SEARCH · Engineering Papers

Results for “Constraint Reduction”

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.

SourceResearch connectionFrom yearThrough year

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

At least 181 records · Page 10

All sources NASA NTRS DOE OSTI

NASA NTRS1998

Comparison of High Aspect Ratio Cooling Channel Designs for a Rocket Combustion Chamber with Development of an Optimized Design

An analytical investigation on the effect of high aspect ratio (height/width) cooling channels, considering different coolant channel designs, on hot-gas-side wall temperature and coolant pressure drop for a liquid hydrogen cooled rocket combustion chamber, was performed. Coolant channel design elements considered were: length of combustion chamber in which high aspect ratio cooling was applied, number of coolant channels, and coolant channel shape. Seven coolant channel designs were investigated using a coupling of the Rocket Thermal Evaluation code and the Two-Dimensional Kinetics code. Initially, each coolant channel design was developed, without consideration for fabrication, to reduce the hot-gas-side wall temperature from a given conventional cooling channel baseline. These designs produced hot-gas-side wall temperature reductions up to 22 percent, with coolant pressure drop increases as low as 7.5 percent from the baseline. Fabrication constraints for milled channels were applied to the seven designs. These produced hot-gas-side wall temperature reductions of up to 20 percent, with coolant pressure drop increases as low as 2 percent. Using high aspect ratio cooling channels for the entire length of the combustion chamber had no additional benefit on hot-gas-side wall temperature over using high aspect ratio cooling channels only in the throat region, but increased coolant pressure drop 33 percent. Independent of coolant channel shape, high aspect ratio cooling was able to reduce the hot-gas-side wall temperature by at least 8 percent, with as low as a 2 percent increase in coolant pressure drop. ne design with the highest overall benefit to hot-gas-side wall temperature and minimal coolant pressure drop increase was the design which used bifurcated cooling channels and high aspect ratio cooling in the throat region. An optimized bifurcated high aspect ratio cooling channel design was developed which reduced the hot-gas-side wall temperature by 18 percent and reduced the coolant pressure drop by 4 percent. Reductions of coolant mass flow rate of up to 50 percent were possible before the hot-gas-side wall temperature reached that of the baseline. These mass flow rate reductions produced coolant pressure drops of up to 57 percent.

Wadel, Mary F.↗

DOE OSTI2021

Concentrated radiative cooling and its constraint from reciprocity

Concentrated radiative cooling, an analogous concept of the concentrated solar power technology, has the potential of amplifying both the cooling power and the temperature reduction. However, concentrators have not yet been systematically optimized. Moreover, a widely used theoretical approach to analyze such systems has neglected a fundamental constraint from reciprocity, which can lead to an overestimate of cooling performance and unclarified limits of amplification factors. Here we develop a theoretical framework addressing these shortcomings. Modeling suggests the optimized shape and geometric dimensions of concentrators, as well as the limiting cooling power and temperature reduction. Using an electroplated Al 2 O 3 emitter and an optimized conical concentrator, we experimentally amplify the nighttime radiative cooling by 26%.

Dong, Minghao↗

NASA NTRS1967

Electric propulsion system reliability.

Graphic optimization technique for mass reduction in solar powered ion propulsion system with acceptable reliability constraint

WEIGHT FACTOR↗

NASA NTRS1984

How to find galaxies at high redshift

The UV and optical spectra presented for the blue star-forming galaxies Mrk 357, Mrk 496, and Mrk 26 have sufficiently large redshifts to make Lyman-alpha observations possible. Weak Lyman-alpha emission is observed in Mrk 357 and Mrk 469, with a flux value comparable to that of H-beta, and no detectable Lyman-alpha emission is detected in the case of Mrk 26. All three objects exhibit strong UV continuum emission, which is characteristic of hot stars, and although little evidence of dust is found, the observational constraints on extinction and neutral hydrogen column densities allow substantial reductions of the Lyman-alpha flux by dust absorption in conjunction with large line optical depths. These results imply that primeval galaxy detection calls for attention to be given to redshifted continuum emission from massive galaxies.

