and the heat sink in different configurations. By placing the energy storage closer to the heat source, the thermal resistance is reduced, which improves the overall thermal performance of the device. Specifically, this work explores the efficacy of two approaches: (1) direct embedding of a PCM within the die for mobile electronics applications and (2) integration of an auxiliary composite PCM/copper thermal energy storage (TES) component in combination with active liquid cooling for high-power power electronics modules.</p><p><br></p><p dir="ltr">The first study explores die-level thermal management for microelectronics using PCMs. Silicon chips with PCM embedded within the die are modeled using ParaPower, a fast-analysis tool, and a genetic algorithm is used to efficiently optimize the distribution of high-conductivity silicon pathways and high thermal capacitance PCM zones. A thermal test vehicle (TTV) of a realistic microelectronics form factor with an embedded PCM layer is first designed, and a process is developed to fabricate such a TTV. This study is the first to successfully fabricate a TTV with fully encapsulated PCM and validate its thermal response across various operational scenarios. For temperature cycling tests (where the TTV temperature fluctuates between predetermined hot and cold setpoints), the embedded-PCM TTVs extend the operational time by up to 2.8x compared to a baseline all-silicon TTV. For duty cycling tests (with a fixed duration of the periodic heating pulses and off times), the embedded-PCM TTVs suppress the hotspot temperature rise by up to 14% and stabilize quasi-steady state temperature fluctuations by up to 65% through repeated PCM melting and solidification cycles. Thermal performance enhancements are observed even for high heat fluxes of ~65W/cm² . Specifically, a TTV with an embedded square-shaped PCM reservoir reduces temperature instability by an average of 40% across a range of cycle durations.</p><p><br></p><p dir="ltr">The second study investigates the effectiveness of different integration strategies for an auxiliary composite PCM/copper TES block integrated alongside a cold plate, for thermal management of high-power power electronics modules, specifically for electric vehicles. These systems are evaluated for realistic drive cycles of various driving intensities. Computational results indicate that this approach is most effective when the composite TES block is positioned directly above the heat-generating silicon carbide dies. This configuration excels at stabilizing transient temperature fluctuations and absorbing thermal shocks, achieving reductions of up to approximately ~33% compared to current thermal management techniques. This strategy is particularly effective for stop-and-go drive cycles characterized by high rates of acceleration and deceleration, low average driving speeds, and frequent stops, typical of driving schedules for public transport buses and mail delivery vehicles.</p><p><br></p><p dir="ltr">The results from both thermal management approaches demonstrate that the integration of a PCM cooling solution in close proximity to the heat source can significantly enhance its effectiveness by absorbing power bursts and limiting temperature instability via repeated melting and solidification. The contributions of this dissertation include the development of an effective optimization strategy for generating optimized PCM distributions, which reduces the maximum temperature and temperature oscillations in a device with significant computational efficiency. (The same optimization strategy can be applied to other thermal management design challenges.) Notably, TTVs of realistic microelectronics form factors with embedded PCM were designed, modeled, fabricated, and validated. With the PCM thermal buffers, the engineered solution demonstrated superior performance compared to a baseline all-silicon TTV. The second study into the integration of composite PCM/copper TES blocks into high-power power electronics modules established trade-offs between different architectures across various performance metrics, and highlighted its effectiveness for drive cycles with varying intensities. These findings offer an important contribution to the development of embedded thermal management techniques for electronic systems design, which will be critical for the advancement of next-generation microelectronics and high-power devices.</p>

<p dir="ltr">Due to the end of Dennard scaling, electronic devices must consume more electrical power for increased functionality. The increased power consumption, combined with diminishing form factors, results in increased power density within the device, leading to increased heat fluxes at the devices surfaces. Without proper thermal management, the increase in heat fluxes can cause device temperatures to exceed operational limits, ultimately resulting in device failure. However, the dissipation of these high heat fluxes often requires pumping or refrigeration of a coolant, which in turn, increases the total energy usage. Data centers, which form the backbone of the cloud infrastructure and the modern economy, account for ~2% of the total US electricity use, of which up to ~40% is spent on cooling needs alone. Thus, it is necessary to optimize the designs of the cooling systems to be able to dissipate higher heat fluxes, but at lower operating powers.