Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “differentiable physics”

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 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Quantifying local and global mass balance errors in physics-informed neural networks

Physics-informed neural networks (PINN) have recently become attractive for solving partial differential equations (PDEs) that describe physics laws. By including PDE-based loss functions, physics laws such as mass balance are enforced softly in PINN. This paper investigates how mass balance constraints are satisfied when PINN is used to solve the resulting PDEs. We investigate PINN’s ability to solve the 1D saturated groundwater flow equations (diffusion equations) for homogeneous and heterogeneous media and evaluate the local and global mass balance errors. We compare the obtained PINN’s solution and associated mass balance errors against a two-point finite volume numerical method and the corresponding analytical solution. We also evaluate the accuracy of PINN in solving the 1D saturated groundwater flow equation with and without incorporating hydraulic heads as training data. We demonstrate that PINN’s local and global mass balance errors are significant compared to the finite volume approach. Tuning the PINN’s hyperparameters, such as the number of collocation points, training data, hidden layers, nodes, epochs, and learning rate, did not improve the solution accuracy or the mass balance errors compared to the finite volume solution. Mass balance errors could considerably challenge the utility of PINN in applications where ensuring compliance with physical and mathematical properties is crucial.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Speeding up and reducing memory usage for scientific machine learning via mixed precision

Scientific machine learning (SciML) has emerged as a versatile approach to address complex computational science and engineering problems. Within this field, physics-informed neural networks (PINNs) and deep operator networks (DeepONets) stand out as the leading techniques for solving partial differential equations by incorporating both physical equations and experimental data. However, training PINNs and DeepONets require significant computational resources, including long computational times and large amounts of memory. In search of computational efficiency, training neural networks using half precision (float16) rather than the conventional single (float32) or double (float64) precision has gained substantial interest, given the inherent benefits of reduced computational time and memory consumed. However, we find that float16 cannot be applied to SciML methods, because of gradient divergence at the start of training, weight updates going to zero, and the inability to converge to a local minima. To overcome these limitations, we explore mixed precision, which is an approach that combines the float16 and float32 numerical formats to reduce memory usage and increase computational speed. Our experiments showcase that mixed precision training not only substantially decreases training times and memory demands but also maintains model accuracy. Here, we also reinforce our empirical observations with a theoretical analysis. The research has broad implications for SciML in various computational applications.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Emulation and detection of physical faults and cyber-attacks on building energy systems through real-time hardware-in-the-loop experiments

The increasing use of remote or mobile access, integrated wearable technologies, data exchange, and cloud-based data analytics in modern smart buildings is steering the building industry towards open communication technologies. The increased connectivity and accessibility could lead to more cyber-attacks in smart buildings. On the other hand, physical faults (e.g., HVAC -heating, ventilation, and air-conditioning faults) may have similar adverse impacts as those from the cyber-attacks on building energy systems, such as occupant discomfort, energy wastage, and equipment downtime. However, current physical behavior-based anomaly detection methods fail to differentiate between cyber-attacks and physical faults in building energy systems. Moreover, the challenge in collecting real-world threat data with ground truth has led researchers to rely on numerical models with user-defined assumptions, which may not accurately reflect real-world conditions due to the lack of in-situ experimental datasets. To address these challenges and gaps, this paper presents a flexible hardware-in-the-loop (HIL) testbed for generating cyber-attack and physical fault datasets and demonstrating threat detection algorithms in a real building automation system (BAS) environment. This testbed combines hardware (i.e., real BAS with local HVAC controllers and a physical network) with software (i.e., high-fidelity models to represent behaviors of building envelope and HVAC energy systems), enabling emulations of realistic threats. Five HIL experiments, including one baseline without any threats, two with physical faults, and two with cyber-attacks, were conducted to generate datasets containing detailed network traffic and system states. A joint classification framework, incorporating a network analyzer and a physical HVAC fault detector, was proposed to automatically detect cyber-physical abnormalities on BAS at both the network and the physical HVAC levels. The network analyzer comprises a conditional random fields (CRF) based command validator and a statistics-based detection strategy. The fault detector employs a weather and schedule-based pattern matching and feature-based principal component analysis (WPM-FPCA) method. Evaluation of the classification using four metrics from the multi-class confusion matrix revealed an average accuracy of 90.2%, recall of 89.7%, precision of 88.5% and F1-score of 89.2%. Finally, these results demonstrate that the proposed joint classification framework can effectively differentiate between specific types of cyber-attacks (e.g., device reinitialization attack, network Denial-of-Service attack) and physical faults (e.g., air handling unit operational fault, cooling coil valve stuck) in real time for improved building energy management.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2026

