SEARCH · Engineering Papers

Results for “Transformer model”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,765 records · Page 98

All sources NASA NTRS DOE OSTI

DOE OSTI2022

Diffraction and thermal effect of a Bessel-Gaussian laser for Ag nanoparticle deposition

Nanoparticles are known to sinter at much lower temperatures than the corresponding bulk or micro size particles. A laser-assisted sintering process is considered in this study to sinter Ag nanoparticles by dispensing Ag paste onto an indium tin oxide-coated Si substrate. The Gaussian beam of a CO 2 laser source is propagated through axicon and biconvex lenses, and the resulting hollow beam is focused on the Ag paste with a hollow parabolic mirror. A Bessel-Gaussian irradiance distribution is obtained at the focal plane of the parabolic mirror due to the interference of the hollow laser cone. The Fresnel diffraction approximation is considered to determine the phasor of the laser and an analytical approach is implemented to calculate the irradiance distribution of the Bessel-Gaussian beam. This irradiance distribution is utilized as a heat source in a heat conduction model and the temperature distribution is analyzed for thin Ag films formed during the laser sintering of Ag nanoparticles. An analytical expression is obtained for the temperature distribution by solving the heat conduction equation using Fourier transform for finite media. The widths of the deposited Ag lines are predicted from the temperature profiles and the model predictions compare well with the experimental results. The isotherms are found to be geometrically noncongruent with convex and concave tips depending on the locally maximum and minimum irradiances of the Bessel-Gaussian beam, respectively. The convex and concave tips, however, appear in the same isotherm for sufficiently high substrate speed relative to the laser beam.

77 NANOSCIENCE AND NANOTECHNOLOGY↗

NASA NTRS1991

Analysis of structural response data using discrete modal filters

The application of reciprocal modal vectors to the analysis of structural response data is described. Reciprocal modal vectors are constructed using an existing experimental modal model and an existing frequency response matrix of a structure, and can be assembled into a matrix that effectively transforms the data from the physical space to a modal space within a particular frequency range. In other words, the weighting matrix necessary for modal vector orthogonality (typically the mass matrix) is contained within the reciprocal model matrix. The underlying goal of this work is mostly directed toward observing the modal state responses in the presence of unknown, possibly closed loop forcing functions, thus having an impact on both operating data analysis techniques and independent modal space control techniques. This study investigates the behavior of reciprocol modal vectors as modal filters with respect to certain calculation parameters and their performance with perturbed system frequency response data.

Freudinger, Lawrence C.↗

DOE OSTI2023

Transformer Leakage Inductance Design Methodology

The leakage inductance exhibited by a transformer depends on its winding geometry, which generally involves the selection of several key design parameters in addition to the winding structure and the interleaving configuration. With few resources explaining the effects of these design choices on the observed leakage inductance, numerous trial-and-error iterations become necessary to realize the desired leakage inductance. This paper explores more than a hundred winding geometries feasible in a 2-winding transformer comprising the same magnetic core, number of turns, and wire gauge, and finds the leakage inductance for each unique design using 2-D finite element method (FEM) simulations in association with the semi-analytical double- 2-D model. These leakage inductances are plotted and further analyzed to understand the effects of different design parameters on the effective leakage inductance. The results presented herein and the conclusions drawn from this research can serve as a valuable resource for future design practitioners from both industry and academia.

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2022

COMPOFF: A Compiler Cost model using Machine Learning to predict the Cost of OpenMP Offloading

