Search NASA⌕ Search

SEARCH · Search NASA

Results for “dynamic inference”

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 217 records · Page 12
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Data-driven models of nonautonomous systems

Nonautonomous dynamical systems are characterized by time-dependent inputs, which complicates the discovery of predictive models describing the spatiotemporal evolution of the state variables of quantities of interest from their temporal snapshots. When dynamic mode decomposition (DMD) is used to infer a linear model, this difficulty manifests itself in the need to approximate the time-dependent Koopman operators. Our approach is to approximate the original nonautonomous system with a modified system derived via a local parameterization of the time-dependent inputs. The modified system comprises a sequence of local parametric systems, which are subsequently approximated by a parametric surrogate model using the DRIPS (dimension reduction and interpolation in parameter space) framework. The offline step of DRIPS relies on DMD to build a linear surrogate model, endowed with reduced-order bases for the observables mapped from training data. The online step interpolates on suitable manifolds to construct a sequence of iterative parametric surrogate models; the target/test parameter points on these manifolds are specified by a local parameterization of the test time-dependent inputs. Here, we use numerical experimentation to demonstrate the robustness of our method and compare its performance with that of deep neural networks.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Identifiability and predictability of integer- and fractional-order epidemiological models using physics-informed neural networks

Here we analyze a plurality of epidemiological models through the lens of physics-informed neural networks (PINNs) that enable us to identify time-dependent parameters and data-driven fractional differential operators. In particular, we consider several variations of the classical susceptible-infectious-removed (SIR) model by introducing more compartments and fractional-order and time-delay models. We report the results for the spread of COVID-19 in New York City, Rhode Island and Michigan states and Italy, by simultaneously inferring the unknown parameters and the unobserved dynamics. For integer-order and time-delay models, we fit the available data by identifying time-dependent parameters, which are represented by neural networks. In contrast, for fractional differential models, we fit the data by determining different time-dependent derivative orders for each compartment, which we represent by neural networks. We investigate the structural and practical identifiability of these unknown functions for different datasets, and quantify the uncertainty associated with neural networks and with control measures in forecasting the pandemic.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2024

Viscosities of the Baryon-Rich Quark-Gluon Plasma from Beam Energy Scan Data

Here, this work presents the first Bayesian inference study of the (3 + 1)D dynamics of relativistic heavy-ion collisions and quark-gluon plasma viscosities using an event-by-event (3 + 1)D hydrodynamics + hadronic transport theoretical framework and data from the Relativistic Heavy Ion Collider Beam energy scan program. Robust constraints on initial state nuclear stopping and the baryon chemical potential-dependent shear viscosity of the produced quantum chromodynamic (QCD) matter are obtained. The specific bulk viscosity of the QCD matter is found to exhibit a preferred maximum around $\sqrt{s_{NN}}$ = 19.6 GeV . This result allows for the alternative interpretation of a reduction (and/or increase) of the speed of sound relative to that of the employed lattice-QCD based equation of state for net baryon chemical potential μ B ~0.2(0.4) GeV .

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2021

Stochastic Gradient-Based Distributed Bayesian Estimation in Cooperative Sensor Networks

Distributed Bayesian inference provides a full quantification of uncertainty offering numerous advantages over point estimates that autonomous sensor networks are able to exploit. However, fully-decentralized Bayesian inference often requires large communication overheads and low network latency, resources that are not typically available in practical applications. In this paper, we propose a decentralized Bayesian inference approach based on stochastic gradient Langevin dynamics, which produces full posterior distributions at each of the nodes with significantly lower communication overhead. We provide analytical results on convergence of the proposed distributed algorithm to the centralized posterior, under typical network constraints. Finally, we also provide extensive simulation results to demonstrate the validity of the proposed approach.

42 ENGINEERING↗
DOE OSTI2021

First M87 Event Horizon Telescope Results. VIII. Magnetic Field Structure near The Event Horizon

Event Horizon Telescope (EHT) observations at 230 GHz have now imaged polarized emission around the supermassive black hole in M87 on event-horizon scales. This polarized synchrotron radiation probes the structure of magnetic fields and the plasma properties near the black hole. Here we compare the resolved polarization structure observed by the EHT, along with simultaneous unresolved observations with the Atacama Large Millimeter/submillimeter Array, to expectations from theoretical models. The low fractional linear polarization in the resolved image suggests that the polarization is scrambled on scales smaller than the EHT beam, which we attribute to Faraday rotation internal to the emission region. We estimate the average density n_e of order 10^4-7 cm-3, magnetic field strength B of order 1-30 G, and electron temperature Te of order (1-12) x 10^10 K of the radiating plasma in a simple one-zone emission model. We show that the net azimuthal linear polarization pattern may result from organized, poloidal magnetic fields in the emission region. In a quantitative comparison with a large library of simulated polarimetric images from general relativistic magnetohydrodynamic (GRMHD) simulations, we identify a subset of physical models that can explain critical features of the polarimetric EHT observations while producing a relativistic jet of sufficient power. The consistent GRMHD models are all of magnetically arrested accretion disks, where near-horizon magnetic fields are dynamically important. We use the models to infer a mass accretion rate onto the black hole in M87 of (3-20) x 10^-4 Msun yr-1.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS1989

