82837 research outputs found
Sort by
Anne Summach - Abstract 16 - Innovate Conference 2025
No full textNPs in Canada are facing significant regulatory changes, including updated
entry-level-competencies established in 2023. The profession is growing rapidly in Alberta, though the province has had limited funding and practice options for NPs. Competency changes triggered degree redesign at the University of Alberta, leading to a proposed novel degree for NPs
Kinetic Study of Steel Corrosion in Supercritical CO₂-Saturated Aqueous Environments
No full textThe long-term reliability of carbon capture, transportation, utilization, and storage (CCTUS) infrastructure depends on the corrosion resistance of steels used in supercritical CO₂ (s-CO₂) transmission systems. While dry s-CO₂ is non-corrosive to steels, the introduction of water, either from the capture processes or at geologic injection sites, creates aggressive corrosion environments that can threaten pipeline integrity.
This study investigates the corrosion behaviour of five steels (X65, P110, 2Cr, P91, and SS316) with varying Cr levels in s-CO₂ saturated saline environments over exposure times up to 1000 h. Corrosion rates were determined through weight loss measurements, and corrosion product layers were characterized using X-ray diffraction (XRD), Raman spectroscopy, and scanning electron microscopy (SEM) with energy dispersive X-ray spectroscopy (EDS). The primary two objectives were to assess the role of chromium (Cr) content, and to investigate the kinetics of the corrosion process in s-CO₂ saturated saline environments.
Corrosion rates consistently decline with increasing Cr content in the studied steels. A critical Cr threshold of approximately 9 wt.% was identified, above which the alloys exhibit satisfactory corrosion resistance in s-CO₂-saturated saline environments. For instance, both P91 and SS316 steels showed exceptional corrosion resistance (<0.01 mm/y), attributed to the formation of dense FeCO₃ layers over Cr-oxide-rich intermediates that effectively inhibited further metal dissolution. Across all steels, corrosion rates decreased exponentially with exposure time from 1 h to 1000 h due to the gradual development of protective FeCO₃ layers.\nCharacterization results indicate that the absence of Cr-oxide formation and the dominance of amorphous iron oxides on the steel surface hinder the development of a stable, protective film. This highlights that Cr content alone is not sufficient to predict corrosion behaviour. The findings suggest that additional factors, such as oxide morphology and environmental conditions, play a significant role in determining corrosion resistance in supercritical CO₂ environments. SS316 and P91 are recommended for applications involving unavoidable water-rich s-CO2 phases, while low-Cr steels may still be suitable for dry s-CO₂ transmission systems where moisture levels can be effectively controlled. The results also reveal that short testing durations tend to overestimate corrosion rates in s-CO₂-saturated aqueous environments. A minimum testing duration of 100 h is recommended to obtain a more accurate assessment of long-term corrosion performance; however, corrosion rates still continue to decrease up to the 1000 h mark
General and Efficient Visual Goal-Conditioned Reinforcement Learning using Object-Agnostic Masks
No full textGoal-conditioned reinforcement learning (GCRL) allows agents to achieve diverse objectives using a single behavior policy. Goal representation and reward formulation play a crucial role in the successful training of GCRL agents. In this thesis, I present a mask-based goal conditioning and reward calculation system that is simple to compute, yet highly effective. My proposed method provides object-agnostic visual signals to the agent, facilitating strong generalization capabilities. In contrast, existing goal representation methods, such as target state images, 3D coordinates, and one-hot vectors, face issues of poor generalization to unseen objects, slow convergence, and require specialized cameras. Masks are dynamically generated throughout the episode, and their position and size change based on the location of the target. As such, they provide rich feedback to the learning system, resulting in faster training compared to static goal representation methods. Furthermore, masks can be exploited to create dense rewards, diminishing the need for complex distance-to-goal computations. Using this system, I achieved 99.9\% accuracy in simulated reaching tasks with both train and test objects. I successfully extended the mask-based system to perform grasp and pick-up, without requiring any positional information of the target object. Moreover, I implemented the mask-based system in the real world by using open vocabulary object detection models for mask generation. I demonstrated strong learning performance in the real world on a learning from scratch task using a physical Franka Panda robotic arm
An Autoethnographic Therapeutic Performative Inquiry on My Lived Experiences as a Black Woman Co-victim of Homicide in Canada
No full text“Anansi Came” was an autoethnographic therapeutic performative inquiry that explored my lived experiences as a homicide co-victim (Center for Victim Research, 2020) of Afro-Caribbean descent and the impact of “cumulative exposure to lifetime adversity” (Silver et al., 2021, p. 5). Informed by my practice as a dramatherapist and registered psychologist and safeguarded by the best practices of trauma work, this research engaged in a process that combined dramatherapy and performance autoethnography with the principles of engaged pedagogy. The script and performance that I created through this process were the focus of my dissertation. Two questions framed this research: (1) What are my lived experiences as an Afro-Caribbean woman who has survived the loss of a loved one to homicide? (2) How might an embodied storytelling approach, adapted from The Story Within (Silverman, 2020) process, assist me in my healing process, with the potential to encourage others to explore their traumatic experiences? Data within my personal stories, the script I developed and performed, and the audiences’ post-performance feedback uncovered how engaging in embodied