Phase-field (PF) models are one of the most powerful tools to simulate microstructural evolution in metallic materials, polymers, and ceramics. However, existing PF approaches rely on rigorous mathematical model development, sophisticated numerical schemes, and high-performance computing for accuracy. Although recently developed surrogate microstructure models employ deep-learning techniques and reconstruction of microstructures from lower-dimensional data, their accuracy is fairly limited as spatio-temporal information is lost in the pursuit of dimensional reduction. Given these limitations, a novel data-driven emulator (DDE) for extrapolation prediction of microstructural evolution is presented, which combines an image-based convolutional and recurrent neural network (CRNN) with tensor decomposition, while leveraging previously obtained PF datasets for training. To assess the robustness of DDE, the emulation sequence and the scaling behavior with phase-field simulations for several noisy initial states are compared. In conclusion, the effectiveness of the microstructure emulation technique is explored in the context of accelerating runtime, along with an emphasis on its trade-off with accuracy.Meanwhile, an interpolation DDE has also been tested, which is based on obtaining a low-dimensional representation of the microstructures via tensor decomposition and subsequently predicting the microstructure evolution in the low-dimensional space using Gaussian process regression (GPR). Once the microstructure predictions are obtained in the low-dimensional space, a hybrid input-output phase retrieval algorithm will be employed to reconstruct the microstructures. As proof of concept, the results on microstructure prediction for spinodal decomposition are presented, although the method itself is agnostic of the material parameters. Results show that GPR-based DDE model are able to predict microstructure evolution sequences that closely resemble the true microstructures (average normalized mean square of 6.78 × 10−7) at time scales half of that employed in obtaining training data. This data-driven microstructure emulator opens new avenues to predict the microstructural evolution by leveraging phase-field simulations and physical experimentation where the time resolution is often quite large due to limited resources and physical constraints, such as the phase coarsening experiments previously performed in microgravity. Future work will also be discussed and demonstrate the intended utilization of these two approaches for 3D microstructure prediction through their combined application.

Existing machine learning and data mining techniques have difficulty in handling three characteristics of real-world data sets altogether in a computationally efficient way: (1) different data types with both categorical data and numeric data, (2) different variable relations in different value ranges of variables, and (3) unknown variable dependency.This dissertation developed a Partial-Value Association Discovery (PVAD) algorithm to overcome the above drawbacks in existing techniques. It also enables the discovery of partial-value and full-value variable associations showing both effects of individual variables and interactive effects of multiple variables. The algorithm is compared with Association rule mining and Decision Tree for validation purposes. The results show that the PVAD algorithm can overcome the shortcomings of existing methods. The second part of this dissertation focuses on knee point detection on noisy data. This extended research topic was inspired during the investigation into categorization for numeric data, which corresponds to Step 1 of the PVAD algorithm. A new mathematical definition of knee point on discrete data is introduced. Due to the unavailability of ground truth data or benchmark data sets, functions used to generate synthetic data are carefully selected and defined. These functions are subsequently employed to create the data sets for this experiment. These synthetic data sets are useful for systematically evaluating and comparing the performance of existing methods. Additionally, a deep-learning model is devised for this problem. Experiments show that the proposed model surpasses existing methods in all synthetic data sets, regardless of whether the samples have single or multiple knee points. The third section presents the application results of the PVAD algorithm to real-world data sets in various domains. These include energy consumption data of an Arizona State University (ASU) building, Computer Network, and ASU Engineering Freshmen Retention. The PVAD algorithm is utilized to create an associative network for energy consumption modeling, analyze univariate and multivariate measures of network flow variables, and identify common and uncommon characteristics related to engineering student retention after their first year at the university. The findings indicate that the PVAD algorithm offers the advantage and capability to uncover variable relationships.

Environmental problems are more abundant because of the rapid increase in urbanization, climate change, and population growth leading to the depletion of natural resources and endangerment of some species. The availability of infrastructure as well as socio-economic factors facilitate the illicit trade of wildlife through supply chain networks, adding further threats to species. Ecosystem conservation and protection of wildlife from illegal trade and poaching is fundamental to guarantee the survival of endangered species. Conservation efforts require a landscape approach that incorporates spatial features for the effective functionality of the selected reserve. This dissertation studies combinatorial optimization problems with application to two classes of societal problems: landscape conservation and disruption of illicit supply chains. The first and second chapter propose a mixed-integer formulation to model the reserve design problem with budget and ecological constraints. The first uses the radius of the smallest circle enclosing the selected areas as a metric of compactness. An extension of the model is proposed to solve the multi reserve design problem and the reserve expansion problem. The solution approach includes warm start heuristic, separation problem and cuts to improve model performance. The enhanced model outperforms the linearized and the equivalent nonlinear model. The second chapter uses the Reock’s metric as a metric of compactness. The solution approach includes warm start heuristic, knapsack based separation problem to inject solutions, and cuts to improve model performance. The enhanced model outperforms the default model. The third chapter proposes an integer programming model to solve the wildlife corridor design problem with minimum width requirement and a budget constraint. A separation algorithm is proposed to identify boundary patches and violations in the corridor width. A branch-and-cut approach is proposed to induce the corridor width and is tested on real-life landscape. The fourth chapter proposes an integer programming formulation to model the disruption of illicit supply chain problem. The proposed model enforces that at least x paths must be disrupted for an Origin-Destination pair to be disrupted and at least y arcs must be disrupted for a path to be disrupted. The proposed model is tested on real-life road networks.

