DSpace@RPI (Rensselaer Polytechnic Institute)
Not a member yet
6809 research outputs found
Sort by
Eliciting Survey Knowledge with Semantic Data Dictionaries
No full textMany countries perform surveys to gather data from their population for supporting decision-making and development of public policies. Questionnaires are possibly the most used type of data acquisition instrument in surveys, although additional kinds may be employed (especially in health-related surveys). In the United States, the NHANES is a national health and nutrition examination survey conducted by the National Center for Health Statistics, designed to collect data on adults' and children's health and nutritional status. Data is organized in several tables, each containing variables to a specific theme, such as demographics, and dietary information. In addition, data dictionaries are available to (sometimes partially) document the tables' contents. While data is mostly provided by survey participants, instruments might be collecting data related to other entities (e.g. from participants' households and families, as well as laboratory results from participants' provided blood and urine samples). All this complex knowledge can often only be elicited by humans when analyzing and understanding the data dictionaries in combination with the data. The representation of this knowledge in a machine-interpretable format could facilitate further use of the data. We detail how Semantic Data Dictionaries (SDDs) have been used to elicit knowledge about surveys, using the publicly available NHANES data and data dictionaries. In SDDs, we formalize the semantics of variables, including entities, attributes, and more, using terminology from relevant ontologies, and demonstrate how they are used in an automated process to generate a rich knowledge graph that enables downstream tasks in support of survey data analysis
Thermal response and applications of graphene-based macrostructures and their composites
No full textDecember 2023Single-layer graphene displays many unique attributes, including record-breaking thermal conductivity, charge carrier mobility, fracture strength, and Young’s modulus, and thus possesses immense potential for a wide range of technological applications in electronics, photonics, nanocomposites, etc. However, it has been challenging to translate the outstanding mechanical, electrical, and thermal properties of single-layer graphene into that of macroscale graphene assemblies, significantly constraining their applicational potentials. In this thesis, several approaches have been developed to fabricate wet-spun graphene macrostructures and their composites, including graphene fibers, hollow graphene tubes, their 3D assemblies, and composites. The graphene macroscopic structures and assemblies display excellent thermomechanical and electrical properties resulting from the inherent properties of single-layer graphene, optimization of their microstructures, and designed material interactions.In particular, to fabricate a high-performance graphene macroscopic structure with both ultrahigh electrical and thermal conductivity, a new electroplating process has been developed to deposit uniform micrometer-level thin copper coatings onto graphene fibers. The electroplating process has also been applied to wet-fused graphene fiber fabrics. The resulting copper-coated graphene fiber fabrics not only gain the high electrical conductivity of copper but also have high flexibility and retain the high thermal conductivity of graphene fibers owing to the conductive junctions between wet-fused graphene fibers.
Building upon the knowledge of fabricating highly thermally conductive graphene fibers, a fast-coagulating coaxial wet-spinning method has been developed to fabricate hollow graphene tubes (HGTs) with high thermal conductivity. By filling the bore of the graphene tubes with 1-octadecanol, a graphene-based phase change composite has been developed, exhibiting extraordinarily high latent heat exceeding that of pure 1-octadecanol. The counterintuitive finding is explained through experimental and literature studies of the interactions between graphene and 1-octadecanol, which results in lattice contraction of the latter, resulting in an abnormally higher phase change enthalpy. This finding contrasts with the conventional wisdom that adding a non-phase changing filler such as graphene into the phase change composite will reduce the latent heat and, thus, heat storage capacity.
A moisture-fusing method of HGTs has been developed based on the wet-fusing of graphene fibers. Without changing the tubular morphology of the HGTs, individual tubes have been layer-by-layer moisture-fused to form an interconnect graphene tube macrostructure with high thermal conductivity. These graphene tube macrostructures have been filled with 1-octadecanol dispersed with multiwalled carbon nanotubes. The resulting phase change composite exhibits high visible light absorption, fast thermal response, and a high heat storage capacity. These traits ensure high solar-thermal conversion and storage efficiency of the composite at low carbon weight percentages, making them suitable for large-scale solar-thermal energy harvesting applications.
