Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “SURROGATE MODELS”

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 1,369 records · Page 76
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Discovering the Unknowns: A First Step

This article aims at discovering the unknown variables in the system through data analysis. The main idea is to use the time of data collection as a surrogate variable and try to identify the unknown variables by modeling gradual and sudden changes in the data. We use Gaussian process modeling and a sparse representation of the sudden changes to efficiently estimate the large number of parameters in the proposed statistical model. The method is tested on a realistic dataset generated using a one-dimensional implementation of a Magnetized Liner Inertial Fusion (MagLIF) simulation model, and encouraging results are obtained.

42 ENGINEERING↗
NASA NTRS2021

Transmission Telescope Optical Dynamic Alignment

The Integrated Radio Optical Communication System (iROC) is designed to transmit data between Mars and Earth by means of radio waves at 32.67 GHz (Ka Band) and laser beam at 1550 nm, both transmitted via a combined teletenna. The iROC terminal will provide ‘beaconless’ operations to allow full function from the outer planets. In order to point without the aid of an uplink beacon, the proof of concept presented here is addressing the need for an accurate determination and control of the relative position of the laser beam with respect to a reference star. The experiment presented simulates a surrogate transmission telescope system in laboratory setting and presents the model used in the correction of the outgoing beam. The results of the model show a nonlinear dependence between the outgoing and the reference beam, indicating the necessity of a minimum of two metrology instruments placed along the optical system for increased pointing precision.

optical communication↗
DOE OSTI2021

Approximating and incorporating model uncertainty in an inversion for seismic source functions: Preliminary results

We present preliminary work on propagating model uncertainty into the estimation of the time domain source time functions of the seismic source. Our method is based on an estimated model covariance function, which we estimate from the data. The model covariance function is then used to construct a suite of surrogate Greens functions which we use in a Monte Carlo type inversion scheme. The result is a probability density function of the six independent source time functions, each of which corresponds to an individual component of the seismic moment tensor. We compare the results of our method with those obtained using a computationally expensive finite difference Monte Carlo method and find that our new method produces results that are deficient in low frequencies. The advantage of our new method, which we term the Karhunen-Loeve Monte Carlo (KLMC) method, is that is several orders of magnitude faster than our current method, which uses a finite difference scheme to produce the suite of forward models.

42 ENGINEERING↗
DOE OSTI2023

Learning macroscopic internal variables and history dependence from microscopic models

This paper concerns the study of history dependent phenomena in heterogeneous materials in a two-scale setting where the material is specified at a fine microscopic scale of heterogeneities that is much smaller than the coarse macroscopic scale of application. Here, we specifically study a polycrystalline medium where each grain is governed by crystal plasticity while the solid is subjected to macroscopic dynamic loads. The theory of homogenization allows us to solve the macroscale problem directly with a constitutive relation that is defined implicitly by the solution of the microscale problem. However, the homogenization leads to a highly complex history dependence at the macroscale, one that can be quite different from that at the microscale. In this paper, we examine the use of machine-learning, and especially deep neural networks, to harness data generated by repeatedly solving the finer scale model to: (i) gain insights into the history dependence and the macroscopic internal variables that govern the overall response; and (ii) to create a computationally efficient surrogate of its solution operator, that can directly be used at the coarser scale with no further modeling. We do so by introducing a recurrent neural operator (RNO), and show that: (i) the architecture and the learned internal variables can provide insight into the physics of the macroscopic problem; and (ii) that the RNO can provide multiscale, specifically FE 2 , accuracy at a cost comparable to a conventional empirical constitutive relation.

36 MATERIALS SCIENCE↗
NASA NTRS2005

Surrogate-based Analysis and Optimization

A major challenge to the successful full-scale development of modem aerospace systems is to address competing objectives such as improved performance, reduced costs, and enhanced safety. Accurate, high-fidelity models are typically time consuming and computationally expensive. Furthermore, informed decisions should be made with an understanding of the impact (global sensitivity) of the design variables on the different objectives. In this context, the so-called surrogate-based approach for analysis and optimization can play a very valuable role. The surrogates are constructed using data drawn from high-fidelity models, and provide fast approximations of the objectives and constraints at new design points, thereby making sensitivity and optimization studies feasible. This paper provides a comprehensive discussion of the fundamental issues that arise in surrogate-based analysis and optimization (SBAO), highlighting concepts, methods, techniques, as well as practical implications. The issues addressed include the selection of the loss function and regularization criteria for constructing the surrogates, design of experiments, surrogate selection and construction, sensitivity analysis, convergence, and optimization. The multi-objective optimal design of a liquid rocket injector is presented to highlight the state of the art and to help guide future efforts.

