46033 research outputs found
Sort by
Autonomous Corrosion Detection In Steel Structures Using Different Non-Contact Techniques
Get PDFMetallic corrosion is an electrochemical process that can occur due to the formation of aqueous adlayers on the metal surface. It is often described as reverse metallurgy, as corrosion drives refined metals back to their lowest energy state, namely an oxidized, ore-like form. Environmental factors that promote corrosion include precipitation, high humidity, and chemical condensation, which can result from the hygroscopic nature of pollutants deposited on the surface. There is a significant impact of metallic corrosion on the U. S. economy including infrastructure, transportation, utilities, production, and manufacturing. The total direct cost of corrosion in the United States has been estimated at $279 billion annually, representing approximately 3.2% of the nation\u27s Gross Domestic Product (GDP). However, timely monitoring can reduce these costs by 15–35%. Applying protective coatings is typically the first line of defense against corrosion, as these coatings act as inhibitors by forming a barrier between the steel surface and the surrounding environment. Current non-destructive inspection methods require the involvement of human inspectors. One of the significant limitations of this approach is the potential inaccessibility of defect-prone areas, which may result in undetected corrosion. Developing an autonomous methodology for corrosion detection became essential to mitigate this issue. This study focused on developing different non-contact and autonomous corrosion identification methods. In this study, a comprehensive cycle of developing a noncontact method has been demonstrated and divided into several steps. At first, a conventional image processing method was used where corrosion features were leveraged in YCbCr color space including other preprocessing steps such as contrast adjustment, histogram equalization, and adaptive histogram equalization. The model segmented 70% of corroded pixels correctly after improving the brightness to an optimum level. However, this conventional image processing method needs user input; hence, it has limited applications, e.g. one set of data from one type of structure. To elevate this method towards an autonomous corrosion detection methodology a deep learning (U-shaped encoder-decoder) network has been trained and tested to detect the corroded pixel from the steel structure’s images. Four types of models such as original UNet and changing the backbone with DenseNet121, EfficientNetB7, and ResNet34 were tested to classify the corroded pixels. Out of all these U-shaped models, the network with ResNet34 as the backbone outperformed the other models by predicting 93.4% corroded pixels precisely with a 90.77% intersection over union (IOU) value. In addition to this, an image classification model, AlexNet was trained and tested in real-time with the help of a customized payload integrated with an Uncrewed Aerial System (UAS). A human-machine interface was introduced to take the inspector’s input in sequential training of this model allowing the model to learn from human expertise. The model was then retrained on the inspector’s output and used in the next inspection. Before retraining 84.78% of images with corrosion were correctly predicted by the model. The result showed that the adapted deep learning model performance improved successfully with more inspection than expected. In particular, the number of reported false calls made by the model has reduced. The accuracy of image classification and semantic segmentation models depends on the expertise in labeling datasets. Moreover, detecting corrosion can be considered the first step in corrosion monitoring. The type of corrosion and corrosion without visual manifestation cannot be recognized with visual sensors. Therefore, the feasibility of hyperspectral imagery (HSI) in detecting early corrosion was investigated. For the bare steel, two types of corrosion tests were performed. The first one is progressive early corrosion without visual manifestation of corrosion, i.e., 2 hours, 4 hours, 6 hours, and 8 hours. The second one is intermittent corrosion, steel specimens have been corroded for different exposure periods such as 6 hours, 16 hours, and 24 hours. For the invisible or early phase of corrosion, the maximum change for the progressive corrosion of 8 hours found in the Near Infrared (NIR) range is 36.82% but for the Visual Near Infrared (VNIR) range, it is 89.09%. Similarly, for the visible corrosion at 24 hours, a 95.4 % change in reflectance was reported in the VNIR range, which is 54.54 % for NIR. After collecting the hyperspectral signature, these samples were coated with primer and topcoat. The performance of the hyperspectral sensor was validated with a Benford model, an empirical model to derive the reflectance value of multilayer coated material. A maximum deviation of 14% from the Benford model calculation was observed as the thickness increased to the range of 220–240 µm. The feasibility of the hyperspectral sensor to identify the corrosion underneath the coating for different exposure times was also investigated. The reflectance value decreased as the coating thickness increased for all samples. The spectral signatures for sound and corroded samples with the same coating thickness were not the same, which showed the feasibility of implementing a hyperspectral sensor in this regard. The first derivative of the spectrum was calculated to locate the maximum change, which was found in the 1350-1400 nm range. Notably, an absorption band for the corroded sample was identified in the 1340-1440nm range. However, this study has certain limitations. The dataset lacked diversity and balance for image-based, non-contact methodologies—an aspect that should be addressed in future research. Additionally, the performance of hyperspectral imaging (HSI) should be further validated using a larger set of corroded samples generated under controlled corrosion chamber conditions
