Tang, Yufei

Short-circuit faults can cause significant damage to power grid infrastructure, resulting in costly maintenance for utility providers. Rapid identification of fault locations can help mitigate these damages and associated expenses. Recent studies have demonstrated that graph neural network (GNN) models, using phasor data from various points in a power grid, can accurately locate fault events by accounting for the grid’s topology—a feature not typically leveraged by other machine learning methods. However, despite their high performance, GNN models are often viewed as ”black-box” systems, making their decision logic difficult to interpret. This thesis demonstrates that explanation methods can be applied to GNN models to enhance their transparency by clarifying the reasoning behind fault location predictions. By systematically benchmarking several explanation techniques for a GNN model trained for fault location detection, we assess and recommend the most effective methods for elucidating fault detection predictions in power grid systems.

Ocean current turbines (OCT) convert the kinetic energy housed within the earth’s ocean currents into electricity. However, due to the harsh environmental conditions that these turbines operate in, their system performance naturally degrades over time. This degradation correlates to high operation and maintenance (O&M) costs, which necessitates the need for robust condition monitoring and fault detection (CMFD). Unfortunately, OCT operational data is not publicly available in large and/or diverse enough quantities to develop such frameworks. Therefore, from an industry-wide perspective, the technologies needed to harvest this energy source are still in their infancy.

Significant reduction in space, weight, power, and cost (SWAP-C) of imaging hardware has induced a paradigm shift in remote sensing where unmanned platforms have become the mainstay. However, mitigating the degraded visual environment (DVE) remains an issue. DVEs can cause a loss of contrast and image detail due to particle scattering and distortion due to turbulence-induced effects. The problem is especially challenging when imaging from unmanned platforms such as autonomous underwater vehicles (AUV) and unmanned ariel vehicles (UAV).
While single-frame image restoration techniques have been studied extensively in recent years, single image capture is not adequate to address the effects of DVEs due to under-sampling, low dynamic range, and chromatic aberration. Significant development has been made to employ multi-frame image fusion techniques to take advantage of spatial and temporal information to aid in the recovery of corrupted image detail and high-frequency content and increasing dynamic range.

Marine and hydrokinetic (MHK) energy systems, including marine current turbines and wave energy converters, could contribute significantly to reducing reliance on fossil fuels and improving energy security while accelerating progress in the blue economy. However, technologies to capture them are nascent in development due to several technical and economic challenges. For example, for capturing ocean flows, the fluid velocity is low but density is high, resulting in early boundary layer separation and high torque. This dissertation addresses critical challenges in modeling, optimization, and control co-design of MHK energy systems, with specific case studies of a variable buoyancy-controlled marine current turbine (MCT). Specifically, this dissertation presents (a) comprehensive dynamic modeling of the MCT, where data recorded by an acoustic Doppler current profiler will be used as the real ocean environment; (b) vertical path planning of the MCT, where the problem is formulated as a novel spatial-temporal optimization problem to maximize the total harvested power of the system in an uncertain oceanic environment; (c) control co-design of the MCT, where the physical device geometry and turbine path control are optimized simultaneously. In a nutshell, the contributions are summarized as follows:

In the past few years, the development of complex dynamical networks or systems has stimulated great interest in the study of the principles and mechanisms underlying the Internet of things (IoT). IoT is envisioned as an intelligent network infrastructure with a vast number of ubiquitous smart devices present in diverse application domains and have already improved many aspects of daily life. Many overtly futuristic IoT applications acquire data gathered via distributed sensors that can be uniquely identified, localized, and communicated with, i.e., the support of sensor networks. Soft-sensing models are in demand to support IoT applications to achieve the maximal exploitation of transforming the information of measurements into more useful knowledge, which plays essential roles in condition monitoring, quality prediction, smooth control, and many other essential aspects of complex dynamical systems. This in turn calls for innovative soft-sensing models that account for scalability, heterogeneity, adaptivity, and robustness to unpredictable uncertainties. The advent of big data, the advantages of ever-evolving deep learning (DL) techniques (where models use multiple layers to extract multi-levels of feature representations progressively), as well as ever-increasing processing power in hardware, has triggered a proliferation of research that applies DL to soft-sensing models. However, many critical questions need to be further investigated in the deep learning-based soft-sensing.

Network embedding or representation learning is important for analyzing many real-world applications and systems, i.e., social networks, citation networks and communication networks. It targets at learning low-dimensional vector representations of nodes with preserved graph structure (e.g., link relations) and content (e.g., texts) information. The derived node representations can be directly applied in many downstream applications, including node classification, clustering and visualization.
In addition to the complex network structures, nodes may have rich non structure information such as labels and contents. Therefore, structure, label and content constitute different aspects of the entire network system that reflect node similarities from multiple complementary facets. This thesis focuses on multifaceted network embedding learning, which aims to efficiently incorporate distinct aspects of information such as node labels and node contents for cooperative low-dimensional representation learning together with node topology.