In addition, a thermally conductive yet mechanically stretchable composite graphene fiber has also been developed by wet spinning a mixture of graphene oxide, bacterial cellulose, and EGaIn liquid metal. A stretchable link is formed between graphene oxide sheets within the composite fiber by grafting thin bacterial cellulose polymers between them. The grafted bacterial cellulose fibrils keep the composite fiber mechanically intact at large tensile strains. The good thermal conductivity of the composite fiber is ensured by its well-aligned graphene sheets. EGaIn provides an electrical percolation path within the composite fiber, allowing for high electrical conductivity. Adding EGaIn also helps reduce Young’s modulus of the composite fiber. At the end of this work, further experiments are proposed to validate the applicational potentials of the stretchable and electrically and thermally conductive composite graphene fibers.Ph
Examination of iron-sulfur clusters in proteins
No full textMay 2024School of ScienceSolar energy utilization is one of the most viable avenues of alternative energy productiondue to efficiency and abundance of the resources. While we can generate electrical energy
using solar photovoltaics, the production of chemicals and fuels using solar energy, water, and
carbon feed stocks is a highly attractive solution. The conversion of solar energy to chemical
energy in a highly efficient manner is performed by photosynthesis. Thus, nature provides a
blueprint for the design of a new generation of devices for the production of solar chemicals
and fuels. The goal of this thesis research was to better understand the tuning of charge
transfer cofactors in natural photosynthesis. There are two types of photosynthetic reaction
centers in nature, Type I (with iron-sulfur clusters serving as the terminal electron acceptors)
and Type II (with quinone molecules acting as the terminal electron acceptors). The Type
I photosynthetic reaction center, Photosystem I (PSI), in higher plants and cyanobacteria,
contains three four iron-four sulfur [4Fe-4S] clusters as charge carriers which play a crucial
role in the efficient electron transfer. In this thesis, I am examining the electronic structure
of iron-sulfur clusters in model proteins and a [4Fe-4S] cluster of PSI using density functional
theory (DFT) methods.M
Tau, heparan sulfate, and cyclophilin a interactions in alzheimer's disease
No full textAugust 2024School of ScienceIn Alzheimer’s disease (AD), tau aggregates form neurofibrillary tangles (NFT), a pathological hallmark of AD. Tau pathology is known to spread in a prion-like manner mediated by tau binding to the cell-surface glycan heparan sulfate (HS), expressed by multiple cell types, including microglia. Tau also interacts with the abundant immunophilin and proline isomerase protein Cyclophilin A (CypA). However, the molecular details of tau−CypA and tau-HS interaction are poorly understood, especially in the context of microglia. Using a multidisciplinary approach, we characterized the interaction between heparan sulfate and tau. We show that tau's proline-rich region 2 (PRR2) domain plays an important role in tau-HS interaction by NMR and SPR. In addition, we characterized various HS analogs for their ability to disrupt tau-HS interaction and inhibit tau uptake in HMC3 microglial cells. Among the analogs, MPS, PPS (semi-synthetic), RPI-27, RPI-28, and RS (marine sulfated polysaccharides) disrupt tau-heparin interaction and cellular uptake in microglia. We further studied the interaction between PRR2 and CypA. Using SPR and NMR, we showed that tau can bind to CypA with nM affinity, and the PPR2 region interacts with CypA with 10 M affinity. CypA can catalyze proline isomerization in PPR2, specifically recognizing regions near P247 and P249 in PRR2. We demonstrated for the first time that HMC3 naturally expresses CypA and produces intracellular reactive oxygen species in response to CypA inhibition and tau uptake. and Our results deepened our understanding of tau, HS, and CypA interaction, which may contribute to the development of novel AD therapeutics.Ph
Deadlock Freedom in Actor Languages
No full textDecember 2023We introduce a framework using session types for denoting the relationships between multiple actors in a system. We prove that ascribing to this framework guarantees deadlock freedom, and demonstrate tests of how an implementation of such would work in the SALSA language
On-chip photonic systems for machine learning applications
No full textApril 2024School of ScienceThis dissertation presents a comprehensive study on the integration of photonic systems within silicon chips to enhance machine learning applications. It explores the potential of on-chip photonic components, such as low-loss silicon nitride delay lines and slow light thermal phase shifters, in overcoming the limitations faced by traditional electronic computing systems. Through detailed design, fabrication, and characterization, this work demonstrates how these photonic components can be utilized in neural network models, offering scalable and energy-efficient solutions for real-time data processing and advanced cognitive tasks. This research marks a significant step toward the realization of next-generation machine learning hardware, leveraging the unique advantages of photonic computing.Ph