Hartmann, L. W.↗

NASA NTRSDate not supplied

Gateway at the Crossroads of Sustainable Lunar Exploration

The Gateway Program has made substantial design and development progress toward delivering a small, human-tended lunar space station purposefully designed to enable sustainable human exploration. The Program integrates partners and providers organizationally and physically as part of the spacecraft. The Power and Propulsion Element (PPE) and the Habitation and Logistics Outpost (HALO) with the European System Providing Refueling, Infrastructure and Telecommunications (ESPRIT) HALO Lunar Communications System (HLCS) have begun manufacturing the long lead components and will be launched first as a Co-Manifested Vehicle (CMV). The International Habitat (I-Hab) and ESPRIT Refueling Module (ERM) are passing life cycle milestones and include capabilities key for human crewmembers, such as windows, private sleeping quarters, and galley functions. The Logistics Module (LM) may provide a variety of services to Gateway depending on each mission. Requirements for the airlock have been developed, including requests that it support the integrated spacecraft with functions like augmenting heat rejection capabilities, and interfaces with new spacesuits will soon be developed in more detail. As a critical element of the architecture for solar system exploration, Gateway implements key tenets and features of international interoperability standards necessary to operate with multiple visiting vehicles and lunar assets, especially avionics, communications, and docking. Specific choices such as software architecture and standards, power standards, and robotics standards make it possible to utilize heritage or proprietary technology, yet still operate as one spacecraft. Engineering teams are evaluating many possible future missions to be executed at or utilizing the Gateway. The system architecture protects for an evolvable, extensible, and flexible capability. Designing systems robust enough to serve as a cornerstone of exploration activities for decades while remaining adaptable is not without its challenges. The detailed integration activities have revealed challenges and the need to mature key technologies. Refueling is a key component of achieving long life for Gateway, with unique operations to plan, safety concerns to mitigate, and risk reduction activities to conduct to better understand the system. The constraints and impacts of the design of visiting vehicles is also an important concern, with orientation constraints, control of attitude and orbit of the Gateway with docked visiting vehicles. Tradeoffs between robust maintainable systems and lightweight, compact systems must be balanced. Opportunities still exist for adding additional advanced capabilities to increase and extend Gateway’s benefits, such as intravehicular robotics, autonomous Guidance Navigation and Control (GN&C), and augmented control propulsion, heat rejection, or other services.

