This thesis presents a thermal gradient gas chromatography (TGGC) system that is implemented on a micro-scale. The GC column is approximately 20 cm long and is fabricated out of silicon with 21 nickel thin-film heaters evenly placed along the length of the column. Computational heat transfer models using ANSYS Mechanical APDL predict heating and cooling rates up to 32,000 deg C/min and 3,600 deg C/min, respectively. These results are verified through testing an experimental silicon channel. A PI controller which uses resistance measurements to calculate thin-film temperature is used for obtaining dynamic thermal gradient control. This controller is shown to possess a characteristic rise time of approximately 0.3 seconds with less than 4% overshoot and precision to within less than a degree. These characteristics present this system as a highly favorable candidate for a micro-GC column with resolution similar to that of conventional GC.

<div><div><div><p>The exceptional long-term performance of concrete is a primary reason that this material represents a significant portion of the construction industry. However, a portion of this construction material is prone to premature deterioration for multi-physical durability issues such as internal frost damage, restrained shrinkage damage, and aggregate susceptibility to fracture. Since each durability issue is associated with a unique damage mechanism, this study aims at investigating the underlying physical mechanisms individually by characterizing the mechanical and thermal properties development and indicating how each unique damage mechanism may compromise the properties development over the design life of the material.</p><p>The first contribution of this work is on the characterization of thermal behavior of porous media (e.g., cement-based material) with a complex solid-fluid coupling subject to thermal cycling. By combining Young-Kelvin-Laplace equation with a computational heat transfer approach, we can calculate the contributions of (i) pore pressure development associated with solidification and melting of pore fluid, (ii) pore size distribution, and (iii) equilibrium phase diagram of multiple phase change materials, to the thermal response of porous mortar and concrete during freezing/thawing cycles. Our first finding indicates that the impact of pore size (and curvature) on freezing is relatively insignificant, while the effect of pore size is much more significant during melting. The fluid inside pores smaller than 5 nm (i.e., gel pores) has a relatively small contribution in the macroscopic freeze-thaw behavior of mortar specimens within the temperature range used in this study (i.e., +24 °C to -35 °C). Our second finding shows that porous cementitious composites containing lightweight aggregates (LWAs) impregnated with an organic phase change material (PCM) as thermal energy storage (TES) agents have the significant capability of improving the freeze-thaw performance. We also find that the phase transitions associated with the freezing/melting of PCM occur gradually over a narrow temperature range (rather than an instantaneous event). The pore size effect of LWA on freezing and melting behavior of PCM is found to be relatively small. Through validation of simulation results with lab-scale experimental data, we then employ the model to investigate the effectiveness of PCMs with various transition temperatures on reducing the impact of freeze-thaw cycling within concrete pavements located in different regions of United States.</p><div><div><div><p>The second contribution of this work is on quantification of mechanical properties development of cementitious composites across multiple length scales, and two damage mechanisms associated with aggregate fracture and restrained shrinkage cracking that lead to compromising the long-term durability of the material. The former issue is addressed by combining finite element method-based numerical tools, computational homogenization techniques, and analytical methods, where we observe a competing fracture mechanism for early- age cracking at two length scales of mortar (meso-level) and concrete (macro-level). When the tensile strength of the cement paste is lower than the tensile strength of the aggregate phase, the crack propagates across the paste. When the tensile strength of the cement paste exceeds that of the aggregate, the cracks begin to deflect and propagate through the aggregates. As such, a critical degree of hydration (associated with a particular time) exists below which the cement paste phase is weaker than the aggregate phase at the onset of hydration. This has implications on the inference of kinetic based parameters from mechanical testing (e.g., activation energy). Next, we focus on digital fabrication of a cement paste structure with controlled architecture to allow for mitigating the intrinsic damage induced by inherent shrinkage behavior followed by extrinsic damage exerted by external loading. Our findings show that the interfaces between the printed filaments tend to behave as the first layer of protection by enabling the structure to accommodate the damage by deflecting the microcrack propagation into the stable configuration of interfaces fabricated between the filaments of first and second layers. This fracture behavior promotes the damage localization within the first layer (i.e., sacrificial layer), without sacrificing the overall strength of specimen by inhibiting the microcrack advancement into the neighboring layers, promoting a novel damage localization mechanism. This study is undertaken to characterize the shrinkage-induced internal damage in 7-day 3D-printed and cast specimens qualitatively using X-ray microtomography (μCT) technique in conjunction with multiple mechanical testing, and finite element numerical modeling. As the final step, the second layer of protection is introduced by offering an enhanced damage resistance property through employing bioinspired Bouligand architectures, promoting a damage delocalization mechanism throughout the specimen. This novel integration of damage localization-delocalization mechanisms allows the material to enhance its flaw tolerant properties and long-term durability characteristics, where the reduction in the modulus of rupture (MOR) of hardened cement paste (hcp) elements with restrained shrinkage racking has been significantly improved by ~ 25% when compared to their conventionally cast hcp counterparts.</p></div></div></div></div></div></div>

