143174 research outputs found
Sort by
αSurf: implicit surface reconstruction for semi-transparent and thin objects with decoupled geometry and opacity
No full textImplicit surface representations such as the signed distance function (SDF) have emerged as a promising approach for image-based surface reconstruction. However, existing optimization methods assume opaque surfaces and therefore cannot properly reconstruct translucent surfaces and sub-pixel thin structures, which also exhibit low opacity due to the blending effect. While neural radiance field (NeRF) based methods can model semi-transparency and synthesize novel views with photo-realistic quality, their volumetric representation tightly couples geometry (surface occupancy) and material property (surface opacity), and therefore cannot be easily converted into surfaces without introducing artifacts. We present α Surf, a novel scene representation with decoupled geometry and opacity for the reconstruction of surfaces with translucent or blending effects. Ray-surface intersections on our representation can be found in closed-form via analytical solutions of cubic polynomials, avoiding Monte-Carlo sampling, and are fully differentiable by construction. Our qualitative and quantitative evaluations show that our approach can accurately reconstruct translucent and extremely thin surfaces, achieving better reconstruction quality than state-of-the-art SDF and NeRF methods
Models and algorithms for safeguarded data-driven decision making
No full textThis thesis focuses on making smart decisions. Early works often equated smartness with the ability to do more with data (e.g., more accurate predictions, more profitable prescriptions), leading to ever progressing algorithms without safeguards. Safeguards are vital, however, to enforce fairness/ethics, interpretability, privacy, and robustness of decisions. A decision can only be truly smart if it is optimized while adhering to these values. With this perspective, this thesis focuses on the following key themes:
(i) Developing safeguarded data-driven decision making models. I formulate safe-guarded data-driven decision making models to address concerns such as privacy and robustness.
(ii) Deriving tractable algorithms to solve safeguarded problems. These safeguards often make the underlying computational tasks intractable; therefore, I derive efficient approximation schemes with rigorous performance guarantees for the optimization problems that arise in such settings.
(iii) Analyzing safeguarding within multi-stage systems. Data-driven decision making involves multiple stages, typically starting with estimating the unseen truth from data and ending with optimizing decisions over the perceived reality. I aim to understand at which stage these safeguards should be imposed for the best outcomes.Open Acces
Decoding coupled mechanical-electrochemical responses in multi-layer batteries via generalized ultrasonic dynamics
No full textCharacterizing and understanding internal battery physics is essential for stability, safety, and recyclability. Ultrasound provides a non-destructive solution by encoding battery dynamics into mechanical waves. However, the complex multi-layer structure and coupled mechanical-electrochemical behaviors of commercial cells hinder standardized and physically interpretable ultrasonic testing. This study presents a unified ultrasonic framework for multi-layer pouch cells, linking wave dynamics to battery structures, materials, and states across frequency and time domains. Inspired by electrochemical impedance spectroscopy, we examine structure- and state-waveform relationships of batteries under various excitation conditions, decoding ultrasonic responses related to mechanical and electrochemical factors in a generalizable manner. Using first-principles modeling and frequency sweep experiments, we identify battery-specific frequency bandstructures and wave modulation signatures tied to cell architecture and cathode chemistry, allowing mechanical discrimination of these factors in electrochemically steady states. In-operando tests demonstrate that changes in localized ultrasonic resonance associated with shifting bandstructure can map variations in battery state of charge, with the evolution of anode material stiffness as a key driving mechanism. This work establishes a physics-grounded foundation for understanding wave-battery interactions and is expected to guide the development of high-sensitivity, task-specific tools and diagnostic strategies across the in-laboratory, post-manufacture, and in-service stages of a battery’s lifecycle
Understanding H₂-X redox flow battery performance through modeling and design of experiments
No full textThis work investigates the impact of operating conditions on the performance of H₂-X redox flow battery (RFB),
using the H₂-V RFB as a case study. A lumped model incorporating electrochemical kinetics and membrane
water transport mechanisms was built and parameterized using polarization tests on a 25 cm2 cell under varying states of charge (SOC), temperatures and catholyte flow rates (v), following a full-factorial design of experiments. The effect of these conditions on voltage efficiency and discharge power density was analyzed. Results revealed that ohmic overpotential, driven by membrane dehydration, dominated total overpotentialeven at 100% relative humidity at the hydrogen inlet. This dehydration resulted from catholyte species uptake and water transport processes, exacerbated at high discharge current due to electro-osmotic drag driving water away from the gas side. Elevated operating temperatures mitigated dehydration, improving performance. At 1500 Am−2, 50% SOC and 70 mL min−1 v, increasing temperature from 24 to 40 ◦C improved peak power