Molly S Anderson↗

NASA NTRSDate not supplied

Gateway at the Crossroads of Sustainable Lunar Exploration

Molly S Anderson↗

DOE OSTI2020

Hadronic contributions to (g - 2) µ

The Muon g-2 Experiment at Fermilab, which recently started running, plans to reduce the uncertain- ties on the already very precisely measured anomalous magnetic moment of the muon by a factor of four. The goal of this effort is to probe the observed difference of more than three standard deviations between Standard-Model theory and experiment, one of the few persistent hints for physics beyond the Standard Model. The Fermilab experiment collected data from their first run last year with statistics comparable to BNL E821. They expect to release the measurement result in 2019. On the theoretical side, because the muon g - 2 arises from quantum- mechanical loop contributions in the Standard Model, it is sensitive to virtual effects of new particles, and places important constraints on Standard-Model extensions. To leverage the anticipated reduction in experimental errors, and determine unambiguously whether or not new-physics effects contribute to this quantity, the theoretical errors must be made more reliable and reduced to a commensurate precision. The Muon g-2 Theory Initiative was created to facilitate this development. The dominant sources of uncertainty in the Standard-Model prediction of the muon g -2 are from the hadronic contributions. The hadronic vacuum polarization (HVP) provides the leading correction followed by hadronic light-by-light (HLbL) scattering. There are a number of complementary theoretical efforts underway to better understand and quantify these contributions, including dispersive and data driven methods, lattice QCD, and effective field theories. Given the precision goals and the phenomenological importance, it is important to have more than one independent method for each of the two hadronic corrections, each with fully quantified uncertainties. Fostering the development of such methods is a prime goal of the initiative, as this will enable critical cross checks, and, upon combination, may yield gains in precision, to maximize the impact of E989. An important aspect of the Muon g-2 Theory Initiative’s activities are providing platforms that facilitate interactions between the different groups, as well as between the theoretical and experimental g - 2 communities. To this end, several workshops were organized in 2017 and 2018. The first meeting, held at Fermilab (June 3–6, 2017, St. Charles, IL, USA), served to kick-off the Initiative’s activities. Two meetings in early 2018 were focused respectively on the HVP and HLbL corrections. The HVP meeting was held at KEK (February 12–14, 2018, Tsukuba, Japan) and the HLbL meeting at the University of Connecticut (March 12–14, 2018, Storrs, CT, USA). The most recent workshop, which served as the second plenary meeting of the Initiative, was held at the University of Mainz (June 18–22, 2018, Mainz, Germany). An important outcome of these meetings are concrete plans for a first white paper, which is currently being written. We aim to post the white paper just prior to the release of the first E989 measurement, to present a clean theoretical prediction. The first white paper is focused on assessing and improving the reliability of the SM prediction. The purpose of the INT workshop in September 2019 is to start the next stage of the Initiative, focusing on the development of strategies to improve the theory uncertainties beyond the current level towards the E989 precision goal. We aim to accelerate theoretical developments on the hadronic contributions to the muon g - 2 so that the Standard-Model theory error can be brought to the needed precision, again in advance of the next release from E989. Hence the workshop will provide crucial theory support for a US experiment with broad impact. The Muon g-2 Theory Initiative relies on input from representatives of all the different communities that are engaged in this effort. It is therefore important that all these areas are properly represented. The funds from this grant will be used to enable more people, especially early-career scientists, to participate and make essential contributions to the workshop discussions. Since the workshop’s main goal is to kick-off the next stage of the theory initiative’s activities, support for this workshop from the DOE will help the theory community provide crucial theoretical support to the Fermilab Muon g-2 Experiment.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2015

TU-AB-303-01: A Feasibility Study for Dynamic Adaptive Therapy of Non-Small Cell Lung Cancer

Purpose: To compare plans for NSCLC optimized using Dynamic Adaptive Therapy (DAT) with conventional IMRT optimization. DAT adapts plans based on changes in the target volume by using dynamic programing techniques to consider expected changes into the optimization process. Information gathered during treatment, e.g. from CBCT, is incorporated into the optimization. Methods and materials: DAT is formulated using stochastic control formalism, which minimizes the total expected number of tumor cells at the end of a treatment course subject to uncertainty inherent in the tumor response and organs-at-risk (OAR) dose constraints. This formulation allows for non-stationary dose distribution as well as non-stationary fractional dose as needed to achieve a series of optimal plans that are conformal to tumor over time. Sixteen phantom cases with various sizes and locations of tumors, and OAR geometries were generated. Each case was planned with DAT and conventional IMRT (60Gy/30fx). Tumor volume change over time was obtained by using, daily MVCT-based, two-level cell population model. Monte Carlo simulations have been performed for each treatment course to account for uncertainty in tumor response. Same OAR dose constraints were applied for both methods. The frequency of plan modification was varied to 1, 2, 5 (weekly), and 29 (daily). The final average tumor dose and OAR doses have been compared to quantify the potential benefit of DAT. Results: The average tumor max, min, mean, and D95 resulted from DAT were 124.0–125.2%, 102.1–114.7%, 113.7–123.4%, and 102.0–115.9% (range dependent on the frequency of plan modification) of those from conventional IMRT. Cord max, esophagus max, lung mean, heart mean, and unspecified tissue D05 resulted from AT were 84–102.4%, 99.8–106.9%, 66.9–85.6%, 58.2–78.8%, and 85.2–94.0% of those from conventional IMRT. Conclusions: Significant tumor dose increase and OAR dose reduction, especially with parallel OAR with mean or dose-volume constraints, can be achieved using DAT.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2013

Retrospective evaluation of dosimetric quality for prostate carcinomas treated with 3D conformal, intensity modulated and volumetric modulated arc radiotherapy