An Explainable AI Framework For Detecting Harmful Algal Blooms (HABs) In Freshwater Lakes
No full textThe increase in economic impacts due to the rapid growth of harmful algal blooms (HABs) in freshwater lakes has prompted the development of monitoring and prediction systems. However, limited target data hinders reliable predictions on HAB formations. To address this limitation, this dissertation focuses on providing solutions through data-driven,interpretable, and explainable machine learning models. In Chapter 3, an interpretable multivariate regression model such as Vector Autoregressive (VAR) was used to analyze the important factors that contribute to the formation of cyanobacteria in lakes. The causality tests performed on multiple water quality data helped to determine Alkalinity, Chlorophyll-a (Chl-a), andWater Temperature as the influential factors. The use of remote sensing satellite Sentinel-2 data for detection of one of the influential factors, Chl-a, using segmentation models such as Otsu and Random Forest was applied in chapter 4. Since microcystins are the most common cyanobacterial toxins found in freshwater bodies, further study in chapters 5 and 6 focused on applying multiple supervised machine learning classification models on microcystin toxicity levels in lakes. The proposal of a threshold detection framework using clustering and explainable AI was used in Chapter 7 to detect HAB hotspots in lakes. A new composite spectral index called AANI was proposed in Chapter 8 that proved to be effective for HAB hotspot detection from satellites. The dissertation also contribute the design of a real-time monitoring dashboard that can be used for HAB hotspot detection. Compared to the existing method of hotspot analysis, which is single-image-based analysis, the current approach provides a robust, explainable, and data-driven solution that can be applied to environmental prediction models
AI-BMS: AI Battery Management System for Safe UAS Operations in High Electric and Magnetic Fields and Uncertain Conditions: Wind, Battery Drain, Temperature, and Payload
Get PDFUnmanned aerial systems/vehicles (UAS/UAVs) are increasingly utilized for inspecting high-voltage (HV) transmission (Tx) lines. However, operating in the vicinity of electric (E) and magnetic (H) fields can affect UAV control and battery performance. Pilots unfamiliar with high E/H field environments often struggle to accurately estimate remaining flight time, resulting in either premature mission terminations or crashes due to depleted batteries. To address these challenges, this thesis focuses on developing an intelligent Battery Management System (BMS) for UAVs to ensure safe and reliable surveillance operations under various uncertain conditions, such as wind gusts, temperature variations, E/H fields, battery drain, and payload variability. A major limitation in this field has been the lack of ground-truth aerial E/H field data from HV Tx lines, which is critical for understanding field distributions and their impact on UAS electronics. To overcome this, a first-of-its-kind large-scale study was conducted, gathering real-time E/H field data across five different transmission lines (69 kVAC, 230 kVAC, 345 kVAC, 500 kVAC, and 250 kVDC), as well as a microwave tower. The data revealed that AC transmission lines exhibited significantly higher E/H field levels compared to DC lines. The study also investigated the influence of high E/H fields on UAV battery performance by measuring power drain under varying field intensities, operational scenarios, and environmental conditions. Analysis of the collected data uncovered critical correlations between battery behavior and proximity to high-voltage infrastructure, offering new insights into the effects of these environments on UAV operations. Building on these findings, a data-driven battery drain forecasting model was developed to enhance flight safety and operational efficiency. Various Hybrid Machine Learning (HML) models were compared, where the hybrid Random Forest-K-Nearest Neighbors (RF-KNN) model demonstrating the best performance by achieving the lowest Mean Absolute Percentage Error (MAPE) among all models tested. The findings of this research make significant contributions to UAV energy management, HV infrastructure inspection, and intelligent flight operation planning. The work not only enables safer and more reliable UAV applications in challenging environments but also aids in implementing FAA rule-making regarding safe operational proximity of UAVs to transmission lines
A Federated Learning Solution For Secure And Scalable Edge Computing In Distributed Environments
No full textThe growing penetration of Distributed Energy Resources (DERs) into the modern power system has introduced significant challenges related to data volume and cybersecurity. Centralized approaches struggle to process the large volume of data generated by inverters or other DER-related assets securely and efficiently. This problem raise concerns about the confidentiality, integrity, availability, and accountability (CIAA) properties of DER assets. Moreover, the large deployment of Internet of Things (IoT) and edge computing devices has introduced new security vulnerabilities. Malicious actors can exploit resource-constrained edge devices for unauthorized activities and resource consumption, threatening the performance and operational capabilities of the edge infrastructure.
