SEARCH · Engineering Papers

Results for “inverse prediction”

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 955 records · Page 53

All sources NASA NTRS DOE OSTI

DOE OSTI2023

Probabilistic-learning-based stochastic surrogate model from small incomplete datasets for nonlinear dynamical systems

We consider a high-dimensional nonlinear computational model of a dynamical system, parameterized by a vector-valued control parameter, in the presence of uncertainties represented by an uncontrolled parameter modeled by a vector-valued random variable, and possibly with stochastic excitation. The objective is to construct a statistical surrogate model where the input is any deterministic value of the control parameter, and the output is a vector-valued observation of the computational model, which is a random vector whose probability measure is updated using a target dataset. To construct this statistical surrogate model, the stochastic response of the computational model must be built, which is a vector-valued time-discretized stochastic process in high dimension, depending on the control parameter. It is assumed that the computational cost of a single evaluation of the deterministic model is high. For the probabilistic updating, we consider a subset of the components of the observation of the computational model, defined as the “identification observation” of the computational model, for which a small target dataset is available. Therefore, the target dataset is associated with partial observability, corresponding to an incomplete data case. Given a prior probability model of the random control and uncontrolled parameters, a training dataset is constructed, consisting of realizations of the random triplet composed of the stochastic response, the random identification observation, and the random control parameter. Since the computational cost of a single evaluation of the deterministic model is assumed to be large, the training dataset is also of small size. The main challenges in this problem are the high dimensionality, partial observability leading to incomplete data in the target dataset for the identification observation of the computational model (which is not sufficient to identify the computational stochastic responses), and the availability of a small training dataset. To address these challenges, we propose a methodology based on statistical methods for constructing necessary reduced representations, direct probabilistic learning under constraints using probabilistic learning on manifolds (PLoM) constrained by the target dataset, and the use of a weak formulation of the Fourier transform of probability measures. Statistical conditioning is also employed to explore the learned dataset. The constructed predictive statistical surrogate model can be implemented in the context of online computation. Here, we apply this approach to a problem of nonlinear stochastic dynamics in high dimensions within the framework of deformable solids mechanics.

Engineering↗

DOE OSTI2020

Quadrupole topological photonic crystals

Quadrupole topological phases, exhibiting protected boundary states that are themselves topological insulators of lower dimensions, have recently been of great interest. Extensions of these ideas from current tight binding models to continuum theories for realistic materials require the identification of quantized invariants describing the bulk quadrupole order. Here we identify the analog of quadrupole order in Maxwell’s equations for a gyromagnetic photonic crystal (PhC) through a double-band-inversion process. The quadrupole moment is quantized by the simultaneous presence of crystalline symmetry and broken time-reversal symmetry, which is confirmed using three independent methods: analysis of symmetry eigenvalues, numerical calculations of the nested Wannier bands and the expectation value of the quadrupole operator. Furthermore, we reveal the boundary manifestations of quadrupole phases as quantized edge polarizations and fractional corner charges. The latter are the consequence of a filling anomaly of energy bands as first predicted in electronic systems.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

DOE OSTI2022

Quasielastic least-squares reverse time migration of PS reflections

Conventional analysis of amplitude variation with offset for elastic PS reflections is based on analytical reflection coefficients in a layered medium, and wave-equation-based PS migration is mainly used to produce a structural image. To overcome this problem, we have developed a least-squares reverse time migration (LSRTM) method for elastic PS reflections based on a quasielastic wave equation. Here, the quasielastic wave equation can accurately model PS reflections with elastic amplitudes under the first-order Born approximation. Our LSRTM method inverts for perturbations of the S-wave velocity and density by minimizing the L 2 norm of the difference between recorded and predicted PS reflections modeled using a quasielastic wave equation. We refer to our new method as quasielastic LSRTM of PS reflections. Numerical tests on synthetic and field data indicate that our method can properly handle the amplitudes of elastic PS reflections and provides an accurate estimate of the perturbations of S-wave velocity and density. Extending the method to the 3D case is not straightforward and might require incorporating certain PS data processing techniques into the inversion itself.