Queipo, Nestor V.↗
NASA NTRSDate not supplied

Fast Machine Learning Lidar Surrogate Simulator: Pristine Clear Sky

The simulations of lidar signals and retrievals rely on a range of optic-physical models, such as radiative transfer models, particle scattering and absorption models, along with the output data from atmospheric physical models. Integrating these different models to represent signals of a lidar system is computationally expensive, and performing backward retrievals can be complex and ambiguous. However, with the advantages of Machine Learning, there is a new potential for building effective lidar signal database linked to corresponding atmospheric profiles. For this project, we are developing a fast pre-trained neural network as the lidar surrogate simulator using simulated data for a CALIPSO-like lidar (355 nm, 532nm, and 1064nm), and a CO2 differential absorption lidar (DIAL) near 1571nm. Specifically, we utilize a long short-term memory (LSTM) model to map the relationships between atmospheric profiles (pressure, temperature, air density and CO2 mixing ratio) and lidar signals. This approach allows us to build machine learning based simulators that can reconstruct lidar signals at specific bands from MERRA reanalysis data, and perform retrievals of atmospheric profiles using lidar signals at various wavelengths. As a first step, the results show the potential of this method to establish a foundational model for sensor signals. This model offers the promise of enabling both accurate predictions and rapid retrievals, providing a more efficient approach to signal processing and analysis.

Shan Zeng↗
DOE OSTI2023

Effect of sample size on the maximum value distribution of fatigue driving forces in metals and alloys

An analytical framework is presented to predict the effects of sample size on the maximum value distribution (MVD) of the driving forces for fatigue crack formation in metals and alloys. The distribution of the maximum driving force for fatigue crack formation over the domain follows the generalized extreme value theory in the limit as the domain size increases to infinity. Here, a simulation-based analysis of microstructure influences on fatigue resistance for polycrystalline metals and alloys is very costly, and reaching those limits is intractable. This work models the MVD of Fatigue Indicator Parameters (FIPs), which serve as surrogate measures for the driving force for fatigue crack formation, at finite sample sizes prior to their convergence to a limiting extreme value distribution. Large-scale crystal plasticity finite element (CPFE) simulations of FCC Al 7075-T6 with microstructure realizations of various sizes are incorporated to calibrate and evaluate the developed framework, and a total of ∼6.5 million grains of Al 7075-T6 are examined. The calibrated analytical solution agrees well with the brute force Monte Carlo simulation results extracted from the CPFE simulations. Furthermore, the developed formulation can predict the MVD of FIPs for different sample sizes using a size-dependent parameter, and it is capable of accurately extrapolating the MVD of FIPs for much larger microstructure sample sizes than the size used for its calibration.