<div> <p>The distribution of local water vapor in the surrounding air has been shown to be the driving mechanism for several phase change phenomena during dropwise condensation and condensation frosting. This thesis uses reduced-order modeling approaches, which account for the effects of the vapor distribution to predict the droplet growth dynamics during dropwise condensation in systems of many droplets. High-fidelity modeling techniques are used to further probe and quantify the heat and mass transport mechanisms that govern the local interactions between a freezing droplet and its surrounding ambient, including neighboring droplets. The relative significance of these transport mechanisms in the propagation of frost are investigated. A reduced-order analytical method is first developed to calculate the condensation rate of each individual droplet within a group of droplets on a surface by resolving the vapor concentration field in the surrounding air. A point sink superposition method is used to account for the interaction between all droplets without requiring solution of the diffusion equation for a full three-dimensional domain. For a simplified scenario containing two neighboring condensing droplets, the rates of growth are studied as a function of the inter-droplet distance and the relative droplet size. Interactions between the pair of droplets are discussed in terms of changes in the vapor concentration field in the air domain around the droplets. For representative systems of condensing droplets on a surface, the total condensation rates predicted by the reduced-order model match numerical simulations to within 15%. The results show that assuming droplets grow as an equivalent film or in a completely isolated manner can severely overpredict condensation rates.</p> <p>The point superposition model is then used to predict the condensation rates measured during condensation experiments. The results indicate that it is critical to consider a large number of interacting droplets to accurately predict the condensation behavior. Even though the intensity of the interaction between droplets decreases sharply with their separation distance, droplets located relatively far away from a given droplet must be considered to accurately predict the condensation rate, due to the large aggregate effect of all such far away droplets. By considering an appropriate number of interacting droplets in a system, the point sink superposition method is able to predict experimental condensation rates to within 5%. The model was also capable of predicting the time-varying condensation rates of individual droplets tracked over time. These results confirm that diffusion-based models that neglect the interactions of droplets located far away, or approximate droplet growth as an equivalent film, overpredict condensation rates.</p> <p>In dropwise condensation from humid air, a full description of the interactions between droplets can be determined by solving the vapor concentration field while neglecting heat transfer across the droplets. In contrast, the latent heat released during condensation freezing processes cause droplet-to-ambient as well as droplet-to-droplet interactions via coupled heat and mas transfer processes that are not well understood, and their relative significance has not been quantified. As a first step in understanding these mechanisms, high-fidelity modeling of the solidification process, along with high-resolution infrared (IR) thermography measurements of the surface of a freezing droplet, are used to quantify the pathways for latent heat dissipation to the ambient surroundings of a droplet. The IR measurements are used to show that the crystallization dynamics are related to the size of the droplet, as the freezing front moves slower in larger droplets. Numerical simulations of the solidification process are performed using the IR temperature data at the contact line of the droplet as a boundary condition. These simulations, which have good agreement with experimentally measured freezing times, reveal that the heat transferred to the substrate through the base contact area of the droplet is best described by a time-dependent temperature boundary condition, contrary to the constant values of base temperature and rates of heat transfer assumed in previous numerical simulations reported in the literature. In further contrast to the highly simplified descriptions of the interaction between a droplet and its surrounding used in previous models, the model developed in the current work accounts for heat conduction, convection, and evaporative cooling at the droplet-air interface. The simulation results indicate that only a small fraction of heat is lost through the droplet-air interface via conduction and evaporative cooling. The heat transfer rate to the substrate of the droplet is shown to be at least one order of magnitude greater than the heat transferred to the ambient air.</p> <p>Subsequently, the droplet-to-droplet interactions via heat and mass exchange between a freezing droplet and a neighboring droplet, for which asymmetries are observed in the final shape of the frozen droplet, are investigated. Side-view infrared (IR) thermography measurements of the surface temperature for a pair of freezing droplets, along with three-dimensional numerical simulations of the solidification process, are used to quantify the intensity and nature of these interactions. Two droplet-to-droplet interaction mechanisms causing asymmetric freezing are identified: (1) non-uniform evaporative cooling on the surface of the freezing droplet caused by vapor starvation in the air between the droplets; and (2) a non-uniform thermal resistance at the contact area of the freezing droplet caused by the heat conduction within the neighboring droplet. The combined experimental and numerical results show that the size of the freezing droplet relative to its neighbor can significantly impact the intensity of the interaction between the droplets and, therefore, the degree of asymmetry. A small droplet freezing in the presence of a large droplet, which blocks vapor from freely diffusing to the surface of the small droplet, causes substantial asymmetry in the solidification process. The droplet-to-droplet interactions investigated in thesis provide insights into the role of heat dissipation in the evaporation of neighboring droplets and ice bridging, and open new avenues for extending this understanding to a system-level description for the propagation of frost.</p> </div> <br>