storytelling can promote self-discovery, resilience, and trauma recovery. Using dramatherapy strategies provided me with the opportunity to develop the necessary skills to identify misinformation and myths society communicated to me about being a Black homicide survivor. The Story Within process was instrumental in reclaiming my voice by building internal resources that helped me manage my somatic (body-based) experiences and process habitual trauma-related responses (Ogden, Paine & Fisher, 2006) within a safe communal healing space. By disrupting the silencing and dismissal of my experiences as a Black co-victim of homicide and fostering a sense of belonging, my embodiment as a practice and pathway to healing has enabled me to regulate a nervous system impacted by years of oppression. While the focus of this study is on understanding my experiences and assisting my healing journey, my larger motivation for this inquiry is my hope that it will empower and inspire others in my community—especially the youth and educators who have experienced trauma—to go on their own journeys of self-exploration, empowering them to sever the cycle of violence and overcome the effects of trauma
Species choice and seed sourcing for forestry field experiments to address climate change across Canada
No full textClimate change adaptation in forestry will need field tested climate-informed seed transfer strategies to improve resilience, preserve genetic diversity and ensure long-term health and productivity of forest ecosystems. This is especially urgent in northern latitudes, such as Canada, where warming trends have been most pronounced. The large-scale DIVERSE research project plans to establish such assisted migration trials at 22 forest management areas across Canada with provincial government and industry participants in British Columbia, Alberta, Ontario, Nova Scotia, Quebec, and Saskatchewan. We contribute an on-line decision support tool to help the DIVERSE researchers and forest managers make climate-informed selections of tree species and seed sources for reforestation. These recommendations include cross-border transfers and can also include introducing new species beyond their current range limits. For the climate-informed seed sourcing recommendations, I used the scaled multivariate Euclidean distance of 12 bioclimatic variables to match seed source’s historic climate to planting site’s new projected future climates, where source and targets were defined by ecosystem delineations for Canada and the US. Climate suitability of a species for a target site in the future was inferred by averaging species’ frequencies of the five ecosystems with the closest climate distance. This resulted in climate matched source ecosystems and species frequencies for the 2020s, 2050s and 20280s for all the ecosystem delineations. This is a lot of information to communicate so a web tool (http://tinyurl.com/DIVERSE-SST) was developed for the forest companies and government stakeholders across Canada that participated in this project. This Euclidean distance ecosystem-based climate matching approach is a fairly basic type of species distribution modeling. However, the simplicity of this approach allowed me to incorporate over 240 of the major tree species in North America in the recommendations. Additionally, the larger geographic scale of the climate matching provided recommendations at a level more in line with current seed sourcing systems making the recommendations more operationally relevant. These recommendations are the first step in the establishment of test plantations to validate whether tree growth, health and survival can be maintained or improved through large scale operational deployment of assisted migration in Canada
Active control of very-large-scale motions using wall deformations and real-time particle image velocimetry
No full textThis work experimentally investigates the active control capabilities of wall-normal surface deformations to attenuate the very-large-scale motions (VLSMs) of a turbulent boundary layer (TBL) at a friction Reynolds number of Reτ = 2600. The control system utilized a feedforward control scheme based on velocity measurements upstream of the actuator obtained using a real-time particle image velocimetry (RT-PIV). The actuator was a soft circular surface with a diameter D roughly equal to the boundary layer height δ that was designed to achieve smooth deformations with a peak amplitude of 0.07δ. The preliminary investigations characterized the frequency response of the developed active surface, followed by assessing the resulting periodic impacts on the TBL when operated at a constant amplitude and frequency. These periodic measurements showed that motions produced by upward deformations were approximately twice as strong as motions produced by downward deformations. Finally, the feedforward control applied an opposition strategy with different gain values to spatio-temporally filtered measurements of streamwise velocity fluctuations using RT-PIV. Opposition control was able to reduce each Reynolds stress component within the logarithmic region. Using the best case within the parametric study, the active surface could reduce the Reynold stresses by 9.3%, 1.7%, and 6.7% for ⟨u2⟩, ⟨v2⟩, ⟨uv⟩, respectively at y+ = 120. Furthermore, when a scale decomposition was applied, it was revealed that all the Reynolds stress reduction occurred in the large-scale structures with minimal effects on the small-scale structures
The Varsity Ski Club: Recreation, Participation, and Leadership in Extracurricular Education at the University of Alberta, 1932–1938
No full textThis thesis examines the social history of the Varsity Ski Club (VSC) at the University of Alberta (U of A) from 1932 to 1938. The student-led club was a driving force for outdoor recreation on campus with primarily English-speaking middle-class members from communities across Alberta. This thesis explores social and cultural influences to understand what shaped skiing, leadership, and volunteerism for women and men in the club. I argue that the VSC was a co-educational site of extracurricular recreation where students cultivated experience and skills in community life and citizenship that were integral to a liberal education in the 1930s.