Magnus method for electronic structure calculations at extreme conditions

We present the application of Magnus based methods to the solution of first order coupled ordinary differential equations in High Energy Density (HED) physics applications. Our focus is on the application to quantum mechanical methods, specifically on the solution of the radial Dirac equation for real and complex energies. HED applications require accurate solutions across a wide range of spatial and energy domains, including regimes where the solutions exhibit pronounced oscillatory behavior. Such cases pose significant computational challenges. We demonstrate that Magnus-based integrators can efficiently and accurately address these challenges. We discuss the implementation of the Magnus method for the solution of the radial Dirac equation, including practical considerations such as the evaluation of matrix exponentials, numerical integration, error estimation, and adaptive step size control. We also discuss the application of these methods to complex energy Green’s function techniques and the efficient approximation of integrals of the solutions relevant to HED electronic structure calculations. Here, we demonstrate the accuracy and robustness of the resulting method in applications to the free-particle case, for which analytic solutions are available for comparison, as well as the challenging case of gold at HED conditions.

general physics↗
DOE OSTI2024

Extending the Takagi-Taupin equations for x-ray nanobeam Bragg coherent diffraction

We present an approach for simulating x-ray nanobeam Bragg coherent diffraction patterns based on the Takagi-Taupin equations. Compared to conventional methods, the current approach can be universally applied to any weakly strained system including semi-infinite crystals that diffract dynamically. It addresses issues such as the curved wave front and redivergence of the focused incident beam. In conclusion, we show excellent agreement with experimental data for a strained La 0.7 ⁢Sr 0.3⁢ MnO 3 thin film on a SrTiO 3 substrate and a path to extracting physical information using automatic differentiation.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2026

Physics vs structure: A systematic benchmark of learning strategies for multi-zone building thermal dynamics

Recent advances in physics-informed and data-driven machine learning promise improved thermal models for advanced building control, yet there is limited quantitative evidence on when added physics structure and architectural complexity are beneficial. Here, this work presents a systematic benchmark of five representative system identification methods for modeling multi-zone building thermal dynamics: linear state-space models, multi-layer perceptrons, neural state-space models, neural ordinary differential equations, and physically-consistent neural networks. The methods are evaluated across multiple data regimes and zone coupling strategies. Using a high-fidelity multi-zone commercial building emulator, we examine short-term and long-term prediction accuracy, computational efficiency, and ease of development. Our results reveal critical trade-offs between prediction performance, model complexity, and physical consistency. We demonstrate that decoupled, nonlinear black-box models consistently outperform coupled physics-constrained architectures in both predictive accuracy and out-of-distribution robustness in majority of the test cases for the building type considered in the study. Our findings quantify the cost of complexity in building thermal modeling and provide concrete, actionable, scenario-based guidelines for selecting model classes for control-oriented applications.