This study examines and compares the dosimetric quality of radiotherapy treatment plans for prostate carcinoma across a cohort of 163 patients treated across five centres: 83 treated with three-dimensional conformal radiotherapy (3DCRT), 33 treated with intensity modulated radiotherapy (IMRT) and 47 treated with volumetric modulated arc therapy (VMAT). Treatment plan quality was evaluated in terms of target dose homogeneity and organs at risk (OAR), through the use of a set of dose metrics. These included the mean, maximum and minimum doses; the homogeneity and conformity indices for the target volumes; and a selection of dose coverage values that were relevant to each OAR. Statistical significance was evaluated using two-tailed Welch's T-tests. The Monte Carlo DICOM ToolKit software was adapted to permit the evaluation of dose metrics from DICOM data exported from a commercial radiotherapy treatment planning system. The 3DCRT treatment plans offered greater planning target volume dose homogeneity than the other two treatment modalities. The IMRT and VMAT plans offered greater dose reduction in the OAR: with increased compliance with recommended OAR dose constraints, compared to conventional 3DCRT treatments. When compared to each other, IMRT and VMAT did not provide significantly different treatment plan quality for like-sized tumour volumes. This study indicates that IMRT and VMAT have provided similar dosimetric quality, which is superior to the dosimetric quality achieved with 3DCRT.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

DOE OSTI2022

Optimal Gas-Electric Energy System Decarbonization Planning

As energy utilities implement climate change mitigation policies, system planners require strategies for achieving affordable emissions reductions. Coordinated planning of electric power and natural gas systems will allow synergistic investments to address cross-sector operational constraints, competing uses for net-zero emissions fuels, and shifts in energy demands across energy carriers. In this study, we develop a novel optimization program that finds the cost-minimizing mix of infrastructure expansion or reduction across gas and electric systems to satisfy sector-specific emissions constraints. Alongside energy supply resources, our framework allows for central-planning of end-use equipment stocks to allow switching between gas and electric appliances upon failure or premature replacement. The proposed model is used to simulate case study scenarios for a benchmark 24-pipe gas network coupled to a 24-node power system test network. We find that electrification of greater than 80% of core gas demands is a component of the least-cost solution for modeled energy systems. Despite this substitution, the gas system is maintained to service difficult-to-electrify customers and to deliver net-zero emissions gas to electricity generators in times of peak electricity demand. Restricting electrification of gas appliances increases reliance on power- to-gas technologies and increases annual costs by 15% in 2040. Neglecting constraints on pipeline blending of hydrogen can produce a misleading result that relies on hydrogen blend fractions of greater than 50%. In all cases, we find the average costs of delivered gas increase nearly 5-fold across the decarbonization transition, highlighting the importance of future work investigating cost-allocation strategies for ensuring an equitable energy transition.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

DOE OSTI2017

Catalytic N 2 Reduction to Silylamines and Thermodynamics of N 2 Binding at Square Planar Fe

The geometric constraints imposed by a tetradentate P4N2 ligand play an essential role in stabilizing square planar Fe complexes with changes in metal oxidation state. A combination of high-pressure electrochemistry and variable temperature UV-vis spectroscopy were used to obtain these thermodynamic measurements, while X-ray crystallography, 57Fe Mössbauer spectroscopy, and EPR spectroscopy were used to fully characterize these new compounds. Analysis of Fe0, FeI, and FeII complexes reveals that the free energy of N2 binding across three oxidation states spans more than 37 kcal mol-1. The square pyramidal Fe0(N2)(P4N2) complex catalyzes the conversion of N2 to N(SiR3)3 (R = Me, Et) at room temperature, representing the highest turnover number (TON) of any Fe-based N2 silylation catalyst to date (up to 65 equiv N(SiMe3)3 per Fe center). Elevated N2 pressures (> 1 atm) have a dramatic effect on catalysis, increasing N2 solubility and the thermodynamic N2 binding affinity at Fe0(N2)(P4N2). Acknowledgment. This research was supported as part of the Center for Molecular Electrocatalysis, an Energy Frontier Research Center funded by the U.S. Department of Energy (DOE), Office of Science, Office of Basic Energy Sciences. EPR experiments were performed using EMSL, a national scientific user facility sponsored by the DOE’s Office of Biological and Environmental Research and located at Pacific Northwest National Laboratory (PNNL). PNNL is operated by Battelle for the U.S. DOE. Computational resources were provided by the National Energy Research Scientific Computing Center (NERSC) at Lawrence Berkeley National Laboratory. The authors thank Prof. Yisong Alex Guo at Carnegie Mellon University for recording Mössbauer data for some complexes and Emma Wellington and Kaye Kuphal for their assistance with the collection of Mössbauer data at Colgate University, Dr. Katarzyna Grubel for X-ray assistance, and Dr. Rosalie Chu for mass spectrometry assistance. The authors also thank Dr. Aaron Appel and Dr. Alex Kendall for helpful discussions.