Crystal plasticity↗
NASA NTRS2014

NASA Numerical and Experimental Evaluation of UTRC Low Emissions Injector

Computational and experimental analyses of a PICS-Pilot-In-Can-Swirler technology injector, developed by United Technologies Research Center (UTRC) are presented. NASA has defined technology targets for near term (called "N+1", circa 2015), midterm ("N+2", circa 2020) and far term ("N+3", circa 2030) that specify realistic emissions and fuel efficiency goals for commercial aircraft. This injector has potential for application in an engine to meet the Pratt & Whitney N+3 supersonic cycle goals, or the subsonic N+2 engine cycle goals. Experimental methods were employed to investigate supersonic cruise points as well as select points of the subsonic cycle engine; cruise, approach, and idle with a slightly elevated inlet pressure. Experiments at NASA employed gas analysis and a suite of laser-based measurement techniques to characterize the combustor flow downstream from the PICS dump plane. Optical diagnostics employed for this work included Planar Laser-Induced Fluorescence of fuel for injector spray pattern and Spontaneous Raman Spectroscopy for relative species concentration of fuel and CO2. The work reported here used unheated (liquid) Jet-A fuel for all fuel circuits and cycle conditions. The initial tests performed by UTRC used vaporized Jet-A to simulate the expected supersonic cruise condition, which anticipated using fuel as a heat sink. Using the National Combustion Code a PICS-based combustor was modeled with liquid fuel at the supersonic cruise condition. All CFD models used a cubic non-linear k-epsilon turbulence wall functions model, and a semi-detailed Jet-A kinetic mechanism based on a surrogate fuel mixture. Two initial spray droplet size distribution and spray cone conditions were used: (1) an initial condition (Lefebvre) with an assumed Rosin-Rammler distribution, and 7 degree Solid Spray Cone; and (2) the Boundary Layer Stripping (BLS) primary atomization model giving the spray size distribution and directional properties. Contour and line plots are shown in comparison with experimental data (where this data is available) for flow velocities, fuel, and temperature distribution. The CFD results are consistent with experimental observations for fuel distribution and vaporization. Analysis of gas sample results, using a previously-developed NASA NOx correlation, indicates that for sea-level takeoff, the PICS configuration is predicted to deliver an EINOx value of about three for the targeted supersonic aircraft. Emissions results at supersonic cruise conditions show potential for meeting the NASA goals with liquid fuel.

Hicks, Yolanda R.↗
NASA NTRS2014

NASA Numerical and Experimental Evaluation of UTRC Low Emissions Injector

Computational and experimental analyses of a PICS-Pilot-In-Can-Swirler technology injector, developed by United Technologies Research Center (UTRC) are presented. NASA has defined technology targets for near term (called "N+1", circa 2015), midterm ("N+2", circa 2020) and far term ("N+3", circa 2030) that specify realistic emissions and fuel efficiency goals for commercial aircraft. This injector has potential for application in an engine to meet the Pratt & Whitney N+3 supersonic cycle goals, or the subsonic N+2 engine cycle goals. Experimental methods were employed to investigate supersonic cruise points as well as select points of the subsonic cycle engine; cruise, approach, and idle with a slightly elevated inlet pressure. Experiments at NASA employed gas analysis and a suite of laser-based measurement techniques to characterize the combustor flow downstream from the PICS dump plane. Optical diagnostics employed for this work included Planar Laser-Induced Fluorescence of fuel for injector spray pattern and Spontaneous Raman Spectroscopy for relative species concentration of fuel and CO2. The work reported here used unheated (liquid) Jet-A fuel for all fuel circuits and cycle conditions. The initial tests performed by UTRC used vaporized Jet-A to simulate the expected supersonic cruise condition, which anticipated using fuel as a heat sink. Using the National Combustion Code a PICS-based combustor was modeled with liquid fuel at the supersonic cruise condition. All CFD models used a cubic non-linear k-epsilon turbulence wall functions model, and a semi-detailed Jet-A kinetic mechanism based on a surrogate fuel mixture. Two initial spray droplet size distribution and spray cone conditions were used: 1) an initial condition (Lefebvre) with an assumed Rosin-Rammler distribution, and 7 degree Solid Spray Cone; and 2) the Boundary Layer Stripping (BLS) primary atomization model giving the spray size distribution and directional properties. Contour and line plots are shown in comparison with experimental data (where this data is available) for flow velocities, fuel, and temperature distribution. The CFD results are consistent with experimental observations for fuel distribution and vaporization. Analysis of gas sample results, using a previously-developed NASA NOx correlation, indicates that for sea-level takeoff, the PICS configuration is predicted to deliver an EINOx value of about 3 for the targeted supersonic aircraft. Emissions results at supersonic cruise conditions show potential for meeting the NASA goals with liquid fuel.

