Search NASA⌕ Search

SEARCH · Search NASA

Results for “Predictive analytics”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,261 records · Page 70

Facilitating better and faster simulations of aerosol-cloud interactions in Earth system models

Focal Area(s): 1. Predictive modeling through the use of AI techniques and AI-derived model components; the use of AI and other tools to design a prediction system comprising a hierarchy of models. 2. Insight gleaned from complex data (both observed and simulated) using AI, big data analytics, and other advanced methods, including explainable AI and physics- or knowledge-guided AI. Science Challenge: One major challenge that Earth system models (ESMs) face in providing credible prediction of the Earth system and its water cycle characteristics (e.g., mean state, variability, and extreme events) is to accurately simulate aerosol-cloud interactions (ACI). The physical, chemical, and dynamical processes affecting ACI are extremely complex and they range from nanoscale to planetary scale. In each model development cycle, scientists spend significant efforts investigating model deficiencies and uncertainties associated with aerosols (e.g., emissions, chemical processes, aerosol microphysics, and transport) and clouds (e.g., macrophysics, microphysics, turbulence, and large-scale circulation) in order to develop improved treatments. However, despite decades of active research, ACI is still a major source of uncertainty in climate projections, even though great progress has been made. Specific scientific challenges include: (i) Parameterizations are developed based on limited data; (ii) The complexity of a parameterization required for accurate predictions is not understood; (iii) Incomplete and unknown physics leads to errors in the fully coupled Earth system; and (iv) Complex physics is computationally too expensive to employ in ESMs.

54 ENVIRONMENTAL SCIENCES↗

Analytic Theory for the Dynamics of Wide Quantum Neural Networks

Here, parametrized quantum circuits can be used as quantum neural networks and have the potential to outperform their classical counterparts when trained for addressing learning problems. To date, much of the results on their performance on practical problems are heuristic in nature. In particular, the convergence rate for the training of quantum neural networks is not fully understood. Here, we analyze the dynamics of gradient descent for the training error of a class of variational quantum machine learning models. We define wide quantum neural networks as parametrized quantum circuits in the limit of a large number of qubits and variational parameters. Then, we find a simple analytic formula that captures the average behavior of their loss function and discuss the consequences of our findings. For example, for random quantum circuits, we predict and characterize an exponential decay of the residual training error as a function of the parameters of the system. Finally, we validate our analytic results with numerical experiments.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Envisioning U.S. Climate Predictions and Projections to Meet New Challenges

In the face of a changing climate, the understanding, predictions, and projections of natural and human systems are increasingly crucial to prepare and cope with extremes and cascading hazards, determine unexpected feedbacks and potential tipping points, inform long-term adaptation strategies, and guide mitigation approaches. Increasingly complex socio-economic systems require enhanced predictive information to support advanced practices. Such new predictive challenges drive the need to fully capitalize on ambitious scientific and technological opportunities. These include the unrealized potential for very high-resolution modeling of global-to-local Earth system processes across timescales, reduction of model biases, enhanced integration of human systems and the Earth Systems, better quantification of predictability and uncertainties; expedited science-to-service pathways, and co-production of actionable information with stakeholders. Enabling technological opportunities include exascale computing, advanced data storage, novel observations and powerful data analytics, including artificial intelligence and machine learning. Looking to generate community discussions on how to accelerate progress on U.S. climate predictions and projections, representatives of Federally-funded U.S. modeling groups outline here perspectives on a six-pillar national approach grounded in climate science that builds on the strengths of the U.S. modeling community and agency goals. This calls for an unprecedented level of coordination to capitalize on transformative opportunities, augmenting and complementing current modeling center capabilities and plans to support agency missions. Tangible outcomes include projections with horizontal spatial resolutions finer than 10 km, representing extremes and associated risks in greater detail, reduced model errors, better predictability estimates, and more customized projections to support next generation climate services.