Development of coupling approach for integrated analysis of thermo-fluid dynamics inside cryogenic tanks
No full textAugust 2024School of EngineeringNumerical analysis of thermo-fluid dynamics for cryogenic propellant storage is vital to future space mission planning. This analysis primarily consists of nodal and computational fluid dynamics (CFD) modeling approaches. While nodal approaches prioritize faster computation over fidelity, CFD approaches promise accuracy and fidelity at the expense of significant computational resources. This dissertation explores the state-of-the-art modeling approaches to simulate cryogenic propellant behaviors under storage conditions including self-pressurization periods and during active pressure control. To have a deeper understanding of the limitations of each numerical analysis code, the accuracy and speed of each code were assessed by simulating cryogenic propellant storage scenarios suitable for each approach.First, the nodal modeling approach was utilized to simulate the tank self-pressurization period of a Multipurpose Hydrogen Test Bed experiment, which represents a prototypical sized cryogenic propellant storage tank of approximately 3 meters in diameter. This was achieved by first identifying the major heat transfer mechanisms in both fluid domains. Then, closure models and constitutive relations were implemented to model the internal flow, and interfacial heat and mass transfer. Results show the predicted pressure evolution of the tank agrees well with experimental data for both 50% fill level and 90% fill level cases. A discrepancy was observed in the vapor temperatures comparison with experimental data. The difference was attributed to the assumption of conduction within the vapor region due to a lack of correlations to model the heat transfer within the vapor region in the presence of thermal stratification.
Next, the CFD approach was used to simulate fast transients such as jet induced mixing during active pressure control. Due to the immense computational requirements and available experimental data, the tank pressure control experiments (TPCE) were identified as a suitable study. Simulation results show the higher fidelity CFD simulation can track the liquid vapor interface and resolve the internal flow using turbulence modeling. In addition, due to thermal considerations, the model was used to perform a parametric study to determine the jet Weber number for penetration of the ullage bubble.
The final part of the dissertation presents a coupling methodology developed to facilitate the simulation of a long-term self-pressurization process of a cryogenic propellant tank. The key highlight of this methodology is the development of a coupling scheme that employs a domain decomposition approach, effectively dividing the computational domains at the liquid-vapor interface. SINDA/FLUINT, the nodal code, is utilized to simulate the liquid region, while ANSYS Fluent, the CFD code, handles the vapor region. An algorithm was proposed to compute the required boundary conditions for the split domains from the local thermodynamic properties of the assumed infinitesimally thin interface. However, due to restricted access to either of the commercial source codes, the data exchange was facilitated through an external routine that computes the interfacial evaporation losses based on temperature and pressure values at the interface. The coupling between the nodal and CFD codes was demonstrated by simulating a self-pressurization process in a small size tank using hydrogen as the working fluid. The effectiveness of the coupling methodology was assessed by comparing the temperature and pressure evolution results from the coupled simulation with those obtained solely from CFD simulation. Additionally, a sensitivity study on the grid sizing and coupling time step was conducted to determine an appropriate spatial resolution to ensure a divergence-free explicit data exchange time step size. Results showed that the interface coupling scheme was successfully implemented, and the coupled simulation agreed well with the CFD simulations. Using the coupling approach, two numerical case studies were considered based on different initial fill levels of the tank to study the pressurization rate. To compare the required computational time, the computational costs associated with both approaches were compared. Lastly, the recommendations for future work was provided in order to improve the accuracy of nodal model and improve upon the current coupled modeling framework in terms of computational speed up and scalability.Ph
A Theoretically Grounded Question Answering Data Set for Evaluating Machine Common Sense
No full textAchieving machine common sense has been a longstanding problem within Artificial Intelligence. Thus far, benchmark data sets that are grounded in a theory of common sense and can be used to conduct rigorous, semantic evaluations of common sense reasoning (CSR) systems have been lacking. One expectation of the AI community is that neuro-symbolic reasoners can help bridge this gap towards more dependable systems with common sense. We propose a novel benchmark, called Theoretically Grounded common sense Reasoning (TG-CSR), modeled as a set of question answering instances, with each instance grounded in a semantic category of common sense, such as space, time, and emotions. The benchmark is few-shot i.e., only a few training and validation examples are provided in the public release to avoid the possibility of overfitting. Results from recent evaluations suggest that TG-CSR is challenging even for state-of-the-art statistical models. Due to its semantic rigor, this benchmark can be used to evaluate the common sense reasoning capabilities of neuro-symbolic systems
