Search NASA⌕ Search

SEARCH · Search NASA

Results for “differentiable simulation”

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
DOE OSTI2025

Boundary-Layer-Coupled and Decoupled Clouds in Global Storm-Resolving Models: Comparisons With the ARM Observations

The accurate representation of interactions between clouds and planetary boundary layer (PBL) is a persistent challenge in climate models, critical for simulating surface energy budget. The emergence of kilometer-grid-scale global storm resolving models (GSRMs) offers the potential for enhanced details of PBL processes in these complex interactions. This study evaluates the representation of PBL-coupled and decoupled clouds in nine GSRM simulations against extensive ground-based observations by the Department of Energy Atmospheric Radiation Measurement (ARM) program, across six sites encompassing diverse regimes such as marine and continental environments in tropics and midlatitude. By differentiating coupling based on the relative positions between cloud bases and PBL tops, our analysis focuses on the simulation of PBL height, cloud frequency, position and vertical extent. The GSRMs generally exhibit commendable agreements with observed cloud structures and PBL diurnal cycles across different ARM sites. In contrast to the relatively consistent representation of decoupled clouds, discrepancies exist between the simulated and the observed coupled clouds, particularly in areas of intense convection, for example, over tropical rainforests and mountainous regions. These biases are probably associated with the models' tendency to underestimate the boundary layer humidity and the frequency of coupled clouds within different ranges of PBL heights. This study underscores the importance for continuous improvements in the representation of boundary layer and convection within these global kilometer-grid-scale models.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

Simulating low-energy neutrino interactions with MARLEY

Monte Carlo event generators are a critical tool for the interpretation of data obtained by neutrino experiments. Several modern event generators are available which are well-suited to the GeV energy scale used in studies of accelerator neutrinos. However, theoretical modeling differences make their immediate application to lower energies difficult. In this paper, I present a new event generator, MARLEY, which is designed to better address the simulation needs of the low-energy (tens of MeV and below) neutrino community. The code is written in C++14 with an optional interface to the popular ROOT data analysis framework. The current release of MARLEY (version 1.2.0) emphasizes simulations of the reaction 40 Ar ( ν e , e − ) 40 K ⁎ but is extensible to other channels with suitable user input. This paper provides detailed documentation of MARLEY's implementation and usage, including guidance on how generated events may be analyzed and how MARLEY may be interfaced with external codes such as Geant4. Further information about MARLEY is available on the official website at http://www.marleygen.org. Program title:MARLEY 1.2.0 CPC Library link to program files:https://doi.org/10.17632/4v7zxnc8j3.1 Developer's respository link:http://github.com/MARLEY-MC/marley Code Ocean capsule:https://codeocean.com/capsule/9868179 Licensing provisions: GNU General Public License 3.0 Programming language: C++14 External routines/libraries used: GNU Scientific Library [1,2] (required), ROOT [3,4] (optional) Nature of problem: Simulation of neutrino-nucleus scattering events at energies of tens-of-MeV and below Solution method: Initial two-to-two scattering kinematics are sampled using the allowed approximation differential cross section and tables of precomputed nuclear matrix elements. Subsequent de-excitations of the remnant nucleus are simulated using a Monte Carlo implementation of the Hauser-Feshbach statistical model and tabulated γ-ray decay schemes for discrete nuclear levels. Additional comments including restrictions and unusual features: Input data are provided with the code that are suitable for producing simulations of the charged-current reaction 40 Ar ( ν e , e − ) 40 K ⁎ , coherent elastic neutrino-nucleus scattering on spin-zero target nuclei, and neutrino-electron elastic scattering on any atomic target. Preparation of new reaction input files (whose format is documented in Appendix B) would enable other reaction channels and nuclear targets to be handled by the existing code framework. Although there is no maximum neutrino energy enforced by the code itself, realistic neutrino-nucleus scattering events may be generated up to roughly 50 MeV. Above this energy, the effects of forbidden nuclear transitions, which are neglected in the current treatment of the cross sections (see section 2.1), become increasingly important.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2023