This research explores a scalable and privacy-preserving solution to these challenges using Federated Learning (FL) within an edge computing environment. For experimentation, an AI-ready edge testbed (AI-TB) was developed at the Center for Cybersecurity Research (C2SR), University of North Dakota (UND). The testbed comprises Nvidia Jetson Nano devices (acted as federated clients) and Jetson AGX devices (acted as federated server), representing real-world scenarios and supporting localized training without the need of transferring raw data.
This thesis covers two use cases. In the first, DER inverter frequency (Hz) data was utilized for anomaly detection using a Long Short-Term Memory (LSTM) Autoencoders trained with Flower FL framework. The frequency data was injected with False Data Injection Attacks (FDIAs) using gaussian, pulse, and sigmoid functions and used to evaluate the model performance. The FL model achieved a Mean Absolute Percentage Error (MAPE) of 0.022, slightly outperforming the Non Federated Learning (NFL) approach with a MAPE of 0.023. In the second use case, edge device security was investigated by monitoring system-level metrics such as CPU, GPU, and memory usage. LSTM and Bidirectional LSTM (BiLSTM) models were trained using two open source FL frameworks including Flower and OpenFL, to detect anomalous behavior during GPU stress test scenarios. The BiLSTM model combined with OpenFL achieved the highest performance, with F1-scores ranging from 0.96 to 0.99 across three attack patterns.
The findings establish FL as a promising solution for anomaly detection in distributed environments. FL offers privacy preservation and enhanced scalability for collaborative model training across edge nodes. The results highlight the effectiveness of FL in safeguarding both DER systems and edge infrastructure through decentralized and secure anomaly detection
Attributions Of Blame In Cases Of Stealthing
Get PDFNonconsensual condom removal (NCCR), also known as “stealthing”, is a form of sexual abuse where an individual removes the condom without their sexual partner’s knowledge. While research regarding prevalence of NCCR has been conducted over the past decade, there are still aspects of previous research on other forms of sexual abuse that did not account for cases of NCCR. One such gap in the current literature are studies that investigate how individuals attribute blame to the victims of NCCR. Previous research has shown that there are differences in how individuals attribute blame towards victims of sexual abuse based on different characteristics, such as gender and sexual orientation. The goal of the current study was to examine attributions of victim blame for NCCR as a function of the victim’s gender and sexual orientation, as well as participant gender. Participants (n = 413) were recruited via the University of North Dakota’s SONA system to complete an online study for course credit. The participants were randomly assigned to one of four vignettes that detailed a case of NCCR, with the manipulation between the different scenarios being the victim’s gender and sexual orientation. Participants were also asked about their knowledge of NCCR. A between-subjects ANOVA was conducted looking at the impact of participant gender and victim gender and sexual orientation. Results indicated that male participants attributed more blame to victims in the NCCR scenario than did female participants, regardless of victim characteristics. There were no significant differences regarding participants’ familiarity with the term “stealthing” nor if participants experienced a case of “stealthing” themselves. However, female participants were more likely than male participants to know someone who had experienced a case of “stealthing”. The findings contribute to the limited research on NCCR and suggest that victim blame may be an issue, especially among male observers. Implications for reducing blame through education are discussed along with suggested for future research
Studentification And Off-Campus Housing: A Case Study Of Grand Forks Near-Campus Neighborhoods