Mass transfer during ice particle collisions in planetary rings

Experimental results are presented from laboratory environment simulations of the ice particle collisional properties defining the structure and dynamical evolution of planetary rings. It is inferred from these data that there is a dependence of the interacting volume on the impact velocity. Although the volume fraction exchanged during a collision is small, the net amount of material transferred can be substantially smaller. Attention is given to the implications of these determinations for planetary ring structure and evolution.

Mcdonald, J. S. B.↗
NASA NTRS2016

Magnetic-Island Contraction and Particle Acceleration in Simulated Eruptive Solar Flares

The mechanism that accelerates particles to the energies required to produce the observed high-energy impulsive emission in solar flares is not well understood. Drake et al. proposed a mechanism for accelerating electrons in contracting magnetic islands formed by kinetic reconnection in multi-layered current sheets (CSs). We apply these ideas to sunward-moving flux ropes (2.5D magnetic islands) formed during fast reconnection in a simulated eruptive flare. A simple analytic model is used to calculate the energy gain of particles orbiting the field lines of the contracting magnetic islands in our ultrahigh-resolution 2.5D numerical simulation. We find that the estimated energy gains in a single island range up to a factor of five. This is higher than that found by Drake et al. for islands in the terrestrial magnetosphere and at the heliopause, due to strong plasma compression that occurs at the flare CS. In order to increase their energy by two orders of magnitude and plausibly account for the observed high-energy flare emission, the electrons must visit multiple contracting islands. This mechanism should produce sporadic emission because island formation is intermittent. Moreover, a large number of particles could be accelerated in each magneto hydro dynamic-scale island, which may explain the inferred rates of energetic-electron production in flares. We conclude that island contraction in the flare CS is a promising candidate for electron acceleration in solar eruptions.

Guidoni, S. E.↗
NASA NTRS2017

Using an Altimeter-Derived Internal Tide Model to Remove Tides from in Situ Data

Internal waves at tidal frequencies, i.e., the internal tides, are a prominent source of variability in the ocean associated with significant vertical isopycnal displacements and currents. Because the isopycnal displacements are caused by ageostrophic dynamics, they contribute uncertainty to geostrophic transport inferred from vertical profiles in the ocean. Here it is demonstrated that a newly developed model of the main semidiurnal (M2) internal tide derived from satellite altimetry may be used to partially remove the tide from vertical profile data, as measured by the reduction of steric height variance inferred from the profiles. It is further demonstrated that the internal tide model can account for a component of the near-surface velocity as measured by drogued drifters. These comparisons represent a validation of the internal tide model using independent data and highlight its potential use in removing internal tide signals from in situ observations.

Zaron, Edward D.↗
DOE OSTI2021

Inference of neutrino flavor evolution through data assimilation and neural differential equations

The evolution of neutrino flavor in dense environments such as core-collapse supernovae and binary compact object mergers constitutes an important and unsolved problem. Its solution has potential implications for the dynamics and heavy-element nucleosynthesis in these environments. In this paper, we build upon recent work to explore inference-based techniques for the estimation of model parameters and neutrino flavor evolution histories. Here, we combine data assimilation, ordinary differential equation solvers, and neural networks to craft an inference approach tailored for nonlinear dynamical systems. Using this architecture, and a simple two-neutrino-beam, two-flavor model, we compare the performances of nine different optimization algorithms and expand upon previous assessments of the efficacy of inference for tackling problems in flavor evolution. We find that employing this new architecture, together with evolutionary optimization algorithms, accurately captures flavor histories in the small-scale model and allows us to quickly explore both model parameters and initial flavor content. In future work we plan to extend these inference techniques to large numbers of neutrinos.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2024

Bayesian analysis of (3 +1)⁢D relativistic nuclear dynamics with the RHIC beam energy scan data

This work presents a Bayesian inference study for relativistic heavy-ion collisions in the beam energy scan program at the BNL Relativistic Heavy-Ion Collider. The theoretical model simulates event-by-event (3+1)-dimensional [(3+1)⁢D] collision dynamics using hydrodynamics and hadronic transport theory. We analyze the model's 20-dimensional posterior distributions obtained using three model emulators with different accuracy and demonstrate the essential role of training an accurate model emulator in the Bayesian analysis. Our analysis provides robust constraints on the quark-gluon plasma's transport properties and various aspects of (3+1)⁢D relativistic nuclear dynamics. By running full model simulations with 100 parameter sets sampled from the posterior distribution, we make predictions for p T -differential observables and estimate their systematic theory uncertainty. Here, a sensitivity analysis is performed to elucidate how individual experimental observables respond to different model parameters, providing useful physics insights into the phenomenological model for heavy-ion collisions.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2025

