Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Machine learning in 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 901 records · Page 50
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Atikokan Digital Twin: Machine learning in a biomass energy system

The Atikokan Generating Station, operated by Ontario Power Generation, has a 200 MW, biomass-fired tower boiler that operates on a dispatch schedule with a five-minute cycle. The boiler is generally operated in the range of 40–100 MW using two of five burner levels. In order to optimize boiler performance, we propose the implementation of a unique digital twin. Our digital twin abstraction couples Bayesian inference from science-based models and from observations (machine learning) with decision theory to predict operating-variable set points that optimize the physical asset (the boiler) in the presence of uncertainty (artificial intelligence). We focus this paper on the continuous Bayesian machine learning part of the Atikokan Digital Twin; we discuss decision theory in a companion paper. We identify and learn about 12 operational, model, and measured-output parameters and their uncertainties from high-fidelity, science-based simulations of the Atikokan boiler and from the observed measurements at the power plant. Since the goal of the Atikokan Digital Twin is to implement it online in real time, we require fast function evaluations for the quantities of interest extracted from the simulations in the Bayesian analysis. We use Gaussian process regression/interpolation to create accurate, robust surrogate models. We define the Bayesian priors and likelihood function and solve for the posterior distributions of the 12 parameters. Here we then propagate these distributions (i.e., parameters with uncertainty) into the predicted distributions of 790 quantities of interest to learn about the relative importance of various sources of error including experimental, model, and operating-parameter errors.

09 BIOMASS FUELS↗
DOE OSTI2022

Density functional theory of water with the machine-learned DM21 functional

The delicate interplay between functional-driven and density-driven errors in density functional theory (DFT) has hindered traditional density functional approximations (DFAs) from providing an accurate description of water for over 30 years. Recently, the deep-learned DeepMind 21 (DM21) functional has been shown to overcome the limitations of traditional DFAs as it is free of delocalization error. To determine if DM21 can enable a molecular-level description of the physical properties of aqueous systems within Kohn–Sham DFT, we assess the accuracy of the DM21 functional for neutral, protonated, and deprotonated water clusters. Furthermore, we find that the ability of DM21 to accurately predict the energetics of aqueous clusters varies significantly with cluster size. Additionally, we introduce the many-body MB-DM21 potential derived from DM21 data within the many-body expansion of the energy and use it in simulations of liquid water as a function of temperature at ambient pressure. We find that size-dependent functional-driven errors identified in the analysis of the energetics of small clusters calculated with the DM21 functional result in the MB-DM21 potential systematically overestimating the hydrogen-bond strength and, consequently, predicting a more ice-like local structure of water at room temperature.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

Matrix-Model Simulations Using Quantum Computing, Deep Learning, and Lattice Monte Carlo

Matrix quantum mechanics plays various important roles in theoretical physics, such as a holographic description of quantum black holes, and it underpins the only practical numerical approach to the study of complex high-dimensional supergravity theories. Understanding quantum black holes and the role of entanglement in a holographic setup is of paramount importance for the realization of a quantum theory of gravity. Moreover, a complete numerical understanding of the holographic duality and the emergence of geometric space-time features from microscopic degrees of freedom could pave the way for new discoveries in quantum information science. Euclidean lattice Monte Carlo simulations are the de facto numerical tool for understanding the spectrum of large matrix models and have been used to test the holographic duality. However, they are not tailored to extract dynamical properties or even the quantum wave function of the ground state of matrix models. Quantum computing and deep learning provide potentially useful approaches to study the dynamics of matrix quantum mechanics. If successful in the context of matrix models, these rapidly improving numerical techniques could become the new Swiss army knife of quantum gravity practitioners. In this paper, we perform the first systematic survey for quantum computing and deep-learning approaches to matrix quantum mechanics, comparing them to lattice Monte Carlo simulations. These provide baseline benchmarks before addressing more complicated problems. In particular, we test the performance of each method by calculating the low-energy spectrum.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
NASA NTRSDate not supplied

Development of Solar Flare and Energetic Particle Prediction Portal (SEP 3 )