Building thermal modeling↗
DOE OSTI2025

Understanding latent timescales in neural ordinary differential equation models of advection-dominated dynamical systems

The neural ordinary differential equation (ODE) framework has shown considerable promise in recent years in developing highly accelerated surrogate models for complex physical systems characterized by partial differential equations (PDEs). For PDE-based systems, state-of-the-art neural ODE strategies leverage a two-step procedure to achieve this acceleration: a nonlinear dimensionality reduction step provided by an autoencoder, and a time integration step provided by a neural-network based model for the resultant latent space dynamics (the neural ODE). This work explores the applicability of such autoencoder-based neural ODE strategies for PDEs in which advection terms play a critical role. More specifically, alongside predictive demonstrations, physical insight into the sources of model acceleration (i.e., how the neural ODE achieves its acceleration) is the scope of the current study. Such investigations are performed by quantifying the effects of both autoencoder and neural ODE components on latent system time-scales using eigenvalue analysis of dynamical system Jacobians. To this end, the sensitivity of various critical training parameters – de-coupled versus end-to-end training, latent space dimensionality, and the role of training trajectory length, for example – to both model accuracy and the discovered latent system timescales is quantified. Furthermore, this work specifically uncovers the key role played by the training trajectory length (the number of rollout steps in the loss function during training) on the latent system timescales: larger trajectory lengths correlate with an increase in limiting neural ODE time-scales, and optimal neural ODEs are found to recover the largest time-scales of the full-order (ground-truth) system. Demonstrations are performed across fundamentally different unsteady fluid dynamics configurations influenced by advection: (1) the Kuramoto–Sivashinsky equations (2) Hydrogen-Air channel detonations (the compressible reacting Navier–Stokes equations with detailed chemistry), and (3) 2D Atmospheric flow.

Advection-dominated dynamical systems↗
DOE OSTI2025

HFBTHO-AD: Differentiation of a nuclear energy density functional code

The HFBTHO code implements a nuclear energy density functional solver to model the structure of atomic nuclei. HFBTHO has previously been used to calibrate energy functionals and perform sensitivity analysis by using derivative-free methods. To enable derivative-based optimization and uncertainty quantification approaches, we must compute the derivatives of HFBTHO outputs with respect to the parameters of the energy functional, which are a subset of all input parameters of the code. Here, we use the algorithmic/automatic differentiation (AD) tool Tapenade to differentiate HFBTHO. We compare the derivatives obtained using AD against finite-difference approximation and examine the performance of the derivative computation.

Algorithmic differentiation↗
DOE OSTI2025

An infrasound source analysis of the OSIRIS-REx sample return capsule hypersonic re-entry

The OSIRIS-REx sample return capsule's hypersonic re-entry into the atmosphere is a rare opportunity to test a variety of sonic boom source models since the projectile dimensions are well characterized. While the as-flown flight path is unknown, the predicted flight path enables a rough approximation of the source Mach number and location. Six infrasound microphones deployed in the boom carpet along the predicted flight path recorded impulsive signals from the OSIRIS-REx re-entry. Using a suite of atmosphere profiles and the geometric acoustics approximation, we estimate locations with uncertainty estimates along the flight path from which the signals were emitted. Acoustic overpressure and signal duration predictions from Whitham's far field theory, Carlson's simplified sonic boom prediction method, and a drag-dominated hypersonic model are analyzed with uncertainty estimates from the location estimate. While the Carlson simplified sonic boom prediction method could be accurate, our preference is for the drag-dominated source model. Using this source model with an inviscid Burgers's equation solver for propagation, we obtained an excellent match to the recorded data. In conclusion, these results will help better inform future sample return capsule re-entry observation campaigns as well as contribute to a better understanding of high altitude infrasonic sources.

58 GEOSCIENCES↗
DOE OSTI2024

Determining Stellar Elemental Abundances from DESI Spectra with the Data-driven Payne