iron, thermodynamics, dinitrogen, square planer, M↗

DOE OSTI2014

Comparative study of beam losses and heat loads reduction methods in MITICA beam source

In negative ion electrostatic accelerators a considerable fraction of extracted ions is lost by collision processes causing efficiency loss and heat deposition over the components. Stripping is proportional to the local density of gas, which is steadily injected in the plasma source; its pumping from the extraction and acceleration stages is a key functionality for the prototype of the ITER Neutral Beam Injector, and it can be simulated with the 3D code AVOCADO. Different geometric solutions were tested aiming at the reduction of the gas density. The parameter space considered is limited by constraints given by optics, aiming, voltage holding, beam uniformity, and mechanical feasibility. The guidelines of the optimization process are presented together with the proposed solutions and the results of numerical simulations.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2013

Sulfur-Modified Zero-Valent Iron for Remediation Applications at DOE Sites - 13600

Many DOE remediation sites have chemicals of concern that are compounds in higher oxidation states, which make them both more mobile and more toxic. The chemical reduction of these compounds both prevents the migration of these chemicals and in some cases reduces the toxicity. It has also been shown that zero-valent iron is a very effective substance to use in reducing oxygenated compounds in various treatment processes. These have included the treatment of halogenated hydrocarbons in the form volatile organic compounds used as solvents and pesticides. Zero-valent iron has also been used to reduce various oxidized metals such as chromium, arsenic, and mercury in order to immobilize them, decrease their toxicity, and prevent further transport. In addition, it has been used to immobilize or break down other non-metallic species such as selenium compounds and nitrates. Of particular interest at several DOE remediation sites is the fact that zero-valent iron is very effective in immobilizing several radioactive metals which are mobile in their oxidized states. These include both technetium and uranium. The main difficulty in using zero-valent iron has been its tendency to become inactive after relatively short periods of time. While it is advantageous to have the zero-valent iron particles as porous as possible in order to provide maximum surface area for reactions to take place, these pores can become clogged when the iron is oxidized. This is due to the fact that ferric oxide has a greater volume for a given mass than metallic iron. When the surfaces of the iron particles oxidize to ferric oxide, the pores become narrower and will eventually shut. In order to minimize the degradation of the chemical activity of the iron due to this process, a modification of zero-valent iron has been developed which prevents or slows this process, which decreases its effectiveness. It is called sulfur-modified iron, and it has been produced in high purity for applications in municipal water treatment applications. Sulfur-modified iron has been found to not only be an extremely economical treatment technology for municipal water supplies, where very large quantities of water must be treated economically, but it has also been demonstrated to immobilize technetium. It has the added benefit of eliminating several other harmful chemicals in water supplies. These include arsenic and selenium. In one large-scale evaluation study an integrated system implemented chemical reduction of nitrate with sulfur-modified iron followed by filtration for arsenic removal. The sulfur-modified iron that was used was an iron-based granular medium that has been commercially developed for the removal of nitrate, co-contaminants including uranium, vanadium and chromium, and other compounds from water. The independent study concluded that 'It is foreseen that the greatest benefit of this technology (sulfur-modified iron) is that it does not produce a costly brine stream as do the currently accepted nitrate removal technologies of ion exchange and reverse osmosis. This investigation confirmed that nitrate reduction via sulfur-modified iron is independent of the hydraulic loading rate. Future sulfur-modified iron treatment systems can be designed without restriction of the reactor vessel dimensions. Future vessels can be adapted to existing site constraints without being limited to height-to-width ratios that would exist if nitrate reduction were to depend on hydraulic loading rate'. Sulfur-modified iron was studied by the Pacific Northwest National Laboratory (PNNL) for its effectiveness in the reduction and permanent sequestration of technetium. The testing was done using Hanford Site groundwater together with sediment. The report stated, 'Under reducing conditions, TcO{sub 4} is readily reduced to TcIV, which forms highly insoluble oxides such at TcO{sub 2}.nH{sub 2}O. However, (re)oxidation of TcIV oxides can lead to remobilization. Under sulfidogenic conditions, most TcIV will be reduced and immobilized as Tc{sub 2}S{sub 7}, which is less readily re-mobilized, even under oxic conditions. This process should be favored by stimulation of sulfidogenic conditions'. The sulfur-modified iron provides the sulfur, together with the iron, to maintain this stable sequestration of technetium. As a result of these and other studies demonstrating the cost-effectiveness of sulfur-modified iron in treating technetium and other hazardous compounds in Hanford Site groundwater and its cost-effectiveness in reducing nitrate, the Richland Operations Office of the Department of Energy issued a change order to the Central Plateau Contractor providing for the testing of sulfur-modified iron in a mobile pilot unit at the Hanford Site. Further testing is anticipated to produce refinements in operating conditions and further optimization of the existing process. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