The ski club brought students together and created a ski hub in the North Saskatchewan River valley near campus. The club fostered a shared sense of community belonging as members worked together to plan Sunday Ski Hikes, ski competitions, and the construction of a ski club cabin. As the club developed, its structure formalized, allowing club executives to practice management, organization, leadership, and social skills that contributed to their growth and education. Club members participated in and promoted the sport of skiing locally and provincially, adding the VSC to the greater structure and organization of skiing in Canada. This microhistory of the VSC combines archival sources, including newspapers, yearbooks, and documents, with secondary sources to trace the Ski Club history. An understand of the club members and extracurricular co-education provides insight into how U of A students reflected the dominant ideals of modern manhood and womanhood in the thirties. Social influences such as amateurism and rational recreation shaped the men’s expression of masculinity and impacted how the VSC developed. A close reading of female member’s life stories suggests that the VSC cultivated female leaders who applied their skills in professional and family life after graduation. Early social influences of sport, girls groups, and family are analyzed to better understand how women in the ski club adopted and resisted ideals of modern womanhood. The VSC story adds to understanding the history of varsity sport clubs, communities, gender ideals, and skiing in western Canada during the 1930s. The VSC reveals the extracurricular scope of liberal education, and a legacy connected to today’s longstanding U of A Outdoor Club
Introduction to Veterinary Terminology
No full textIntroducton: Medical terminology is commonly used in most medical settings, including veterinary clinics. This language is based on Greek and Latin terms and veterinary terminology is a subset of medical terminology. Just like learning a new language, it can be a daunting task. The primary objective of this OER is to introduce you to the medical terms in a simple and easy to understand format. This resource was designed as an introductory medical terminology course for the Veterinary Office Assistants at NorQuest College; however, it is likely to be useful for individuals in other veterinary medical settings as well. To assist with learning the complex language of veterinary medical terminology, each chapter has embedded H5P activities, including a final chapter review. We have also included content on abbreviations, common procedures and species-specific terms. We hope that this textbook will provide a thorough overview of not only medical terminology, but also introduce connections to other veterinary topics
Enhancing Methane Emission Estimates through Deep Learning-Based Data Interpolation
No full textMethane is a potent greenhouse gas with significant implications for climate change, making accurate monitoring and analysis of its spatial distribution critical for environmental management. Among the various observational sources, the Sentinel-5P satellite is widely used due to its ability to provide high-resolution, global measurements of atmospheric methane. However, such datasets often suffer from missing or incomplete data, limiting their utility for detailed analysis.
This study presents a novel deep learning-based approach for spatial interpolation of methane concentration data, addressing gaps in traditional methods and advancing capabilities in environmental monitoring. The proposed model, Methane Interpolation Network (MINet), demonstrates superior performance compared to established interpolation methods such as Inverse Distance Weighting (IDW), Ordinary Kriging, Nearest Neighbor, and Radial Basis Function (RBF), achieving significant improvements in metrics like RMSE, MAE, and . Additionally, it outperforms baseline deep neural network models, including U-Net and U-Net with Partial Convolution, in terms of accuracy and robustness. An ablation study was conducted to evaluate the effectiveness of the attention mechanism and normalization layers. By systematically removing or replacing these modules, we assessed their individual contributions to the overall performance. Two types of missing data patterns, large sparse masks and small dense masks, were analyzed to evaluate the model's adaptability. Results showed that the proposed model effectively reconstructs missing spatial data, capturing fine-grained spatial details and trends that are often overlooked by traditional methods. To further assess real-world applicability, we selected 20 super-emitter events from the Kayrros Methane Watch dataset. In these cases, methane leakage regions were partially masked to simulate incomplete observations. The model was applied to reconstruct the missing portions of the plumes. The results demonstrate that the model successfully restores the spatial extent and shape of the emissions, revealing detailed patterns that are critical for identifying and characterizing super emitters in regions with sparse or noisy data.
Despite its promising results, the study is limited by its exclusive reliance on data from Sentinel-5P. Future work could incorporate additional remote sensing datasets to provide a more comprehensive perspective. Adding features such as meteorological conditions, land use patterns, and industrial emission sources may further improve the model’s performance and generalizability. In addition, developing region-specific models that integrate both spatial and temporal dynamics could better capture evolving emission patterns and support real-time forecasting. These advancements would enhance the role of deep learning in environmental monitoring by providing high-resolution insights, reducing reliance on costly ground-based sensor networks, and supporting more effective policy-making and mitigation strategies.
This work underscores the potential of deep learning to transform methane monitoring, offering a scalable and robust solution for addressing data gaps and improving the understanding of methane dynamics. By bridging observational gaps and enabling detailed spatial and temporal analysis, this approach lays the foundation for more effective monitoring and management of greenhouse gas emissions