Get PDFWith national trends in higher education experiencing dwindling enrollments and personal economies increasingly strained, the University of North Dakota (UND) has grown to become the biggest public higher education institution in North Dakota, having effectively bounced back from the COVID-19 pandemic drop in enrollment. Grand Forks, North Dakota, is the host of UND and home to a student body of over 14,000 as of 2023 fall semester. This study focuses on; to what extent is studentification taking place in Grand Forks, ND; and how the observed housing- and demographic patterns in Grand Forks align with, or diverge from, the broader theoretical conceptualization of studentification. First, a spatial analysis, applying spatial autocorrelation methods Global Moran’s I and Local Moran’s I in ArcGIS Pro is conducted. The analysis reveals an increase in clustering across the study period, concentrating in the near-campus neighborhoods. Second, Multiple Linear Regression and Geographically Weighted Regression identify the driving factors to student housing choices based on the effects of studentification identified in the literature, identifying affordability, rental availability, and newer housing stock as deciding factors to student clustering. The results contribute to furthering the conceptualization of studentification and student geographies in nonmetropolitan college towns in the United States
Investigating The Role Of Keratin 6 And SOX2 In Regulating Growth, Stemness, Basal/Squamous Expression, And Chemotherapy Response In Muscle-Invasive Urothelial Carcinoma
Get PDFUrothelial carcinoma (UC) is a prevalent and aggressive malignancy arising in urothelial cells, with arsenic exposure recognized as a significant environmental risk factor. It is categorized into two primary subtypes based on the extent of cancer progression beyond the bladder’s muscle wall: non-muscle-invasive urothelial carcinoma (NMIUC) and muscle-invasive urothelial carcinoma (MIUC). While NMIUC is associated with a relatively favorable prognosis, progression to MIUC is more aggressive and leads to significantly worse outcomes, including poor treatment response and increased metastasis. Squamous differentiation (SD) in UC is an emerging hallmark of aggressive UC, with patients often having poor prognoses and developing chemoresistance, particularly in the MIUC subtype. SD is marked by the expression of basal keratins, such as KRT6. Although KRT6 is recognized as a biomarker for SD, its functional role in driving disease progression in UC remains inadequately understood. Furthermore, SOX2, a transcription factor implicated in cancer stem cell properties and resistance to chemotherapy, has been associated with recurrence, metastasis, and poor prognosis in UC. This study will investigate SOX2’s role in regulating SD and the therapeutic potential of inhibiting SOX2 expression. Despite progress, the interplay between SOX2, cancer stem cell properties, and chemoresistance in UC progression remains inadequately understood. We hypothesize that the knockdown of KRT6 and SOX2 in UC cells will disrupt key signaling pathways involved in tumor growth, differentiation, and chemoresistance. Targeting either KRT6, the end marker of SD, and SOX2, a potential upstream regulator, may enhance the therapeutic response and reduce tumor growth in UC. This study is the first to investigate the functional repercussions of KRT6 and SOX2 knockdown in UC with SD models, marking a substantial milestone and paving the way for a better understanding of their pathological role in UC. The outcomes of this study have the potential to guide future therapeutic interventions targeting KRT6 and SOX2 to enhance our comprehension of their significance in UC research. The outcomes of this study may lead to the development of novel treatment strategies that target KRT6 and SOX2, potentially improving patient outcomes for those with aggressive and chemoresistant forms of UC
Integrating Seismic Attributes And Machine Learning For Accurate 3D Petrophysical And Geomechanical Modeling In Shale Plays
Get PDFThe increasing reliance on shale plays for hydrocarbon production necessitates accurate 3D petrophysical and geomechanical models to optimize resource recovery. However, conventional methods struggle with shale reservoir complexities such as low permeability, anisotropy, and data heterogeneity, leading to uncertainties in property predictions. This dissertation integrates seismic attributes with machine learning (ML) techniques, including Long Short-Term Memory (LSTM) networks, Temporal Convolutional Networks (TCNs), and Wasserstein Generative Adversarial Networks with Gradient Penalty (WGAN-GP), to enhance predictive accuracy.LSTMs and TCNs effectively capture temporal dependencies, while WGAN-GP generates high-quality synthetic data to mitigate data scarcity. The study critically evaluates existing 3D modeling methodologies, identifying key limitations in data integration and resolution enhancement. Essential seismic attributes are selected based on strong correlations with