Despite the central importance of porous electrodes to any advanced electrochemical system, there is no clear answer to “<i>How to make the best electrode</i>?”. The source of ambiguity lies in the incomplete understanding of convoluted material interactions at smaller – difficult to observe length and timescales. Such mesoscopic interactions, however, abide by the fundamental physical principles such as mass conservation. The porous electrodes are investigated in such a physics-based setting to comprehend the interplay among structural arrangement and off-equilibrium processes. As a result, a synergistic approach exploiting the complementary characteristics of controlled experiments and theoretical analysis emerges to allow mechanistic insights into the associated mesoscopic phenomena. The potential of this philosophy is presented by investigating three distinct electrochemical systems with their unique peculiarities.

Heat actuated adsorption heat pumps offer the opportunity to improve overall energy efficiency in waste heat applications by eliminating shaft work requirements accompanying vapor compression cycles. The coefficient of performance (COP) in adsorption heat pumps is generally low. The objective of this thesis is to model the adsorption system to gain critical insight into how its performance can be improved. Because adsorption heat pumps are intermittent devices, which induce cooling by adsorbing refrigerant in a sorption bed heat/mass exchanger, transient models must be used to predict performance. In this thesis, such models are developed at the adsorbent particle level, heat/mass exchanger component level and system level. Adsorption heat pump modeling is a coupled heat and mass transfer problem. Intra-particle mass transfer resistance and sorption bed heat transfer resistance are shown to be significant, but for very fine particle sizes, inter-particle resistance may also be important. The diameter of the adsorbent particle in a packed bed is optimized to balance inter- and intra-particle resistances and improve sorption rate. In the literature, the linear driving force (LDF) approximation for intra-particle mass transfer is commonly used in place of the Fickian diffusion equation to reduce computation time; however, it is shown that the error in uptake prediction associated with the LDF depends on the working pair, half-cycle time, adsorbent particle radius, and operating temperatures at hand. Different methods for enhancing sorption bed heat/mass transfer have been proposed in the literature including the use of binders, adsorbent compacting, and complex extended surface geometries. To maintain high reliability, the simple, robust annular-finned-tube geometry with packed adsorbent is specified in this work. The effects of tube diameter, fin pitch and fin height on thermal conductance, metal/adsorbent mass ratio and COP are studied. As one might expect, many closely spaced fins, or high fin density, yields high thermal conductance; however, it is found that the increased inert metal mass associated with the high fin density diminishes COP. It is also found that thin adsorbent layers with low effective conduction resistance lead to high thermal conductance. As adsorbent layer thickness decreases, the relative importance of tube-side convective resistance rises, so mini-channel sized tubes are used. After selecting the proper tube geometry, an overall thermal conductance is calculated for use in a lumped-parameter sorption bed simulation. To evaluate the accuracy of the lumped-parameter approach, a distributed parameter sorption bed simulation is developed for comparison. Using the finite difference method, the distributed parameter model is used to track temperature and refrigerant distributions in the finned tube and adsorbent layer. The distributed-parameter tube model is shown to be in agreement with the lumped-parameter model, thus independently verifying the overall UA calculation and the lumped-parameter sorption bed model. After evaluating the accuracy of the lumped-parameter model, it is used to develop a system-level heat pump simulation. This simulation is used to investigate a non-recuperative two-bed heat pump containing activated carbon fiber-ethanol and silica gel-water working pairs. The two-bed configuration is investigated because it yields a desirable compromise between the number of components (heat exchangers, pumps, valves, etc.) and steady cooling rate. For non-recuperative two-bed adsorption heat pumps, the average COP prediction in the literature is 0.39 for experiments and 0.44 for models. It is important to improve the COP in mobile waste heat applications because without high COP, the available waste heat during startup or idle may be insufficient to deliver the desired cooling duty. In this thesis, a COP of 0.53 is predicted for the non-recuperative, silica gel-water chiller. If thermal energy recovery is incorporated into the cycle, a COP as high as 0.64 is predicted for a 90, 35 and 7.0°C source, ambient and average evaporator temperature, respectively. The improvement in COP over heat pumps appearing in the literature is attributed to the adsorbent particle size optimization and careful selection of sorption bed heat exchanger geometry.