The HPC industry is inexorably moving towards an era of extremely heterogeneous architectures, with more devices configured on any given HPC platform and potentially more kinds of devices, some of them highly specialized. Writing a separate code suitable for each target system for a given HPC application is not practical. The better solution is to use directive-based parallel programming models such as OpenMP. OpenMP provides a number of options for offloading a piece of code to devices like GPUs. To select the best option from such options during compilation, most modern compilers use analytical models to estimate the cost of executing the original code and the different offloading code variants. Building such an analytical model for compilers is a difficult task that necessitates a lot of effort on the part of a compiler engineer. Recently, machine learning techniques have been successfully applied to build cost models for a variety of compiler optimization problems. In this paper, we present COMPOFF, a cost model which uses the multi-layer perceptrons to statically estimates the Cost of OpenMP OFFloading. We used six different transformations on a parallel code of Wilson Dslash Operator to support GPU offloading, and we predicted their cost of execution on different GPUs using COMPOFF during compile time. Our results show that this model can predict offloading costs with a root mean squared error in prediction of less than 0.5 seconds. Our preliminary findings indicate that this work will make it much easier and faster for scientists and compiler developers to port legacy HPC applications that use OpenMP to new heterogeneous computing environment.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2022

Standard model physics and the digital quantum revolution: thoughts about the interface

Advances in isolating, controlling and entangling quantum systems are transforming what was once a curious feature of quantum mechanics into a vehicle for disruptive scientific and technological progress. Pursuing the vision articulated by Feynman, a concerted effort across many areas of research and development is introducing prototypical digital quantum devices into the computing ecosystem available to domain scientists. Through interactions with these early quantum devices, the abstract vision of exploring classically-intractable quantum systems is evolving toward becoming a tangible reality. Beyond catalyzing these technological advances, entanglement is enabling parallel progress as a diagnostic for quantum correlations and as an organizational tool, both guiding improved understanding of quantum many-body systems and quantum field theories defining and emerging from the standard model. Here, from the perspective of three domain science theorists, this article compiles thoughts about the interface on entanglement, complexity, and quantum simulation in an effort to contextualize recent NISQ-era progress with the scientific objectives of nuclear and high-energy physics.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

DOE OSTI2025

Neural refinement of sample weights

Monte Carlo simulations are an essential tool in particle physics data analysis. Events are typically generated alongside weights that redistribute the cross section of the simulated process across the phase space. These weights can be negative, and several post hoc methods have been developed to eliminate or mitigate the negative values. All of these methods share the common strategy of approximating the average weight as a function of phase space. We introduce an alternative approach, which, instead of reweighting to the average, refines the initial weights with a scaling transformation, utilizing a phase space-dependent factor. Since this new refinement method does not need to model the full weight distribution, it can be more accurate. High-dimensional and unbinned phase space is processed using neural networks for the refinement method. In addition to the refinement method, we introduce a new resampling protocol, which can be used in conjunction with any weight transformation to not only preserve the average weight but also the statistical uncertainties of the initial distribution. Using both realistic and synthetic examples, we show that the new neural refinement method is able to match or exceed the accuracy of similar weight transformations and that the new resampling protocol is simpler in implementation than previous methods while exhibiting equivalent statistical properties.

Artificial neural networks↗

NASA NTRS1989

Perception for rugged terrain

A three-dimensional perception system for building a geometrical description of rugged terrain environments from range image data is presented with reference to the exploration of the rugged terrain of Mars. An intermediate representation consisting of an elevation map that includes an explicit representation of uncertainty and labeling of the occluded regions is proposed. The locus method used to convert range image to an elevation map is introduced, along with an uncertainty model based on this algorithm. Both the elevation map and the locus method are the basis of a terrain matching algorithm which does not assume any correspondences between range images. The two-stage algorithm consists of a feature-based matching algorithm to compute an initial transform and an iconic terrain matching algorithm to merge multiple range images into a uniform representation. Terrain modeling results on real range images of rugged terrain are presented. The algorithms considered are a fundamental part of the perception system for the Ambler, a legged locomotor.

