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 1,189 records · Page 66

Organo-montmorillonite Barrier Layers Formed by Combustion: Nanostructure and Permeability

Self-assembly of nanoparticles into barrier layers has been the most cited theoretical explanation for the significant reduction in flammability often noted for nanocomposites formed from polymers and montmorillonite organoclays. Both mass and heat transport reductions have been credited for such improvements, and in most cases a coupled mechanism is expected. To provide validation for early models, new model barrier layers were produced from organoclays, and these barrier layers subjected to novel permeability analysis to obtain a flux. The effects of surfactant, temperature and pressure on barrier layer structure were examined. XRD versus TGA results suggest that chemical degradation of four different organoclays and physical collapse on heating are not correlated. Addition of pressure as low as 7kPa also altered the structure produced. Permeability of Ar through the ash was found to be sensitive to structural change/self assembly of high aspect ratio MMT nanoparticles. Actual fluxes ranged from 0.139 to 0.151 mol(m2.sec)-1, values which will provide useful limits in verifying models for the coupled contribution of mass and heat transfer to flammability parameters such as peak heat release rate.

36 MATERIALS SCIENCE↗

Bayesian sequential optimal experimental design for nonlinear models using policy gradient reinforcement learning

We present a mathematical framework and computational methods for optimally designing a finite sequence of experiments. This sequential optimal experimental design (sOED) problem is formulated as a finite-horizon partially observable Markov decision process (POMDP) under a Bayesian setting and with information-theoretic utilities. The formulation is general and may accommodate continuous random variables, non-Gaussian posteriors, and nonlinear forward models. The sOED design policy incorporates elements of feedback and lookahead simultaneously, and we show it to generalize the commonly-used batch and greedy design strategies. We solve for the sOED policy using the policy gradient (PG) method from reinforcement learning, and provide a derivation for the PG expression in the sOED context. Adopting an actor-critic approach, the policy and value functions are parameterized using deep neural networks and improved via PG estimates produced from simulated episodes of designs and observations. The new PG-sOED algorithm is first validated on a linear-Gaussian benchmark, and then compared against other design baselines on a sensor movement problem for contaminant source inversion in a convection-diffusion field. As a result, we provide explanation for the policy behaviors using knowledge of the underlying physical process.

97 MATHEMATICS AND COMPUTING↗

Influences of O{sub 2} concentration on NO reduction and N{sub 2}O formation in thermal deNOx process

Flow reactor experiment and chemical kinetic modeling were carried out to evaluate the influences of oxygen concentration on NO reduction and N{sub 2}O formation during thermal deNOx process for temperatures of 1073-1473 K. Increasing oxygen concentration in the reactant gas from 0.3 to 9.3% not only reduces the NO reduction potential and N{sub 2}O emission, but also lowers and widens the ''temperature window'' for NO conversion and N{sub 2}O formation. Trace amount of impurity O{sub 2} leads to the lowest NO emission at a high temperature of 1373 K with low N{sub 2}O emission. Test reveals that high molar ratio NH{sub 3}/NO suppresses the reverse effect of oxygen and brings out high percentage of NO reduction at leaner environment. Chemical kinetic simulations using 4 detailed mechanisms display differing trends in the NO and N{sub 2}O dependence on O{sub 2} and temperature. Rota 2000 model gives the best description of our experiments, but some assumptions made in this mechanism are not fully realistic. This study provides new experimental data on the less noticed N{sub 2}O, and raises a question about the understanding of the detailed elementary chemistry on the thermal deNOx process. Further work is still required for reasonable kinetic explanations. (author)

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

A first principles study of commonly observed planar defects in Ti/TiB system

Here, TiB exhibits a hexagonal cross-section with growth faults on (1 0 0) planes and contains B27-B f bicrystals. The hexagonal cross-section is presently explained by surface free energy minimization principle. We show that interfacial energy calculations explain the longer (1 0 0) facet compared to (1 0 1) type facets whereas free surface energy arguments do not provide the true picture. No quantitative explanation of stacking faults and B27-B f interfaces in TiB exists. We show that the low formation energy of stacking faults and B27-B f interfaces explain their abundance. The low energy barrier for B f formation is shown to be responsible for their presence in TiB.