Flow physics and control for improved tailless vehicle aerodynamics via leading-edge vortex manipulation
No full textDecember 2023The use of active flow control using finite-span synthetic jet actuators to affect the aerodynamicloads on a generic tailless chined forebody delta wing was examined experimentally in multiple
stages performed in two different wind tunnels. First, the flowfield around the generic model
having a chined forebody and a simple delta wing configuration was measured at different angles
of attack, with or without a yaw angle. This was done using Oil Flow Visualizations (OFV) and
Stereoscopic Particle Image Velocimetry (SPIV) at a mean chord-based Reynolds number of 2.3 x
10^5. The detailed flowfield measurements, using SPIV, were conducted for the cases where the
angles of attack were 20° and 30°, and yaw angles of 0° and 5°. The flowfield over the model was
seen to exhibit pairs of leading edge vortices similar to the flowfield around a double delta wing
except for the influence of the fuselage, particularly in the forebody region where the chine
vortex formed over the convex portion and followed its curvature. The development and
interaction of the chine and wing vortices were measured and analyzed. It was found that the
downstream evolution of these vortices and their interaction depended on the angle of attack
and yaw angle. Increasing the angle of attack resulted in wake-like vortices while increasing the
yaw angle yielded a wake-like vortex on the windward side and a jet-like vortex on the leeward
side. The interaction of the windward side vortex with the physical barrier of the forebody surface
was observed to greatly affect its behavior and the interaction downstream. In some
combinations of angle of attack and yaw angle, the merged vortex exhibited a breakdown. The
analysis of the flowfield and vortex dynamics served to provide insight into manipulating the
flowfield using physics-based flow control.
Based on the full-model flowfield, the effect of flow control was examined using a half-model
with removable forebodies, each of which was equipped with a pair of synthetic jets located as
close as possible to the leading edge. Three different jet orientations were explored, one
employing surface-normal SJs, one employing horizontal synthetic jets, and one employing SJs
Angled 45° away from the leading edge. Apart from a baseline for comparison taken with the jets'
orifices covered, six cases were explored with actuation, one with each jet individually actuated
and one with both jets actuated together with and without the pulse modulation at the helical
mode frequency. In all cases, the synthetic jets were activated with ?b = 1.667 (?mu = 7.15 ∗
10^-5 per jet). Tuft visualizations were used to qualitatively compare the flowfield with that of
the full model. Then, aerodynamic load measurements were conducted to explore the effects of
flow control. These measurements were followed by detailed flow field measurements using
SPIV to shed light on the reasons for these effects. It was found that the surface-normal jets had
a much larger effect on the aerodynamic coefficients, especially the lift, than the other two
orientations. The maximum calculated increase in drag for the surface-normal jets was around
16% from ? = 28° to ? = 32° whereas modest, single-digit percentages in lift occur for the other
two orientations. Therefore, the chined forebody with the surface normal synthetic jets was
chosen for detailed flow measurements. Since the increased lift is a consequence of increased
vortex lift, it was accompanied by an increase in vortex drag. The reason for the difference
between the jet orientations was also investigated. It was observed that, despite causing a local
increase in the forebody vortex circulation, the horizontal jets blew close enough to the chine
that they negatively affected the feeding of the vortex by the shear layer, causing a
low-velocity region over the leading edge that decreased the vortex lift. On the other hand, the
upwards-oriented jets presented the opposite behavior, blowing too close to the vortex core and
not strengthening the vortex to the same degree. This also caused the SJ to impinge on the core,
pushing it away from the surface and also decreasing the vortex lift. The surface-normal jets
yielded the largest performance enhancement by adding circulation to the vortex while
simultaneously reducing its distance to the surface and the leading edge. The jets acted in a quasisteady
manner, and since the chine vortex acts as an oscillator at the helical mode frequency, the
interaction between the jet and the chine vortex locks to this frequency, causing the effects to
superpose. The enhanced chine vortex induced a larger velocity on the wing vortex, causing an
earlier merge of the two vortices, and also a more jet-like merged vortex downstream. This
resulted in an increase in nose-down pitching moment, increased lift, and increased drag, all of
which are desired for takeoff and landing.