54 ENVIRONMENTAL SCIENCES↗

Modeling Marangoni Convection using a Velocity-Time Relationship

The Waste Treatment and Immobilization Plant (WTP) at the U.S. Department of Energy’s (DOE’s) Handford site is designed to treat 56 million gallons of radioactive waste through vitrification. The melting vessels physical and chemical integrity is essential to maintain, as failure has a high safety and financial penalty. Monofrax® K-3 refractory, the material widely used for the melting vessel, undergoes significant corrosion over time, concentrated at the triple-point junction between the gaseous atmosphere, refractory, and glass melt, producing a characteristic neck-like profile that significantly limits refractory service life. Accurate prediction of the corrosion within the melter can be used to prevent this and extend the lifetime of the melters. A computational fluid dynamics (CFD) model was previously developed, and predicted subsurface refractory corrosion, but failed to predict the neck-line profile observed experimentally. To address this limitation, the glass-melt meniscus geometry at the glass-refractory interface was estimated from experimentally measured neck profiles and wettability experiments. A curve was fitted to the neck profile to identify the triple point location, and the meniscus shape was characterized by a contact angle of 1°. A velocity profile accounting for Marangoni convection-driven corrosion was derived from the estimated meniscus geometry. Surface tension measurements from LORPM14R1 glasses of varying composition were coupled with K-3 coupon dissolution rates to establish a surface tension–time relationship. This analytical velocity-time model was integrated into the existing CFD framework to improve prediction of the neck corrosion profile. Work was done in a computational setting with national collaboration, resulting in professional development through presentational and technical writing growth.

36 - MATERIALS SCIENCE↗

Modeling and analysis of fission product transport in the AGR-3/4 experiment - 18693

In this work we describe the ongoing modeling and analysis efforts in support of the AGR-3/4 experiment. AGR-3/4 is intended to provide data to assess fission product retention and transport (e.g., diffusion coefficients) in fuel matrix and graphite materials. We describe a set of pre-test predictions that incorporate the results of detailed thermal and fission product release models into a coupled 1D radial diffusion model of the experiment, using diffusion coefficients reported in the literature for Ag, Cs, and Sr. We make some comparisons of the predicted Cs profiles to preliminary measured data and find these to be reasonable, in most cases within an order of magnitude. Our ultimate objective is to refine the diffusion coefficients using AGR-3/4 data, so we develop an analytical method for doing so and demonstrate its efficacy via a series of numerical experiments using the model predictions. (authors)

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Is Radioactive Tank Waste a Bingham Fluid?

Slurries and sludges across the United States Department of Energy (DOE) complex rank among the most rheologically interesting. Their composition is heterogeneous, spanning a very broad range of particle sizes, densities, and interparticle forces. All exhibit shear thinning, some have yield stresses, and many are thixotropic. Despite the variety, these complex fluids are often represented using the historic Bingham fluid model, which fits higher shear rate data to a simple straight line. The intercept provides a yield stress, which has been a key design parameter in construction of large-scale waste processing facilities. However, many radioactive wastes are simply not Bingham fluids, which extrapolate but poorly across low to intermediate shear rates that are characteristic of typical processing conditions. Indeed, processing shear rates as high as 200 1/s, which has been a typical minimum shear rate used in fitting, are seldom encountered in nuclear waste processing. Therefore, more realistic rheological models are necessary to accurately predict waste processing performance. PNNL recently re-evaluated the rheology of reconstituted Hanford REDOX sludge waste against a wide variety of rheological models including the Bingham, Cross, Cross with yield stress, Carreau, biviscous, Herschel-Bulkley (which includes a power law dependence), Casson, and Gay models. They found that all of the models provided a closer fit than the Bingham model and that the biviscous model and Cross with yield stress model were convincing. However, reconstituted Hanford REDOX sludge waste is but one type of DOE waste and a direct contrast and comparison of these three models against undiluted, unmixed tank waste (actual not simulant) has not been performed previously. Therefore, the purpose of this paper is to evaluate the rheology of actual tank waste with these more accurate rheological models. In this paper, we evaluate select rheological data for slurry samples from Hanford’s AZ-101, AZ-102, and SY-101 waste tanks. In each of these cases, we find that Cross’ model with yield stress and the biviscous model significantly outperform the Bingham fluid model. Furthermore, the AZ-101 data also shows that the shear stress peak that startup significantly exceeds the Bingham yield stress, which is commonly observed in the incipient moments of rheological measurements on simulants. In contrast, Cross’ model may empirically accommodate an initial spike in shear stress at modest shear rates. These are important observations because computational and analytical fluid dynamics simulations rely on rheological constitutive models for accurately and conservatively predict waste processing performance. These findings suggest the need for better rheological modeling of and validation against radioactive waste.

Rheology, Radioactive Waste↗

Analytic Expressions for Extracting Far Scrape-Off Layer Transport Coefficients from the Plasma-Wall Interaction of Divertor Tokamaks

Magnetic-confinement DT pilot plants and DEMO-class devices will employ thin breeding blanket first walls with substantially lower steady-state power-handling margins than ITER-like thick walls, making reliable prediction of main-chamber particle and heat loads increasingly critical. This report addresses that need by developing analytic tools to interpret far scrape-off layer (far-SOL) plasma–wall interaction measurements in divertor tokamaks, with the goal of extracting effective cross-field transport coefficients from “window-frame” experiments and related diagnostics.

