21793 research outputs found
Sort by
Investigating Surrogate-based Models for Holistic Building Performance Assessment and Retrofit Solutions
No full textExisting commercial buildings in Québec are responsible for a major share of GHG emissions within the local building sector, yet limited progress has been achieved in their operational transformations over the past decades. Building retrofit is recognized as a key approach to improving energy efficiency while considering the intertwining economic and environmental effects of the applied retrofit measures. Balancing these objectives transforms the problem into a high-dimensional, multi-objective, and challenging task requiring iterative simulations. When conventional physics-based approaches are used, such analyses become computationally prohibitive, with the challenge further intensifying when analyzing multiple buildings or extending the scope to future building performance.
This thesis addresses the highlighted challenge by developing surrogate models that allow the investigation of high-dimensional retrofit analyses. The study establishes a holistic building retrofit framework for post-industrialized buildings within the Montréal region to rigorously identify feasible retrofit measures specific to the studied building typology. It further develops robust surrogate models for energy consumption, embodied carbon, and investment costs that aid in decision-making across a wide array of measures. The methodology is further extended to integrate future building performance through feature extraction methods. Similarly, the study extends the generalizability of the surrogates to a multi-building analysis using a bottom-up approach, which integrates additional training data samples. Given the prevalence of data scarcity in commercial buildings, the methodology integrates building archetypes while identifying the modelling approach that accurately represents the studied building typology.
The findings demonstrate the robustness of the surrogate-based approach in building retrofit analyses, highlighting a significant improvement in computational efficiency. This thesis introduces a comprehensive methodological framework that advances building performance assessment using surrogate-based methods, identifying their limitations and effectiveness in the building retrofit research domain
Grid and Inverter-Fed Induction Machine Emulation
Get PDFPower-hardware-in-the-loop (PHIL) based machine emulation is increasingly being used as an effective approach for simplifying the testing of electric drive systems. The PHIL-based machine emulator systems control the power converters or power amplifiers in order to mimic machine behavior. Control of the machine emulator is achieved with the help of a machine model running on a real-time controller. Utilizing a motor emulator allows performance testing of the physical machine and its controller before the manufacturing process, reducing risks and costs. This methodology enables testing of different drive inverter faults, including diode rectifier faults, transistor switch faults, and line-to-line faults, without risking damage to the machine. Moreover, it can be applied to test various faulty machines and grid faults, such as open circuits, short circuits, and unbalanced faults. Therefore, this approach finds extensive application in diverse industrial sectors, including military, aerospace, and electric vehicles.
Various open research challenges exist in this domain, all of which aim to improve the accuracy and utility of this test methodology. Emulation accuracy depends on various factors, namely the detailed mathematical model of the electrical machine, the design and selection of the control for the emulation system, and the selection of emulator hardware technologies. Since PHIL-based machine emulation is a relatively novel testing methodology, there is a need to develop a high-performance, highly accurate test bench. Therefore, this PhD work aims to develop an accurate motor emulation test bench to truly validate the performance of the motor, drive inverter, and drive inverter control prior to manufacturing the prototype motor.
A reliable and flexible grid is crucial for PHIL-based machine emulation system test benches. To achieve this, grid emulators are introduced to enhance the reliability and flexibility of motor emulation systems. Currently, power amplifiers are popular choices for emulators due to their high bandwidth, compact size, and ease of control. However, amplifiers are designed for specific voltage ratings, and when higher voltage is required, a step-up transformer is added to the amplifier output. The connection of the transformer at the output of the amplifier can impact the performance and characteristics of the grid emulator system. Therefore, this research work aims to characterize the grid emulator system with an amplifier and transformer under various operating conditions.
Machine emulation accuracy depends on various factors, with one of the key factors being the selection of emulator hardware technology. Thus, this research compares three types of machine emulation systems: one utilizing a conventional IGBT-based hard switching power converter, another employing a high-bandwidth soft switching power converter, and the third using a linear amplifier. Detailed configurations and control descriptions for these systems are provided in this work. Comparative experimental studies are conducted, and the obtained results are then compared with those of the physical induction motor (IM) to verify the emulation performance.
The typical voltage source inverter is employed as an emulator converter in motor emulation systems. However, this converter introduces various harmonics into the motor emulation system, primarily attributed to dead time, switching components, and control signals. These harmonics can deteriorate motor emulation accuracy. Therefore, it is important to investigate and compensate for emulator converter harmonics in the motor emulation system. Hence, this PhD work will also explore this topic.