Hicks, Yolanda R.↗
DOE OSTI2020

Assessment of various tar and soot treatment methods and a priori analysis of the steady laminar flamelet model for use in coal combustion simulation

Soot formed in the fuel-rich regions of coal flames is known to have a significant impact on the radiative heat transfer characteristics of the flame. It is therefore important to account for the effect of soot radiation when simulating coal combustion systems. Several recent coal combustion studies have used acetylene as a surrogate for devolatilized tar and neglect soot formation altogether, while others have utilized empirical models for tar and soot or even chemical mechanisms which include tar and soot as species. In this work, we investigate several methods of treating coal-derived tar and soot and the impact each treatment has on predicted furnace temperature, particle mass loss, and other quantities of interest. Furthermore, we perform a priori comparisons of simulation results to calculations from a steady laminar flamelet model in order to assess the accuracy of a flamelet approach for each tar and soot treatment considered. Data used as a basis for the study is generated through One-Dimensional Turbulence (ODT) simulation of a coal combustion system in which high- fidelity models are used for gas phase chemistry and devolatilization, where char combustion is neglected.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Modeling DOE Standard Canister Configurations with Updated Surface Chemistry

Road-ready and final disposition packaging configurations for the advanced test reactor (ATR) fuel currently specifies storage within helium backfilled DOE sealed standard canisters. The aluminum cladding of the ATR fuel contains an oxyhydroxide layer of boehmite/bayerite that generates hydrogen when subjected to irradiation. Understanding the effect of this hydrogen buildup over time to important for long term storage considerations. Previous modeling efforts have built a coupled CFD-chemical model to simulate the temperature gas phase concentrations within the DOE sealed standard canisters. These models have coupled the temperature conditions to both the gas phase radiolysis chemistry and the surface chemistry associated with the oxyhydroxide layer. Continued experimental work has identified trends for aluminum surrogate samples with oxyhydroxide layers for tests done at higher dose rates. At low initial doses a fast generation rate of hydrogen occurs which starts to roll over to a lower generation rate as the total dose applied in increased. A previous study created a small-scale chemical model was built to replicate a mini-canister surrogate system at SRNL as well as for the smaller capsule tests performed at INL. The previous iteration of the model utilized step function for its G-values for the hydrogen generation over a 200-year period. This model makes an update to the surface chemistry to account for theorized surface chemistry reactions allowing for oxygen to remain bounded to the oxyhydroxide layer. As additional experiments continue, the kinetic fits may be adjusted to adapt to new data. Three canister configurations are modeled – the base 18-inch, 15-foot DOR standard canister with 30 ATR fuel elements, an 18-inch, 10-foot DOE standard canister with 32 ATR fuel elements and a 24-inch, 10-foot DOE standard canister with 40 ATR fuel elements. In previous reports, the primary sensitivity of the canister conditions was identified as the decay heat of the fuel and the dried condition of the fuel. Only these parameters are studied in the report. For the nominal case with dried fuel, the hydrogen generation is only about 2%, so it is even less than the flammability limit if it were exposed to oxygen. For the densely packed 18-inch case the total hydrogen concentration is only around 4% with the nominal decay heat. For the densely packed 24-inch case the total hydrogen concentration is only 2.5%, roughly 20% higher than the original packing design, and still under flammability conditions if exposed to oxygen.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
NASA NTRS2006

Beam Heating of Samples: Modeling and Verification

Energy absorbed from the X-ray beam by the sample requires cooling by forced convection (i.e. cryostream) to minimize temperature increase and the damage caused to the sample by the X-ray heating. In this presentation we will first review the current theoretical models and recent studies in the literature, which predict the sample temperature rise for a given set of beam parameters. It should be noted that a common weakness of these previous studies is that none of them provide actual experimental confirmation. This situation is now remedied in our investigation where the problem of x-ray sample heating is taken up once more. We have theoretically investigated, and at the same time, in addition to the numerical computations, performed experiments to validate the predictions. We have modeled, analyzed and experimentally tested the temperature rise of a 1 mm diameter glass sphere (sample surrogate) exposed to an intense synchrotron X-ray beam, while it is being cooled in a uniform flow of nitrogen gas. The heat transfer, including external convection and internal heat conduction was theoretically modeled using CFD to predict the temperature variation in the sphere during cooling and while it was subjected to an undulator (ID sector 19) X-ray beam at the APS. The surface temperature of the sphere during the X-ray beam heating was measured using the infrared camera measurement technique described in a previous talk. The temperatures from the numerical predictions and experimental measurements are compared and discussed. Additional results are reported for the two different sphere sizes and for two different supporting pin orientations.