36 MATERIALS SCIENCE↗

Understanding and control of Zener pinning via phase field and ensemble learning

Zener pinning refers to the dispersion of fine particles which influences grain size distribution via movement of grain boundaries in a polycrystalline material. Grain size distribution in polycrystals has a significant impact on their properties including physical, chemical, mechanical, and optical to name a few. We explore the use of Phase-field modeling and machine-learning techniques to understand and improve the control of grain size distribution via Zener pinning in polycrystalline materials. We develop a machine learning model that determines the relative importance of various parameters to exercise microstructure control via Zener pinning. Our workflow combines high-throughput phase-field simulations and machine learning to address the computational bottlenecks associated with large-scale simulations as well as identify features necessary for microstructure control in polycrystals. A random forest (RF) regression model was developed to predict grain sizes based on five Phase-field model parameters, achieving an average prediction error of 0.72 nm for the training data and 1.44 nm for the test data. The importance of the input parameters is analyzed using the SHapley Additive exPlanations (SHAP) approach which reveals that diffusivity, volume fraction, and particle diameter are the most important parameters in determining the final grain size. These findings will allow us to select the best second-phase particles, optimize grain size distributions and thus design microstructures with the desired properties. The developed method is a highly versatile and generalizable approach that can be used to assess the combined effects of individual features in the presence of multiple variables.

36 MATERIALS SCIENCE↗

Explainable machine learning reveals that local structural motifs encode the thermodynamic state across the CuZr metallic glass-forming range

Metallic glasses derive their properties from the statistics of local atomic motifs rather than from long-range order, yet a quantitative, chemistry-specific link between motif populations and the underlying glassy state has remained elusive. In this work we combine large-scale molecular dynamics, Voronoi tessellation, deep neural networks, and SHapley Additive exPlanations (SHAP) to identify which local structural motifs define the glassy state of Cu—Zr metallic glasses. A dataset of 17,180 atomistic configurations spanning ten compositions (Cu 20 Zr 80 –Cu 80 Zr 20 ) and four quench rates (10 9 –10 12 K/s) is used to train a feed-forward neural network that regresses temperature across the 50–2000 K liquid–supercooled–glass range, achieving a mean absolute error of 19.89 K and R 2 = 0.9974, confirming that the local structural state is faithfully encoded in motif-level structure. SHAP analysis then reveals that a tightly coupled near-icosahedral family of motifs (coordination numbers (CN) 11–13, including the full icosahedron 001200 and its single-atom-perturbation sibling 10930) collectively encodes the thermodynamic state of the system across the full glass-forming range. The CN = 11–13 ordered members carry negative SHAP values at high populations, tracking the most deeply-quenched configurations, while 10930 shows the reversed signature consistent with its role as a soft-spot host whose population shrinks as the icosahedral network deepens. The analysis demonstrates that explainable machine learning can isolate the minimal motif vocabulary defining the glassy state and recovers the near-icosahedral building blocks previously identified by data-driven analyses of Cu—Zr. The approach provides a general, chemistry-specific route for characterizing the structural state of disordered materials.

36 MATERIALS SCIENCE↗

Tree-level carbon stock estimations across diverse species using multi-source remote sensing integration

Forests are critical carbon sinks, and remote sensing has been increasingly widely used for forest monitoring and biomass estimations. However, species-specific tree-level studies remain limited. In this study, we demonstrated the feasibility of integrating UAV-based LiDAR with high-resolution optical satellite imagery (0.5 m) to estimate biomass for individual trees across different species. The proposed method accurately estimated biomass for 53 trees (R² = 0.82, rRMSE = 0.44), with species-specific datasets, showing an average 25.2% increase in R² and a 14.8% reduction in rRMSE. A novel vegetation index combining forest structure parameters with vegetation indices (VIs) was developed using high-resolution multispectral satellite data (3 m) to explore its relationship with individual tree biomass. Combining forest structural parameters with VIs further improved estimation accuracy, achieving an R²of 0.89 and an rRMSE of 0.34. Species-specific datasets show an 11.6% increase in R²compared to methods without VIs, and a 22.2% improvement over methods using only VIs. SHapley Additive exPlanations (SHAP) analysis shows that the volume feature played a key role in model performance and remained stable throughout the training process. Altogether, the proposed approach enhances individual tree biomass and carbon sink estimations, showing great potential for large-scale precise forest carbon monitoring using multi-source remote sensing data.

59 BASIC BIOLOGICAL SCIENCES↗

Drying shrinkage of alkali activated binders cured at room temperature

Alkali activated binders tend to exhibit higher drying shrinkage movements when compared with Portland cement. An experimental investigation was undertaken in order to assess the drying shrinkage of a basic alkali activated binder versus Portland cement, and to evaluate the effectiveness of different additives in reducing the drying shrinkage of the alkali activated binder. The results indicated that the drying shrinkage of the basic alkali activated binder was about twice that of Type I Portland cement. Some of the additives used to refine the alkali activated binder formulation were found to lower the drying shrinkage of the alkali activated binder to levels that were comparable with that of Portland cement. Explanations were presented for the mechanisms through which these additives enhance the dimensional stability of the alkali activated binder. SEM, TG/DTG and heat of hydration techniques were used to gain more insight into the mechanisms of action of different additives in alkali activated binders.