The induction motor drive with an LC filter configuration enables smooth pulse width modulation (PWM) output voltages of the motor drives. Moreover, employing high-performance, high-bandwidth class D power amplifiers as emulator converters enhances the system's bandwidth, thereby ensuring accurate machine emulation. Another open research topic that could improve motor emulation systems is testing electric drives with filters. Testing such a machine drive using conventional emulator structures would be challenging. Therefore, this PhD work will also investigate this topic
Craft & Craftivism: A Biographical Dictionary of Contemporary Ceramic, Fibre and Class Artists in Canada: Volume 2: Fibres
Get PDFContemporary work in ceramics, fibres and glass is commanding unprecedented attention in the visual arts world. Many Canadian artists have gained recognition for their aesthetic proficiency, conceptual relevance, and skill-based expertise in provincial, national, and international public institutions using these materials. The visual arts community in Canada and elsewhere seeks information concerning the historical importance and significance to contemporary Canadian craft artists of these materials and skills. Craft and Craftivism, edited by Loren Lerner, Janice Anderson, Shannon Stride, and Karine Antaki, is a biographical dictionary of artists working in these media, created as a free e-publication to fill this need. The material is readily available for both pedagogical purposes and the general public’s use. It aims to encourage scholarly interest while acknowledging the artists’ contributions to Canadian visual arts, a path currently encouraged in many Canadian post-secondary art education institutions. -- Introduction.
Volume Two of Craft & Craftivism is dedicated to Fibres. It includes entries for 137 artists
Life is plastic, it’s fantastic: Environmental and population variation shapes maternal life history responses in a salmonid fish
Get PDFLife-history theory is a central component of evolutionary biology, which predicts that organisms trade-off investing their finite energy into growth or reproduction to maximize their fitness. Yet, experimental studies seldomly assess the extent to which maternal life-histories vary across populations, life-stages or their interactions. Brook trout (Salvelinus fontinalis) are an excellent model system to examine such variation, as diverse populations can be isolated despite occurring within small scales (25km2) in nature. Using two common garden experiments over three years, we assessed the influence of environmental and population variation on maternal and early life history responses. First, we captive-reared four wild brook trout populations under four feed treatments to assess five maternal life history responses for 403 females. As early juveniles, we reared fish on a low-variability, high-food or a high variability, low-food treatment. As late juveniles until maturation, we switched half the individuals to the opposite food treatment, while the other half remained unchanged. In a second experiment, we reared offspring produced by the crosses from the previous experiment under a thermal stress gradient to investigate transgenerational plasticity of early life history traits. We found that females grew faster in resource-rich environments, consistent with the life history theory that they would mature at larger body sizes, producing many, small eggs compared to females in resource-poor environments. Moreover, energy allocation into reproductive trade-offs varied substantially by population and life-stage, suggesting populations are locally adapted to their environments. Remarkably, manipulations to the maternal environment elicited limited transgenerational plasticity in offspring exposed to thermal variability and stress. Our study highlights the extent to which maternal phenotypic plasticity is moderated by population effects across life-stages and environmental gradients
Weathering on Coarse Gravel Roads: Modelling Carbon Dioxide Removal in Weathering Processes on Gravel Roads
Get PDFAs global atmospheric carbon dioxide levels rise, negative emissions strategies are being researched to rapidly capture and store these emissions. Enhanced rock weathering as a negative emissions strategy has a strong focus on research on agricultural applications, however, the construction and maintenance of gravel roadways represents an under-researched potential form of weathering. This research examines the contributions of gravel roadways in West Bolton, QC to the reduction of atmospheric carbon dioxide levels through chemical weathering. A model was produced using Stella modeling software to simulate the atmospheric carbon dioxide transformation through weathering of the roadway material in the study area. A sieving analysis was performed to determine both the change in rock particle size over time and for the calculation of the available surface area of the road material. A weathering analysis was performed using road material sampled from the roadway and rainwater captured within the study area. Alkalinity readings as the concentration of CaCO3 were obtained during the weathering analysis and using particle size distribution for the surface area along with the annual rainfall data, weathering rates for all three sample sites were calculated. The weathering rates were found to be 1.36x10-6 mol/m2, 2.25x10-6 mol/m2, and 4.11x10-6 mol/m2. Of critical influence on the weathering rate was the volume of annual rainfall. Model simulations using the results of the sieving and weathering analysis found that the carbon transformation using the analysis results corresponded to the modeled results using reference rates. Factors influencing the modeled carbon transformation were found to be mechanical fracturing from annual daily car passage and maintenance frequency over timescales of ten years. This study begins the research on gravel roads as a negative emissions strategy and produces a model that can inform road management decisions
Development of Stimuli-Responsive Carbon Nanodot Loaded E-spun Nanofibers for Dual Functional Wound Dressings
Get PDFAbstract
Timely and accurate assessment of wounds during the healing process is crucial for proper diagnosis and treatment. Conventional wound dressings lack both real-time monitoring capabilities and active therapeutic functionalities, limiting their effectiveness in dynamic wound environments. Currently, antibiotics are mostly used to treat bacterial infections, negatively resulting in the emergence of numerous drug-resistant bacteria demanding the development of alternative strategies.