Efficient Neural Network Approaches for Conditional Optimal Transport with Applications in Bayesian Inference

In this work, we present two neural network approaches that approximate the solutions of static and dynamic conditional optimal transport (COT) problems. Both approaches enable conditional sampling and conditional density estimation, which are core tasks in Bayesian inference—particularly in the simulation-based (“likelihood-free”) setting. Our methods represent the target conditional distribution as a transformation of a tractable reference distribution. Obtaining such a transformation, chosen here to be an approximation of the COT map, is computationally challenging even in moderate dimensions. To improve scalability, our numerical algorithms use neural networks to parameterize candidate maps and further exploit the structure of the COT problem. Our static approach approximates the map as the gradient of a partially input convex neural network. It uses a novel numerical implementation to increase computational efficiency compared to state-of-the-art alternatives. Our dynamic approach approximates the conditional optimal transport via the flow map of a regularized neural ODE; compared to the static approach, it is slower to train but offers more modeling choices and can lead to faster sampling. We demonstrate both algorithms numerically, comparing them with competing state-of-the-art approaches, using benchmark datasets and simulation-based Bayesian inverse problems.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Bayesian Gaussian process inference for neutron spin echo measurement

Neutron spin echo (NSE) spectroscopy provides unique access to microscopic dynamics, but its application is often constrained by low neutron flux, long acquisition times, and significant noise. Here, we present a Bayesian inference approach based on Gaussian process regression (GPR) to reconstruct high-quality spin echo signals from sparse and noisy data by exploiting correlations in reciprocal space. Benchmarks on synthetic datasets and validation with experimental NSE measurements of dendrimers show that GPR suppresses noise, interpolates missing intensity values, and accommodates irregular observations. The method improves accuracy, shortens acquisition times, and enables high-throughput and real-time studies. Beyond NSE, the framework is broadly applicable to other low signal-to-noise ratio scattering techniques, thereby extending the scope of neutron spectroscopy.

Tung, Chi-Huan [Oak Ridge National Laboratory (ORN↗
DOE OSTI2022

Dictionary Learning with Accumulator Neurons

The Locally Competitive Algorithm (LCA) uses local competition between non-spiking leaky integrator neurons to infer sparse representations, allowing for potentially real-time execution on massively parallel neuromorphic architectures such as Intel's Loihi processor. Here, we focus on the problem of inferring sparse representations from streaming video using dictionaries of spatiotemporal features optimized in an unsupervised manner for sparse reconstruction. Non-spiking LCA has previously been used to achieve unsupervised learning of spatiotemporal dictionaries composed of convolutional kernels from raw, unlabeled video. We demonstrate how unsupervised dictionary learning with spiking LCA (\hbox{S-LCA}) can be efficiently implemented using accumulator neurons, which combine a conventional leaky-integrate-and-fire (\hbox{LIF}) spike generator with an additional state variable that is used to minimize the difference between the integrated input and the spiking output. We demonstrate dictionary learning across a wide range of dynamical regimes, from graded to intermittent spiking, for inferring sparse representations of both static images drawn from the CIFAR database as well as video frames captured from a DVS camera. On a classification task that requires identification of the suite from a deck of cards being rapidly flipped through as viewed by a DVS camera, we find essentially no degradation in performance as the LCA model used to infer sparse spatiotemporal representations migrates from graded to spiking. We conclude that accumulator neurons are likely to provide a powerful enabling component of future neuromorphic hardware for implementing online unsupervised learning of spatiotemporal dictionaries optimized for sparse reconstruction of streaming video from event based DVS cameras.

artificial intelligence↗
DOE OSTI2024

Learning Nonlinear Reduced Models from Data with Operator Inference

This review discusses Operator Inference, a nonintrusive reduced modeling approach that incorporates physical governing equations by defining a structured polynomial form for the reduced model, and then learns the corresponding reduced operators from simulated training data. The polynomial model form of Operator Inference is sufficiently expressive to cover a wide range of nonlinear dynamics found in fluid mechanics and other fields of science and engineering, while still providing efficient reduced model computations. The learning steps of Operator Inference are rooted in classical projection-based model reduction; thus, some of the rich theory of model reduction can be applied to models learned with Operator Inference. This connection to projection-based model reduction theory offers a pathway toward deriving error estimates and gaining insights to improve predictions. Furthermore, through formulations of Operator Inference that preserve Hamiltonian and other structures, important physical properties such as energy conservation can be guaranteed in the predictions of the reduced model beyond the training horizon. This review illustrates key computational steps of Operator Inference through a large-scale combustion example.

