Progressive miniaturization in electronics demands advanced materials with excellent energy conversion and transport properties. Opportunities exist in novel material morphologies such as hierarchical structures, multi-functional composites and nanoscale architectures which may offer mechanical, thermal and electronic properties tailored to a wide range of applications (e.g., aerospace, robotics, biomedical etc.). However, the manufacturing capabilities have always posed a grand challenge in realizing the advanced material morphologies. Furthermore, the multi-scale modeling of complex material architectures has been extremely challenging owing to the limitations in computation methodologies and lack of understanding in nano-/micro-meter scale physics. To address these challenges, this work considers the morphology effect on carbon nanotube (CNT)-based composites, CNT fibers and thermoelectric (TE) materials. First, this work reports additively manufacturable TE morphologies and analyzes the thermo-electric transport behavior. This research introduces innovative honeycomb TE architectures that showed ~26% efficiency increase and ~25% density reduction compared to conventional rectangular TE architectures. Moreover, this work presents 3D printable compositionally segmented TE architecture which provides record-high efficiencies (up to 8.7%) over wide temperature ranges if the composition and aspect ratio of multiple TE materials are optimized within a single TE device. Next, this research proposes computationally efficient two-dimensional (2D) finite element model (FEM) to study the electrical and thermal properties in CNT based composites by simultaneously considering the stochastic CNT distributions, CNT fractions (upto 80%) and interfacial resistances. The FEM allows to estimate the theoretical maximum possible conductivities with corresponding interfacial resistances if the CNT morphologies are carefully controlled, along with appreciable insight into the energy transport physics. Then, this work proposes a data-driven surrogate model based on convolutional neural networks to rapidly approximate the composite conductivities in a second with accuracy > 98%, compared to FEM taking >100 minutes per simulation. Finally, this research presents a pseudo 2D FEM to approximate the electrical and thermal properties in CNT fibers at various CNT aspect ratios (up to 10,000) by simultaneously considering CNT-CNT interfacial effects along with the stochastic distribution of inter-bundle voids.

High-entropy alloys (HEAs) is a new class of materials which have been studied heavily due to their special mechanical properties. HEAs refers to alloys with multiple equimolar or nearly equimolar elements. HEAs show exceptional and attractive properties currently absent from conventional alloys, which make them the center of intense investigation. HEAs obtain their properties from four core effects that they exhibit and most of the work on them have been dedicated to study their mechanical properties. In contrast, little or no research have gone into studying the functional or even thermal properties of HEAs. Some HEAs have also shown exceptional or very high melting points. According to the definition of HEAs, Si-Ge-Sn alloys with equal or comparable concentrations of the three group IV elements belong to the category of HEAs. Thus, the equimolar components of Si-Ge-Sn alloys probably allow their atomic structures to display the same fundamental effects of metallic HEAs. The experimental fabrication of such alloys has been proven to be very difficult, which is mainly due to differences between the properties of their constituent elements, as indicated from their binary phase diagrams. However, previous computational studies have shown that SiGeSn HEAs have some very interesting properties, such as high electrical conductivity, low thermal conductivity and semiconducting properties. In this work, going for a complete characterization of the SiGeSn HEA properties, the melting point of this alloy is studied using classical molecular dynamics (MD) simulations and density functional theory (DFT) calculations. The aim is to investigate the effects of high Sn content in this alloy on the melting point compared with the traditional SiGe alloys. Classical MD simulations results strongly indicates that none of the available empirical potentials is able to predict accurate or reasonable melting points for SiGeSn HEAs and most of its subsystems. DFT calculations results show that SiGeSn HEA have a melting point which represent the mean value of its constituent elements and that no special deviations are found. This work contributes to the study of SiGeSn HEA properties, which can serve as guidance before the successful experimental fabrication of this alloy.

Magnetic liquids called ferrofluids have been used in applications ranging from audio speaker cooling and rotary pressure seals to retinal detachment surgery and implantable artificial glaucoma valves. Recently, ferrofluids have been investigated as a material for use in magnetically controllable liquid droplet robotics. Liquid droplet robotics is an emerging technology that aims to apply control theory to manipulate fluid droplets as robotic agents to perform a wide range of tasks. Furthermore, magnetically controlled micro-robotics is another popular area of study where manipulating a magnetic field allows for the control of magnetized micro-robots. Both of these emerging fields have potential for impact toward medical applications: liquid characteristics such as being able to dissolve various compounds, be injected via a needle, and the potential for the human body to automatically filter and remove a liquid droplet robot, make liquid droplet robots advantageous for medical applications; while the ability to remotely control the torques and forces on an untethered microrobot via modulating the magnetic field and gradient is also highly advantageous. The research described in this dissertation explores applications and methods for the electromagnetic control of ferrofluid droplet robots. First, basic electrical components built from fluidic channels containing ferrofluid are made remotely tunable via the placement of ferrofluid within the channel. Second, a ferrofluid droplet is shown to be fully controllable in position, stretch direction, and stretch length in two dimensions using proportional-integral-derivative (PID) controllers. Third, control of a ferrofluid’s position, stretch direction, and stretch length is extended to three dimensions, and control gains are optimized via a Bayesian optimization process to achieve higher accuracy. Finally, magnetic control of both single and multiple ferrofluid droplets in two dimensions is investigated via a visual model predictive control approach based on machine learning. These achievements take both liquid droplet robotics and magnetic micro-robotics fields several steps closer toward real-world medical applications such as embedded soft electronic health monitors, liquid-droplet-robot-based drug delivery, and automated magnetically actuated surgeries.