36 MATERIALS SCIENCE↗

Can exascale computing and explainable artificial intelligence applied to plant biology deliver on the United Nations sustainable development goals?

Human population growth and accelerated climate change necessitate agricultural improvements using designer crop ideotypes (idealized plants that can grow in niche environments). Diverse and highly skilled research groups must integrate efforts to bridge the gaps needed to achieve international goals toward sustainable agriculture. Given the scale of global agricultural needs and the breadth of multiple types of omics data needed to optimize these efforts, explainable artificial intelligence (AI with a decipherable decision making process that provides a meaningful explanation to humans) and exascale computing (computers that can perform 1018 floating-point operations per second, or exaflops) are crucial. Accurate phenotyping and daily-resolution climatype associations are equally important for refining ideotype production to specific environments at various levels of granularity. In this article, we review advances toward tackling technological hurdles to solve multiple United Nations Sustainable Development Goals and discuss a vision to overcome gaps between research and policy.

59 BASIC BIOLOGICAL SCIENCES↗

Developing an Integrated History and future of People on Earth (IHOPE)

The challenges of modeling future socio-ecological states numerically suggest that qualitative understanding of system behavior should be further developed. The objective here is to identify and develop general principles of socio-ecological system behavior, supported by empirical evidence drawn from long records of regional environmental change. A few are deduced from historical case studies. But for others, we can take principles (often aphorisms) from theoretical ecology and complexity science and test their validity in the real world by comparison with historical records. Past records not only provide longer timescales than are conventionally available for observations, but also provide a larger array of socio-ecological systems than currently exist. As a basic starting point, here are some principles drawn from the literature and discussions at IHOPE meetings that could be tested (or tested further) with archeological/ palaeoecological/historical records. The bullet points note variants on each theme, further explanations and implications.

54 ENVIRONMENTAL SCIENCES↗

A finite element procedure for radio-frequency sheath–plasma interactions based on a sheath impedance model

A finite element code that solves self-consistent radio-frequency (RF) sheath-plasma interaction problems is improved by incorporating a generalized sheath boundary condition in the macroscopic solution scheme. This sheath boundary condition makes use of a complex sheath impedance including both the sheath capacitance and resistance, which enables evaluation of not only the RF voltage across the sheath but also the power dissipation in the sheath. The newly developed finite element procedure is applied to cases where the background magnetic field is perpendicular to the sheath surface in one- and two-dimensional domains filled by uniform low- and high-density plasmas. The numerical results are compared with those obtained by employing the previous capacitive sheath model at a typical frequency for ion cyclotron heating used in fusion experiments. It is shown that for sheaths on the order of 100 V in a high-density plasma, localized RF power deposition can reach a level which causes material damage. It is also shown that the sheath-plasma wave resonances predicted by the capacitive sheath model do not occur when parameters are such that the generalized sheath impedance model substantially modifies the capacitive character of the sheath. Here, possible explanations for the difference in the maximum RF sheath voltage depending on the plasma density are also discussed.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

STORM: Scrape-off layer turbulence in tokamak fusion reactors

The scrape-off layer of a tokamak fusion reactor carries the plasma exhaust from the hot core plasma to the material surfaces of the reactor vessel. The heat loads imposed by the exhaust are a critical limit on the performance of fusion power plants. Turbulent transport of the plasma regulates the width of the scrape-off layer plasma and must be modelled to understand the intensity of these heat loads. STORM is a plasma turbulence code capable of simulating three dimensional turbulence across the full scrape-off layer of a tokamak fusion reactor, using a drift reduced, collisional fluid model. STORM uses mostly finite difference schemes, with a staggered grid in the direction parallel to the magnetic field. We describe the model, geometry and initialisation options used by STORM, as well as the numerical methods, which are implemented using the BOUT++ plasma simulation framework. BOUT++ has been enhanced alongside the development of STORM, providing better support for staggered grid methods. We summarise these enhancements, including a detailed explanation of the parallel derivative methods, which underwent a major update for version 4 of BOUT++.