58 GEOSCIENCES↗

DOE OSTI2023

Unveiling the Significance of Correlations in K-Space and Configuration Space for Drift Wave Turbulence in Tokamaks

Turbulence and transport phenomena play a crucial role in the confinement and stability of tokamak plasmas. Turbulent fluctuations in certain physical quantities, such as density or temperature fluctuations, can have a wide range of spatial scales, and understanding their correlation length is important for predicting and controlling the behavior of the plasma. The correlation length in the radial direction is identified as the critical length in real space. The dynamics in real space are of significant interest because transport in configuration space is primarily focused on them. When investigating transport caused by the E×B drift, the correlation length in real space represents the size of E×B whirls. It was numerically discovered that in drift wave turbulence, this length is inversely proportional to the normalized mode number of the fastest growing mode relative to the drift frequency. Considerable time was required before a proper analytical derivation of this condition was accomplished. Therefore, a connection has been established between phenomena occurring in real space and those occurring in k-space. Although accompanied by a turbulent spectrum in k-space with a substantial width, transport in real space is uniquely determined by the correlation length, allowing for accurate transport calculations through the dynamics of a single mode. Naturally, the dynamics are subject to nonlinear effects, with resonance broadening in frequency being the most significant nonlinear effect. Thus, mode number space is once again involved. Resonance broadening leads to the detuning of waves from particles, permitting a fluid treatment. It should be emphasized that the consideration here involves the total electric field, including the induction part, which becomes particularly important at higher beta plasmas.

Weiland, Jan (ORCID:0000000253496025)↗

NASA NTRS1981

Intergalactic shells at large redshift

The intergalactic shells produced by galactic explosions at large redshift, whose interiors cool by inverse Compton scattering off the cosmic background radiation, have a characteristic angular size of about 1 arcmin at peak brightness. At z values lower than 2, the shells typically have a radius of 0.5 Mpc, a velocity of about 50 km/sec, a metal abundance of about 0.0001 of cosmic values, and strong radiation in H I(Lyman-alpha), He II 304 A, and the IR fine-structure lines of C II and Si II. The predicted extragalactic background emission from many shells, strongly peaked toward the UV, sets an upper limit to the number of exploding sources at z values of about 10. Shell absorption lines of H I, C II, Si II, and Fe II, which may be seen at more recent epochs in quasar spectra, may probe otherwise invisible explosions in the early universe.

Shull, J. M.↗

NASA NTRS1985

Bidirectional Reflectance Modeling of Non-homogeneous Plant Canopies

The objective of this research is to develop a 3-dimensional radiative transfer model for predicting the bidirectional reflectance distribution function (BRDF) for heterogeneous vegetation canopies. The model (named BIGAR) considers the angular distribution of leaves, leaf area index, the location and size of individual subcanopies such as widely spaced rows or trees, spectral and directional properties of leaves, multiple scattering, solar position and sky condition, and characteristics of the soil. The model relates canopy biophysical attributes to down-looking radiation measurements for nadir and off-nadir viewing angles. Therefore, inversion of this model, which is difficult but practical should provide surface biophysical pattern; a fundamental goal of remote sensing. Such a model also will help to evaluate atmospheric limitations to satellite remote sensing by providing a good surface boundary condition for many different kinds of canopies. Furthermore, this model can relate estimates of nadir reflectance, which is approximated by most satellites, to hemispherical reflectance, which is necessary in the energy budget of vegetated surfaces.

Norman, J. M.↗

NASA NTRS1994

The Uranian aurora and its relationship to the magnetosphere

About 32 h of Voyager Ultraviolet Spectrometer (UVS) observations of Uranus H2 band airglow emission (875 less than or equal to lambda less than or equal to 1115 A) have been analyzed using the singular value decomposition (SVD) approach to inversion, producing an intensity map showing aurora at both magnetic poles. An H Lyman alpha aurora may also be present but is difficult to separate from scattered solar and local interstellar medium components. SVD analysis of variance shows that the intensity estimate is significantly larger than the error estimate over both Uranographic poles and part of the equatorial region, fortuitously including both magnetic polar regions. The Goddard Space Flight Center Q(sub 3) magnetic field model correctly predicts that the aurora should be larger in area and emit more power at the weaker N magnetic pole than at the stronger S magnetic pole. However, the auroral emissions are quite localized in magnetic longitude and so do not form complete auroral ovals. The brightest auroral emission at each magnetic pole is confined to a range of approximately 90 deg of magnetic longitude centered on the magnetotail direction, at moderate magnetic L parameter (5 less than or equal to L less than or equal to 10), but some emission at each pole is distributed over a range of more than 180 deg of longitude. The magnetic longitudes of the aurora are completely inconsistent with the 'windshield wiper' effect for either ions or electrons, indicating that some other effect, such as rapid depletion of the population of precipitating particles of highly localized strong pitch-angle diffusion, may be acting to localize emission. The low apparent L of the precipitating particles indicates that their energies may be less than or equal to 10 keV. Hence magnetospheric convection is likely to be important, and thus particles exciting the aurora may not remain on constant L shells. The precipitating particles may be a relatively low-energy population at high L that is heated to aurora-exciting energy by adiabatic compression during convection to low L. We estimate that the total auroral power output at H Lyman alpha and shorter wavelengths is about 3 x 10(exp 9) to 7 x 10(exp 9) W, requiring about 10 times that much power for excitation.