Polylactic Acid (PLA), a thermoplastic polymer is well-known for its biocompatibility, making it ideal for the manufacturing of biomedical devices. However, the current applications of PLA are commonly limited by its intrinsic polymer characteristics, such as low modulus, mechanical strength, and thermal conductivity. To enhance these physical properties, a biocompatible nanodiamond enhanced PLA filament has been studied. Thermogravimetric analysis was performed to unveil the composition of nanodiamond in the composite. Four printing parameters: nozzle temperature, layer height, infill pattern and printing speed were considered and the Taguchi L9 orthogonal array was implemented for the design of experiments. Fused deposition modeling (FDM) technique was utilized to 3D print the PLA/Nanodiamond samples by altering the four printing parameters considered and were tested according to the standards for tensile strength, flexural strength, and thermal conductivity. Using the Taguchi optimization approach and analysis of variance (ANOVA), the generated experimental data was used to find the optimum set of printing parameters. Finally, cell studies were performed to demonstrate the biocompatibility of PLA/Nanodiamond. All these results could aid in determining the working ranges for FDM fabrication of PLA/Nanodiamond for biomedical applications.

Autonomous systems inevitably must interact with other surrounding systems; thus, algorithms for intention/behavior estimation are of great interest. This thesis dissertation focuses on developing passive and active model discrimination algorithms (PMD and AMD) with applications to set-valued intention identification and fault detection for uncertain/bounded-error dynamical systems. PMD uses the obtained input-output data to invalidate the models, while AMD designs an auxiliary input to assist the discrimination process. First, PMD algorithms are proposed for noisy switched nonlinear systems constrained by metric/signal temporal logic specifications, including systems with lossy data modeled by (m,k)-firm constraints. Specifically, optimization-based algorithms are introduced for analyzing the detectability/distinguishability of models and for ruling out models that are inconsistent with observations at run time. On the other hand, two AMD approaches are designed for noisy switched nonlinear models and piecewise affine inclusion models, which involve bilevel optimization with integer variables/constraints in the inner/lower level. The first approach solves the inner problem using mixed-integer parametric optimization, whose solution is included when solving the outer problem/higher level, while the second approach moves the integer variables/constraints to the outer problem in a manner that retains feasibility and recasts the problem as a tractable mixed-integer linear programming (MILP). Furthermore, AMD algorithms are proposed for noisy discrete-time affine time-invariant systems constrained by disjunctive and coupled safety constraints. To overcome the issues associated with generalized semi-infinite constraints due to state-dependent input constraints and disjunctive safety constraints, several constraint reformulations are proposed to recast the AMD problems as tractable MILPs. Finally, partition-based AMD approaches are proposed for noisy discrete-time affine time-invariant models with model-independent parameters and output measurement that are revealed at run time. Specifically, algorithms with fixed and adaptive partitions are proposed, where the latter improves on the performance of the former by allowing the partitions to be optimized. By partitioning the operation region, the problem is solved offline, and partition trees are constructed which can be used as a `look-up table' to determine the optimal input depending on revealed information at run time.

With the abundance of increasingly large datasets, the ability to predict the phase of high-entropy alloys (HEAs) based solely on elemental composition could become a reliable tool for the discovery of new HEAs. However, as the amount of data expands so does the computational time and resources required to train predictive classical machine learning models. Quantum computers, which use quantum bits (qubits), could be the solution to overcoming these demands. Their ability to use quantum superposition and interference to perform calculations could be the key to handling large amounts of data. In this work, a hybrid quantum-classical machine learning algorithm is implemented on both quantum simulators and quantum processors to perform the supervised machine learning task. Their feasibility as a future tool for HEA discovery is evaluated based on the algorithm’s performance. An artificial neural network (ANN), run by classical computers, is also trained on the same data for performance comparison. The accuracy of the quantum-classical model was found to be comparable to the accuracy achieved by the classical ANN with a slight decrease in accuracy when ran on quantum hardware due to qubit susceptibility to decoherence. Future developments in the applied quantum machine learning method are discussed.