Kazmierczak, Michael↗
DOE OSTI2025

Material Fracturing and Failure Simulation Datasets

Fracturing is a fundamental physics phenomena with broad relevance across multiple domains, ranging from infrastructure integrity, aerospace durability, reservoir production, and seismic events. We present a diverse dataset of simulated fracture evolution and material failure generated from two numerical solvers: the phase-field method and the combined finite-discrete element method (FDEM). These solvers differ in formulation, physical fidelity, and computational efficiency. The dataset includes five materials: PBX, anisotropic shale, tungsten, aluminum, and steel. For each, phase-field simulations span 400,000 cases: 200,000 under uniaxial tension and 200,000 under biaxial tension. The computationally expensive FDEM simulations include 90,000 split evenly among PBX, shale, and tungsten under uniaxial loading. All simulations begin with randomized initial fracture patterns. Each entry includes temporal data capturing fracture propagation dynamics. This comprehensive dataset is designed to support the development of foundational or surrogate machine learning approaches for predicting material failure. While no such models are introduced here, the dataset lays a robust foundation for advancing future research and innovation in these areas.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Accelerating discoveries at DIII-D with the Integrated Research Infrastructure

DIII-D research is being accelerated by leveraging high performance computing (HPC) and data resources available through the National Energy Research Scientific Computing Center (NERSC) Superfacility initiative. As part of this initiative, a high-resolution, fully automated, whole discharge kinetic equilibrium reconstruction workflow was developed that runs at the NERSC for most DIII-D shots in under 20 min. This has eliminated a long-standing research barrier and opened the door to more sophisticated analyses, including plasma transport and stability. These capabilities would benefit from being automated and executed within the larger Department of Energy Advanced Scientific Computing Research program’s Integrated Research Infrastructure (IRI) framework. The goal of IRI is to empower researchers to meld DOE’s world-class research tools, infrastructure, and user facilities seamlessly and securely in novel ways to radically accelerate discovery and innovation. For transport, we are looking at producing flux matched profiles and also using particle tracing to predict fast ion heat deposition from neutral beam injection before a shot takes place. Our starting point for evaluating plasma stability focuses on the pedestal limits that must be navigated to achieve better confinement. This information is meant to help operators run more effective experiments, so it needs to be available rapidly inside the DIII-D control room. So far this has been achieved by ensuring the data is available with existing tools, but as more novel results are produced new visualization tools must be developed. In addition, all of the high-quality data we have generated has been collected into databases that can unlock even deeper insights. This has already been leveraged for model and code validation studies as well as for developing AI/ML surrogates. The workflows developed for this project are intended to serve as prototypes that can be replicated on other experiments and can be run to provide timely and essential information for ITER, as well as next stage fusion power plants.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
NASA NTRS1999

Measuring Directional Wave Spectra and Wind Speed with a Scanning Radar Altimeter

The geometry for the NASA Scanning Radar Altimeter (SRA) is shown. It transmits a 8-ns duration pulse at Ka-band (8.3 mm) and measures time of flight as it scans a 1 degree (two-way) beam from left to right across the aircraft ground track. The most recent configuration determines the surface elevation at 64 points spaced at uniform angular intervals of about 0.7 across a swath whose width is about 0.8 times the aircraft altitude. The system generates these raster lines of the surface topography beneath the aircraft at about a 10 Hz rate. In postflight processing the SRA wave topographic data are transformed with a two-dimensional Fast Fourier Transformation (FFT) and Doppler corrected to produce directional wave spectra. The SRA is not absolutely calibrated in power, but by measuring the relative fall-off of backscatter with increasing incidence angle, the SRA can also determine the mean square slope (mss) of the sea surface, a surrogate for wind speed. For the slope-dependent specular point model of radar sea surface scattering, an expression approximated by a geometric optics form, for the relative variation with incidence angle of the normalized backscatter radar cross section would be sigma (sup 0) (sub rel) = sec (exp 4) theta exp (-tan squared theta/mss) where theta is the off-nadir incidence angle.