Stangeby, P. C. [Univ. of Toronto, ON (Canada)]↗

Report of improved performance in Talbot–Lau phase-contrast computed tomography

Purpose: Many expectations have been raised since the use of conventional x-ray tubes on grating-based x-ray phase-contrast imaging. Despite a reported increase in contrast-to-noise ratio (CNR) in many publications, there is doubt on whether phase-contrast computed tomography (CT) is advantageous in clinical CT scanners in vivo. The aim of this paper is to contribute to this discussion by analyzing the performance of a phase-contrast CT laboratory setup. Methods: A phase-contrast CT performance analysis was done. Projection images of a phantom were recorded, and image slices were reconstructed using standard filtered back projection methods. The resulting image slices were analyzed by determining the CNRs in the attenuation and phase image. These results were compared to analytically calculated expectations according to the already published phase-contrast CT performance analysis by Raupach and Flohr [Med. Phys. 39, 4761–4774 (2012)]. There, a severe mistake was found leading to wrong predictions of the performance of phase-contrast CT. The error was corrected and with the new formulae, the experimentally obtained results matched the analytical calculations. Results: The squared ratios of the phase-contrast CNR and the attenuation CNR obtained in the authors’ experiment are five- to ten-fold higher than predicted by Raupach and Flohr [Med. Phys. 39, 4761–4774 (2012)]. The effective lateral spatial coherence length deduced outnumbers the already optimistic assumption of Raupach and Flohr [Med. Phys. 39, 4761–4774 (2012)] by a factor of 3. Conclusions: The authors’ results indicate that the assumptions made in former performance analyses are pessimistic. The break-even point, when phase-contrast CT outperforms attenuation CT, is within reach even with realistic, nonperfect gratings. Further improvements to state-of-the-art clinical CT scanners, like increasing the spatial resolution, could change the balance in favor of phase-contrast computed tomography even more. This could be done by, e.g., quantum-counting pixel detectors with four-fold smaller pixel pitches.

60 APPLIED LIFE SCIENCES↗

Analytical ab initio hessian from a deep learning potential for transition state optimization

Identifying transition states—saddle points on the potential energy surface connecting reactant and product minima—is central to predicting kinetic barriers and understanding chemical reaction mechanisms. In this work, we train a fully differentiable equivariant neural network potential, NewtonNet, on thousands of organic reactions and derive the analytical Hessians. By reducing the computational cost by several orders of magnitude relative to the density functional theory (DFT) ab initio source, we can afford to use the learned Hessians at every step for the saddle point optimizations. We show that the full machine learned (ML) Hessian robustly finds the transition states of 240 unseen organic reactions, even when the quality of the initial guess structures are degraded, while reducing the number of optimization steps to convergence by 2–3× compared to the quasi-Newton DFT and ML methods. All data generation, NewtonNet model, and ML transition state finding methods are available in an automated workflow.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Physiologically-based pharmacokinetic model for Fentanyl in support of the development of Provisional Advisory Levels

Provisional Advisory Levels (PALs) are tiered exposure limits for toxic chemicals in air and drinking water that are developed to assist in emergency responses. Physiologically-based pharmacokinetic (PBPK) modeling can support this process by enabling extrapolations across doses, and exposure routes, thereby addressing gaps in the available toxicity data. Here, we describe the development of a PBPK model for Fentanyl – a synthetic opioid used clinically for pain management – to support the establishment of PALs. Starting from an existing model for intravenous Fentanyl, we first optimized distribution and clearance parameters using several additional IV datasets. We then calibrated the model using pharmacokinetic data for various formulations, and determined the absorbed fraction, F, and time taken for the absorbed amount to reach 90% of its final value, t90. For aerosolized pulmonary Fentanyl, F = 1 and t90 < 1 min indicating complete and rapid absorption. The F value ranged from 0.35 to 0.74 for oral and various transmucosal routes. Oral Fentanyl was absorbed the slowest (t90 ∼ 300 min); the absorption of intranasal Fentanyl was relatively rapid (t90 ∼ 20–40 min); and the various oral transmucosal routes had intermediate absorption rates (t90 ∼ 160–300 min). Based on these results, for inhalation exposures, we assumed that all of the Fentanyl inhaled from the air during each breath directly, and instantaneously enters the arterial circulation. We present model predictions of Fentanyl blood concentrations in oral and inhalation scenarios relevant for PAL development, and provide an analytical expression that can be used to extrapolate between oral and inhalation routes for the derivation of PALs. - Highlights: • We develop a Fentanyl PBPK model for relating external dose to internal levels. • We calibrate the model to oral and inhalation exposures using > 50 human datasets. • Model predictions are in good agreement with the available pharmacokinetic data. • The model can be used for extrapolating across routes, doses and exposure durations. • We illustrate how the model can be used for developing Provisional Advisory Levels.