Reliability growth is not a new topic in either engineering or statistics and has been a major focus for the past few decades. The increasing level of high-tech complex systems and interconnected components and systems implies that reliability problems will continue to exist and may require more complex solutions. The most heavily used experimental designs in assessing and predicting a systems reliability are the "classical designs", such as full factorial designs, fractional factorial designs, and Latin square designs. They are so heavily used because they are optimal in their own right and have served superbly well in providing efficient insight into the underlying structure of industrial processes. However, cases do arise when the classical designs do not cover a particular practical situation. Repairable systems are such a case in that they usually have limitations on the maximum number of runs or too many varying levels for factors. This research explores the D-optimal design criteria as it applies to the Poisson Regression model on repairable systems, with a number of independent variables and under varying assumptions, to include the total time tested at a specific design point with fixed parameters, the use of a Bayesian approach with unknown parameters, and how the design region affects the optimal design. In applying experimental design to these complex repairable systems, one may discover interactions between stressors and provide better failure data. Our novel approach of accounting for time and the design space in the early stages of testing of repairable systems should, theoretically, in the final engineering design improve the system's reliability, maintainability and availability.

Assembly lines are low-cost production systems that manufacture similar finished units in large quantities. Manufacturers utilize mixed-model assembly lines to produce customized items that are not identical but share some general features in response to consumer needs. To maintain efficiency, the aim is to find the best feasible option to balance the lines efficiently; allocating each task to a workstation to satisfy all restrictions and fulfill all operational requirements in such a way that the line has the highest performance and maximum throughput. The work to be done at each workstation and line depends on the precise product configuration and is not constant across all models. This research seeks to enhance the subject of assembly line balancing by establishing a model for creating the most efficient assembly system. Several realistic characteristics are included into efficient optimization techniques and mathematical models to provide a more comprehensive model for building assembly systems. This involves analyzing the learning growth by task, employing parallel line designs, and configuring mixed models structure under particular constraints and criteria. This dissertation covers a gap in the literature by utilizing some exact and approximation modeling approaches. These methods are based on mathematical programming techniques, including integer and mixed integer models and heuristics. In this dissertation, heuristic approximations are employed to address problem-solving challenges caused by the problem's combinatorial complexity. This study proposes a model that considers learning curve effects and dynamic demand. This is exemplified in instances of a new assembly line, new employees, introducing new products or simply implementing engineering change orders. To achieve a cost-based optimal solution, an integer mathematical formulation is proposed to minimize the production line's total cost under the impact of learning and demand fulfillment. The research further creates approaches to obtain a comprehensive model in the case of single and mixed models for parallel lines systems. Optimization models and heuristics are developed under various aspects, such as cycle times by line and tooling considerations. Numerous extensions are explored effectively to analyze the cost impact under certain constraints and implications. The implementation results demonstrate that the proposed models and heuristics provide valuable insights.