Performance analysis of functionally graded multifunctional piezoelectric energy harvesting microgyroscopes

Functionally graded materials (FGMs) are composite materials with varying material properties in one or more directions. These materials possess distinct characteristics compared to their constituent components. The ability to control material distribution and compositions in FGMs offers improved constraints over the natural frequencies of the system, making them highly suitable for energy harvesting applications. In this study, we focus on a piezoelectric FGM composed of Platinum (Pt) and Lead Titanate Zirconate (PZT). By utilizing FGMs for energy harvesting, we simplify the system from a multilayer structure to a single-layer system, thereby increasing power density by reducing volume. In this work, a power law distribution is used to model the material variation throughout the thickness of the single-layered beam. The governing equations of motion and boundary conditions for FG energy harvesting microgyroscope are derived using Euler-Bernoulli beam theory, the constitutive piezoelectric principle, and the extended Hamilton's principle. To simulate the nonlinear motion of the FG energy harvester, we employ the differential quadrature method (DQM). Subsequently, an investigation is carried out to examine the effects of various factors including material distribution, platinum percentage, base rotation, DC voltage, and electrical load resistance on the energy harvesting microgyroscope with functionally graded materials. Here, the findings suggest that functionally graded energy harvesting gyroscopes can be adjusted to obtain effective energy harvesting and sensing capabilities by carefully selecting the material distribution, input voltage, and electrical load resistance.

42 ENGINEERING↗
DOE OSTI2024

Learning nonlinear operators in latent spaces for real-time predictions of complex dynamics in physical systems

Abstract Predicting complex dynamics in physical applications governed by partial differential equations in real-time is nearly impossible with traditional numerical simulations due to high computational cost. Neural operators offer a solution by approximating mappings between infinite-dimensional Banach spaces, yet their performance degrades with system size and complexity. We propose an approach for learning neural operators in latent spaces, facilitating real-time predictions for highly nonlinear and multiscale systems on high-dimensional domains. Our method utilizes the deep operator network architecture on a low-dimensional latent space to efficiently approximate underlying operators. Demonstrations on material fracture, fluid flow prediction, and climate modeling highlight superior prediction accuracy and computational efficiency compared to existing methods. Notably, our approach enables approximating large-scale atmospheric flows with millions of degrees, enhancing weather and climate forecasts. Here we show that the proposed approach enables real-time predictions that can facilitate decision-making for a wide range of applications in science and engineering.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Molecular dynamics on quantum annealers

Abstract In this work we demonstrate a practical prospect of using quantum annealers for simulation of molecular dynamics. A methodology developed for this goal, dubbed Quantum Differential Equations (QDE), is applied to propagate classical trajectories for the vibration of the hydrogen molecule in several regimes: nearly harmonic, highly anharmonic, and dissociative motion. The results obtained using the D-Wave 2000Q quantum annealer are all consistent and quickly converge to the analytical reference solution. Several alternative strategies for such calculations are explored and it was found that the most accurate results and the best efficiency are obtained by combining the quantum annealer with classical post-processing (greedy algorithm). Importantly, the QDE framework developed here is entirely general and can be applied to solve any system of first-order ordinary nonlinear differential equations using a quantum annealer.

74 ATOMIC AND MOLECULAR PHYSICS↗
DOE OSTI2024

The DEHVILS in the details: Type Ia supernova Hubble residual comparisons and mass step analysis in the near-infrared