</p><p dir="ltr">The design optimization of various thermal management components such as cold plates, heat sinks, and heat exchangers relies on accurate prediction of flow heat transfer and pressure drop. During the iterative design process, the heat transfer and pressure drop is typically either computed numerically or obtained using geometry-specific correlations for Nusselt number (<i>Nu</i>) and friction factor (<i>f</i>). Numerical approaches are accurate for evaluation of a single design but become computationally expensive if many design iterations are required (such as during formal optimization processes). Moreover, traditional empirical correlations are highly geometry dependent and assume functional forms that could introduce inaccuracies. To overcome these limitations, this thesis introduces accurate and continuous-valued machine-learning (ML)-based surrogate models for predicting Nusselt number and friction factor on various heat exchange surfaces. These surrogate models, which are applicable to more geometries than traditional correlations, enable flexible and computationally inexpensive design optimization. The utility of these surrogate models is first demonstrated through the optimization of single-phase liquid cold plates under specific boundary conditions. Subsequently, their effectiveness is further showcased in the more practical challenge of designing liquid-to-liquid heat exchangers by integrating the surrogate models with a homogenization-based topology optimization framework. As topology optimization relies heavily on accurate predictions of pressure drop and heat transfer at every point in the domain during each iteration, using ML-based surrogate models greatly reduces the computational cost while enabling the development of high-performance, customized heat exchange surfaces. Thus, this work contributes to the advancement of thermal management by leveraging machine learning techniques for efficient and flexible design optimization processes.</p><p dir="ltr">First, artificial neural network (ANN)-based surrogate correlations are developed to predict <i>f</i> and <i>Nu</i> for fully developed internal flow in channels of arbitrary cross section. This effectively collapses all known correlations for channels of different cross section shapes into one correlation for <i>f</i> and one for <i>Nu</i>. The predictive performance and generality of the ANN-based surrogate models is verified on various shapes outside the training dataset, and then the models are used in the design optimization of flow cross sections based on performance metrics that weigh both heat transfer and pressure drop. The optimization process leads to novel shapes outside the training data, the performance of which is validated through numerical simulations. Although the ML model predictions lose accuracy outside the training set for these novel shapes, the predictions are shown to follow the correct trends with parametric variations of the shape and therefore successfully direct the search toward optimized shapes.</p><p dir="ltr">The success of ANN-aided shape optimization of constant cross-section internal flow channels serves as a compelling proof-of-concept, highlighting the potential of ML-aided optimization in thermal-fluid applications. However, to address the complexities of widely used thermal management devices such as cold plates and heat exchangers, known for their intricate surface geometries beyond constant cross-section channels, a strategic shift is imperative. With the goal of crafting ML models specifically tailored for practical design optimization algorithms like topology optimization, the thesis next delves into diverse micro-pin fin arrangements commonly employed in applications like cold plates and heat exchangers. This study on pin fins includes the exploration of hydrodynamic and thermal developing effects, as well as the impact of pin fin cross section shape and orientation. The ML-based predictive models are trained on numerically simulated synthetic data. The large amounts of accurate synthetic data required to train machine learning models are generated using a custom-developed simulation automation framework. With this framework, numerical flow and heat transfer simulations can be run on thousands of geometries and boundary conditions with minimal user intervention. The proposed models provide accurate predictions of <i>f</i> and <i>Nu</i>, with a near exact match to the training data as well as on unseen testing data. Furthermore, the outputs of the ANNs are inspected to propose new analytical correlations to estimate the hydrodynamic and thermal entrance lengths for flow through square pin fin arrays. The ML models are also shown to be useable for fluids other than water, employing physics-based, Prandtl-number-dependent scaling relations.</p><p dir="ltr">The thesis further demonstrates the utility of the ML surrogate models to facilitate the design optimization of thermal management components through their integration in the topology optimization (TO) framework for heat exchanger design. Topology optimization is a computational design methodology for determining the optimal material distribution within a design space based on given constraints. The use of topology optimization in the design of heat exchangers and other thermal management devices has been gaining significant attention in recent years, particularly with the widespread availability of additive manufacturing techniques that offer geometric design flexibility. Particularly advantageous for heat exchanger design is the homogenization approach to topology optimization, which represents partial densities in the design domain using a physical unit cell structure to achieve sub-grid resolution features. This approach requires geometry-specific, correlations for <i>f</i> and <i>Nu</i> to simulate the performance of designs and evaluate the objective function during the optimization process. Topology optimized pin fin-based component designs rely on additive manufacturing, posing production scalability challenges with current technologies. Furthermore, the demand for flow and thermal anisotropy in several applications adds complexity to the design requirements. To address these challenges, the focus is shifted to traditional heat exchanger surface geometries that can be manufactured using conventional techniques, and which also exhibit pronounced anisotropy in flow and heat transfer characteristics. Traditionally, these geometries are distributed uniformly across heat exchange surfaces. However, incorporating