Search NASA⌕ Search

SEARCH · Search NASA

Results for “Explanability”

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 685 records · Page 38

Understanding oxidation of Fe-Cr-Al alloys through explainable artificial intelligence

Abstract The oxidation resistance of FeCrAl based on alloying composition and oxidizing conditions is predicted using a combinatorial experimental and artificial intelligence approach. A neural network (NN) classification model was trained on the experimental FeCrAl dataset produced at GE Research. Furthermore, using the SHapley Additive exPlanations (SHAP) explainable artificial intelligence (XAI) tool, we explore how the NN can showcase further material insights that are unavailable directly from a black-box model. We report that high Al and Cr content forms protective oxide layer, while Mo in FeCrAl creates thick unprotective oxide scale that is vulnerable to spallation due to thermal expansion. Graphical abstract

Materials Science↗

A model of the effects of flow fluctuations on fall Chinook salmon spawning habitat availability in the Columbia River

Previously we reported that about 30% to 60% of the area predicted to be used by fall Chinook salmon (Oncorhynchus tshawytscha) for spawning in the Hanford Reach of the Columbia River did not contain redds. One explanation for the overprediction of habitat was that our model did not incorporate streamflow fluctuation. Daily fluctuation in flow caused by load-following operations (power generation to meet short-term electrical demand) at Priest Rapids Dam, situated at the upper end of the Hanford Reach, changes the hydraulic characteristics to which fish respond in selecting redd sites. The purpose of the study described here was to examine the effect of flow changes on spawning habitat modeling and, in particular, to look at the connection between spawning and the variability and persistence of habitat variables caused by rapid changes in flow resulting from load-following operations at Priest Rapids Dam. We found that spawning habitat use by fall Chinook salmon was consistent with previous fall Chinook salmon studies in the Reach. Dynamic variables that were based on hourly time series were used to account for the variability in habitat as a result of flow fluctuations. The analysis showed that the proportion of velocities that fell within the range of 1.0 to 2.5 m/s differed significantly between locations that were predicted to be spawning by the logistic regression model where spawning actually occurred and locations that were predicted to be spawning where spawning did not occur. However, the resulting sequential logistic regression model that incorporated the dynamic variables did not provide significant improvement in the percentage of errors for areas predicted to be spawning; the model’s overprediction errors still ranged from 63% to 78%. We suggest that while flow fluctuation may affect spawning habitat and individual fish behavior, the high correlation between time-averaged velocities and the proportion of hourly velocities that fell within the most favorable range negated any improvements in model predictions.

54 ENVIRONMENTAL SCIENCES↗

Soil Carbon Response to Rising Temperature

The primary objective of this study was to investigate soil carbon response to rising temperatures by generalizing a linear four-pool soil carbon model introduced by Thornley and Cannell [9]. Currently, all of the major models predict that soil respiration will increase more than net primary productivity, Pn , as temperatures rise [9]. However, some have questioned these predictions, pointing to a lack of evidence of decreasing soil carbon in some parts of the world as annual mean temperatures have risen. The current explanation for this contradiction is that soil respiration’s temperature sensitivity has been overestimated in the models. However, the soil respiration data are conflicting, making this explanation unsatisfactory.

Montz, A.↗

The Surprising Occurrence of Non-Steady-State Growth of Divergent Lamellar Decomposition Products in Uranium-Niobium Alloys: A Preliminary Report

Metastable materials are all expected to decompose sooner or later, but sometimes the kinetic paths they travel confound standard textbook explanations. Such is the case for supersaturated uranium-niobium alloy microstructures that transform to lamellar products in the 300-647 ºC temperature range. An experimental study of these isothermal transformation kinetics is reported here for U-5.5Nb and U-7.5Nb alloys. About 85 specimens from each alloy were isothermally aged at eight temperatures for durations spanning five orders of magnitude, up to two years. The volume fractions and growth rates of fine-spacing discontinuous precipitation (DP) and coarse-spacing discontinuous coarsening (DC) lamellar reaction products were directly measured by quantitative metallography. This study was enhanced by explicit consideration of the DP initiation site during measurements: (1) prior matrix grain boundaries (the usual site found in every alloy that produces DP) and (2) carbide inclusions (a rarely-encountered site). This constitutes the first systematic study of inclusion-nucleated DP in any alloy system. The growth rates were measured by the robust method of Cahn and Hagel (1963). The measurements were taken across the entire span of fraction transformed, giving the best opportunity to discriminate between typical vs. anomalous growth behavior. In both alloys the growth rates for DP and DC are constant at the lowest aging temperatures (300-400ºC), in line with standard explanations. The growth rates, however, noticeably diminish with reaction time at the higher aging temperatures (450-625ºC), dropping by an order of magnitude or more. These results corroborate as-yet unpublished data of divergent lamellar spacings (i.e., increase of the spacing with reaction time) observed at the same higher aging temperatures. These surprising results are rationalized in terms of kinetic competition for chemical driving force between the lamellar and precursor nonlamellar reactions on comparable time scales. In previous studies divergent lamellar products occurred only in transformations involving diffusion of both interstitial and substitutional elements. Since the same phenomena are now seen in a purely substitutional system like U-Nb, it is likely that divergent and time-varying lamellar reactions may be more prevalent in other substitutional alloy systems than previously thought.