reservoir properties. A robust data processing workflow, including seismic-to-well calibration and Principal Component Analysis (PCA), ensures optimal feature selection without information loss. Advanced ML models predict petrophysical and geomechanical properties, improving shale formation characterization. Shapley values enhance explainability, increasing trust in ML-driven predictions. Results demonstrate that integrating seismic attributes with ML improves model accuracy, reduces uncertainties, and supports better hydraulic fracturing and well placement strategies. This novel framework leveraging ML bridges critical gaps in traditional modeling, leading to reducing operational risks and supporting sustainable shale development in the oil and gas industry
Designing Professional Development In Pedagogy And Technology: A Phenomenological Study Of Faculty Members From Myanmar Universities
Get PDFEducation is a fundamental pillar of national development, with faculty development playing a crucial role in shaping instructional quality and enhancing students’ access to academics. However, for over seven decades in Myanmar, faculty members have frequently lacked access to professional development opportunities related to pedagogy and technology due to institutional constraints such as limited funding, resources, and support. This study examines the lived experiences of faculty members from public universities in Myanmar, focusing on their perceptions of professional development, the challenges they encounter, and the transformative changes they observe after participating in these programs. Utilizing a qualitative phenomenological research design, this study gathers data through semi-structured interviews with fifteen faculty members, department chairs, and senior administrators from public universities in Myanmar. The findings reveal that faculty members consider instructional technology and pedagogical knowledge the most critical aspects of professional development. The results indicate that participation in professional development fosters behavioral changes, enhances motivation, improves confidence, and increases self-efficacy. Furthermore, the findings underscore a significant need for mandatory evaluations to assess the effectiveness of professional development programs. Based on these results, this study proposes a 60-hour faculty development program focused on technology and pedagogy. It also presents pre- and post-surveys, questions, and a student evaluation for instructors. Moreover, this study contributes to the existing literature by applying adult learning and transformative learning theories to faculty development in a resource-constrained environment like Myanmar
Addressing The Impact Of Unlicensed And Under-Licensed Teachers In PK-12 Schools: How School Administrators Offer Support
Get PDFAs the impact of the teacher shortage grows, more teachers are entering classrooms with incomplete teacher licensure. Related literature affirms that new-to-profession teachers experience what Danielson (1999) described as a “sink or swim in the deep end of the pool” culture as they attempt to navigate the beginning of their teaching career with an incomplete license or no license entirely (p. 251). The influx of partially-licensed teachers engaging students is a dilemma created by the growth of the teacher shortage. The purpose of this study was to address the impact of unlicensed and under-licensed teachers on student growth and achievement in pursuit of assisting school administrators to better understand effective measures for supporting unlicensed or under-licensed teachers. The results of this study are intended to supplement the gap in practice that unlicensed and under-licensed teachers are experiencing in Minnesota classrooms. The final purpose of this study was to add substantive literature to an area of need, as well as influencing the effective support of new-to-profession teachers who are actively working as educators while missing important qualifications for success.As part of the research process, a qualitative study comprised of 31 individually structured interviews was conducted. Interviews were conducted with 17 individuals who began their teaching careers in an un/under-licensed capacity, working through a Tier 1 or Tier 2 Minnesota teaching license. Interviews with 14 school leaders were also conducted. A professional development presentation and corresponding slide deck were then created to explain a proven list of six key strategies that school leaders can utilize to support unlicensed and under-licensed teachers