Abstract Stellar abundances for a large number of stars provide key information for the study of Galactic formation history. Large spectroscopic surveys such as the Dark Energy Spectroscopic Instrument (DESI) and LAMOST take median-to-low-resolution (R≲ 5000) spectra in the full optical wavelength range for millions of stars. However, the line-blending effect in these spectra causes great challenges for elemental abundance determination. Here we employDD-Payne, a data-driven method regularized by differential spectra from stellar physical models, to the DESI early data release spectra for stellar abundance determination. Our implementation delivers 15 labels, including effective temperatureT eff , surface gravity log g , microturbulence velocityv mic , and the abundances for 12 individual elements, namely C, N, O, Mg, Al, Si, Ca, Ti, Cr, Mn, Fe, and Ni. Given a spectral signal-to-noise ratio of 100 per pixel, the internal precisions of the label estimates are about 20 K forT eff , 0.05 dex for log g , and 0.05 dex for most elemental abundances. These results agree with the theoretical limits from the Crámer–Rao bound calculation within a factor of 2. The majority of the accreted halo stars contributed by the Gaia–Enceladus–Sausage are discernible from the disk and in situ halo populations in the resultant [Mg/Fe]–[Fe/H] and [Al/Fe]–[Fe/H] abundance spaces. We also provide distance and orbital parameters for the sample stars, which spread over a distance out to ∼100 kpc. The DESI sample has a significantly higher fraction of distant (or metal-poor) stars than the other existing spectroscopic surveys, making it a powerful data set for studying the Galactic outskirts. The catalog is publicly available.

Astronomy & Astrophysics↗
DOE OSTI2024

Scientific machine learning for closure models in multiscale problems: A review

Here, closure problems are omnipresent when simulating multiscale systems, where some quantities and processes cannot be fully prescribed despite their effects on the simulation's accuracy. Recently, scientific machine learning approaches have been proposed as a way to tackle the closure problem, combining traditional (physics-based) modeling with data-driven (machine-learned) techniques, typically through enriching differential equations with neural networks. This paper reviews the different reduced model forms, distinguished by the degree to which they include known physics, and the different objectives of a priori and a posteriori learning. The importance of adhering to physical laws (such as symmetries and conservation laws) in choosing the reduced model form and choosing the learning method is discussed. The effect of spatial and temporal discretization and recent trends toward discretization-invariant models are reviewed. In addition, we make the connections between closure problems and several other research disciplines: inverse problems, Mori-Zwanzig theory, and multi-fidelity methods. In conclusion, much progress has been made with scientific machine learning approaches for solving closure problems, but many challenges remain. In particular, the generalizability and interpretability of learned models is a major issue that needs to be addressed further.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Weak baselines and reporting biases lead to overoptimism in machine learning for fluid-related partial differential equations

One of the most promising applications of machine learning in computational physics is to accelerate the solution of partial differential equations (PDEs). The key objective of machine-learning-based PDE solvers is to output a sufficiently accurate solution faster than standard numerical methods, which are used as a baseline comparison. Here, we first perform a systematic review of the ML-for-PDE-solving literature. Out of all of the articles that report using ML to solve a fluid-related PDE and claim to outperform a standard numerical method, we determine that 79% (60/76) make a comparison with a weak baseline. Second, we find evidence that reporting biases are widespread, especially outcome reporting and publication biases. We conclude that ML-for-PDE-solving research is overoptimistic: weak baselines lead to overly positive results, while reporting biases lead to under-reporting of negative results. To a large extent, these issues seem to be caused by factors similar to those of past reproducibility crises: researcher degrees of freedom and a bias towards positive results. We call for bottom-up cultural changes to minimize biased reporting as well as top-down structural reforms to reduce perverse incentives for doing so.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Nonperturbative random matrix model of N = 2 JT supergravity

It is shown how to nonperturbatively define a random matrix model that captures key physics of N = 2 Jackiw-Teitelboim supergravity, going well beyond the perturbative topological expansion defined recently by Turiaci and Witten. A decomposition into an infinite family of certain multicritical models is derived, leading to the definition of a nonlinear ordinary differential equation from which the physics may be computed. Bogomol’nyi-Prasad-Sommerfield (BPS) states are naturally described by the model. The nonperturbative completions of the spectral densities for non-BPS multiplets are readily extracted. Published by the American Physical Society 2024