Solar activity is a primary factor determining the state of the Earth’s space environment, geomagnetic and ionospheric disturbances, and radiation hazards. In the current state of knowledge, machine learning (ML) methods provide essential tools for processing data, investigating relationships among various physical properties and characteristics, uncovering hidden connections, and predicting hazardous solar events. The primary difficulty in developing and applying modern machine-learning tools in heliophysics is that the essential data are scattered among over a hundred data repositories developed by instrument teams of space missions and ground-based observatories. In addition, statistical and ML methods require long time series of homogeneous measurements. To facilitate ML-ready data preparation and access, we have developed an interactive database of solar flares integrating the most essential datasets (https://solarflare.njit.edu/). The database performs an initial data processing and is automatically updated. In addition, we are developing the Solar Energetic Particle Prediction Portal (SEP3, https://sun.njit.edu/SEP3), which hosts web applications that allow users to retrieve the database records. The Portal has a search page for browsing the events from the most widely used catalogs and a dedicated space to share the most recent achievements of the team. The interactive widget can display soft X-ray and proton flux time series from GOES satellites and the flare records. The data portal has been used to evaluate the forecasts of solar proton events and investigate machine-learning approaches to SEP prediction.

SMD↗
DOE OSTI2021

Validation of Subhourly Clipping Loss Error Corrections

Under-performance of solar PV systems is an important issue that increases risks for stakeholders, including developers, investors and operators. Recently some attention has focused on underestimation of inverter clipping losses as a possible source of over-prediction where sub-hourly solar variability is high. Several models and data sets have been analyzed over the past few years, with the aim of quantifying, predicting, and correcting underestimated clipping loss errors for systems with high DC/AC ratio and solar variability. In this research, we apply a machine learning model developed at NREL to two physical PV systems, to correct for subhourly clipping losses. For each system, we compare overall AC power output for the model taken at 1-minute intervals to AC power output taken at 1-hour intervals with the addition of the subhourly clipping correction. Our findings consistently show that the addition of the clipping loss correction lead to a reduction in mean bias error of 0.8% and 1.2% for systems A and B, respectively, with no additional filtering applied. When examining high solar variability periods where clipping is more pronounced, system A and B experienced a 1.8% and 2.7% reduction in mean bias error, respectively, when the clipping correction was applied.

machine learning↗
DOE OSTI2021

Validation of Subhourly Clipping Loss Error Corrections: Preprint

Under-performance of solar PV systems is an important issue that increases risks for stakeholders, including developers, investors and operators. Recently some attention has focused on underestimation of inverter clipping losses as a possible source of over-prediction where sub-hourly solar variability is high. Several models and data sets have been analyzed over the past few years, with the aim of quantifying, predicting, and correcting underestimated clipping loss errors for systems with high DC/AC ratio and solar variability. In this research, we apply a machine learning model developed at NREL to two physical PV systems, to correct for subhourly clipping losses. For each system, we compare overall AC power output for the model taken at 1-minute intervals to AC power output taken at 1-hour intervals with the addition of the subhourly clipping correction. Our findings consistently show that the addition of the clipping loss correction lead to a reduction in mean bias error of 0.6\% and 1.1\% for systems A and B, respectively, with no additional filtering applied. When examining high solar variability periods where clipping is more pronounced, system A and B experienced a 1.4\% and 2.5\% reduction in mean bias error, respectively, when the clipping correction was applied.

machine learning↗
DOE OSTI2024

Non-resonant anomaly detection with background extrapolation

Complete anomaly detection strategies that are both signal sensitive and compatible with background estimation have largely focused on resonant signals. Non-resonant new physics scenarios are relatively under-explored and may arise from off-shell effects or final states with significant missing energy. In this paper, we extend a class of weakly supervised anomaly detection strategies developed for resonant physics to the non-resonant case. Machine learning models are trained to reweight, generate, or morph the background, extrapolated from a control region. A classifier is then trained in a signal region to distinguish the estimated background from the data. The new methods are demonstrated using a semi-visible jet signature as a benchmark signal model, and are shown to automatically identify the anomalous events without specifying the signal ahead of time.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
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 OSTI2020

