Search NASA⌕ Search

SEARCH · Search NASA

Results for “generalization bounds”

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 307 records · Page 17
All sourcesNASA NTRSDOE OSTI
NASA NTRS2005

Uncertainty Evaluation of Computational Model Used to Support the Integrated Powerhead Demonstration Project

NASA and the U.S. Air Force are working on a joint project to develop a new hydrogen-fueled, full-flow, staged combustion rocket engine. The initial testing and modeling work for the Integrated Powerhead Demonstrator (IPD) project is being performed by NASA Marshall and Stennis Space Centers. A key factor in the testing of this engine is the ability to predict and measure the transient fluid flow during engine start and shutdown phases of operation. A model built by NASA Marshall in the ROCket Engine Transient Simulation (ROCETS) program is used to predict transient engine fluid flows. The model is initially calibrated to data from previous tests on the Stennis E1 test stand. The model is then used to predict the next run. Data from this run can then be used to recalibrate the model providing a tool to guide the test program in incremental steps to reduce the risk to the prototype engine. In this paper, they define this type of model as a calibrated model. This paper proposes a method to estimate the uncertainty of a model calibrated to a set of experimental test data. The method is similar to that used in the calibration of experiment instrumentation. For the IPD example used in this paper, the model uncertainty is determined for both LOX and LH flow rates using previous data. The successful use of this model is then demonstrated to predict another similar test run within the uncertainty bounds. The paper summarizes the uncertainty methodology when a model is continually recalibrated with new test data. The methodology is general and can be applied to other calibrated models.

Steele, W. G.↗
NASA NTRS2009

Dynamic Stability Testing of the Mars Science Laboratory Entry Capsule

Results from a 26 shot ballistic range test of the Mars Science Laboratory (MSL) entry capsule are presented. The supersonic pitch damping properties of the MSL capsule were characterized between Mach 1.35 and Mach 3.5 and total angles-of-attack from 0 to 30 degrees. In flight, the MSL entry capsule will utilize a radial center-of-gravity offset to produce a non-zero trim angle-of-attack. This offset trim angle will produce lift, enabling the capsule to fly a guided entry and reducing the landing footprint dimensions to within 10 km of the desired landing site. A lifting configuration could not be tested at the ballistic range used for this test as the models would swerve into the range walls, possibly damaging cameras, the coordinate reference system or other facility assets. Ballistic (non-lifting) data was extracted and will be implemented in a conservative fashion to ensure that the dynamic stability characteristics of the flight vehicle are bounded. A comparison between the MSL pitch damping results and the dynamic model of the Mars Exploration Rover capsule shows generally close agreement with no significant differences in damping characteristics due to the change in backshell geometry. Dynamic moments are also compared to the MSL reaction control system (RCS) control authority to show the controller has sufficient margin to easily damp any dynamic stability effects.

Schroenenberger, Mark↗
DOE OSTI2025

Feature learning and generalization in deep networks with orthogonal weights

Fully-connected deep neural networks with weights initialized from independent Gaussian distributions can be tuned to criticality, which prevents the exponential growth or decay of signals propagating through the network. However, such networks still exhibit fluctuations that grow linearly with the depth of the network, which may impair the training of networks with width comparable to depth. We show analytically that rectangular networks with tanh activations and weights initialized from the ensemble of orthogonal matrices have corresponding preactivation fluctuations which are independent of depth, to leading order in inverse width. Moreover, we demonstrate numerically that, at initialization, all correlators involving the neural tangent kernel (NTK) and its descendants at leading order in inverse width—which govern the evolution of observables during training—saturate at a depth of ~20, rather than growing without bound as in the case of Gaussian initializations. We speculate that this structure preserves finite-width feature learning while reducing overall noise, thus improving both generalization and training speed in deep networks with depth comparable to width. We provide some experimental justification by relating empirical measurements of the NTK to the superior performance of deep non-linear orthogonal networks trained under full-batch gradient descent on the MNIST and CIFAR-10 classification tasks.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1987

WHAMII - An enumeration and insertion procedure with binomial bounds for the stochastic time-constrained traveling salesman problem