Kweon, In SO↗

NASA NTRS1997

CMOTT Turbulence Module for NPARC

This is a user's manual of the CMOTT turbulence module, version 2.0, developed for the NPARC code. The module is written in a self-contained manner so that the user can use any turbulence model in the module without concern as to how it is implemented and solved. Three two-equation turbulence models have been built into the module: Chien, Shih-Lumley and CMOTT models, and all of them have both the low Reynolds number and wall function options. Unlike Chien's model, both the Shih-Lumley and CMOTT models do not involve the dimensionless wall distance y(sup +) in the low Reynolds number approach, an advantage for separated flow calculations. The Van Driest transformation is used so that the wall functions can be applied to both incompressible and compressible flows. The manual gives the details of the turbulence models used and their numerical implementation. It also gives two application examples, one for subsonic and the other for transonic flow, for demonstration. The module can be easily linked to the NPARC code for practical applications.

Zhu, J.↗

NASA NTRS1994

Surface matching for correlation of virtual models: Theory and application

Virtual reality can enable a robot user to off line generate and test in a virtual environment a sequence of operations to be executed by the robot in an assembly cell. Virtual models of objects are to be correlated to the real entities they represent by means of a suitable transformation. A solution to the correlation problem, which is basically a problem of 3-dimensional adjusting, has been found exploiting the surface matching theory. An iterative algorithm has been developed, which matches the geometric surface representing the shape of the virtual model of an object, with a set of points measured on the surface in the real world. A peculiar feature of the algorithm is to work also if there is no one-to-one correspondence between the measured points and those representing the surface model. Furthermore the problem of avoiding convergence to local minima is solved, by defining a starting point of states ensuring convergence to the global minimum. The developed algorithm has been tested by simulation. Finally, this paper proposes a specific application, i.e., correlating a robot cell, equipped for biomedical use with its virtual representation.

Caracciolo, Roberto↗

DOE OSTI2024

How Earth System Models Can Inform Key Dimensions of Marine Food Security in the Alaskan Arctic

The Arctic is home to several groups of Indigenous Peoples, each with distinct ways of interacting with their environment and ways of life. Arctic, Indigenous Peoples’ food sovereignty is tightly linked with food security. Subsistence harvesting activities provide nutritious and culturally vital foods for Alaska Native households and communities. Climate change is causing rapid and more unpredictable shifts in environmental conditions that impact three of the key aspects of food security, availability, stability, and accessibility. While communities monitor the abundance and health of food webs and environments as part of subsistence harvest practices, anticipating major transformations and changes in these systems is challenging. We explored the potential of Earth System Model output in helping anticipate or project physical or ecosystem changes relevant to Alaska Indigenous peoples’ food security needs. Through examples of model products, that provide measures of accessibility and availability of marine resources, we show that modern models, such as the Energy Exascale Earth System Model presented here, can provide estimates of a broad suite of variables relevant to food security. We investigate how Earth System Model output could contribute to exploring questions related to aspects of Arctic food security such as accessibility and availability and highlight present model shortcomings that, if addressed, would move Earth System Models closer to being a useful tool for understanding environmentally driven changes to the availability and accessibility of harvestable food resources. Our example model-derived food security indicators illustrate how Earth System Model output could be combined with relevant, non-model, information sources; These model products are meant only as a starting point and a tool for engaging community members and to present, in an accessible way, the model’s potential utility, or current lack thereof, to rights holders and stakeholders concerned about food security. We are hopeful that with example products in hand, additional model development efforts will have a higher likelihood of success in achieving an iterative discussion with stakeholders regarding feasible and desired products.

Gibson, Georgina A.↗

DOE OSTI2026

A roadmap toward scaling, reasoning and self-evolving foundation models for nuclear and particle physics