DOE OSTI2021

Budget constrained machine learning for early prediction of adverse outcomes for COVID-19 patients

Abstract The combination of machine learning (ML) and electronic health records (EHR) data may be able to improve outcomes of hospitalized COVID-19 patients through improved risk stratification and patient outcome prediction. However, in resource constrained environments the clinical utility of such data-driven predictive tools may be limited by the cost or unavailability of certain laboratory tests. We leveraged EHR data to develop an ML-based tool for predicting adverse outcomes that optimizes clinical utility under a given cost structure. We further gained insights into the decision-making process of the ML models through an explainable AI tool. This cohort study was performed using deidentified EHR data from COVID-19 patients from ProMedica Health System in northwest Ohio and southeastern Michigan. We tested the performance of various ML approaches for predicting either increasing ventilatory support or mortality. We performed post hoc analysis to obtain optimal feature sets under various budget constraints. We demonstrate that it is possible to achieve a significant reduction in cost at the expense of a small reduction in predictive performance. For example, when predicting ventilation, it is possible to achieve a 43% reduction in cost with only a 3% reduction in performance. Similarly, when predicting mortality, it is possible to achieve a 50% reduction in cost with only a 1% reduction in performance. This study presents a quick, accurate, and cost-effective method to evaluate risk of deterioration for patients with SARS-CoV-2 infection at the time of clinical evaluation.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2023

Complete quantification of parametric uncertainties in ( d , p ) transfer reactions

Deuteron-induced transfer reactions are a popular probe in nuclear structure and nuclear astro physics studies. The interpretation of these transfer measurements relies on reaction theory that takes as input effective interactions between the nucleons and the target nucleus. Previous work quantified the uncertainty associated with the optical potentials between the nucleons and the target. In this study, we extend that work by also including the parameters of the mean field associated with the overlap function of the final bound state, thus obtaining the full parametric uncertainty on transfer observables. We use Bayesian Markov Chain Monte Carlo simulations to obtain parameter posterior distributions. We use elastic-scattering cross sections to constrain the optical potential parameters and use the asymptotic normalization coefficient of the final state to constrain the bound state interaction. We then propagate these posteriors to the transfer angular distributions and obtain confidence intervals for this observable. We study (d,p) reactions on 14 C, 16 O, and 48 Ca at energies in the range $E$ $d$ = 7 - 24 MeV. Our results show a strong reduction in uncertainty by using the asymptotic normalization coefficient as a constraint, particularly for those reactions most sensitive to ambiguities in the mean field. For those reactions, the importance of constraining the bound state interaction is equal to that of constraining the optical potentials. Furthermore, the case of 14 C is an outlier because the cross section is less sensitive to the nuclear interior. When minimal constraints are used on the parameters of the nucleon-target interaction, the 1σ uncertainties on the differential cross sections are large (~ 500-100%). However, if elastic-scattering data and the asymptotic normalization coefficient are used in the analysis, with an error of 10% (5%), this uncertainty reduces to ~ 30% (~ 15%).