Heat transfer and solidification models were developed for use in a numerical model of a continuous caster to provide a means of predicting how the developing shell would react under variable operating conditions. Measurement data of the operating conditions leading up to a breakout occurrence were provided by an industrial collaborator and were used to define the model boundary conditions. Steady-state and transient simulations were conducted, using boundary conditions defined from time-averaged measurement data. The predicted shell profiles demonstrated good agreement with thickness measurements of a breakout shell segment – recovered from the quarter-width location. Further examination of the results with measurement data suggests pseudo-steady assumption may be inadequate for modeling shell and flow field transition period following sudden changes in casting speed. An adaptive mesh refinement procedure was established to increase refinement in areas of predicted shell growth and to remove excess refinement from regions containing only liquid. A control algorithm was developed and employed to automate the refinement procedure in a proof-of-concept simulation. The use of adaptive mesh refinement was found to decrease the total simulation time by approximately 11% from the control simulation – using a static mesh.

A key component of research in the aerospace industry constitutes hypersonic flights (M>5) which includes the design of commercial high-speed aircrafts and development of rockets. Computational analysis becomes more important due to the difficulty in performing experiments and reliability of its results at these harsh operating conditions. There is an increasing demand from the industry for the accurate prediction of wall-shear and heat transfer with a low computational cost. Direct Numerical Simulations (DNS) create the standard for accuracy, but its practical usage is difficult and limited because of its high cost of computation. The usage of Reynold's Averaged Navier Stokes (RANS) simulations provide an affordable gateway for industry to capitalize its lower computational time for practical applications. However, the presence of existing RANS turbulence closure models and associated wall functions result in poor prediction of wall fluxes and inaccurate solutions in comparison with high fidelity DNS data. In recent years, machine learning emerged as a new approach for physical modeling. This thesis explores the potential of employing Machine Learning (ML) to improve the predictions of wall fluxes for hypersonic turbulent boundary layers. Fine-grid RANS simulations are used as training data to construct a suitable machine learning model to improve the solutions and predictions of wall quantities for coarser meshes. This strategy eliminates the usage of wall models and extends the range of applicability of grid sizes without a significant drop in accuracy of solutions. Random forest methodology coupled with a bagged aggregation algorithm helps in modeling a correction factor for the velocity gradient at the first grid points. The training data set for the ML model extracted from fine-grid RANS, includes neighbor cell information to address the memory effect of turbulence, and an optimal set of parameters to model the gradient correction factor. The successful demonstration of accurate predictions of wall-shear for coarse grids using this methodology, provides the confidence to build machine learning models to use DNS or high-fidelity modeling results as training data for reduced-order turbulence model development. This paves the way to integrate machine learning with RANS to produce accurate solutions with significantly lesser computational costs for hypersonic boundary layer problems.