Nonlinear sparse Bayesian learning for physics-based models

This paper addresses the issue of overfitting while calibrating unknown parameters of over-parameterized physics-based models with noisy and incomplete observations. Here, a semi-analytical Bayesian framework of nonlinear sparse Bayesian learning (NSBL) is proposed to identify sparsity among model parameters during Bayesian inversion. NSBL offers significant advantages over machine learning algorithm of sparse Bayesian learning (SBL) for physics-based models, such as 1) the likelihood function or the posterior parameter distribution is not required to be Gaussian, and 2) prior parameter knowledge is incorporated into sparse learning (i.e. not all parameters are treated as questionable). NSBL employs the concept of automatic relevance determination (ARD) to facilitate sparsity among questionable parameters through parameterized prior distributions. The analytical tractability of NSBL is enabled by employing Gaussian ARD priors and by building a Gaussian mixture-model approximation of the posterior parameter distribution that excludes the contribution of ARD priors. Subsequently, type-II maximum likelihood is executed using Newton's method whereby the evidence and its gradient and Hessian information are computed in a semi-analytical fashion. We show numerically and analytically that SBL is a special case of NSBL for linear regression models. Subsequently, a linear regression example involving multimodality in both parameter posterior pdf and model evidence is considered to demonstrate the performance of NSBL in cases where SBL is inapplicable. Next, NSBL is applied to identify sparsity among the damping coefficients of a mass-spring-damper model of a shear building frame. These numerical studies demonstrate the robustness and efficiency of NSBL in alleviating overfitting during Bayesian inversion of nonlinear physics-based models.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2022

Machine learning on FPGA for event selection

Real-time data processing is a frontier field in experimental particle physics. The application of FPGAs at the trigger level is used by many current and planned experiments (CMS, LHCb, Belle2, PANDA). Usually they use conventional processing algorithms. LHCb has implemented Machine Learning (ML) elements for real-time data processing with a triggered readout system that runs most of the ML algorithms on a computer farm. The work described in this article aims to test the ML-FPGA algorithms for streaming data acquisition. Herein, there are many experiments working in this area and they have a lot in common, but there are many specific solutions for detector and accelerator parameters that are worth exploring further. This report describes the purpose of the work and progress in evaluating the ML-FPGA application.

47 OTHER INSTRUMENTATION↗
DOE OSTI2022

Interpretable machine learning for knowledge generation in heterogeneous catalysis

Most applications of machine learning in heterogeneous catalysis thus far have used black-box models to predict computable physical properties (descriptors), such as adsorption or formation energies, that can be related to catalytic performance (that is, activity or stability). Here, extracting meaningful physical insights from these black-box models has proved challenging, as the internal logic of these black-box models is not readily interpretable due to their high degree of complexity. Interpretable machine learning methods that merge the predictive capacity of black-box models with the physical interpretability of physics-based models offer an alternative to black-box models. In this Perspective, we discuss the various interpretable machine learning methods available to catalysis researchers, highlight the potential of interpretable machine learning to accelerate hypothesis formation and knowledge generation, and outline critical challenges and opportunities for interpretable machine learning in heterogeneous catalysis.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Efficient and Selective Chemical Transformations in Highly Charged and Confined Nanodroplets

The acceleration of chemical reaction rates and the increased product selectivity in microdroplets compared to that in bulk solutions has become a topic of increasing interest that has been extensively characterized by electrospray ionization mass spectrometry (ESI-MS). However, the sources of this acceleration and the detailed relationships between droplet properties and resulting reaction rate acceleration are still under debate. Moreover, droplet properties are governed by multiple interrelated experimental parameters, i.e., electrospray voltage, solution flow rate, etc., which makes it difficult and time-consuming to explore this diverse parameter space using traditional manual experimental or computational approaches. In this work, we developed an automated experimental platform integrating reactions in controlled charged microdroplet environments with ESI-MS characterization and sequential hybrid Bayesian modeling, as well as an optimal experimental design framework, to achieve multidimensional parameter optimization for higher reaction turnover rates, based on a model reaction of tetraethylenepentamine (TEPA) with carbon dioxide. With the current platform, we have achieved automated scans with a range of electrospray voltages and solution flow rates, and determined and optimized parameter settings to achieve increased reaction turnovers. We have also linked this platform to the underlying properties of droplets via a hybrid model incorporating physics, high-level theoretical calculations, and machine learning (ML) approaches. The autonomous platform is broadly applicable to a range of chemical reactions relevant to DOE’s mission in chemical separations, catalysis, and materials synthesis.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