In addition, the synthetic jets were actuated using a pulse-modulated waveform, where the
modulation frequency was near the helical mode frequency. This made the actuation much more
energy-efficient as, opposite to the non-modulated actuation, the pulse modulation near the
helical mode frequency caused the helical mode to lock onto the actuation frequency. This meant
the addition of momentum from the jets to the vortex is approximately the same as without
pulse modulation, despite the jets being off during parts of the actuation cycle. Furthermore,
when using pulse modulation, actuating using either jet (or both jets) produced the same overall
result. This was because the excited helical mode is a global instability mode, and the slightly
different location of the jets is unimportant to the overall effect of the jets. Performing triple
decomposition of the velocity vectors to estimate the vorticity transport equation using the timeresolved
and phase-averaged data showed that this addition of circulation, core velocity, and
vortex lift primarily affected the mean values. The spectral behavior of the modes associated with
the vortical motion was seen to be unaffected beyond the appearance of a peak at the pulse
modulation frequency and the actuation decreased the wandering.
In parallel, the flowfield associated with the interaction of a synthetic jet with an isolated induced
vortex over a flat plate was explored to aid in understanding the mechanisms at play in the chined
forebody-delta wing model. It was seen that the train of vortex rings produced by the SJ at
skewed and pitched angles to the crossflow broke down into smaller-scale structures that
interact with each other to generate a single streamwise vortex downstream. The vortical
structures generated by the synthetic jet were seen to strengthen the induced streamwise vortex
(generated by a vortex generator) passing over the jet at the correct location and distance. In
addition, it was seen that a streamwise vortex passing over the middle of the angled jet could be
pulled closer to the wall by the induced velocities caused by the vortex rings passing under it.
This was also seen over the second jet on the actuated chined forebody.
The present work focused on active flow control via synthetic jets. Future work can explore the
use of passive control via surface-mounted, low-aspect-ratio cantilevered circular pins.
Therefore, the flowfield associated with chamfered pins when immersed in a laminar boundary
layer was analyzed. These pins were originally considered for implementation into the chined
forebody as an option to the finite-span synthetic jets but were not selected in the end. Two
chamfered pins, where the chamfer encompassed either half of the pin's planform or its full
planform, were analyzed with the chamfer at various skew angles with respect to the freestream
and were compared to a pin without a chamfer. All pins exhibited a complex flowfield, including
an array of streamwise vortical structures. The chamfered pins resulted in two additional
counter-rotating streamwise vortices, named Chamfered Induced Vortices (CIVs). It was shown
that changing the skew angle resulted in a change in the strength of these vortical structures,
their direction of rotation, and as a result, net circulation produced. Comparing the two
chamfered pins, the pin where the chamfer encompassed half of its planform produced stronger
CIVs. These effects are discussed in detail to provide insight into a future use of these pins as flow
control devices.Ph
Social Convos: Capturing Agendas and Emotions on Social Media
No full textSocial media platforms are popular tools for disseminating targeted information during major public events like elections or pandemics. Systematic analysis of the message traffic can provide valuable insights into prevailing opinions and social dynamics among different segments of the population. We are specifically interested in influence spread, and in particular whether more deliberate influence operations can be detected. However, filtering out the essential messages with telltale influence indicators from the extensive and often chaotic social media traffic is a major challenge. In this paper we present a novel approach to extract influence indicators from messages circulating among groups of users discussing particular topics. We build upon the concept of a convo to identify influential authors who are actively promoting some particular agenda around that topic within the group. We focus on two influence indicators: the (control of) agenda and the use of emotional language