density from 2325 to 3123 W m−2 and voltage efficiency from 81.3% to 85.1%. Avoiding operating the cell at 10% or lower SOC reduced ohmic overpotential associated with membrane resistance caused by higher catholyte species uptake at low SOC. At 1500 A m−2, 32 ◦C and 45 mL min−1 v, voltage efficiency increased from 80.7% to 82.9% as SOC was reduced from 90% to 50%, but dropped to 76.3% at 10% SOC. Catholyte flow rate had a smaller impact compared to temperature or SOC, primarily affecting concentration overpotential
Photonic hybrid quantum computing
No full textPhotons are a ubiquitous carrier of quantum information: they are fast, suffer minimal decoherence, and do not require huge cryogenic facilities. Nevertheless, their intrinsically weak photon-photon interactions remain a key obstacle to scalable quantum computing. This review surveys hybrid photonic quantum computing, which exploits multiple photonic degrees of freedom to combine the complementary strengths of discrete and bosonic encodings, thereby significantly mitigating the challenge of weak photon-photon interactions. We first outline the basic principles of discrete-variable, native continuous-variable, and bosonic-encoding paradigms. We then summarize recent theoretical advances and state-of-the-art experimental demonstrations with a particular emphasis on the hybrid approach. Its unique advantages, such as efficient generation of resource states and nearly ballistic (active-feedforward-free) operations, are highlighted alongside the remaining technical challenges. To facilitate a clear comparison, we explicitly present the error thresholds and resource overheads required for fault-tolerant quantum computing. Our work offers a focused overview that clarifies how the hybrid approach enables scalable and compatible architectures for quantum computing
Goal-driven multi-fidelity approaches for military vehicle system-level design
No full textThe AVT-331 Research Task Group studied the potential benefits of Multi-Fidelity (MF) methods in vehicle design and documented the relative strengths and weaknesses of different MF methods using a common benchmark suite developed by the team. Benchmarks were established at three levels of complexity for air and sea vehicles and two levels of complexity for a space vehicle configuration. The use of benchmarks at different levels of complexity enabled broad collaboration and knowledge exchange while critically testing method practicality. MF methods were generally found to provide cost benefit for tasks related to the exploration of vehicle parameter spaces or in the optimization of vehicle designs. Cost decrease of one order of magnitude was observed for the sea benchmarks and the intermediate-complexity air benchmark. Numerous opportunities exist for larger and more reliable cost savings through further method development, e.g., improved dimensionality reduction and gradient-based methods. Studies were conducted by 21 appointed team members from nine countries, along with several additional collaborators. The strong collaboration within AVT-331 and the research relevance of the task group’s activities are reflected by over 50 peer-reviewed articles, conference papers, and theses, many of which involve partnerships between AVT-331 members
The role of hormonal and reproductive events on cognitive aging in a cohort of female civil servants: The Whitehall II Study
No full textBackground
Sex differences in dementia prevalence and cognitive decline have been widely reported, yet the role of sex-specific risk factors is still unclear. We aimed to investigate the relationship between hormonal/reproductive events and cognitive performance in women.
Methods
We analyzed data from 1,398 female civil servants (aged 45-68 years) participating in the Whitehall II study. We assessed the longitudinal associations between baseline hormonal and reproductive events on cognitive performance including memory (immediate recall) and fluency (phonemic fluency test) over a 25-year follow-up (1997-2015). Mixed models including an interaction term to model change with time, were used to analyse repeated cognitive scores (standardized to baseline scores), adjusted for demographic (age, ethnicity), socioeconomic (education, income, occupation, marital status), vascular (body mass index, cardiovascular disease, hypertension, diabetes, stroke), lifestyle (smoking, alcohol, sleep, diet) and other factors (depression, anxiety, cancer).
Results
Steeper declines in phonemic fluency were observed in women who had been pregnant (adjusted mean difference: -0.05, 95% confidence interval (95% CI)-0.05,-0.04) (reference: had not been pregnant), had two (-0.03, 95% CI -0.03,-0.02) or more children (-0.05, 95% CI -0.05,-0.05) (reference: no children), had their first child at 30-34 years (-0.06, 95%CI -0.07, -0.04) (reference: aged 20-29 years), experienced cycle lengths of ≤23 days (-0.02, 95%CI -0.03,-0.009) (reference: cycle length 23-35 days), underwent menopause (-0.1, 95%CI -0.1,-0.1) (reference: still menstruating), or stopped menstruating at 40-49 years (-0.06, 95% CI -0.07, -0.06) (reference: stopped menstruating aged ≥50 years). In contrast, having children at age 35+ years (0.05, 95% CI 0.03,0.06), oral contraceptive use (0.07, 95% CI 0.06,0.08), and cycle lengths of 35+ days (0.1, 95% CI 0.1,0.1) were associated with less declines over 25 years. Similar steeper memory declines were observed among women who had been pregnant (-0.06, 95% CI -0.06,-0.05), had 3+ children (-0.1, 95% CI -0.2,-0.1), had a first pregnancy at 30-34 years (-0.05, 95% CI -0.06,-0.04) and underwent menopause (-0.07, 95% CI -0.07,-0.06).