Measurements of type Ia supernovae (SNe Ia) in the near-infrared (NIR) have been used both as an alternate path to cosmology compared to optical measurements and as a method of constraining key systematics for the larger optical studies. With the DEHVILS sample, the largest published NIR sample with consistent NIR coverage of maximum light across three NIR bands ( Y, J , and H ), we check three key systematics: (i) the reduction in Hubble residual scatter as compared to the optical, (ii) the measurement of a “mass step” or lack thereof and its implications, and (iii) the ability to distinguish between various dust models by analyzing slopes and correlations between Hubble residuals in the NIR and optical. We produce SN Ia simulations of the DEHVILS sample and find that it is harder to differentiate between various dust models than previously understood. Additionally, we find that fitting with the current SALT3-NIR model does not yield accurate wavelength-dependent stretch-luminosity correlations, and we propose a limited solution for this problem. From the data, we see that (i) the standard deviation of Hubble residual values from NIR bands treated as standard candles are 0.007–0.042 mag smaller than those in the optical, (ii) the NIR mass step is not constrainable with the current sample size of 47 SNe Ia from DEHVILS, and (iii) Hubble residuals in the NIR and optical are correlated in the data. We test a few variations on the number and combinations of filters and data samples, and we observe that none of our findings or conclusions are significantly impacted by these modifications.

Astronomy & Astrophysics↗
DOE OSTI2022

The Impulse Approximation Scattering Function and Its Use in Monte Carlo Photon Transport Simulations

The viability of using the impulse approximation scattering function in Monte Carlo photon transport simulations is explored. This scattering function can be constructed from the double differential incoherent scattering cross section developed by Ribberfors and Berggren. A commonly used method for modeling photon Doppler broadening, which is referred to as the hybrid Doppler broadening method, can also be derived from this cross section. A new photon Doppler broadening method, called the consistent Doppler broadening method, is derived and discussed. This method eliminates some of the commonly employed approximations in the hybrid Doppler broadening method, in part, by using the impulse approximation scattering function. Integrated incoherent cross sections generated using the impulse approximation scattering function and the widely used Waller-Hartree scattering function are in good agreement above 20 keV. Below 20 keV, differences as high as 70% are observed, which differs from the roughly 5% differences observed by Ribberfors for some of the materials. Integral and spectral quantities for two problems are also generated using the Monte Carlo photon transport capabilities of the Framework for REsearch in Nuclear ScIence and Engineering (FRENSIE). Due to the small, relative result differences observed when using the impulse approximation scattering function, it is considered a viable alternative to the Waller-Hartree scattering function. In addition, some small, but expected, differences in spectral fluxes at low energies can be avoided by adopting the consistent Doppler broadening method.

42 ENGINEERING↗
DOE OSTI2024

Measurement of jet substructure in boosted $t\overline{t}$ events with the ATLAS detector using 140 fb -1 of 13 TeV $pp$ collisions

Measurements of the substructure of top-quark jets are presented, using 140 fb -1 of 13 TeV pp collision data recorded with the ATLAS detector at the LHC. Top-quark jets reconstructed with the anti-k t algorithm with a radius parameter R = 1.0 are selected in top-quark pair ($t\overline{t}$) events where one top quark decays semileptonically and the other hadronically, or where both top quarks decay hadronically. The top-quark jets are required to have transverse momentum p T > 350 GeV, yielding large samples of data events with jet p T values between 350 and 600 GeV. One- and two-dimensional differential cross sections for eight substructure variables, defined using only the charged components of the jets, are measured in a particle-level phase space by correcting for the smearing and acceptance effects induced by the detector. The differential cross sections are compared with the predictions of several Monte Carlo simulations in which top-quark pair-production quantum chromodynamic matrix-element calculations at next-to-leading-order precision in the strong coupling constant α S are passed to leading-order parton shower and hadronization generators. The Monte Carlo predictions for measures of the broadness, and also the two-body structure, of the top-quark jets are found to be in good agreement with the measurements, while variables sensitive to the three-body structure of the top-quark jets exhibit some tension with the measured distributions.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2021

Application of Quantum Machine Learning to High Energy Physics Analysis at LHC using IBM Quantum Computer Simulators and IBM Quantum Computer Hardware