Herbert, Floyd↗

NASA NTRS2013

Downscaling Satellite Precipitation with Emphasis on Extremes: A Variational 1-Norm Regularization in the Derivative Domain

The increasing availability of precipitation observations from space, e.g., from the Tropical Rainfall Measuring Mission (TRMM) and the forthcoming Global Precipitation Measuring (GPM) Mission, has fueled renewed interest in developing frameworks for downscaling and multi-sensor data fusion that can handle large data sets in computationally efficient ways while optimally reproducing desired properties of the underlying rainfall fields. Of special interest is the reproduction of extreme precipitation intensities and gradients, as these are directly relevant to hazard prediction. In this paper, we present a new formalism for downscaling satellite precipitation observations, which explicitly allows for the preservation of some key geometrical and statistical properties of spatial precipitation. These include sharp intensity gradients (due to high-intensity regions embedded within lower-intensity areas), coherent spatial structures (due to regions of slowly varying rainfall),and thicker-than-Gaussian tails of precipitation gradients and intensities. Specifically, we pose the downscaling problem as a discrete inverse problem and solve it via a regularized variational approach (variational downscaling) where the regularization term is selected to impose the desired smoothness in the solution while allowing for some steep gradients(called 1-norm or total variation regularization). We demonstrate the duality between this geometrically inspired solution and its Bayesian statistical interpretation, which is equivalent to assuming a Laplace prior distribution for the precipitation intensities in the derivative (wavelet) space. When the observation operator is not known, we discuss the effect of its misspecification and explore a previously proposed dictionary-based sparse inverse downscaling methodology to indirectly learn the observation operator from a database of coincidental high- and low-resolution observations. The proposed method and ideas are illustrated in case studies featuring the downscaling of a hurricane precipitation field.

Hurricanes↗

NASA NTRS2015

Under-Track CFD-Based Shape Optimization for a Low-Boom Demonstrator Concept

The detailed outer mold line shaping of a Mach 1.6, demonstrator-sized low-boom concept is presented. Cruise trim is incorporated a priori as part of the shaping objective, using an equivalent-area-based approach. Design work is performed using a gradient-driven optimization framework that incorporates a three-dimensional, nonlinear flow solver, a parametric geometry modeler, and sensitivities derived using the adjoint method. The shaping effort is focused on reducing the under-track sonic boom level using an inverse design approach, while simultaneously satisfying the trim requirement. Conceptual-level geometric constraints are incorporated in the optimization process, including the internal layout of fuel tanks, landing gear, engine, and crew station. Details of the model parameterization and design process are documented for both flow-through and powered states, and the performance of these optimized vehicles presented in terms of inviscid L/D, trim state, pressures in the near-field and at the ground, and predicted sonic boom loudness.

Wintzer, Mathias↗

NASA NTRS2023

Time Series of Magnetic Field Parameters Extracted from Merged Space-Weather MDI/HMI Active Region Patches as Potential Tool for Solar Flare Forecasting

Space-Weather MDI Active Region Patches (SMARPs) and Space-Weather HMI Active Region Patches (SHARPs) are two recently developed data products, which have been used for solar flare prediction studies. The present work is an effort to expand the application of SMARP and SHARP summary heliomagnetic parameters to the forecasting of solar flares. A new data product was derived by filtering, rescaling, and merging the SMARP and SHARP summary parameter data series, which were further converted into two-dimensional arrays by selecting time slices corresponding to R-value maxima, where R-value is a measure of the unsigned magnetic flux near polarity inversion lines. The resulting combined MDI-HMI time series currently span the period between April 4, 1996 and December 13, 2022, and can be extended to a more recent date, providing an opportunity to correlate and compare them with other solar activity parameters, such as the daily solar flare index, which is computed as a sum of the product of GOES X-ray flare magnitude and flare duration, for all M- and X-class flares during a day. Preliminary results demonstrate a significant overall correlation, with Pearson coefficients between 0.339 and 0.627. In addition, an oscillating pattern is seen in the daily-averaged sliding-window correlation coefficient. Time-lagged cross-correlation indicates that a leader-follower dynamic exists in some parameters, especially R-value, where they lead the flare index by at least several days, which may have potential for further application in space weather forecasting.