36 MATERIALS SCIENCE↗

Solar Leasing Summary, Houston Texas

A relatively new option for homeowners looking to add solar to their home is the solar lease. At present, the solar lease option can be found in California, Arizona, Texas, Colorado, Hawaii, New York and Oregon. The most active companies currently offering solar leases are NRG Energy, Sungevity, Solar City and Sun Run. With the uncertainty and/or lack of subsidies the states participating in these programs have ebbed and flowed over the last few years. However, there is an expectation that in the current market solar leasing will make solar viable without the utility and federal subsidies. NRG Energy is currently testing this expectation in Houston, TX where currently no subsidies or incentives beyond the federal tax incentives, exist. Following is an explanation on the state of solar leasing in Houston, TX and explanation of the current financing options.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Pu doped with Hydrogen: Magnetic Properties

The magnetic properties of plutonium (Pu) metal remain an enigmatic problem based on our understanding of common magnetic compounds and understanding of coordination chemistry that explains and predicts the magnetic state of metals. The 5ƒ occupancy of Pu is believed to be fluctuating at very fast (fempto-second) time scales. As a result of occupancy fluctuations the average number of electrons to occupy the 5ƒ energy level of Pu is expected to be more than 5.3 electrons on average. This explanation provides a solution to the non-magnetic state observed in pure Pu metal. Despite the intellectual effort that has been deployed to provide an explanation to the “missing magnetism” in Pu metal it has been known since the 1950s that Uranium Hydride (UH 2 ) has a ferromagnetic ground state and similarly since the 1960s, that PuH 2 has a ferromagnetic ground state at moderate to low temperatures. The contrast in non-magnetic ground state observed in Pu metal and the ferromagnetic ground state of PuH 2 may provide further insight into the behavior of Pu metal and its electronic configuration and fluctuations.

36 MATERIALS SCIENCE↗

Development and Deployment of a Short Rotation Woody Crops Harvesting System Based on a Case New Holland Forage Harvester and SRC Woody Crop Header

Biomass for biofuels, bioproducts and bioenergy can be sourced from forests, agricultural crops, various residue streams, and dedicated woody or herbaceous crops. Short rotation woody crops (SRWC), like willow and hybrid poplar, are perennial cropping systems that produce a number of environmental and economic development benefits in addition to being a renewable source of biomass that can be produced on marginal land. Both hybrid poplar and willow have several characteristics that make them an ideal feedstock for biofuels, bioproducts, and bioenergy; these include high yields that can be obtained in three to four years, ease of cultivar propagation from dormant cuttings, a broad underutilized genetic base, ease of breeding, ability to resprout after multiple harvests, and feedstock composition similar to other sources of woody biomass. Despite the range of benefits associated with SRWC systems, their deployment has been restricted by high costs, low market acceptance associated with inconsistent chip quality (see below for further explanation), and misperceptions about other feedstock characteristics (see below for further explanation). Harvesting of SRWC is the largest single cost factor (~1/3 of the final delivered cost) in the feedstock supply system. Harvesting is also the second largest input of primary fossil energy in the system after commercial N fertilizer, accounting for about one third of the input. Therefore, improving the efficiency of the harvesting system has the potential to reduce both cost and environmental impact. At the start of this project, we projected that improving the overall efficiency of the harvesting system by 25% would reduce the delivered cost of SRWC by approximately $0.50/MMBtu (or about $7.50/dry ton). This goal was exceeded over the duration of this project, as noted below.