BOUT++↗

Determination of the collisional energy transfer distribution responsible for the collision-induced dissociation of NO 2 with Ar

Collisional energy transfer is an essential aspect of chemical reactivity and maintenance of thermal equilibrium. We report the shape (energy-dependence) of the collisional energy transfer probability function for collisions of vibrationally excited NO 2 entrained in a molecular beam and photoexcited to within 40 cm -1 of its dissociation threshold. The internally excited molecules undergo collisions with Ar atoms in a crossed beam apparatus. Dissociative collisions rapidly produce the NO(J) fragment, which is observed by velocity-mapped ion imaging and REMPI techniques. The measured collisional energy transfer function is obtained via energy conservation and is compared with the results of classical trajectory calculations. Good agreement between the theory and experiment is found for collisions that transfer small amounts of energy, but the theory predicts a higher likelihood of energetic collisions than is observed experimentally. We explore possible explanations for this discrepancy in the dynamics of the collision excitation process.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Adsorption of NO on Fe 3 O 4 (111)

Adsorption of NO on Fe 3 O 4 (111) is studied by density functional theory (DFT) calculations. NO is prefer-ably adsorbed atop the octahedral site which has a clearly higher adsorption energy than the tetrahedral site. The difference in adsorption energy correlates with differences in adsorption geometries and NO stretch vibrations. Here, the results are in good agreement with temperature programmed desorption (TPD)and reflection-adsorption IR spectroscopy (RAIRS) measurements and provide an explanation of the observation of only one vibrational mode despite two distinct TPD features.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

The Empirical Case For 10 GeV Dark Matter

In this article, I summarize and discuss the body of evidence which has accumulated in favor of dark matter in the form of approximately 10 GeV particles. This evidence includes the spectrum and angular distribution of gamma rays from the Galactic Center, the synchrotron emission from the Milky Way's radio filaments, the diffuse synchrotron emission from the Inner Galaxy (the 'WMAP Haze') and low-energy signals from the direct detection experiments DAMA/LIBRA, CoGeNT and CRESST-II. This collection of observations can be explained by a relatively light dark matter particle with an annihilation cross section consistent with that predicted for a simple thermal relic (sigma v ~ 10^-26 cm^3/s) and with a distribution in the halo of the Milky Way consistent with that predicted from simulations. Astrophysical explanations for the gamma ray and synchrotron signals, in contrast, have not been successful in accommodating these observations. Similarly, the phase of the annual modulation observed by DAMA/LIBRA (and now supported by CoGeNT) is inconsistent with all known or postulated modulating backgrounds, but are in good agreement with expectations for dark matter scattering. This scenario is consistent with all existing indirect and collider constraints, as well as the constraints placed by CDMS. Consistency with xenon-based experiments can be achieved if the response of liquid xenon to very low-energy nuclear recoils is somewhat suppressed relative to previous evaluations, or if the dark matter possesses different couplings to protons and neutrons.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Two Emission Mechanisms in the Fermi Bubbles: A Possible Signal of Annihilating Dark Matter

We study the variation of the spectrum of the Fermi Bubbles with Galactic latitude. Far from the Galactic plane (|b| > 30 degrees), the observed gamma-ray emission is nearly invariant with latitude, and is consistent with arising from inverse Compton scattering of the interstellar radiation field by cosmic-ray electrons with an approximately power-law spectrum. The same electrons in the presence of microgauss-scale magnetic fields can also generate the the observed microwave "haze". At lower latitudes (b < 20 degrees), in contrast, the spectrum of the emission correlated with the Bubbles possesses a pronounced spectral feature peaking at 1-4 GeV (in E^2 dN/dE) which cannot be generated by any realistic spectrum of electrons. Instead, we conclude that a second (non-inverse-Compton) emission mechanism must be responsible for the bulk of the low-energy, low-latitude emission. This second component is spectrally similar to the excess GeV emission previously reported from the Galactic Center (GC), and also appears spatially consistent with a luminosity per volume falling approximately as r^-2.4, where r is the distance from the GC. We argue that the spectral feature visible in the low-latitude Bubbles is the extended counterpart of the GC excess, now detected out to at least 2-3 kpc from the GC. The spectrum and angular distribution of the signal is consistent with that predicted from ~10 GeV dark matter particles annihilating to leptons, or from ~50 GeV dark matter particles annihilating to quarks, following a distribution similar to the canonical Navarro-Frenk-White (NFW) profile. We also consider millisecond pulsars as a possible astrophysical explanation for the signal, as observed millisecond pulsars possess a spectral cutoff at approximately the required energy. Any such scenario would require a large population of unresolved millisecond pulsars extending at least 2-3 kpc from the GC.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