This paper presents an algorithm (WHAMII) designed to solve the Artificial Intelligence Design Challenge at the 1987 AIAA Guidance, Navigation and Control Conference. The problem under consideration is a stochastic generalization of the traveling salesman problem in which travel costs can incur a penalty with a given probability. The variability in travel costs leads to a probability constraint with respect to violating the budget allocation. Given the small size of the problem (eleven cities), an approach is considered that combines partial tour enumeration with a heuristic city insertion procedure. For computational efficiency during both the enumeration and insertion procedures, precalculated binomial probabilities are used to determine an upper bound on the actual probability of violating the budget constraint for each tour. The actual probability is calculated for the final best tour, and additional insertions are attempted until the actual probability exceeds the bound.

Dahl, Roy W.↗
DOE OSTI2024

Diagnostics of Mixed-State Topological Order and Breakdown of Quantum Memory

Topological quantum memory can protect information against local errors up to finite error thresholds. Such thresholds are usually determined based on the success of decoding algorithms rather than the intrinsic properties of the mixed states describing corrupted memories. Here we provide an intrinsic characterization of the breakdown of topological quantum memory, which both gives a bound on the performance of decoding algorithms and provides examples of topologically distinct mixed states. We employ three information-theoretical quantities that can be regarded as generalizations of the diagnostics of ground-state topological order, and serve as a definition for topological order in error-corrupted mixed states. We consider the topological contribution to entanglement negativity and two other metrics based on quantum relative entropy and coherent information. In the concrete example of the two-dimensional (2D) Toric code with local bit-flip and phase errors, we map three quantities to observables in 2D classical spin models and analytically show they all undergo a transition at the same error threshold. This threshold is an upper bound on that achieved in any decoding algorithm and is indeed saturated by that in the optimal decoding algorithm for the Toric code. Published by the American Physical Society 2024

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2025

Reduced Order Modeling conditioned on monitored features for response and error bounds estimation in engineered systems

Reduced Order Models (ROMs) form essential tools across engineering domains by virtue of their function as surrogates for computationally intensive digital twinning simulators. Although purely data-driven methods are available for ROM construction, schemes that allow to retain a portion of the physics tend to enhance the interpretability and generalization of ROMs. However, physics-based techniques can adversely scale when dealing with nonlinear systems that feature parametric dependencies. This study introduces a generative physics-based ROM that is suited for nonlinear systems with parametric dependencies and is additionally able to provide numerical error bounds associated with the respective estimates. A main contribution of this work is the conditioning of these parametric ROMs to features that can be derived from monitoring measurements, feasibly in an online fashion. This is contrary to most existing ROM schemes, which remain restricted to the prescription of the physics-based, and usually a priori unknown, system parameters. Our work utilizes conditional Variational Autoencoders to continuously map the required reduction bases to a feature vector extracted from limited output measurements, while additionally allowing for a probabilistic assessment of the ROM-estimated Quantities of Interest. An auxiliary task using a neural network-based parametrization of suitable probability distributions is introduced to re-establish the link with physical model parameters. We verify the proposed scheme on a series of simulated case studies incorporating effects of geometric and material nonlinearity under parametric dependencies related to system properties and input load characteristics.

Conditional VAEs↗
DOE OSTI2024

Photovoltaic Inverter Failure Mechanism Estimation Using Unsupervised Machine Learning and Reliability Assessment

This article introduces a data-driven approach to assessing failure mechanisms and reliability degradation in outdoor photovoltaic (PV) string inverters. The manufacturer's stated PV inverter lifetime can vary due to the impact of operating site conditions. To address limitations in degradation estimation through accelerated testing, condition monitoring, or degradation modeling, we propose a machine learning (ML) oriented approach. Utilizing data from a 1.4 MW PV power plant operational since 2016, with 46 string PV inverters tied to the grid, we employ the unsupervised one-class support vector machine ML technique to analyze inverter and sensor data, capable of classifying humidity cycling and temperature fluctuations as dominant failure mechanisms. Utilizing the anomaly alert relationship and alert details specific to the inverter, the level of PV inverter output is considered as its availability or available reliability. Subsequently, a continuous Markov model is applied to six-month alert data, revealing an average stated reliability of 20% after 20 years of continuous operation. These results support recommendations for time-bound preventive measures to enhance PV inverter reliability under diverse outdoor conditions. Furthermore, the approach provides a nondestructive, top–down, and generalized method for analyzing any commercial PV inverter exposed to outdoor conditions, contingent on the availability of relevant data.