09 BIOMASS FUELS↗

First Search for the EMC Effect and Nuclear Shadowing in Neutrino Nucleus Deep Inelastic Scattering at MINERvA

Decades of research in electron-nucleus deep inelastic scattering (DIS) have provided a clear picture of nuclear physics at high momentum transfer. While these effects have been clearly demonstrated by experiment, the theoretical explanation of their origin in some kinematic regions has been lacking. Particularly, the effects in the intermediate regions of Bjorken-x, anti-shadowing and the EMC effect have no universally accepted quantum mechanical explanation. In addition, these effects have not been measured systematically with neutrino-nucleus deep inelastic scattering, due to experiments lacking multiple heavy targets.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

ARM West Antarctic Radiation Experiment (AWARE) Science Plan

West Antarctica is one of the most rapidly warming regions on Earth, and this warming is closely connected with global sea level rise. The discovery of rapid climate change on the West Antarctic Ice Sheet (WAIS) has challenged previous explanations of Antarctic climate change that focused on strengthening of circumpolar westerlies in response to the positive polarity trend in the Southern Annular Mode. West Antarctic warming does not yet have a comprehensive explanation: dynamical mechanisms may vary from one season to the next, and these mechanisms very likely involve complex teleconnections with subtropical and tropical latitudes. The prime motivation for this proposal is that there has been no substantial atmospheric science or climatological field work on West Antarctica since the 1957 International Geophysical Year and that research continued for only a few years. Direct meteorological information on the WAIS has been limited to a few automatic weather stations for several decades, yet satellite imagery and meteorological reanalyses indicate that West Antarctica is highly susceptible to advection of warm and moist maritime air with related cloud cover, depending on the location and strength of low pressure cells in the Amundsen, Ross, and Bellingshausen Seas. There is a need to quantify the role of these changing air masses on the surface energy balance, including all surface energy components and cloud-radiative forcing. More generally, global climate model simulations are known to perform poorly over the Antarctic and Southern Oceans, and the marked scarcity of cloud information at southern high latitudes has so far inhibited significant progress. Fortunately, McMurdo Station, where the Atmospheric Radiation Measurement Facility’s (ARM’s) most advanced cloud and aerosol instrumentation is situated, has a meteorological relationship with the WAIS via circulation patterns in the Ross and Amundsen Seas. We can therefore gather sophisticated data with cloud radars and high spectral resolution lidar and a complete aerosol suite at McMurdo that have relevance to the WAIS as well. At the same time, we will send basic radiometric, surface energy balance, and upper air equipment directly to the WAIS to make the first well calibrated climatological suite of measurements seen in this extremely remote but globally critical region in more than 40 years.

Lubin, D↗

Parsimonious Inference Information-Theoretic Foundations for a Complete Theory of Machine Learning (CIS-LDRD Project 218313 Final Technical Report)

This work examines how we may cast machine learning within a complete Bayesian framework to quantify and suppress explanatory complexity from first principles. Our investigation into both the philosophy and mathematics of rational belief leads us to emphasize the critical role of Bayesian inference in learning well-justified predictions within a rigorous and complete extended logic. The Bayesian framework allows us to coherently account for evidence in the learned plausibility of potential explanations. As an extended logic, the Bayesian paradigm regards probability as a notion of degrees of truth. In order to satisfy critical properties of probability as a coherent measure, as well as maintain consistency with binary propositional logic, we arrive at Bayes' Theorem as the only justifiable mechanism to update our beliefs to account for empiracle evidence. Yet, in the machine learning paradigm, where explanations are unconstrained algorithmic abstractions, we arrive at a critical challenge: Bayesian inference requires prior belief. Conventional approaches fail to yield a consistent framework in which we could compare prior plausibility among the infinities of potential choices in learning architectures. The difficulty of articulating well-justified prior belief over abstract models is the provinence of memorization in traditional machine learning training practices. This becomes exceptionally problematic in the context of limited datasets, when we wish to learn justifiable predictions from only a small amount of data.

97 MATHEMATICS AND COMPUTING↗

Review of "ML for Surface Complexation Model Development"

The presentation given by Jadallah Zouabe on “ML for Surface Complexation Model Development”was extremely informative, but lacked some explanation of details that made the latter half hard to follow as someone outside of the field. He began his presentation with an anecdote to spilling milk or ink on the floor and seeing it spread, cleverly connecting that to the spread of contamination in the real world. He then explained the importance of being able to model contamination, as it allows specialists to see the potential effects on different areas and environments in the coming years. Heal so made it clear that the main issue trying to be addressed here is how expensive modelling is and that their goal is to make computationally cheaper models. The introduction was overall very well formulated and attention-grabbing, but it seemed to take up too much of the presentation time, as after this point the presentation lacked details and explanations.