60 APPLIED LIFE SCIENCES↗

SuperCDMS Background Models for Low-Mass Dark Matter Searches

An abundance of astrophysical and cosmological evidence indicates the existence of a non-luminous, non-baryonic form of matter, called dark matter, that is approximately a quarter of all energy in the universe. One promising candidate for dark matter is the Weakly Interacting Massive Particle (WIMP) which interacts with baryonic matter at most on the scale of the weak force. The Cryogenic Dark Matter Search (CDMS) experiment aims to detect the nuclear recoils induced by the elastic scattering of WIMPs off of germanium nuclei. This is a rare signal and difficult to detect, especially the low-energy recoils that are produced by low-mass dark matter. The CDMS project operated at the Soudan Underground Laboratory from 2003--2015, with an upgrade to the SuperCDMS experiment in 2012. The germanium detectors were operated at 50~mK and able to measure both the ionization and athermal phonons produced in a particle interaction. Measuring two signals enables discrimination between electron recoil and nuclear recoil events. An alternative operating mode for the detectors is called the CDMS low ionization threshold experiment (CDMSlite), where a higher bias was applied to the detectors and only the phonon signal analyzed. This method increased sensitivity to low-mass dark matter interactions, but sacrificed discrimination capability. The CDMSlite spectrum had a large contribution from electron recoil background events. From the information gained during the first two CDMSlite Runs, a background model was developed for the third and final CDMSlite Run. Analytical descriptions were identified for those backgrounds that were theoretically known, e.g. tritium $\beta$-spectrum, and Geant simulations were used to understand and predict the low-energy spectra from other sources, e.g. Compton scattering. Multiple new models were developed for detectors operated in CDMSlite at Soudan. These include the analytical formula for Compton scattering, and empirical models for surface backgrounds from $^{210}$Pb contamination of the germanium crystals and detector housing. In order to accurately describe the surface events, a new detector response function was developed that included information about the electric field and energy resolution of the detector. These models were essential to the implementation of a profile likelihood analysis of the CDMSlite Run 3 data, which improved on the sensitivity to dark matter over the Run 2 optimum interval analysis for WIMP masses above 2.5~GeV/$c^2$. This demonstrated a successful application of a likelihood analysis to the high-voltage operating mode, and the potential for these analyses in the future SuperCDMS SNOLAB experiment. For the SuperCDMS SNOLAB experiment, the change in background rates from radiogenic neutrons was considered as additional towers of detectors were added, and the feasibility of an active neutron veto as a potential upgrade for large payloads was studied. This veto could be constructed of plastic scintillator with layers of gadolinium resin, and would aid in reducing the nuclear recoil single scatter background that is indistinguishable from the WIMP signal.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Multiscale modeling and nonlinear model predictive control for flue gas desulfurization

The primary source of sulfur dioxide (SO 2 ) emissions is flue gas from fossil fuels-based power plants. SO 2 emissions are known to not only cause health issues, but also have an adverse effect on the environment in various ways. Several Flue Gas Desulfurization (FGD) technologies have been incorporated in power plants. The most popular technology is Wet FGD, where a limestone slurry is used to absorb SO 2 from the flue gas. A detailed droplet scale model describing the instantaneous and finite rate chemistry is developed. The ill-posed Differential-Algebraic Equation (DAE) droplet model is reformulated to a well-posed index-1 DAE through index reduction. The droplet model is integrated with the bulk phase by incorporating gas-liquid mass transfer, and an oxidation reactor model to simulate the dynamic operation of the counter-current spray scrubber. As a result, the model has a well-conditioned Jacobian and overcomes the modeling challenges of previous works and enables numerical solution without requiring carefully selected initialization or specialized solution procedures. The model is successfully validated using power plant measurements, and nonlinear model predictive control (NMPC) studies are demonstrated to optimize recycle stream flowrates to minimize pumping costs.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Derivation of a crack opening deflection relationship for fibre reinforced concrete panels using a stochastic model: Application for predicting the flexural behaviour of round panels using stress crack opening diagrams