My MSc research focuses on a proof-of-concept approach exploring the unique emission properties and antimicrobial activities of carbon nanodots (CNDs), for simultaneous detection of wound healing progress and treatment of bacterial infections. This approach centers on the fabrication of well-defined CND-embedded PVA e-spun nanofibrous mats, which are crosslinked with degradable boronic ester (BE) crosslinks. The BE-CND/PVA mats exhibit stimuli-responsive degradation to pHs and hydrogen peroxide and pH-responsive release of CNDs allowing for both localized antibacterial action and real-time optical detection. Promisingly, the mats turn out to be biocompatible with skin cells and exhibit antimicrobial activities against both Gram-positive and Gram-negative bacteria. Furthermore, they showcase great potential for real-time monitoring of wound pH to assess the wound status.
Overall, these results suggest that BE-CND/PVA mats could significantly enhance wound healing by providing localized therapeutic action, reducing the risk of bacterial resistance and enabling non-invasive monitoring of wound progress
Forest elephant habitat use and interactions with humans in the Campo-Ma’an landscape, southern Cameroon
Get PDFHabitat loss from forest conversion to agriculture threatens tropical biodiversity. While wildlife typically avoids human-mediated risk, some species may adopt riskier strategies to access food in human-dominated landscapes. The recent conversion of a protected area into an agro-industrial plantation near Campo-Ma’an National Park, Cameroon, coincides with increased forest elephant sightings near human settlements, suggesting a change in habitat use. Camera traps were deployed and reconnaissance walks were conducted between camera trap stations to examine the influence of human activity on forest elephant habitat use. Households were interviewed to assess local experiences with human-elephant conflict and attitudes toward elephant conservation. Reconnaissance walks provided a greater amount of data than camera traps. Elephants tended to avoid sites with higher human activity and were less active during the midday peak in human activity. However, their proximity to settlements suggests an overall risk-taking behavior in their habitat use which, combined with site-level avoidance and distinct activity patterns that reduce encounters, points to a complex trade-off between human-mediated risk and resource access. Most households reported increased elephant presence and crop damage, often attributing it to the agro-industrial plantation. Vulnerability to crop damage better predicted attitudes toward elephant conservation than actual damage, and past interactions with wildlife and conservation authorities also appeared to influence perceptions. Despite increased human-elephant conflict, attitudes toward elephant conservation remained generally positive but could deteriorate if conflicts persist. Continued monitoring of forest elephants in human-dominated landscapes, effective land-use planning, and incorporating local perspectives into conservation strategies are crucial for sustainable coexistence between humans and wildlife
Digital Interactivity in Space AI-Augmented Eco-Didactic Experience in Public Realm
No full textThe ecological crisis is advancing rapidly, and it is crucial to spread awareness, create dialogues, and educate about environment and sustainability to encourage behavioral change and eco-action. The public realm, characterized by its high human circulation, serves as an ideal space to initiate, and foster these environmental conversations. Public artworks stand out as one of the most prevalent and impactful approaches for engaging with the society. In the environmental context, eco-art facilitates sharing of eco-messages, fosters community dialogues around critical issues and solutions, and motivates individuals to take meaningful action. Although public eco-art installations significantly engage audiences, the potential to amplify this impact through Artificial Intelligence (AI) augmented interactive gamification within a didactic framework remains largely unexplored. AI technologies are rapidly infiltrating both professional and personal spheres, significantly influencing consumer behavior through the advertising industry shaping visions for future urban landscapes, as evidenced in conceptual designs for smart cities. However, the high energy consumption associated with AI raises environmental concerns, even as its adoption in daily life becomes inevitable. This research explores how AI technologies can enhance environmental learning by developing an AI-augmented, eco-didactic interactive game. The goal is to support the dissemination of the United Nations’ (UN) Sustainable Development Goals (SDGs) related to the built environment. By integrating AI, interactivity, and gamification with an artistic approach, the study seeks to transform urban spaces into creative and interactive hubs. These engaging experiences aim to spark curiosity, inspire