Over the past few decades, medical imaging is becoming important in medicine for disease diagnosis, prognosis, treatment assessment and health monitoring. As medical imaging has progressed, imaging biomarkers are being rapidly developed for early diagnosis and staging of disease. Detecting and segmenting objects from images are often the first steps in quantitative measurement of these biomarkers. While large objects can often be automatically or semi-automatically delineated, segmenting small objects (blobs) is challenging. The small object of particular interest in this dissertation are glomeruli from kidney magnetic resonance (MR) images. This problem has its unique challenges. First of all, the size of glomeruli is extremely small and very similar with noises from images. Second, there are massive of glomeruli in kidney, e.g. over 1 million glomeruli in human kidney, and the intensity distribution is heterogenous. A third recognized issue is that a large portion of glomeruli are overlapping and touched in images. The goal of this dissertation is to develop computational algorithms to identify and discover glomeruli related imaging biomarkers. The first phase is to develop a U-net joint with Hessian based Difference of Gaussians (UH-DoG) blob detector. Joining effort from deep learning alleviates the over-detection issue from Hessian analysis. Next, as extension of UH-DoG, a small blob detector using Bi-Threshold Constrained Adaptive Scales (BTCAS) is proposed. Deep learning is treated as prior of Difference of Gaussian (DoG) to improve its efficiency. By adopting BTCAS, under-segmentation issue of deep learning is addressed. The second phase is to develop a denoising convexity-consistent Blob Generative Adversarial Network (BlobGAN). BlobGAN could achieve high denoising performance and selectively denoise the image without affecting the blobs. These detectors are validated on datasets of 2D fluorescent images, 3D synthetic images, 3D MR (18 mice, 3 humans) images and proved to be outperforming the competing detectors. In the last phase, a Fréchet Descriptors Distance based Coreset approach (FDD-Coreset) is proposed for accelerating BlobGAN’s training. Experiments have shown that BlobGAN trained on FDD-Coreset not only significantly reduces the training time, but also achieves higher denoising performance and maintains approximate performance of blob identification compared with training on entire dataset.

Sequential event prediction or sequential pattern mining is a well-studied topic in the literature. There are a lot of real-world scenarios where the data is released sequentially. People believe that there exist repetitive patterns of event sequences so that the future events can be predicted. For example, many companies build their recommender system to predict the next possible product for the users according to their purchase history. The healthcare system discovers the relationships among patients’ sequential symptoms to mitigate the adverse effect of a treatment (drugs or surgery). Modern engineering systems like aviation/distributed computing/energy systems diagnosed failure event logs and took prompt actions to avoid disaster when a similar failure pattern occurs. In this dissertation, I specifically focus on building a scalable algorithm for event prediction and extraction in the aviation domain. Understanding the accident event is always the major concern of the safety issue in the aviation system. A flight accident is often caused by a sequence of failure events. Accurate modeling of the failure event sequence and how it leads to the final accident is important for aviation safety. This work aims to study the relationship of the failure event sequence and evaluate the risk of the final accident according to these failure events. There are three major challenges I am trying to deal with. (1) Modeling Sequential Events with Hierarchical Structure: I aim to improve the prediction accuracy by taking advantage of the multi-level or hierarchical representation of these rare events. Specifically, I proposed to build a sequential Encoder-Decoder framework with a hierarchical embedding representation of the events. (2) Lack of high-quality and consistent event log data: In order to acquire more accurate event data from aviation accident reports, I convert the problem into a multi-label classification. An attention-based Bidirectional Encoder Representations from Transformers model is developed to achieve good performance and interpretability. (3) Ontology-based event extraction: In order to extract detailed events, I proposed to solve the problem as a hierarchical classification task. I improve the model performance by incorporating event ontology. By solving these three challenges, I provide a framework to extract events from narrative reports and estimate the risk level of aviation accidents through event sequence modeling.

Additive manufacturing consists of successive fabrication of materials layer upon layer to manufacture three-dimensional items. Several key problems such as poor quality of finished products and excessive operational costs are yet to be addressed before it becomes widely applicable in the industry. Retroactive/offline actions such as post-manufacturing inspections for defect detection in finished products are not only extremely expensive and ineffective but are also incapable of issuing corrective action signals during the building span. In-situ monitoring and optimal control methods, on the other hand, can provide viable alternatives to aid with the online detection of anomalies and control the process. Nevertheless, the complexity of process assumptions, unique structure of collected data, and high-frequency data acquisition rate severely deteriorates the performance of traditional and parametric control and process monitoring approaches. Out of diverse categories of additive manufacturing, Large-Scale Additive Manufacturing (LSAM) by material extrusion and Laser Powder Bed Fusion (LPBF) suffer the most due to their more advanced technologies and are therefore the subjects of study in this work. In LSAM, the geometry of large parts can impact the heat dissipation and lead to large thermal gradients between distance locations on the surface. The surface's temperature profile is captured by an infrared thermal camera and translated to a non-linear regression model to formulate the surface cooling dynamics. The surface temperature prediction methodology is then combined into an optimization model with probabilistic constraints for real-time layer time and material flow control. On-axis optical high-speed cameras can capture streams of melt pool images of laser-powder interaction in real-time during the process. Model-agnostic deep learning methods offer a great deal of flexibility when facing such unstructured big data and thus are appealing alternatives to their physical-related and regression-based modeling counterparts. A configuration of Convolutional Long-Short Term Memory (ConvLSTM) auto-encoder is proposed to learn a deep spatio-temporal representation from sequences of melt pool images collected from experimental builds. The unfolded bottleneck tensors are then further mined to construct a high accuracy and low false alarm rate anomaly detection and monitoring procedure.