Johnson, Clifford V. (ORCID:0000000189645830)↗
DOE OSTI2026

Multiscale Neural Networks for Approximating Green’s Functions

Neural networks (NNs) have been widely used to solve partial differential equations (PDEs) in the applications of physics, biology, and engineering. One effective approach for solving PDEs with a fixed differential operator is learning Green’s functions. However, Green’s functions are notoriously difficult to learn due to their poor regularity, which typically requires larger NNs and longer training times. In this work, we address these challenges by leveraging multiscale NNs to learn Green’s functions. Through theoretical analysis using multiscale Barron space methods and experimental validation, we show that the multiscale approach significantly reduces the necessary NN size and accelerates training.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

A tale of two planets: Disparate evolutionary models for Mars inferred from radiogenic isotope compositions of Martian meteorites

The radiogenic isotopic compositions of basaltic Martian meteorites (shergottites) and clinopyroxene/olivine cumulate meteorites (nakhlite/chassignites) are used to define the global evolution of Mars. However, the two main groups of meteorites demonstrate that their sources underwent divergent styles of magmatic evolution. The shergottites portray a planet that differentiated ~4.52 billion years ago via solidification of a magma ocean, producing incompatible element-depleted and -enriched reservoirs that remained isolated until melt production. In contrast, the reservoir from which the nakhlite/chassignites derive may have formed earlier, produced melts that fractionated Sm/Nd and Hf/W differently, was compositionally less variable, and experienced a significantly more complex history following primordial differentiation than the shergottite sources. The disparate histories recorded by these two groups of meteorites elucidate important questions that could be addressed by acquiring additional samples. Obtaining samples that shared the isotopic systematics of the shergottites would provide confidence that extrapolating the primordial differentiation history of Mars from shergottite radiogenic isotope systematics is reasonable. Returned samples from Mars will also constrain the physical locations of the meteorite source regions, providing insights into the general structure of the Martian mantle. In addition, they will help constrain the phases present in the martian mantle during melting and the conditions under which they are stable. Finally, identifying an evolved lithology that satisfies the geochemical and isotopic constraints placed on the incompatible element-enriched endmember observed in the shergottites would define the nature of magmatic evolution on Mars and whether it is more akin to processes on the Earth or the Moon.

Mars↗
DOE OSTI2025

SDA: a symbolic differential algebra package in C++

Truncated Power Series Algebra (TPSA), or Differential Algebra (DA), is a well-established tool in accelerator physics, commonly used for generating high-order maps of dynamic systems, as well as in symplectic tracking, normal form analysis, verified integration, optimization, and fast multipole methods. This package is the first to perform symbolic DA computations, enabling traceability of initial condition contributions and runtime reduction for repeated DA calculations, potentially expanding DA’s applications.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Causality-respecting adaptive refinement for PINNs: enabling precise interface evolution in phase field modeling

Physics-informed neural networks (PINNs) have emerged as a powerful tool for solving physical systems described by partial differential equations (PDEs). However, their accuracy in dynamical systems, particularly those involving sharp moving boundaries with complex initial morphologies, remains a challenge. Here, this study introduces an approach combining residual-based adaptive refinement (RBAR) with causality-informed training to enhance the performance of PINNs in solving spatio-temporal PDEs. Our method employs a three-step iterative process: initial causality-based training, RBAR-guided domain refinement, and subsequent causality training on the refined mesh. Applied to the Allen-Cahn equation, a widely-used model in phase field simulations, our approach demonstrates significant improvements in solution accuracy and computational efficiency over traditional PINNs. Notably, we observe an ‘overshoot and relocate’ phenomenon in dynamic cases with complex morphologies, showcasing the method’s adaptive error correction capabilities. This synergistic interaction between RBAR and causality training enables accurate capture of interface evolution, even in challenging scenarios where traditional PINNs fail. Our framework not only resolves the limitations of uniform refinement strategies but also provides a generalizable methodology for solving a broad range of spatio-temporal PDEs. The enhanced performance of the RBAR–causality combined framework demonstrates its strong potential for advancing PINN-based modeling of physical systems characterized by complex, evolving interfaces.

Allen-Cahn equations↗