Disordered many-body systems are ubiquitous in condensed matter physics, materials science and biological systems. Examples include amorphous and glassy states of matter, granular materials, and tissues composed of packings of cells in the extra-cellular matrix (ECM). Understanding the collective emergent properties in these systems is crucial to improving the capability for controlling, engineering and optimizing their behaviors, yet it is extremely challenging due to their complexity and disordered nature. The main theme of the thesis is to address this challenge by characterizing and understanding a variety of disordered many-body systems via unique statistical geometrical and topological tools and the state-of-the-art simulation methods. Two major topics of the thesis are modeling ECM-mediated multicellular dynamics and understanding hyperuniformity in 2D material systems. Collective migration is an important mode of cell movement for several biological processes, and it has been the focus of a large number of studies over the past decades. Hyperuniform (HU) state is a critical state in a many-particle system, an exotic property of condensed matter discovered recently. The main focus of this thesis is to study the mechanisms underlying collective cell migration behaviors by developing theoretical/phenomenological models that capture the features of ECM-mediated mechanical communications in vitro and investigate general conditions that can be imposed on hyperuniformity-preserving and hyperuniformity-generating operations, as well as to understand how various novel transport physical properties arise from the unique hyperuniform long-range correlations.

This research seeks to answer the question if there is a singular relationship between stishovite nucleation and the atomistic structure of the preshocked amorphous SiO$_2$. To do this a stishovite manufacturing method is developed in which 1,152 samples were produced. The majority of these samples did crystallize. The method was produced through two rounds of experiments and fine-tuning with the pressure damp, temperature damp, shock pressure using an NPHug fix, and sample origin. A new random atomic insertion method was used to generate a new and different SiO$_2$ amorphous structure not before seen within the research literature. The optimal values for shock were found to be 60~GPa for randomly atom insertion samples and 55~GPa for quartz origin samples. Temperature damp appeared to have a slight effect optimizing at 0.05~ps and the pressure damp had no visible effect, testing was done with temperature damp from 0.05 to 0.5~ps and pressure damp from 0.1 to 10.0~ps. There appeared to be significant randomness in crystallization behavior. The preshocked and postnucleated samples were transformed into Gaussian fields of crystal, mass, and charge. These fields were divided and classified using a cut-off method taking the number of crystals produced in portions of each simulation and classifying each potion as nucleated or non-nucleated. Data in which some nucleation but not a critical amount was present was removed constituting 2.6\% to 20.3\% of data in all tests. A max method was also used which takes only the maximum portions of each simulation to classify as nucleating. There are three other variables tested within this work, a sample size of 18,000 or 72,728~atoms, Gaussian variance of 1 or 4~\AA, and Convolutional neural network (CNN) architecture of a garden verity or all convolution along with the portioning classification method, sample origination, and Gaussian field type. In total 64 tests were performed to try every combination of variable. No significant classifications were made by the CNNs to nucleation or non-nucleation portions. The results clearly confirmed that the data was not abstracting to atomistic structure and was random by all classifications of the CNNs. The all convolution CNN testing did show smoother outcomes in training with less fluctuations. 59\% of all validation accuracy was held at 0.5 for a random state and 84\% was within $\pm0.02$ of 0.5. It is conclusive that prenucleation structure is not the sole predictor of nucleation behavior. It is not conclusive if prenucleation structure is a partial or non-factor within nucleation of stishovite from amorphous SiO$_2$.

A Compact Linear Fresnel Reflector (CLFR) is a simple, cost-effective, and scalable option for generating solar power by concentrating the sun rays. To make a most feasible application, design parameters of the CLFR, such as solar concentrator design parameters, receiver design parameters, heat transfer, power block parameters, etc., should be optimized to achieve optimum efficiency. Many researchers have carried out modeling and optimization of CLFR with various numerical or analytical methods. However, often computational time and cost are significant in these existing approaches. This research attempts to address this issue by proposing a novel computational approach with the help of increased computational efficiency and machine learning. The approach consists of two parts: the algorithm and the machine learning model. The algorithm has been created to fulfill the requirement of the Monte Carlo Ray tracing method for CLFR collector simulation, which is a simplified version of the conventional ray-tracing method. For various configurations of the CLFR system, optical losses and optical efficiency are calculated by employing these design parameters, such as the number of mirrors, mirror length, mirror width, space between adjacent mirrors, and orientation angle of the CLFR system. Further, to reduce the computational time, a machine learning method is used to predict the optical efficiency for the various configurations of the CLFR system. This entire method is validated using an existing approach (SolTrace) for the optical losses and optical efficiency of a CLFR system. It is observed that the program requires 6.63 CPU-hours of computational time are required by the program to calculate efficiency. In contrast, the novel machine learning approach took only seconds to predict the optical efficiency with great accuracy. Therefore, this method can be used to optimize a CLFR system based on the location and land configuration with reduced computational time. This will be beneficial for CLFR to be a potential candidate for concentrating solar power option.