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2025

Understanding and Estimating Error Propagation in Neural Networks for Scientific Data Analysis

Neural networks are increasingly integrated into scientific discovery, where input data reduction and model quantization play a key role in accelerating inference. However, understanding and mitigating the impact of these techniques on output error is critical for ensuring reliable results, particularly in tasks demanding high numerical precision. This paper introduces a comprehensive framework for optimizing neural network inference in scientific computing by combining data reduction and weight quantization while maintaining error-controlled outcomes. We develop theoretical analyses to bound error propagation under these reductions and propose a framework that balances computational performance with error constraints. Evaluation on real-world learning-based combustion simulations and satellite image classification demonstrates that our derived error bounds accurately predict observed errors while enabling significant computational speedup under our framework. This work highlights the potential for further leveraging advancements in modern lossy compression algorithms and hardware accelerators that support lower-precision formats.

He, Weiming [New Jersey Institute of Technology]↗

DOE OSTI2014

Road Map for Development of Crystal-Tolerant High Level Waste Glasses

The U.S. Department of Energy (DOE) is building a Tank Waste Treatment and Immobilization Plant (WTP) at the Hanford Site in Washington to remediate 55 million gallons of radioactive waste that is being temporarily stored in 177 underground tanks. Efforts are being made to increase the loading of Hanford tank wastes in glass while meeting melter lifetime expectancies and process, regulatory, and product quality requirements. This road map guides the research and development for formulation and processing of crystal-tolerant glasses, identifying near- and long-term activities that need to be completed over the period from 2014 to 2019. The primary objective is to maximize waste loading for Hanford waste glasses without jeopardizing melter operation by crystal accumulation in the melter or melter discharge riser. The potential applicability to the Savannah River Site (SRS) Defense Waste Processing Facility (DWPF) will also be addressed in this road map. The planned research described in this road map is motivated by the potential for substantial economic benefits (significant reductions in glass volumes) that will be realized if the current constraints (T 1% for WTP and T L for DWPF) are approached in an appropriate and technically defensible manner for defense waste and current melter designs. The basis of this alternative approach is an empirical model predicting the crystal accumulation in the WTP glass discharge riser and melter bottom as a function of glass composition, time, and temperature. When coupled with an associated operating limit (e.g., the maximum tolerable thickness of an accumulated layer of crystals), this model could then be integrated into the process control algorithms to formulate crystal-tolerant high-level waste (HLW) glasses targeting high waste loadings while still meeting process related limits and melter lifetime expectancies. The modeling effort will be an iterative process, where model form and a broader range of conditions, e.g., glass composition and temperature, will evolve as additional data on crystal accumulation are gathered. Model validation steps will be included to guide the development process and ensure the value of the effort (i.e., increased waste loading and waste throughput). A summary of the stages of the road map for developing the crystal-tolerant glass approach, their estimated durations, and deliverables is provided.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

NASA NTRS1972

The new civil aviation within our grasp.

Noise and congestion present the two main technological constraints on air-transportation growth. Although some of the noise reduction will come with improved flight-path control and steep approaches, the main requirement remains quiet propulsion systems. Higher engine temperatures will compensate for efficiency losses due to noise suppression. Composite structures can reduce structural weight by 20%. New developments in rotorcraft transports are discussed together with advanced subsonic transports of the 1980s and the possibilities for further evolution of the SSTs, leading to a hypersonic aircraft.

Cortright, E. M.↗