Heliophysics↗

DOE OSTI2024

Machine learning-based discovery of molecular descriptors that control polymer gas permeation

While machine learning has found increasing use in predicting the properties of polymeric materials with only a knowledge of chain architecture, determining the molecular factors underpinning properties (“interpretable AI”) has remained less well explored. We show that encoding chain chemistry in commonly employed formats, e.g., binary-valued fingerprints, leads to uniqueness issues during the hashing process to save storage space. This is because the hashing algorithm can map several chemical moieties into the same bit. These issues carry over into the ML algorithms, especially for “inverse” design and interpretable AI, and cannot be avoided by changing the length of the fingerprint. Using MACCS key featurizations of monomer repeats resolves some of these issues, and we show that a few substructures consistently appear in top features for maximizing permeability across several gases and ML models. These are carbon–carbon double bonds (as in polyacetylenes) especially when they are associated with methyl groups (found in branching architectures). Here these results, derived from the limited data set of ~ 500 polymers with experimental gas permeation data, are in agreement with physical insight and thus provide a robust foundation which could further enable study of these material classes through detailed experiments and simulations.

36 MATERIALS SCIENCE↗

DOE OSTI2023

Hexadecapole strength in the rare isotopes 74,76 Kr

In the Ge-Sr mass region, isotopes with neutron number N ≤ 40 are known to feature rapid shape changes with both nucleon number and angular momentum. To gain new insights into their structure, inelastic proton scattering experiments in inverse kinematics were performed on the rare isotopes 74,76 Kr. This work focuses on observables related to the J π = $4$$^{+}_{1}$ states of the Kr isotopes and, in particular, on the hexadecapole degree of freedom. By performing coupled-channels calculations, hexadecapole deformation parameters β 4 were determined for the J π = $4$$^{+}_{1}$ states of 74,76 Kr from inelastic proton scattering cross sections. Two possible coupled-channels solutions were found. A comparison to predictions from nuclear energy density functional theory, employing both non-relativistic and relativistic functionals, clearly favors the large, positive β 4 solutions. These β 4 values are unambiguously linked to the well deformed prolate configuration. Given the β 2 - β 4 trend, established in this work, it appears that β 4 values could provide a sensitive measure of the nuclear shell structure.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

NASA NTRS1983

Comparison of model high-lititude electron densities with Millstone Hill observations

The predictions of a high-latitude ionospheric model are compared with the diurnal variations of plasma convection velocities and electron densities observed at Millstone Hill on a geomagnetically moderately active day near equinox. The observed convection pattern was consistent with a two-cell, asymmetric pattern with enhanced plasma flow in the dusk sector, with flow speeds reaching 1.5 km/s. In the dusk strong convection cell, the falloff of the magnetospheric potential with latitude was proportional to the inverse of the sine of colatitude to the fourth power. On the dayside, a region of high density occurred at 500 km in the 1000-1900 LT sector. The nocturnal midlatitude trough was deepest and widest and reached its most equatorward position in the morning sector. The model, which is based on average auroral precipitation fluxes, can describe the gross features of the enhanced densities in the auroral zone.

Sojka, J. J.↗

NASA NTRS1993

Trajectory optimization for the National Aerospace Plane

The objective of this second phase research is to investigate the optimal ascent trajectory for the National Aerospace Plane (NASP) from runway take-off to orbital insertion and address the unique problems associated with the hypersonic flight trajectory optimization. The trajectory optimization problem for an aerospace plane is a highly challenging problem because of the complexity involved. Previous work has been successful in obtaining sub-optimal trajectories by using energy-state approximation and time-scale decomposition techniques. But it is known that the energy-state approximation is not valid in certain portions of the trajectory. This research aims at employing full dynamics of the aerospace plane and emphasizing direct trajectory optimization methods. The major accomplishments of this research include the first-time development of an inverse dynamics approach in trajectory optimization which enables us to generate optimal trajectories for the aerospace plane efficiently and reliably, and general analytical solutions to constrained hypersonic trajectories that has wide application in trajectory optimization as well as in guidance and flight dynamics. Optimal trajectories in abort landing and ascent augmented with rocket propulsion and thrust vectoring control were also investigated. Motivated by this study, a new global trajectory optimization tool using continuous simulated annealing and a nonlinear predictive feedback guidance law have been under investigation and some promising results have been obtained, which may well lead to more significant development and application in the near future.