Foundation models have revolutionized artificial intelligence, with Large Language Models demonstrating unprecedented capabilities in multimodal understanding, reasoning and tool use. Nuclear and particle physics stands at a critical juncture where similar transformative potential awaits realization. The field generates exabytes of experimental data, exascale simulations, and decades of theoretical insights — yet these remain largely disconnected from modern Artifical Intelligence (AI) capabilities, with most physics AI applications confined to narrow, task-specific models that suffer from domain shifting when applied to real experimental data. We present a roadmap for FM4NPP (Foundation Model for Nuclear and Particle Physics), systematically scaling from current proof-of-concept models to trillion-parameter architectures capable of autonomous discovery. Our approach advances three critical frontiers: unified data infrastructure integrating detector data, scientific knowledge and computational tools across global facilities; multi-facility foundation models enabling cross-experiment knowledge transfer and accelerated discovery; and agentic AI capabilities for reasoning and autonomous tool use. The resulting self-evolving FM4NPP will transform physics research by converting time-intensive data analysis, theory derivation and computational bottlenecks into rapid AI–human collaborative discovery. This paradigm shift promises to fundamentally accelerate scientific progress in nuclear and particle physics, enabling researchers to focus on high-level insights while AI handles routine analysis and explores vast parameter spaces beyond human capacity.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

NASA NTRS2020

Toward Coupled Data Assimilation in NASA’s GEOS: Developments in the Ocean Context

The Global Modeling & Assimilation Office (GMAO) at NASA GSFC produces analyses and predictions of the Earth system using various configurations of the Goddard Earth Observing System (GEOS) model and assimilation system. The current sub-seasonal-to-seasonal prediction system (GEOS-S2S) is based on a coupled atmosphere-ocean-land-ice configuration of GEOS which includes the Modular Ocean Model version 5 (MOM5) run at approximately 50-km resolution and a de-coupled OI-based ocean analysis that uses an initialization of MOM5 forced by the MERRA-2 reanalysis. GMAO will soon implement an updated GEOS-S2S system that will run at 25-km resolution and adopt aspects of the hybrid four-dimensional ensemble-variational (H4DEnVar) system already running in the production-version atmospheric analysis system, including a Local Ensemble Transform Kalman Filter (LETKF) to provide initial conditions for the oceanic state. This presentation will focus on developments to sustain the GMAO's systems on longer time horizons, where more radical transformations will be required to adapt to advanced computing environments, higher resolution and more diverse model components, and new observations for the Earth system. Results will describe progress toward a version of the GEOS coupled system that will be based around the Joint Effort for Data assimilation Integration (JEDI) framework being developed within Joint Center for Satellite Data Assimilation (JCSDA) and include an updated ocean model, MOM6. Discussion will focus specifically on the use of a Unified Forward Operator (UFO) for simulating observations and the Object Oriented Prediction System (OOPS) for providing the state estimate. These features are being developed as a multi-agency effort under the auspices of the JCSDA and are being adopted in the GMAO for all its applications of coupled data assimilation including S2S, numerical weather prediction, and reanalysis.

Mahajan, Rahul↗

DOE OSTI2025

Transformers and Long Short-Term Memory Transfer Learning for GenIV Reactor Temperature Time Series Forecasting

Automated monitoring of the coolant temperature can enable autonomous operation of generation IV reactors (GenIV), thus reducing their operating and maintenance costs. Automation can be accomplished with machine learning (ML) models trained on historical sensor data. However, the performance of ML usually depends on the availability of large amount of training data, which is difficult to obtain for GenIV, as this technology is still under development. We propose the use of transfer learning (TL), which involves utilizing knowledge across different domains, to compensate for this lack of training data. TL can be used to create pre-trained ML models with data from small-scale research facilities, which can then be fine-tuned to monitor GenIV reactors. In this work, we develop pre-trained Transformer and long short-term memory (LSTM) networks by training them on temperature measurements from thermal hydraulic flow loops operating with water and Galinstan fluids at room temperature at Argonne National Laboratory. The pre-trained models are then fine-tuned and re-trained with minimal additional data to perform predictions of the time series of high temperature measurements obtained from the Engineering Test Unit (ETU) at Kairos Power. The performance of the LSTM and Transformer networks is investigated by varying the size of the lookback window and forecast horizon. The results of this study show that LSTM networks have lower prediction errors than Transformers, but LSTM errors increase more rapidly with increasing lookback window size and forecast horizon compared to the Transformer errors.