Conclusion
These findings demonstrate potential long-term associations between hormonal and reproductive events on cognitive aging. Further research is needed on the underlying mechanisms and the potential interactions with risk factors that exacerbate these associations
Graphene biosensors using microwave methods
No full textThis thesis presents the first demonstration of a liquid-gated graphene field-effect biosensor transduced by a microwave cavity. To achieve this, a microwave cavity was developed to measure the sheet resistance of graphene with low uncertainty and high stability over extended periods. Specifically, the Allan variance in the Q-factor of ~10^4 was found to be ~0.1 over an averaging time of 1 s, and ~0.5 over an averaging time of 2000 s. The biosensor was demonstrated to detect bovine serum albumin in 0.01 × PBS with a limit of detection below 10^-12 M, which is competitive with other sensors reported in the literature. The adsorption dynamics at different concentrations of BSA were investigated, and detection of the clinically relevant CA 15-3 breast cancer antigen was demonstrated. This study establishes a proof-of-concept for a novel graphene biosensor that may be further explored in future research.
This thesis also presents analytical and finite element modelling using COMSOL to demonstrate and assess the feasibility of a system for molecular mass detection that employs a graphene drum, with its frequency transduced by near-field scanning microwave microscopy (NSMM). It was established that the frequency of the graphene drum modes could be measured using an NSMM, and that if the frequency could be determined with a resolution of 100 Hz, a mass resolution of ~10^-22 kg could be achieved. If the frequency could be determined with a resolution of 1 Hz, a mass resolution of ~10^-24 kg could be achieved. Such a system, if demonstrated experimentally, would offer a molecular mass resolution exceeding that of the best silicon beam resonators reported to date.Open Acces
Effect of infill architecture on structural performance and sustainability of 3D-printed reinforced concrete columns
No full textThis study presents a holistic design-to-performance investigation of 3D-printed reinforced concrete (RC) columns featuring sinusoidal infill patterns, targeting both structural optimisation and sustainability. The primary contribution lies in the development and fabrication of functionally graded, lightweight RC columns using additive manufacturing, integrated with conventional reinforcement and cast mortar. Among four parametrically designed configurations (CS0 – non-peak sinusoidal infill, CS4 – 4-peak sinusoidal infill, CS8 – 8-peak sinusoidal infill, CS12 – 12-peak sinusoidal infill), the CS12 column is selected for physical fabrication and testing under uniaxial compression. The experimental results provide insights into load-bearing capacity and failure mechanisms, which are then used to calibrate a validated finite element model. The model enables an extensive parametric study comparing the four infill designs, focusing on structural efficiency, stress distribution, and damage progression. A cradle-to-gate life cycle assessment (LCA), conducted following the parametric analysis, quantifies the environmental impact of each configuration per unit mechanical strength. Results reveal that columns with CS8 and CS12 infill geometries achieve superior capacity-to-mass ratios and reduce global warming potential (GWP), demonstrating that infill architecture can be strategically tuned to optimise both performance and sustainability. This work advances digital concrete fabrication by providing a scalable framework for performance-driven and environmentally responsive column design
Solidification - microstructure - performance relationships in sn-ag and sn-ag-cu solders
No full textElectronic solder joints are sub-millimetre tin-based alloys whose microstructure is created by solidification in an undercooled melt. This thesis studies the effect of undercooling on the microstructure and mechanical performance of Sn-Ag and Sn-Ag-Cu solders, and develops new insights into their solidification process.
The microstructures of 500 m diameter Sn-Ag balls with different Ag contents and β-Sn nucleation undercoolings are studied and developed into a solidification microstructure selection map. Both deeper undercooling and larger Ag content are found to increase the chance of β-Sn cyclic twinning. Besides the classical single-grain, beach ball, partially interlaced and fully interlaced microstructures, a microstructure with large fully eutectic regions is observed in hypereutectic Sn-5Ag balls, which is explained by competitive nucleation between -Sn and Ag3Sn and coupled zone theory. An orientation study on Sn-Ag balls suggested that solid-state processes happened after solidification and resulted in microstructural features like subgrains and cores.
An immersed thermocouple experiment is performed to study the solidification of SAC305 balls. The microstructures of SAC305/Cu joints with different undercoolings are studied. Deeper undercooling is found to promote β-Sn cyclic twinning, interlacing and refine the joint microstructure. The dendrite growth direction is confirmed for β-Sn, and a dendrite tip splitting phenomenon is observed to help β-Sn grow out of its {001} plane.
Isothermal shear fatigue, shear creep and microhardness test are conducted on SAC305/Cu joints of known undercooling. Deeper undercooling benefits the joint performance under all the tests. Finer eutectic microstructure and the β-Sn interlaced microstructure are suggested to be the major contributors to the improved performance in joints that solidified at deeper undercooling.
This thesis builds new understanding of the solidification – microstructure – performance relationships of Sn-based solder balls and joints, and gives new insight on how to improve the solder performance by controlling the solidification and microstructure.Open Acces