<p>The main accomplishments of this research are </p> <p>(1) developed a high fidelity computational methodology based on large eddy simulation to capture lean blowout (LBO) behaviors of different fuels; </p> <p>(2) developed fundamental insights into the combustion processes leading to the flame blowout and fuel composition effects on the lean blowout limits; </p> <p>(3) developed artificial intelligence-based models for early detection of the onset of the lean blowout in a realistic complex combustor. </p> <p>The methodologies are demonstrated by performing the lean blowout (LBO) calculations and statistical analysis for a conventional (A-2) and an alternative bio-jet fuel (C-1).</p> <p>High-performance computing methodology is developed based on the large eddy simulation (LES) turbulence models, detailed chemistry and flamelet based combustion models. This methodology is employed for predicting the combustion characteristics of the conventional fuels and bio-derived alternative jet fuels in a realistic gas turbine engine. The uniqueness of this methodology is the inclusion of as-it-is combustor hardware details such as complex hybrid-airblast fuel injector, thousands of tiny effusion holes, primary and secondary dilution holes on the liners, and the use of highly automated on the fly meshing with adaptive mesh refinement. The flow split and mesh sensitivity study are performed under non-reacting conditions. The reacting LES simulations are performed with two combustion models (finite rate chemistry and flamelet generated manifold models) and four different chemical kinetic mechanisms. The reacting spray characteristics and flame shape are compared with the experiment at the near lean blowout stable condition for both the combustion models. The LES simulations are performed by a gradual reduction in the fuel flow rate in a stepwise manner until a lean blowout is reached. The computational methodology has predicted the fuel sensitivity to lean blowout accurately with correct trends between the conventional and alternative bio-jet fuels. The flamelet generated manifold (FGM) model showed 60% reduction in the computational time compared to the finite rate chemistry model. </p> <p>The statistical analyses of the results from the high fidelity LES simulations are performed to gain fundamental insights into the LBO process and identify the key markers to predict the incipient LBO condition in swirl-stabilized spray combustion. The bio-jet fuel (C-1) exhibits significantly larger CH₂O concentrations in the fuel-rich regions compared to the conventional petroleum fuel (A-2) at the same equivalence ratio. It is observed from the analysis that the concentration of formaldehyde increases significantly in the primary zone indicating partial oxidation as we approach the LBO limit. The analysis also showed that the temperature of the recirculating hot gases is also an important parameter for maintaining a stable flame. If this temperature falls below a certain threshold value for a given fuel, the evaporation rates and heat release rated decreases significantly and consequently leading to the global extinction phenomena called lean blowout. The present study established the minimum recirculating gas temperature needed to maintain a stable flame for the A-2 and C-1 fuels. </p> The artificial intelligence (AI) models are developed based on high fidelity LES data for early identification of the incipient LBO condition in a realistic gas turbine combustor under engine relevant conditions. The first approach is based on the sensor-based monitoring at the optimal probe locations within a realistic gas turbine engine combustor for quantities of interest using the Support Vector Machine (SVM). Optimal sensor locations are found to be in the flame root region and were effective in detecting the onset of LBO ~20ms ahead of the event. The second approach is based on the spatiotemporal features in the primary zone of the combustor. A convolutional autoencoder is trained for feature extraction from the mass fraction of the OH ( data for all time-steps resulting in significant dimensionality reduction. The extracted features along with the ground truth labels are used to train the support vector machine (SVM) model for binary classification. The LBO indicator is defined as the output of the SVM model, 1 for unstable and 0 for stable. The LBO indicator stabilized to the value of 1 approximately 30 ms before complete blowout.

<div>A growing global population and improved standard of living in developing countries have resulted in an unprecedented increase in energy demand over the past several decades. While renewable energy sources are increasing, a huge portion of energy is still converted into useful work using heat engines. The combustion process in diesel and petrol engines releases carbon dioxide and other greenhouse gases as an unwanted side-effect of the energy conversion process. By improving the efficiency of internal combustion engines, more chemical energy stored in petroleum resources can be realized as useful work and, therefore, reduce global emissions of greenhouse gases. This research focused on improving the thermal efficiency of opposed-piston engines, which, unlike traditional reciprocating engines, do not use a cylinder head. The cylinder head is a major source of heat loss in reciprocating engines. Therefore, the opposed-piston engine has the potential to improve overall engine efficiency relative to inline or V-configuration engines.</div><div><br></div>The objective of this research project was to further improve the design of opposed-piston engines by using computational fluid dynamics (CFD) modeling to optimize the engine geometry. The CFD method investigated the effect of intake port geometry and exhaust piston lead angle on the scavenging process and in-cylinder turbulence. After the CFD data was analyzed, scavenging efficiency was found insensitive to transfer port geometry and exhaust piston lead angle with a maximum change of 0.61%. Trapping efficiency was altered exclusively by exhaust piston lead angle and changed from 18% to 26% as the lead angle was increased. The in-cylinder turbulence parameters of the engine (normalized swirl circulation, normalized tumble circulation, and normalized TKE) experienced more complex relationships. All turbulence parameters were sensitive to changing transfer port geometry and exhaust piston lead angle. Some examples of trends seen during the analysis include: an increase in normalized swirl circulation from 0.01 to 4.45 due to changes in swirl angle, a change in normalized tumble circulation from -28.52 to 21.11 as swirl angle increased, and an increase in normalized tumble circulation from 14.20 to 33.68 as exhaust piston lead angle was increased. Based on the present work, an optimum configuration was identified for a swirl angle of 15°, a tilt angle of 10°, and an exhaust piston lead angle of 20°. Future work includes expanding the numerical model’s domain to support a complete cylinder-port configuration, adding combustion products to the diffusivity equation in the UDF, and running additional test cases to describe the entire input space for the sensitivity analysis.<br>