One of the major objectives of the experimental programs at the LHC is the discovery of new physics. This requires the identification of rare signals in immense backgrounds. Using machine learning algorithms greatly enhances our ability to achieve this objective. With the progress of quantum technologies, quantum machine learning could become a powerful tool for data analysis in high energy physics. In this study, using IBM gate-model quantum computing systems, we employ the quantum variational classifier method and the quantum kernel estimator method in two recent LHC flagship physics analyses: $t\bar{t}H$ (Higgs boson production in association with a top quark pair) and $H\rightarrow\mu\mu$ (Higgs boson decays to two muons). We have obtained early results with 10 qubits on the IBM quantum simulator and the IBM quantum hardware. On the quantum simulator, the quantum machine learning methods perform similarly to classical algorithms such as SVM (support vector machine) and BDT (boosted decision tree), which are often employed in LHC physics analyses. On the quantum hardware, the quantum machine learning methods have shown promising discrimination power, comparable to that on the quantum simulator. This study demonstrates that quantum machine learning has the ability to differentiate between signal and background in realistic physics datasets.

Chan, Jay↗
DOE OSTI2021

Hierarchical Fractional Advection-Dispersion Equation (FADE) to Quantify Anomalous Transport in River Corridor over a Broad Spectrum of Scales: Theory and Applications

Fractional calculus-based differential equations were found by previous studies to be promising tools in simulating local-scale anomalous diffusion for pollutants transport in natural geological media (geomedia), but efficient models are still needed for simulating anomalous transport over a broad spectrum of scales. This study proposed a hierarchical framework of fractional advection-dispersion equations (FADEs) for modeling pollutants moving in the river corridor at a full spectrum of scales. Applications showed that the fixed-index FADE could model bed sediment and manganese transport in streams at the geomorphologic unit scale, whereas the variable-index FADE well fitted bedload snapshots at the reach scale with spatially varying indices. Further analyses revealed that the selection of the FADEs depended on the scale, type of the geomedium (i.e., riverbed, aquifer, or soil), and the type of available observation dataset (i.e., the tracer snapshot or breakthrough curve (BTC)). When the pollutant BTC was used, a single-index FADE with scale-dependent parameters could fit the data by upscaling anomalous transport without mapping the sub-grid, intermediate multi-index anomalous diffusion. Pollutant transport in geomedia, therefore, may exhibit complex anomalous scaling in space (and/or time), and the identification of the FADE’s index for the reach-scale anomalous transport, which links the geomorphologic unit and watershed scales, is the core for reliable applications of fractional calculus in hydrology.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Idealized simulation study of the relationship of disdrometer sampling statistics with the precision of precipitation rate measurement

Due to the discretized nature of rain, the measurement of a continuous precipitation rate by disdrometers is subject to statistical sampling errors. Here, Monte Carlo simulations are employed to obtain the precision of rain detection and rate as a function of disdrometer collection area and compared with World Meteorological Organization guidelines for a 1 min sample interval and 95 % probability. To meet these requirements, simulations suggest that measurements of light rain with rain rates R ≤ 0.50 mm h-1 require a collection area of at least 6 cm × 6 cm, and for R = 1 mm h -1 , the minimum collection area is 13 cm × 13 cm. For R = 0.01 mm h -1 , a collection area of 2 cm × 2 cm is sufficient to detect a single drop. Simulations are compared with field measurements using a new hotplate device, the Differential Emissivity Imaging Disdrometer. The field results suggest an even larger plate may be required to meet the stated accuracy, likely in part due to non-Poissonian hydrometeor clustering.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Network-Aware and Welfare-Maximizing Dynamic Pricing for Energy Sharing: Preprint

The proliferation of behind-the-meter (BTM) distributed energy resources (DER) within the electrical distribution network presents significant supply and demand flexibilities, but also introduces operational challenges such as voltage spikes and reverse power flows. In response, this paper proposes a network-aware dynamic pricing framework tailored for energy-sharing coalitions that aggregate small, but ubiquitous, BTM DER downstream of a distribution system operator's (DSO) revenue meter that adopts a generic net energy metering (NEM) tariff. By formulating a Stackelberg game between the energy-sharing market leader and its prosumers, we show that the dynamic pricing policy induces the prosumers toward a network-safe operation and decentrally maximizes the energysharing social welfare. The dynamic pricing mechanism involves a combination of a locational ex-ante dynamic price and an ex-post allocation, both of which are functions of the energy sharing's BTM DER. The ex-post allocation is proportionate to the price differential between the DSO NEM price and the energy sharing locational price. Simulation results using real DER data and the IEEE 13-bus test systems illustrate the dynamic nature of network-aware pricing at each bus, and its impact on voltage.