14 SOLAR ENERGY↗
DOE OSTI2026

Emergent pair density wave order across a Lifshitz transition

We numerically investigate the telltale signs of pair-density-wave order (PDW) in the Kondo-Heisenberg chain by focusing on the momentum resolved spectrum in different parameter regimes. Density matrix renormalization group calculations reveal that this phase is characterized by a dispersion with two minima and four Fermi points, indicating the emergence of an effective next-nearest-neighbor hopping that arises as a second-order effect to avoid magnetic frustration. The pairs appear in the spectrum as in-gap bound states with weight concentrated in the hole pockets. The low-energy physics can be understood by means of a generalized t−J model with next-nearest-neighbor hopping. Our results offer a guide for searching for experimental signatures, and for other models that can realize PDW physics.

Yang, Luhang [University of Tennessee Knoxville]↗
NASA NTRS1974

Three-body Coulomb systems using generalized angular-momentum S states

An expansion of the three-body Coulomb potential in generalized angular-momentum eigenfunctions developed earlier by one of the authors is used to compute energy eigenvalues and eigenfunctions of bound S states of three-body Coulomb systems. The results for He, H(-), e(-)e(+)e(-), and pmu(-)p are compared with the results of other computational approaches.

Whitten, R. C.↗
NASA NTRS1979

Phosphine absorption in the 5-micron window of Jupiter

Since the original suggestion by Gillett et al. (1969) it has generally been assumed that the region of partial transparency near 5 micron in Jupiter's atmosphere (the 5-micron window) is bounded by the nu sub 4 NH3 at 6.1 micron and the nu sub 3 CH4 band at 3.3 micron. New measurements of Jupiter and of laboratory phosphine (PH3) samples show that PH3 is a significant contributor to the continuum opacity in the window and in fact defines its short-wavelength limit. This has important implications for the use of 5-micron observations as a means to probe the deep atmospheric structure of Jupiter. The abundance of PH3 which results from a comparison of Jovian and laboratory spectra is about 3 to 5 cm-am. This is five to eight times less than that found by Larson et al. (1977) in the same spectral region, but is in good agreement with the result of Tokunaga et al. (1979) from 10-micron observations.

Beer, R.↗
NASA NTRS1989

An analysis of curvature effects for the control of wall-bounded shear flows

The Reynolds stress transport equations are used to predict the effects of simultaneous and sequential combinations of distortions on turbulent boundary layers. The equations are written in general orthogonal curvilinear coordinates, with the curvature terms expressed in terms of the principal radii of curvature of the respective coordinate surfaces. Results are obtained for the cases of two-dimensional and three-dimensional flows in the limit where production and pressure-strain redistribution dominate over diffusion effects.

Gatski, T. B.↗
NASA NTRS1990

Bounded energy states in homogeneous turbulent shear flow: An alternative view

The equilibrium structure of homogeneous turbulent shear flow is investigated from a theoretical standpoint. Existing turbulence models, in apparent agreement with physical and numerical experiments, predict an unbounded exponential time growth of the turbulent kinetic energy and dissipation rate; only the anisotropy tensor and turbulent time scale reach a structural equilibrium. It is shown that if vortex stretching is accounted for in the dissipation rate transport equation, then there can exist equilibrium solutions, with bounded energy states, where the turbulence production is balanced by its dissipation. Illustrative calculations are present for a k-epsilon model modified to account for vortex stretching. The calculations indicate an initial exponential time growth of the turbulent kinetic energy and dissipation rate for elapsed times that are as large as those considered in any of the previously conducted physical or numerical experiments on homogeneous shear flow. However, vortex stretching eventually takes over and forces a production-equals-dissipation equilibrium with bounded energy states. The validity of this result is further supported by an independent theoretical argument. It is concluded that the generally accepted structural equilibrium for homogeneous shear flow with unbounded component energies is in need of re-examination.