LSTM↗

DOE OSTI2020

Quantification of High-Temperature Transition Al2O3 and Their Phase Transformations

High temperature exposure of ?-Al2O3 can lead to a series of polymorphic transformations, including the formation of ?-Al2O3 and ?-Al2O3. Quantification of the microstructure in the ?/?-Al2O3 formation range represents a formidable challenge as both phases accommodate a high degree of structural disorder. In this work, we explore the use of XRD recursive stacking formalism for quantification of high temperature transition aluminas. We formulate the recursive stacking methodology for modelling of disorder in ?-Al2O3 and twinning in ?-Al2O3 and show that explicitly accounting for the disorder is necessary to reliably model the XRD patterns of high temperature transition alumina. In the second part, we use the recursive stacking approach to study phase transformation during high temperature (1050 ºC) treatment. We show that the two different intergrowth modes of ?-Al2O3 have different transformation characteristics, and that a significant portion of ?-Al2O3 is stabilized with ?-Al2O3 even after prolonged high-temperature exposures. In discussions, we outline the limitation of the current XRD approach and discuss a possible multimodal XRD and NMR approach which can improve analysis of complex transition aluminas. This work was performed in the Wiley Environmental Molecular Sciences Laboratory (EMSL), a national scientific user facility sponsored by DOEs Office of Biological and Environmental Research and located at PNNL. The work was supported by the U.S. Department of Energy (DOE), Office of Basic Energy Sciences, Division of Chemical Sciences, Geosciences, and Biosciences.

Kovarik, Libor↗

NASA NTRS1985

A tomographic technique for aerodynamics at transonic speeds

Computer aided tomography (CAT) provides a means of noninvasively measuring the air density distribution around an aerodynamic model. This technique is global in that a large portion of the flow field can be measured. A test of the applicability of CAT to transonic velocities was studied. A hemispherical-nose cylinder afterbody model was tested at a Mach number of 0.8 with a new laser holographic interferometer at the 2- by 2-Foot Transonic Wind Tunnel. Holograms of the flow field were taken and were reconstructed into interferograms. The fringe distribution (a measure of the local densities) was digitized for subsequent data reduction. A computer program based on the Fourier-transform technique was developed to convert the fringe distribution into three-dimensional densities around the model. Theoretical aerodynamic densities were calculated for evaluating and assessing the accuracy of the data obtained from the tomographic method.

Lee, G.↗

NASA NTRS1985

A tomographic technique for aerodynamics at transonic speeds

Lee, G.↗

NASA NTRS2000

Real-Time Parameter Estimation in the Frequency Domain

A method for real-time estimation of parameters in a linear dynamic state-space model was developed and studied. The application is aircraft dynamic model parameter estimation from measured data in flight. Equation error in the frequency domain was used with a recursive Fourier transform for the real-time data analysis. Linear and nonlinear simulation examples and flight test data from the F-18 High Alpha Research Vehicle were used to demonstrate that the technique produces accurate model parameter estimates with appropriate error bounds. Parameter estimates converged in less than one cycle of the dominant dynamic mode, using no a priori information, with control surface inputs measured in flight during ordinary piloted maneuvers. The real-time parameter estimation method has low computational requirements and could be implemented

Morelli, Eugene A.↗

DOE OSTI2022

The energy system transformation needed to achieve the US long-term strategy

The authors designed and executed the integrated assessment modeling for the United States long-term strategy. They bring diverse expertise to the modeling and analysis of United States decarbonization. Russell Horowitz, Matthew Binsted, and Haewon McJeon are scientists at the Joint Global Change Research Institute, a partnership between Pacific Northwest National Laboratory and the University of Maryland. Allen Fawcett, James McFarland, and Morgan Browning are economists at the Environmental Protection Agency’s Climate Economics Branch. Claire Henly is White House Fellow at the Office of the US Special Presidential Envoy for Climate. Nathan Hultman is the Director of the Center for Global Sustainability at the University of Maryland.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