Mechanics↗
NASA NTRS2011

Large Eddy Simulation Study for Fluid Disintegration and Mixing

A new modeling approach is based on the concept of large eddy simulation (LES) within which the large scales are computed and the small scales are modeled. The new approach is expected to retain the fidelity of the physics while also being computationally efficient. Typically, only models for the small-scale fluxes of momentum, species, and enthalpy are used to reintroduce in the simulation the physics lost because the computation only resolves the large scales. These models are called subgrid (SGS) models because they operate at a scale smaller than the LES grid. In a previous study of thermodynamically supercritical fluid disintegration and mixing, additional small-scale terms, one in the momentum and one in the energy conservation equations, were identified as requiring modeling. These additional terms were due to the tight coupling between dynamics and real-gas thermodynamics. It was inferred that if these terms would not be modeled, the high density-gradient magnitude regions, experimentally identified as a characteristic feature of these flows, would not be accurately predicted without the additional term in the momentum equation; these high density-gradient magnitude regions were experimentally shown to redistribute turbulence in the flow. And it was also inferred that without the additional term in the energy equation, the heat flux magnitude could not be accurately predicted; the heat flux to the wall of combustion devices is a crucial quantity that determined necessary wall material properties. The present work involves situations where only the term in the momentum equation is important. Without this additional term in the momentum equation, neither the SGS-flux constant-coefficient Smagorinsky model nor the SGS-flux constant-coefficient Gradient model could reproduce in LES the pressure field or the high density-gradient magnitude regions; the SGS-flux constant- coefficient Scale-Similarity model was the most successful in this endeavor although not totally satisfactory. With a model for the additional term in the momentum equation, the predictions of the constant-coefficient Smagorinsky and constant-coefficient Scale-Similarity models were improved to a certain extent; however, most of the improvement was obtained for the Gradient model. The previously derived model and a newly developed model for the additional term in the momentum equation were both tested, with the new model proving even more successful than the previous model at reproducing the high density-gradient magnitude regions. Several dynamic SGS-flux models, in which the SGS-flux model coefficient is computed as part of the simulation, were tested in conjunction with the new model for this additional term in the momentum equation. The most successful dynamic model was a "mixed" model combining the Smagorinsky and Gradient models. This work is directly applicable to simulations of gas turbine engines (aeronautics) and rocket engines (astronautics).

Bellan, Josette↗
NASA NTRS1990

Conditions for the formation of prominences as inferred from optical observations

It is suggested that three conditions, if dynamically maintained for a sufficient period of time, will result in the formation of a solar prominence. These conditions are the presence of a coronal arcade that connects the magnetic fields on opposite sides of a prominence; the existence of a long-term (hours to days) converging flow of small patches of opposite-polarity magnetic flux toward a common-polarity inversion zone; and the cancellation of encountered patches of magnetic flux of opposite polarity at a photospheric polarity inversion boundary.

Martin, Sara F.↗
NASA NTRS2005

Seismic Study of the Solar Interior: Inferences from SOI/MDI Observations During Solar Activity

Work on the structure, asphericity and dynamics of the solar interior from p-mode frequencies and frequency splittings was carried out primarily in collaboration with Dr. Eff-Darwich (University of La Laguna, Tenerife). This ongoing collaboration produced new results for the inversion of the internal solar rotation rate and further development in inversion methodologies. It also resulted in inferences on the solar stratification. Substantial progress towards the characterization of high-degree p-modes has been achieved. In collaboration with Drs. Rabello-Soares and Schou (Stanford University), we have gained a clear conceptual understanding of the various elements that affect the leakage matrix of the SOI/MDI instrument. This work has precise implications on the properties and the characterization of the HMI instrument being developed for the SDO mission.

Korzennik, Sylvain G.↗
NASA NTRS1997

Stochastic-Dynamical Modeling of Space Time Rainfall

The focus of this research work is the elucidation of the physical origins of the observed extreme-rainfall variability over tropical oceans. The quantitative results of this work may be used to establish links between deterministic models of the mesoscale and synoptic scale with statistical descriptions of the temporal variability of local tropical oceanic rainfall. In addition, they may be used to quantify the influence of measurement error in large-scale forcing and cloud scale observations on the accuracy of local rainfall variability inferences, important for hydrologic studies. A simple statistical-dynamical model, suitable for use in repetitive Monte Carlo experiments, is formulated as a diagnostic tool for this purpose. Stochastic processes with temporal structure and parameters estimated from observed large-scale data represent large-scale forcing.

Georgankakos, Konstantine P.↗