This study is aimed at proposing a simple analytical model to investigate the post-cracking behaviour of FRC panels, using an arbitrary tension softening, stress crack opening diagram, as the input. A new relationship that links the crack opening to the panel deflection is proposed. Due to the stochastic nature of material properties, the random fibre distribution, and other uncertainties that are involved in concrete mix, this relationship is developed from the analysis of beams having the same thickness using the Monte Carlo simulation (MCS) technique. The softening diagrams obtained from direct tensile tests are used as the input for the calculation, in a deterministic way, of the mean load displacement response of round panels. A good agreement is found between the model predictions and the experimental results.

36 MATERIALS SCIENCE↗

Matisse: A Visual Analytics System for Exploring Emotion Trends in Social Media Text Streams

Dynamically mining textual information streams to gain real-time situational awareness is especially challenging with social media systems where throughput and velocity properties push the limits of a static analytical approach. In this paper, we describe an interactive visual analytics system, called Matisse, that aids with the discovery and investigation of trends in streaming text. Matisse addresses the challenges inherent to text stream mining through the following technical contributions: (1) robust stream data management, (2) automated sentiment/emotion analytics, (3) interactive coordinated visualizations, and (4) a flexible drill-down interaction scheme that accesses multiple levels of detail. In addition to positive/negative sentiment prediction, Matisse provides fine-grained emotion classification based on Valence, Arousal, and Dominance dimensions and a novel machine learning process. Information from the sentiment/emotion analytics are fused with raw data and summary information to feed temporal, geospatial, term frequency, and scatterplot visualizations using a multi-scale, coordinated interaction model. After describing these techniques, we conclude with a practical case study focused on analyzing the Twitter sample stream during the week of the 2013 Boston Marathon bombings. The case study demonstrates the effectiveness of Matisse at providing guided situational awareness of significant trends in social media streams by orchestrating computational power and human cognition.

data visualization↗

A semi-analytical model for laser induced densification in fused silica at elevated temperatures

A semi-analytical model for deterministic surface smoothing of fused silica below the evaporation temperature threshold is presented. Using this model, a method for calculating deterministic track depths for laser polishing fused silica is derived and validated with experiments. In this work, model predictions were in good agreement with experiment across varying laser preheating temperature and power. With the semi-analytical method presented here, deterministic laser smoothing can be achieved in the densification regime to close the process loop of laser polishing.

36 MATERIALS SCIENCE↗

MODELING AND ANALYSIS OF FISSION PRODUCT TRANSPORT IN THE AGR-3/4 EXPERIMENT

In this work we describe the ongoing modeling and analysis efforts in support of the AGR-3/4 experiment. AGR-3/4 is intended to provide data to assess fission product retention and transport (e.g., diffusion coefficients) in fuel matrix and graphite materials. We describe a set of pre-test predictions that incorporate the results of detailed thermal and fission product release models into a coupled 1D radial diffusion model of the experiment, using diffusion coefficients reported in the literature for Ag, Cs, and Sr. We make some comparisons of the predicted Cs profiles to preliminary measured data for Cs and find these to be reasonable, in most cases within an order of magnitude. Our ultimate objective is to refine the diffusion coefficients using AGR-3/4 data, so we identify an analytical method for doing so and demonstrate its efficacy via a series of numerical experiments using the model predictions. Finally, we discuss development of a post-irradiation examination plan informed by the modeling effort and simulate some of the heating tests that are tentatively planned.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

The L-CAPE Project at FNAL

The controls system at FNAL records data asynchronously from several thousand Linac devices at their respective cadences, ranging from 15Hz down to once per minute. In case of downtimes, current operations are mostly reactive, investigating the cause of an outage and labeling it after the fact. However, as one of the most upstream systems at the FNAL accelerator complex, the Linac’s foreknowledge of an impending downtime as well as its duration could prompt downstream systems to go into standby, potentially leading to energy savings. The goals of the Linac Condition Anomaly Prediction of Emergence (L-CAPE) project that started in late 2020 are (1) to apply data-analytic methods to improve the information that is available to operators in the control room, and (2) to use machine learning to automate the labeling of outage types as they occur and discover patterns in the data that could lead to the prediction of outages. We present an overview of the challenges in dealing with time-series data from 2000+ devices, our approach to developing an ML-based automated outage labeling system, and the status of augmenting operations by identifying the most likely devices predicting an outage.

43 PARTICLE ACCELERATORS↗