energy communities↗
DOE OSTI2020

DPM: A deep learning PDE augmentation method with application to large-eddy simulation

A framework is introduced that leverages known physics to reduce overfitting in machine learning for scientific applications. The partial differential equation (PDE) that expresses the physics is augmented with a neural network that uses available data to learn a description of the corresponding unknown or unrepresented physics. Training within this combined system corrects for missing, unknown, or erroneously represented physics, including discretization errors associated with the PDE's numerical solution. For optimization of the network within the PDE, an adjoint PDE is solved to provide high-dimensional gradients, and a stochastic adjoint method (SAM) further accelerates training. Additionally, the approach is demonstrated for large-eddy simulation (LES) of turbulence. High-fidelity direct numerical simulations (DNS) of decaying isotropic turbulence provide the training data used to learn sub-filter-scale closures for the filtered Navier–Stokes equations. Out-of-sample comparisons show that the deep learning PDE method outperforms widely-used models, even for filter sizes so large that they become qualitatively incorrect. It also significantly outperforms the same neural network when a priori trained based on simple data mismatch, not accounting for the full PDE. Measures of discretization errors, which are well-known to be consequential in LES, point to the importance of the unified training formulation's design, which without modification corrects for them. For comparable accuracy, simulation runtime is significantly reduced. A relaxation of the typical discrete enforcement of the divergence-free constraint in the solver is also successful, instead allowing the DPM to approximately enforce incompressibility physics. Since the training loss function is not restricted to correspond directly to the closure to be learned, training can incorporate diverse data, including experimental data.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2025

Acceleration of Power System Dynamic Simulations Using a Deep Equilibrium Layer and Neural ODE Surrogate

The dominant paradigm for power system dynamic simulation is to build system-level simulations by combining physics-based models of individual components. The sheer size of the system along with the rapid integration of inverter-based resources exacerbates the computational burden of running time domain simulations. Here, in this paper, we propose a data-driven surrogate model based on implicit machine learningspecifically deep equilibrium layers and neural ordinary differential equationsto learn a reduced order model of a portion of the full underlying system. The data-driven surrogate achieves similar accuracy and reduction in simulation time compared to a physics-based surrogate, without the constraint of requiring detailed knowledge of the underlying dynamic models. This work also establishes key requirements needed to integrate the surrogate into existing simulation workflows; the proposed surrogate is initialized to a steady state operating point that matches the power flow solution by design.

Neural ordinary differential equations↗
DOE OSTI2025

Predictive Modeling and Uncertainty Quantification in Condition Monitoring of Active Components: A Reactor Coolant Pump Use Case

This work develops data-driven models for onset of thermal barrier leakage in reactor coolant pumps. It incorporates uncertainty quantification to enhance the reliability and robustness of pre- dictions. Using synthetic data generated by the Generic Pressurized Water Reactor simulator, realistic degradation scenarios were simulated across lifecycle stages—beginning, middle, and end of life. Key variables, including differential pressure, flow rate, vibration, and temperatures, were analyzed using machine learning framework. The fully connected neural network models demonstrated exceptional performance, achieving R2 scores exceeding 0.99 and root mean square errors as low as around 8.23 × 10-2 gallon per minute (gpm) for the three stages of the lifecy- cle. UQ analysis further validated the model’s robustness, with narrow uncertainty bounds during steady-state operations and appropriately wider bounds during transitional phases, reflecting the physical behavior of the system. This work addresses important gaps in real-time condition moni- toring and regulatory compliance by integrating advanced condition monitoring technologies with UQ into IST programs. The ability to detect thermal barrier leakage early and quantify prediction reliability supports optimizing maintenance strategies while ensuring nuclear power plants’ safe and reliable operation.

