Dissertation Defense Announcements

While continuous improvement initiatives such as benchmarking have a history of utilization for general business objectives, their successful utilization in the built environment industries, such as construction and facilities management is not nearly as well documented or researched. This project identifies how the built environment fields are using continual improvement initiatives, evaluates how effectively these initiatives are being utilized, and identifies critical success factors for improving and leveraging these techniques to achieve the sustained continuous improvement initiatives that will be necessary to meet long -term sustainability goals in relation to the operations of the built environment. This project takes place in three parts; a case study of a novel way to benchmark and identify areas for improvement, a large-scale survey of how facility managers are using benchmarking and their involvement in benchmarking networks, and an analysis of the relationship of organizational learning culture and the role that it plays in facilitating and supporting benchmarking initiatives. This research provides the first-of its-kind survey and assessment of how practitioners in the built environment are utilizing benchmarking. The results of this project serve to assist facility practitioners in developing, leveraging, and strengthening their continuous improvement initiatives to sustain ongoing change critical for the success of long-term organizational goals related to the built environment lifecycle.

The prediction of RNA secondary structure is complex. The biomolecule can adopt or sample numerous stable conformations, can change structure in response to a stimulus such as a binding event, and does not simply obey thermodynamically favorable folding rules. Due to this, estimation of structure based on the primary sequence is unreliable and misleading. Probing RNA structure dramatically improves computational prediction. The examination of both, ex vivo and in cell RNA can provide important information regarding structural stability, the RNA interactome, and refolding effects.
Current structural probes for RNA selective 2′-hydroxy acylation analyzed by primer extension (SHAPE) rely on a reaction with the 2′-OH on the ribose sugar of residues that are not base paired. These flexible residues can then be determined using gel electrophoresis or quantified using next generation sequencing (NGS) and mutational profiling (MaP) to prepare a library of probing data. The advent of SHAPE technologies led to a rapid increase in the accessibility of RNA structural data. Several successful SHAPE probes have been previously demonstrated, but arguments regarding reactivity and cell permeability remain. In this work, the design and application of novel, variably reactive SHAPE probes is shown ex vivo and a novel in cell probe is demonstrated.

In this thesis, a combined approach based on the Finite Element (FE)  and Smoothed Particle Hydrodynamics (SPH) methods is proposed to model turning operations.  The approach exploits the advantages of each method and leads to high-fidelity coupled FE-SPH machining models that are significantly more numerically efficient and are on par with the models based on each of the two methods alone. Both two-dimensional and three-dimensional models are developed and validated by comparing predicted forces and chip morphologies with experimental results. Parametric studies are carried out to fine-tune the model-based parameters in order to avoid numerical stability issues. The three-dimensional models are extended to include modulated tool path (MTP) machining which is a technique for breaking chips during machining by modulating the motion of the tool. The MTP model predictions are shown to agree with the results from an existing analytical model.  With this model, various tool paths can be simulated to choose an optimal path that decreases tool-wear without sacrificing productivity. Preliminary results from a three-dimensional turning model incorporating machining dynamics through a spring-damper system are also presented. This model has the potential to be used for studying machining stability for a given set of machining conditions.

In addition to the above, another significant contribution of this thesis is the determination of Johnson-Cook material model parameters for a given material using an inverse method and experimental values of cutting forces and workpiece temperatures. The methodology described in the present work identifies the non-uniqueness of the solution to the inverse problem and proposes an approach that eliminates the non-uniqueness.

The main goal of learning analytics and early detection systems is to extract knowledge from student data to understand students' trends of activities towards success and risk and, therefore, design intervention methods to improve learning performance and experience. However, many factors contribute to the challenge of designing and building effective learning analytics systems. Because of the complexity of heterogeneous student data, models designed to analyze it frequently neglect temporal correlations in the interest of convenience. Moreover, the performance descriptions gained from the student data model or prediction results from the analytical models do not always help explain the "why" and "how" behind it. Furthermore, domain specialists are unable to participate in the knowledge discovery process since it necessitates significant data science abilities, and an analytical model is a black box to them.

This research aims to develop analytical models that enable domain experts to study their students' performance behavior and explore trustworthy sources of information with the help of explanations on the analytics. Our work demonstrates various approaches to using the temporal aspect of heterogeneous student data to build analytical models: weighted network analysis, unsupervised cluster analysis, and recurrent neural network analytics. The description, implementation process, and findings of each method are presented as technical contributions to the temporal analysis of student data. We experiment with all these analytical models that highlight the complexity of heterogeneous-temporal data, model building, decision-making tasks, and the need for a more in-depth focus on visual information of analytics with state-of-art explainable AI tools and techniques.

Our work underscores a need for developing a robust way to integrate the possibilities inherent within each approach. To achieve this goal, we present a comprehensive yet flexible and empirical framework to support the design and development of analytical models to extract meaningful insights about students' academic performance and identify early actionable interventions to improve the learning experience. We illustrate our framework on three applications (e.g., student network model, unsupervised clustering model, and recurrent neural network analytics) to demonstrate the value of this framework in addressing the challenges of using student data for learning analytics. These applications present vast opportunities to benefit students' learning experience by implementing flexible educational data representations, fitting different predictive models, and extracting insights for designing prescriptive analytics and building strategies to overcome perceived limitations.

An academic institution's culture drives its ability to accept, leverage, and deploy predictive and prescriptive analytics to enhance the workflow of maximizing pedagogical outcomes. We believe that our work will aid in the future development or refinement of a set of design standards for learning analytics systems.