such geometries into the topology optimization framework merges the strengths of both approaches, yielding mathematically optimized heat exchange surfaces with conventionally manufacturable designs. Offset strip fins, one such commonly used geometry, is chosen to be the physical unit cell structure to demonstrate the integration of ML-based surrogate models into the topology optimization framework. The large amount of data required to develop robust machine learning-based surrogate <i>f</i> and <i>Nu</i> models for axial and cross flow of water through offset strip fins are generated through numerical simulations performed for convective flows through these geometries. The data generated are compared against in-house-measured experimental data as well as against data from literature. To facilitate the integration of ML models into topology optimization, a discrete adjoint method was developed to calculate the sensitivities during topology optimization, to circumvent the absence of the analytical gradients.</p><p dir="ltr">Successful integration of the machine learning-based surrogate models into the topology optimization framework was demonstrated through the design optimization of a counterflow heat exchanger. The topology optimized design outperformed the benchmarks that used uniform, parametrically optimized offset strip fin arrays. The topology optimized design exhibited domain-specific enhancements such as peripheral flow paths for enhanced heat transfer and open channels to minimize pressure drops. This integration showcases the potential of combining ML models with topology optimization, providing a flexible framework that can be extended to a wide range of enhanced surface structure types and geometric configurations for which ML models can be trained. Thus, by enabling spatially localized optimization of enhanced surface structures using ML models, and consequently offering a pathway for expanding the design space to include many more surface structures in the topology optimization framework than previously possible, this thesis lays the foundation for advancing design optimization of thermal-fluid components and systems, using both additively and conventionally manufacturable geometries.</p>

The thermal efficiency of gas turbines increases with the temperature of the gas entering its turbine component. To enable high inlet temperatures, even those that far exceed the melting point of the turbine materials, the turbine must be cooled. One way is by internal cooling, where cooler air passes through U-ducts embedded inside turbine vanes and blades. Since the flow and heat transfer in these ducts are highly complicated, computational fluid dynamics (CFD) based on RANS have been used extensively to explore and assess design concepts. However, RANS have been found to be unreliable – giving accurate results for some designs but not for others. In this study, large-eddy simulations (LES) were performed for a U-duct with a trapezoidal cross section to assess four widely used RANS turbulence models: realizable k-ε (k-ε), shear-stress transport (SST), Reynolds stress model with linear pressure strain (RSM-LPS), and the seven-equation stress-omega full Reynolds stress model (RSM).<div><br></div><div>When examining the capability of steady RANS, two versions of the U-duct were examined, one with a staggered array of pin fins and one without pin fins. Results obtained for the heat-transfer coefficient (HTC) were compared with experimental measurements. The maximum relative error in the predicted “averaged” HTC was found to be 50% for k-ε and RSM-LPS, 20% for SST, and 30% for RSM-τω when there are no pin fins and 25% for k-ε, 12% for the SST and RSM-τω when there are pin fins. When there are no pin fins, all RANS models predicted a large separated flow region downstream of the turn, which the experiment does show to exist. Thus, all models predicted local distributions poorly. When there were pin fins, they behaved like guide vanes in turning the flow and confined the separation around the turn. For this configuration, all RANS models predicted reasonably well.<br></div><div><br></div><div>To understand why RANS cannot predict the HTC in the U-duct after the turn when there are no pin fins, LES were performed. To ensure that the LES is benchmark quality, verification and validation were performed via LES of a straight duct with square cross section where data from experiments and direct numerical simulation (DNS) are available. To ensure correct inflow boundary condition is provided for the U-duct, a concurrent LES is performed of a straight duct with the same trapezoidal cross section and flow conditions as the U-duct. Results obtained for the U-duct show RANS models to be inadequate in predicting the separation due to their inability to predict the unsteady separation about the tip of the turn. To investigate the limitations of the RANS models, LES results were generated for the turbulent kinetic energy, Reynolds-stresses, pressure-strain rate, turbulent diffusion, pressure diffusion, turbulent transport, and velocity-temperature correlations with focus on understanding their behavior induced by the turn region of the U-duct. As expected, the Boussinesq assumption was found to be incorrect, which led to incorrect predictions of Reynolds stresses. For RSM-τω, the modeling of the pressure-strain rate was found to match LES data well, but huge error was found on modeling the turbulent diffusion. This huge error indicates that the two terms in the turbulent diffusion – pressure diffusion and turbulent transport – should be modeled separately. Since the turbulent transport was found to be ignorable, the focus should be on modeling the pressure diffusion. On the velocity-temperature correlations, the existing eddy-diffusivity model was found to be over simplified if there is unsteady separation with shedding. The generated LES data could be used to provide the guidance for a better model.<br></div>