54 ENVIRONMENTAL SCIENCES↗

Verification of Data-Driven Models of Physical Phenomena using Interpretable Approximation

Machine-learned models, specifically neural networks, are increasingly used as “closures” or “constitutive models” in engineering simulators to represent fine-scale physical phenomena that are too computationally expensive to resolve explicitly. However, these neural net models of unresolved physical phenomena tend to fail unpredictably and are therefore not used in mission-critical simulations. In this report, we describe new methods to authenticate them, i.e., to determine the (physical) information content of their training datasets, qualify the scenarios where they may be used and to verify that the neural net, as trained, adhere to physics theory. We demonstrate these methods with neural net closure of turbulent phenomena used in Reynolds Averaged Navier-Stokes equations. We show the types of turbulent physics extant in our training datasets, and, using a test flow of an impinging jet, identify the exact locations where the neural network would be extrapolating i.e., where it would be used outside the feature-space where it was trained. Using Generalized Linear Mixed Models, we also generate explanations of the neural net (à la Local Interpretable Model agnostic Explanations) at prototypes placed in the training data and compare them with approximate analytical models from turbulence theory. Finally, we verify our findings by reproducing them using two different methods.

42 ENGINEERING↗

Understanding the Origin of the Hadron Mass within the Standard Model

Understanding the origin of the hadron mass, which constitutes 99% of our visible universe, is one of the central goals of nuclear physics. Although the Higgs mechanism provides mass for the fundamental building blocks of matter, it can only contribute less than 2% of the proton mass. The vast majority of the proton mass is believed to come from the strong force that tightly binds quarks and gluons (collectively called partons) together as described by Quantum Chromodynamics (QCD). The mass that emerges as a consequence of the strong interactions within QCD is commonly denoted as Emergent Hadronic Mass (EHM). Understanding how the nucleon mass emerges in QCD is a prerequisite to an explanation of how the Universe came into being, therefore it is of utmost importance and one of the key questions to be addressed by the future Electron-Ion Collider (EIC). When it comes to light mesons, particularly pions, the mass decomposition is drastically different. Since the pion is naturally massless in the chiral limit, the majority of its observed mass needs to come from other mechanisms within QCD. Any successful explanation for the EHM needs to be able to reconcile both the heavy proton mass and the very light pion mass (~15% of proton mass) simultaneously. The EHM theories have direct measurable implications on the description of the internal structure of the hadron, i.e., how the partons distribute inside the hadron. Precise measurement of the parton distribution functions (PDFs) will provide necessary experimental verifications and constraints of potential EHM mechanisms. We propose to carry out a comprehensive study of the poorly known pion PDFs at the AMBER experiment at CERN. The measurement will provide vital input to constrain the global analysis of the pion PDFs, which are still based on limited data obtained more than 30 years ago. The proposed pion measurement is the only direct measurement achievable within this decade, which could lead to a future major meson structure program parallel to the EIC’s proton structure measurement.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Flexoelectricity and New Phenomena

It is easy to miss the scientific implications of our recent work on Triboelectricity. Everyone knows that rubbing and contact can produce static electricity; less appreciated is that the thermodynamic driver has been an open question since static electricity was first observed by Thales of Miletus around 585 BC. People think they understand it, for instance one common explanation that can still be found in the current literature is that differences in the work function drives charge transfer, often called the Volta-Helmholtz hypothesis. As summarized in 1967 by Harper, this fails to explain many experimental observations, for instance that charging can occur when two pieces of the same material are rubbed against each other. We are the first to place triboelectricity on a solid foundation rooted in quantum mechanics – the flexoelectric effect. We were able to explain a significant number of previously unexplained phenomena: 1) Bipolar tribocurrents associated with stick-slip, due to the change in sign of the strain gradients. 2) A one-third power scaling of tribocurrents with indentation force. 3) Tribocharging when two identical materials are used, these being due to local variations in the asperities so there are usually local potential differences. 4) Inhomogeneous charging of insulators, related to the statistical nature of asperities. 5) An experimentally observed reversal in the sign of charge transfer for negative and positive curvature, which is related to a change in the sign of the strain gradient. Exploiting our DOE prior funded work on flexoelectricity, we obtained semi-quantitative matching to existing experimental measurements of the surface charge in triboelectric experiments. The work has been well received in the literature. The work has been the focus of a number of popular science press articles, and also formed the basis for a Podcast for children 6-10 “The Rise and Fall of Static Man” posted in December 2019 by NPR as part of their “Wow in the World” series. I was also briefly interviewed by the Chicago PBS station in January 2020. This work are significant for a wide range of energy applications; to quote from an independent source: "Triboelectric power has plenty of potential, says Wenzhuo Wu, an assistant professor of engineering at Purdue. If the basics of static electricity are better understood, we could maximize the efficiency of wind or wave power generators, Wu says. The body's own movement could be used to power internal medical devices. Imagine being able to create a roof shaped to harness the power of a raindrop — the friction of the rain passing over the surface — to generate triboelectricity, powering the building below it." This is the start of new science, some of which we already partially understand such as the role of band bending in charge transfer. We need to understand charge transfer combining elasticity, quantum mechanics, band bending and defect states. These directly involve several of the DOE Grand Challenges: "How do we control material processes at the level of electrons? How do remarkable properties of matter emerge from complex correlations of the atomic or electronic constituents and how can we control these properties? How do we characterize and control matter away—especially very far away— from equilibrium?" I will argue that this work truly falls into the class of disruptive science; it is not just a simple extension, linear science. Not everyone will accept the approach. Since we explain far more about triboelectricity than anyone before, the preponderance of evidence supports the model. The feedback I have received is that many agree with the work, to quote: "The model makes sense, says Michael McAlpine, a professor of engineering at the University of Minnesota. "It's such a simple explanation, I was surprised I didn't put my finger on that," McAlpine says." The proposal received strong reviews. It was also publicized on the Department of Energy Web Site.