Lu, Ping↗

NASA NTRS2010

Success Stories in Control: Nonlinear Dynamic Inversion Control

NASA plays an important role in advancing the state of the art in flight control systems. In the case of Nonlinear Dynamic Inversion (NDI) NASA supported initial implementation of the theory in an aircraft and demonstration in a space vehicle. Dr. Dale Enns of Honeywell Aerospace Advanced Technology performed this work in cooperation with NASA and under NASA contract. Honeywell and Lockheed Martin were subsequently contracted by AFRL to create "Design Guidelines for Multivariable Control Theory". This foundational work directly contributed to the advancement of the technology and the credibility of the control law as a design option. As a result Honeywell collaborated with Lockheed Martin to produce a Nonlinear Dynamic Inversion controller for the X-35 and subsequently Lockheed Martin did the same for the production Lockheed Martin F-35 vehicle. The theory behind NDI is to use a systematic generalized approach to controlling a vehicle. Using general aircraft nonlinear equations of motion and onboard aerodynamic, mass properties, and engine models specific to the vehicle, a relationship between control effectors and desired aircraft motion can be formulated. Using this formulation a control combination is used that provides a predictable response to commanded motion. Control loops around this formulation shape the response as desired and provide robustness to modeling errors. Once the control law is designed it can be used on a similar class of vehicle with only an update to the vehicle specific onboard models.

Bosworth, John T.↗

NASA NTRS2005

Experimental and Analytical Determination of the Geometric Far Field for Round Jets

An investigation was conducted at the NASA Glenn Research Center using a set of three round jets operating under unheated subsonic conditions to address the question: "How close is too close?" Although sound sources are distributed at various distances throughout a jet plume downstream of the nozzle exit, at great distances from the nozzle the sound will appear to emanate from a point and the inverse-square law can be properly applied. Examination of normalized sound spectra at different distances from a jet, from experiments and from computational tools, established the required minimum distance for valid far-field measurements of the sound from subsonic round jets. Experimental data were acquired in the Aeroacoustic Propulsion Laboratory at the NASA Glenn Research Center. The WIND computer program solved the Reynolds-Averaged Navier-Stokes equations for aerodynamic computations; the MGBK jet-noise prediction computer code was used to predict the sound pressure levels. Results from both the experiments and the analytical exercises indicated that while the shortest measurement arc (with radius approximately 8 nozzle diameters) was already in the geometric far field for high-frequency sound (Strouhal number >5), low-frequency sound (Strouhal number <0.2) reached the geometric far field at a measurement radius of at least 50 nozzle diameters because of its extended source distribution.

Koch, L. Danielle↗

NASA NTRSDate not supplied

Combining Data with Physical Knowledge for Uncertainty Quantification in Certification and Reliability Analysis

Unifying empirical data with predictive models can enable engineering cost-savings through certification by analysis and reliability-based design. Both concepts require rigorous uncertainty quantification (UQ) and robust understanding and treatment of relevant physics. Combining sampling-based UQ algorithms with high-fidelity simulations creates a computational bottleneck that is often alleviated through the use of machine learning (ML). ML can be used to create computationally efficient surrogates for simulations of complex or high-dimensional physical interactions (e.g., multi-phase interactions associated with melt pools in laser powder bed fusion or spatially-dependent material properties in functionally graded materials). However, negative side effects of ML may include a lack of interpretability and negative correlation between event rarity and simulation accuracy due to a lack of training data. As such, it is important to infuse ML algorithms with physics-based guardrails to provide confidence in their predictions. This talk will provide a brief review of recent NASA research at this intersection of physics-based simulation, ML, and UQ with a focus on certification and reliability analysis.

uncertainty quantification↗

DOE OSTI2025

An adaptive and stability-promoting layerwise training approach for sparse deep neural network architecture

This work presents a two-stage adaptive framework for progressively developing deep neural network (DNN) architectures that generalize well for a given training data set. In the first stage, a layerwise training approach is adopted where a new layer is added each time and trained independently by freezing parameters in the previous layers. We impose desirable structures on the DNN by employing manifold regularization, sparsity regularization, and physics-informed terms. We introduce a ε – δ – stability-promoting concept as a desirable property for a learning algorithm and show that employing manifold regularization yields a ε – δ stability-promoting algorithm. Further, we also derive the necessary conditions for the trainability of a newly added layer and investigate the training saturation problem. In the second stage of the algorithm (post-processing), a sequence of shallow networks is employed to extract information from the residual produced in the first stage, thereby improving the prediction accuracy. Numerical investigations on prototype regression and classification problems demonstrate that the proposed approach can outperform fully connected DNNs of the same size. Moreover, by equipping the physics-informed neural network (PINN) with the proposed adaptive architecture strategy to solve partial differential equations, we numerically show that adaptive PINNs not only are superior to standard PINNs but also produce interpretable hidden layers with provable stability. As a result, we also apply our architecture design strategy to solve inverse problems governed by elliptic partial differential equations.

42 ENGINEERING↗