EFIT-Prime: Probabilistic and physics-constrained reduced-order neural network model for equilibrium reconstruction in DIII-D

We introduce EFIT-Prime, a novel machine learning surrogate model for EFIT (Equilibrium FIT) that integrates probabilistic and physics-informed methodologies to overcome typical limitations associated with deterministic and ad hoc neural network architectures. EFIT-Prime utilizes a neural architecture search-based deep ensemble for robust uncertainty quantification, providing scalable and efficient neural architectures that comprehensively quantify both data and model uncertainties. Physically informed by the Grad–Shafranov equation, EFIT-Prime applies a constraint on the current density J tor and a smoothness constraint on the first derivative of the poloidal flux, ensuring physically plausible solutions. Furthermore, the spatial location of the diagnostics is explicitly incorporated in the inputs to account for their spatial correlation. Extensive evaluations demonstrate EFIT-Prime's accuracy and robustness across diverse scenarios, most notably showing good generalization on negative-triangularity discharges that were excluded from training. Timing studies indicate an ensemble inference time of 15 ms for predicting a new equilibrium, offering the possibility of plasma control in real-time, if the model is optimized for speed.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2022

From lab to lamp: Understanding downconverter degradation in LED packages

Downconverters, primarily inorganic phosphors, are critical components in white solid-state LED-based lighting and liquid crystal display backlights. Research efforts have led to a fundamental understanding of a downconverter's absorption, photoluminescence, and efficiency as a function of composition, structure, and processing conditions. However, considerably less work has focused on the reliability of phosphors once they are incorporated into LED packages. Solving these issues is often the final step before the commercialization of new materials, but the significant resources and time required to evaluate and mitigate materials failure are rarely discussed in the literature. In this Perspective, we discuss the need for conducting downconverter reliability testing and the potential of accelerating, screening, and understanding downconverter failure modes. Our focus highlights the mechanisms of failure and discusses how this influences materials selection and the design of different LED packages. We also stress the potential for accelerated reliability testing protocols and note the potential role first-principles calculations and data-driven models could play in establishing the compositional-processing trends for different aspects of downconverter reliability. In conclusion, we close with possible research directions that could improve downconverter reliability and emphasize the importance of assessing a material's (chemical) stability where multiple manufacturing and processing steps can dictate system performance.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2025

Learning robust parameter inference and density reconstruction in flyer plate impact experiments

Estimating physical parameters or material properties from experimental observations is a common objective in many areas of physics and material science. In many experiments, especially in shock physics, radiography is the primary means of observing the system of interest. However, radiography does not provide direct access to key state variables, such as density, which prevents the application of traditional parameter estimation approaches. Here we focus on flyer plate impact experiments on porous materials, and resolving the underlying parameterized equation of state (EoS) and crush porosity model parameters given radiographic observation(s). We use machine learning as a tool to demonstrate with high confidence that using only high impact velocity data does not provide sufficient information to accurately infer both EoS and crush model parameters, even with fully resolved density fields or a dynamic sequence of images. We thus propose an observable data set consisting of low and high impact velocity experiments/simulations that capture different regimes of compaction and shock propagation, and proceed to introduce a generative machine learning approach which produces a posterior distribution of physical parameters directly from radiographs. We demonstrate the effectiveness of the approach in estimating parameters from simulated flyer plate impact experiments, and show that the obtained estimates of EoS and crush model parameters can then be used in hydrodynamic simulations to obtain accurate and physically admissible density reconstructions. Finally, we examine the robustness of the approach to model mismatches, and find that the learned approach can provide useful parameter estimates in the presence of out-of-distribution radiographic noise and previously unseen physics, thereby promoting a potential breakthrough in estimating material properties from experimental radiographic images.

97 MATHEMATICS AND COMPUTING↗