Walsh, E. J.↗
NASA NTRSDate not supplied

Structured Covariance Gaussian Networks for Orion Crew Module Aerodynamic Uncertainty Quantification

In this paper we propose a new approach for nonlinear regression and uncertainty quantification. The method is based on a pair of neural networks which parameterize mean and dense covariance functions of a multivariate Gaussian process, trained together to maximize the log-likelihood of observing the given data. The covariance matrix is made positive definite at every input by construction. We also propose a sampling approach that produces viable surrogate function realizations from the Gaussian process. We call the proposed model a Structured Covariance Gaussian Network (SCGN). We illustrate the use of SCGNs for learning an aerodynamic response surface with built-in uncertainty for the Orion crew module. We find that SCGN provides an efficient and systematic way to learn nonlinear functional relationships and dense covariances. We compare results to a baseline Gaussian process regressor and observe that the SCGN provides comparable uncertainty descriptions with improved scalability to dataset size. The sample functions generated by SCGN are fast to evaluate online and are therefore convenient for use in trajectory simulations. These results suggest that SCGN may be a viable computational method for aerodynamic uncertainty quantification.

machine learning↗
NASA NTRSDate not supplied

Structured Covariance Gaussian Networks for Orion Crew Module Aerodynamic Uncertainty Quantification

In this paper we propose a new approach for nonlinear regression and uncertainty quantification. The method is based on a pair of neural networks which parameterize mean and dense covariance functions of a multivariate Gaussian process, trained together to maximize the log-likelihood of observing the given data. The covariance matrix is made positive definite at every input by construction. We also propose a sampling approach that produces viable surrogate function realizations from the Gaussian process. We call the proposed model a Structured Covariance Gaussian Network (SCGN). We illustrate the use of SCGNs for learning an aerodynamic response surface with built-in uncertainty for the Orion crew module. We find that SCGN provides an efficient and systematic way to learn nonlinear functional relationships and dense covariances. We compare results to a baseline Gaussian process regressor and observe that the SCGN provides comparable uncertainty descriptions with improved scalability to dataset size. The sample functions generated by SCGN are fast to evaluate online and are therefore convenient for use in trajectory simulations. These results suggest that SCGN may be a viable computational method for aerodynamic uncertainty quantification.

machine learning↗
DOE OSTI2025

Ensemble cure kinetics network (ECK-Net): A method to derive cure kinetics of thermosetting resin

This paper introduces an Ensemble Cure Kinetics Network (ECK-Net), a neural network (NN)–based framework for modeling the cure kinetics of thermosetting resins within a phenomenological context. ECK-Net replaces traditional analytic models, which require extensive chemical insight and multiple isothermal/non-isothermal experiments, with a data-driven surrogate that maps nonlinear relationships between temperature, degree of cure, and reaction rate from differential scanning calorimetry data. The proposed approach predicts input-dependent kinetic coefficients of a generalized nth-order reaction equation rather than reaction rates directly, enabling a single unified model to represent various epoxy systems without relying on iso-conversional analysis or predefined functional forms. To ensure robustness, multiple independently trained networks under different random initializations are blended through an ensemble strategy, effectively mitigating the stochastic variability inherent to neural networks. The framework is validated using experimental datasets from multiple resin systems, including aerospace-grade materials (Toray 3900-2, Cycom 5320-1, and Hexcel 8552) and a windmill-grade resin (RIMR 035c). The model accurately reproduces the temporal evolution of the degree of cure under manufacturers’ recommended cure cycles across all tested resins systems, yielding Pearson’s correlation coefficients of 0.992, 0.994, 0.993, 0.997, respectively. To demonstrate process-level applicability, the trained network was implemented within the Abaqus environment to simulate out-of-autoclave (OOA) curing process of the CFRP panel composed of Toray T830H-6K/3900-2D prepreg. The simulation results showed excellent agreement with experimental temperature response (maximum peak temperature, simulation: 189.6 °C, experiment: 188.5 °C) and the final degree of cure (simulation: 0.948, experiment: 0.960 ± 0.013), confirming ECK-Net’s capability as a reliable alternative to conventional cure kinetics modeling methods.

Composite curing↗