enthusiasm, and encourage proactive eco-friendly actions within the public realm. This research advances studies in creative AI, autonomous AI, and social AI focusing on their integration within physical environments, eco-didactic spaces and public domains in design, fine arts, architecture, urban studies, and environmental fields. It contributes directly to development of engaging environmental and educational public space experiences both in Canada and globally. The findings promise to be broadly applicable, offering a pioneering framework for interactive eco-didactic design practices and providing unique insights for future advancements. By illustrating the benefits of AI in fostering ecological awareness and sustainable engagement, this research supports the dissemination of SDGs, particularly in educating for sustainability within urban environments
Extreme Views: 3DGS Filter for Novel View Synthesis from Out-of-Distribution Camera Poses
Get PDF3D reconstruction is a foundational component in robotics and autonomous systems, enabling machines to perceive and interpret their environment for tasks such as navigation, obstacle avoidance, and motion planning. As these systems increasingly operate in real-time and dynamic environments, the need for efficient and robust scene understanding becomes paramount. Among recent innovations, 3D Gaussian Splatting (3DGS) has gained attention for its ability to reconstruct photorealistic 3D scenes with high efficiency and support for real-time novel view synthesis. Unlike traditional mesh- or voxel-based representations, 3DGS models scenes using a compact set of anisotropic Gaussians, each carrying spatial and appearance information, which are then splatted into the image plane during rendering.
However, a persistent challenge in such learned representations is uncertainty due to insufficient, ambiguous, or occluded data in the input views i.e. epistemic uncertainty. This can result in visual artifacts, especially when rendering novel viewpoints that diverge significantly from the training set. Existing methods such as BayesRays, designed for NeRF-based models, address this issue via probabilistic ray-based sampling and post-hoc filtering, but require retraining.
In this work, we propose a gradient sensitivity-based filtering framework for 3DGS that mitigates epistemic artifacts in real-time without the need for model retraining. Specifically, we introduce a novel sensitivity score that quantifies the directional gradient of pixel color with respect to spatial perturbations at the point of ray-Gaussian intersection. This score captures the local instability in the rendering process caused by insufficient coverage or ambiguity in training views. By computing this score directly within the existing rendering pipeline, we enable on-the-fly filtering of Gaussians whose contributions are deemed unstable or unreliable.
Our approach can be applied to any pre-trained 3DGS model, making it highly practical for deployment in real-time systems. We evaluate our method on challenging indoor and outdoor scenes, including those from the Deep Blending and NeRF-On-the-Go datasets, and show that it effectively suppresses rendering artifacts. Notably, our filtering substantially improves visual quality, realism, and consistency compared to BayesRays, while avoiding the overhead of additional training or scene-specific tuning. This makes our method particularly suited for robotics and AR/VR applications where rapid adaptation and robustness to viewpoint changes are critical
Design of Multi-Sine Watermark using Power Spectral Analysis for Replay Attack Detection
Get PDFDesign of multi-sine watermark using power spectral analysis for replay attack detection
Sunitha George
Replay attacks are a critical security concern in cyber-physical systems (CPS), where adversaries record legitimate data transmissions and maliciously retransmit them later to disrupt normal system operations. These attacks are particularly dangerous because they often replay legitimate data, making them difficult to detect using traditional intrusion detection systems. As CPS continue to integrate deeper into critical infrastructure such as power systems, industrial automation, and transportation networks, the need for better safety measures becomes increasingly urgent.
One promising line of defense involves watermarking techniques, in particular, using multi-sine watermarks with switching frequencies. This thesis studies the problem of choosing the parameters of multi-sine watermarks to achieve replay attack detection with desired level of confidence. The proposed method is derived from a power spectral analysis of the output of the plant in both normal (no attack) and during attack operation.
A flow control process involving a tank is utilized as an illustrative example. Through this example, the effectiveness of the proposed method is validated, showing its capability to design a watermark that can successfully detect replay attacks and thus enhance the security of the control system