99 - GENERAL AND MISCELLANEOUS↗
DOE OSTI2025

A compartmentalized model of multiphase chemical kinetics

There are significant challenges in predicting multiphase chemical kinetics due to the complex coupling of reaction and mass transport across a phase boundary (i.e., interface). Here, we describe a framework for predicting multiphase kinetics that embeds the elementary kinetic steps of reaction, solvation, and diffusion into a coarse grain spatial description of two phases. The model is constructed to bridge the short-timescale interfacial dynamics observed in molecular simulations with the longer timescales observed in kinetic experiments. A simple set of governing differential equations is derived, which, when solved numerically or analytically, yield accurate predictions of multiphase kinetics in microdroplets. Although the equations are formulated for gas-liquid reactions, the underlying conceptual framework is general and can be applied to transformations in other two-phase systems (solid-liquid, liquid-liquid, etc.).

Chemical kinetics and dynamics↗
DOE OSTI2024

Differentiable stochastic halo occupation distribution

ABSTRACT In this work, we demonstrate how differentiable stochastic sampling techniques developed in the context of deep reinforcement learning can be used to perform efficient parameter inference over stochastic, simulation-based, forward models. As a particular example, we focus on the problem of estimating parameters of halo occupation distribution (HOD) models that are used to connect galaxies with their dark matter haloes. Using a combination of continuous relaxation and gradient re-parametrization techniques, we can obtain well-defined gradients with respect to HOD parameters through discrete galaxy catalogue realizations. Having access to these gradients allows us to leverage efficient sampling schemes, such as Hamiltonian Monte Carlo, and greatly speed up parameter inference. We demonstrate our technique on a mock galaxy catalogue generated from the Bolshoi simulation using a standard HOD model and find near-identical posteriors as standard Markov chain Monte Carlo techniques with an increase of ∼8× in convergence efficiency. Our differentiable HOD model also has broad applications in full forward model approaches to cosmic structure and cosmological analysis.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2023

Massively parallel modeling and inversion of electrical resistivity tomography data using PFLOTRAN

Abstract. Electrical resistivity tomography (ERT) is a broadly accepted geophysical method for subsurface investigations. Interpretation of field ERT data usually requires the application of computationally intensive forward modeling and inversion algorithms. For large-scale ERT data, the efficiency of these algorithms depends on the robustness, accuracy, and scalability on high-performance computing resources. In this regard, we present a robust and highly scalable implementation of forward modeling and inversion algorithms for ERT data. The implementation is publicly available and developed within the framework of PFLOTRAN, an open-source, state-of-the-art massively parallel subsurface flow and transport simulation code. The forward modeling is based on a finite-volume discretization of the governing differential equations, and the inversion uses a Gauss–Newton optimization scheme. To evaluate the accuracy of the forward modeling, two examples are first presented by considering layered (1D) and 3D earth conductivity models. The computed numerical results show good agreement with the analytical solutions for the layered earth model and results from a well-established code for the 3D model. Inversion of ERT data, simulated for a 3D model, is then performed to demonstrate the inversion capability by recovering the conductivity of the model. To demonstrate the parallel performance of PFLOTRAN's ERT process model and inversion capabilities, large-scale scalability tests are performed by using up to 131 072 processes on a leadership class supercomputer. These tests are performed for the two most computationally intensive steps of the ERT inversion: forward modeling and Jacobian computation. For the forward modeling, we consider models with up to 122 ×106 degrees of freedom (DOFs) in the resulting system of linear equations and demonstrate that the code exhibits almost linear scalability on up to 10 000 DOFs per process. On the other hand, the code shows superlinear scalability for the Jacobian computation, mainly because all computations are fairly evenly distributed over each process with no parallel communication.

58 GEOSCIENCES↗