Bernard, Peter S.↗
NASA NTRS1992

Electron acceleration by young supernova remnant blast waves

Some general considerations regarding relativistic particle acceleration by young supernova remnants are reviewed. Recent radio observations of supernova remnants apparently locate the bounding shock and exhibit large electron density gradients which verify the presence of strong particle scattering. The radio 'rim' in Tycho's remnant has been found to contain a predominantly radial magnetic field. This may be attributable to an instability of the shock surface and a progress report on an investigation of the stability of strong shocks in partially ionized media is presented.

Blandford, R. D.↗
NASA NTRS1996

Application of Design Methodologies for Feedback Compensation Associated with Linear Systems

The work that follows is concerned with the application of design methodologies for feedback compensation associated with linear systems. In general, the intent is to provide a well behaved closed loop system in terms of stability and robustness (internal signals remain bounded with a certain amount of uncertainty) and simultaneously achieve an acceptable level of performance. The approach here has been to convert the closed loop system and control synthesis problem into the interpolation setting. The interpolation formulation then serves as our mathematical representation of the design process. Lifting techniques have been used to solve the corresponding interpolation and control synthesis problems. Several applications using this multiobjective design methodology have been included to show the effectiveness of these techniques. In particular, the mixed H 2-H performance criteria with algorithm has been used on several examples including an F-18 HARV (High Angle of Attack Research Vehicle) for sensitivity performance.

Smith, Monty J.↗
NASA NTRS2001

Perspectives: Black Holes

When asked to discuss Cyg XR-1, E. E. Salpeter once concluded, 'A black hole in Cyg X(R)-1 is the most conservative hypothesis.' Recent observations now make it likely that a black hole in Cyg XR-1 is the only hypothesis tenable. Chandrasekhar first showed that compact stars - those with the inward force of gravity on their outer layers balanced by the pressure generated by the Pauli exclusion principle acting on its electrons (in white dwarfs) or nucleons (in neutron stars) - have a maximum mass. Equilibrium is achieved at a minimum of the total energy of the star, which is the sum of the positive Fermi energy and the negative gravitational energy. The maximum mass attainable in equilibrium is found by setting E = 0: M(max) = 1.5 M(Sun). If the mass of the star is larger than this, then E can be decreased without bound by decreasing the star's radius and increasing its (negative) gravitational energy. No equilibrium value of the radius exist, and general relativity predicts that gravitational collapse to a point occurs. This point singularity is a black hole.

Dolan, Joseph F.↗
NASA NTRS2005

Solar-System Tests of Gravitational Theories

We are engaged in testing gravitational theory, mainly using observations of objects in the solar system and mainly on the interplanetary scale. Our goal is either to detect departures from the standard model (general relativity) - if any exist within the level of sensitivity of our data - or to support this model by placing tighter bounds on any departure from it. For this project, we have analyzed a combination of observational data with our model of the solar system, including planetary radar ranging, lunar laser ranging, and spacecraft tracking, as well as pulsar timing and pulsar VLBI measurements. In the past year, we have added to our data, primarily lunar laser ranging measurements, but also supplementary data concerning the physical properties of solar-system objects, such as the solar quadrupole moment, planetary masses, and asteroid radii. Because the solar quadrupole moment contributes to the classical precession of planetary perihelia, but with a dependence on distance from the Sun that differs from that of the relativistic precession, it is possible to estimate effects simultaneously. However, our interest is mainly in the relativistic effect, and we find that imposing a constraint on the quadrupole moment from helioseismology studies, gives us a dramatic (about ten-fold) decrease in the standard error of our estimate of the relativistic component of the perihelion advance.

Shapiro, Irwin I.↗
NASA NTRS2005

Solar-System Tests of Gravitational Theories

This research is aimed at testing gravitational theory, primarily on an interplanetary scale and using mainly observations of objects in the solar system. Our goal is either to detect departures from the standard model (general relativity) - if any exist within the level of sensitivity of our data - or to support this model by placing tighter bounds on any departure from it. For this project, we have analyzed a combination of observational data with our model of the solar system, including planetary radar ranging, lunar laser ranging, and spacecraft tracking, as well as pulsar timing and pulsar VLBI measurements.

Shapiro, Irwin I.↗