16 TIDAL AND WAVE POWER↗

Review of the Proton Size Puzzle

The proton charge radius puzzle exists because the proton charge radius appears different when measured with electrons than when measured with muons. "Ordinary" explanations include problems with the extrapolations or theoretical corrections involved in the measurements, and "exotic" explanations include such beyond the standard model ideas as a breakdown of electron-muon universality. We will discuss possibilities in both categories.

Carlson, Carl↗

X-Ray Generation for Diffraction Experiments on Z

Explanation of the concept and benefit of X-ray diffraction in the context of material sciences on Z. Explanation why we use ZBL and ZPW for these experiments and what the progress on X-ray generation for this purpose is.

Geissel, Matthias↗

On The Anomalous Fast Ion Energy Diffusion in Toroidal Plasmas Due to Cavity Modes

An enormous wave-particle diffusion coefficient along paths suitable for alpha channeling had been deduced in mode converted ion Bernstein wave experiments on Tokamak Fusion Test Reactor (TFTR) the only plausible explanation advanced for such a large diffusion coefficient was the excitation of internal cavity modes which induce particle diffusion along identical diffusion paths, but at much higher rates. Although such a mode was conjectured, it was never observed. However, recent detailed observations of high frequency compressional Alfven eigenmodes (CAEs) on the National Spherical torus Experiment (NSTX) indirectly support the existence of the related conjectured modes on TFTR. The eigenmodes responsible for the high frequency magnetic activity can be identified as CAEs through the polarization of the observed magnetic field oscillations in NSTX and through a comparison with the theoretically derived freuency dispersion relation. Here, we show how these recent observations of high frequency CAEs lend support to this explanation of the long-standing puzzle of anomalous fast ion energy diffusion on TFTR. The support of the conjecure that these internal modes could have caused the remarkable ion energy diffusion on TFTR carries significant and favorable implications for the possibilities in achieving the alpha channeling effect with small injected power in a tokamak reactor.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Energy-Efficiency Improvement Opportunities for the Textile Industry

The textile industry is one of the most complicated manufacturing industries because it is a fragmented and heterogeneous sector dominated by small and medium enterprises (SMEs). Energy is one of the main cost factors in the textile industry. Especially in times of high energy price volatility, improving energy efficiency should be a primary concern for textile plants. There are various energy-efficiency opportunities that exist in every textile plant, many of which are cost-effective. However, even cost-effective options often are not implemented in textile plants mostly because of limited information on how to implement energy-efficiency measures, especially given the fact that a majority of textile plants are categorized as SMEs and hence they have limited resources to acquire this information. Know-how on energy-efficiency technologies and practices should, therefore, be prepared and disseminated to textile plants. This guidebook provides information on energy-efficiency technologies and measures applicable to the textile industry. The guidebook includes case studies from textile plants around the world and includes energy savings and cost information when available. First, the guidebook gives a brief overview of the textile industry around the world, with an explanation of major textile processes. An analysis of the type and the share of energy used in different textile processes is also included in the guidebook. Subsequently, energy-efficiency improvement opportunities available within some of the major textile sub-sectors are given with a brief explanation of each measure. The conclusion includes a short section dedicated to highlighting a few emerging technologies in the textile industry as well as the potential for the use of renewable energy in the textile industry.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