Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Surrogates”

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 1,495 records · Page 83
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Towards accelerating particle-resolved direct numerical simulation with neural operators

In this paper, we present our ongoing work aimed at accelerating a particle-resolved direct numerical simulation model designed to study aerosol–cloud–turbulence interactions. The dynamical model consists of two main components—a set of fluid dynamics equations for air velocity, temperature, and humidity, coupled with a set of equations for particle (i.e., cloud droplet) tracing. Rather than attempting to replace the original numerical solution method in its entirety with a machine learning (ML) method, we consider developing a hybrid approach. We exploit the potential of neural operator learning to yield fast and accurate surrogate models and, in this study, develop such surrogates for the velocity and vorticity fields. We discuss results from numerical experiments designed to assess the performance of ML architectures under consideration as well as their suitability for capturing the behavior of relevant dynamical systems.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

A data-driven framework for permeability prediction of natural porous rocks via microstructural characterization and pore-scale simulation

Understanding the microstructure–property relationships of porous media is of great practical significance, based on which macroscopic physical properties can be directly derived from measurable microstructural informatics. However, establishing reliable microstructure–property mappings in an explicit manner is difficult, due to the intricacy, stochasticity, and heterogeneity of porous microstructures. In this paper, a data-driven computational framework is presented to investigate the inherent microstructure–permeability linkage for natural porous rocks, where multiple techniques are integrated together, including microscopy imaging, stochastic reconstruction, microstructural characterization, pore-scale simulation, feature selection, and data-driven modeling. A large number of 3D digital rocks with a wide porosity range are acquired from microscopy imaging and stochastic reconstruction techniques. A broad variety of morphological descriptors are used to quantitatively characterize pore microstructures from different perspectives, and they compose the raw feature pool for feature selection. Here high-fidelity lattice Boltzmann simulations are conducted to resolve fluid flow passing through porous media, from which reliable permeability references are obtained. The optimal feature set that best represents permeability is identified through a performance-oriented feature selection process, upon which a cost-effective surrogate model is rapidly fitted to approximate the microstructure-permeability mapping via data-driven modeling. This surrogate model exhibits great advantages over empirical/analytical formulas in terms of prediction accuracy and generalization capacity, which can predict reliable permeability values spanning four orders of magnitude. Besides, feature selection also greatly enhances the interpretability of the data-driven prediction model, from which new insights into the mechanism of how microstructural characteristics determine intrinsic permeability are obtained.

58 GEOSCIENCES↗
DOE OSTI2024

Embedded symmetric positive semi-definite machine-learned elements for reduced-order modeling in finite-element simulations with application to threaded fasteners

Here, we present a machine-learning strategy for finite element analysis of solid mechanics wherein we replace complex portions of a computational domain with a data-driven surrogate. In the proposed strategy, we decompose a computational domain into an “outer” coarse-scale domain that we resolve using a finite element method (FEM) and an “inner” fine-scale domain. We then develop a machine-learned (ML) model for the impact of the inner domain on the outer domain. In essence, for solid mechanics, our machine-learned surrogate performs static condensation of the inner domain degrees of freedom. This is achieved by learning the map from displacements on the inner-outer domain interface boundary to forces contributed by the inner domain to the outer domain on the same interface boundary. We consider two such mappings, one that directly maps from displacements to forces without constraints, and one that maps from displacements to forces by virtue of learning a symmetric positive semi-definite (SPSD) stiffness matrix. We demonstrate, in a simplified setting, that learning an SPSD stiffness matrix results in a coarse-scale problem that is well-posed with a unique solution. We present numerical experiments on several exemplars, ranging from finite deformations of a cube to finite deformations with contact of a fastener-bushing geometry. We demonstrate that enforcing an SPSD stiffness matrix drastically improves the robustness and accuracy of FEM–ML coupled simulations, and that the resulting methods can accurately characterize out-of-sample loading configurations with significant speedups over the standard FEM simulations.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Data-driven wind turbine wake modeling via probabilistic machine learning

Wind farm design primarily depends on the variability of the wind turbine wake flows to the atmospheric wind conditions and the interaction between wakes. Physics-based models that capture the wake flow field with high-fidelity are computationally very expensive to perform layout optimization of wind farms, and, thus, data-driven reduced-order models can represent an efficient alternative for simulating wind farms. In this work, we use real-world light detection and ranging (LiDAR) measurements of wind-turbine wakes to construct predictive surrogate models using machine learning. Specifically, we first demonstrate the use of deep autoencoders to find a low-dimensional latent space that gives a computationally tractable approximation of the wake LiDAR measurements. Then, we learn the mapping between the parameter space and the (latent space) wake flow fields using a deep neural network. Additionally, we also demonstrate the use of a probabilistic machine learning technique, namely, Gaussian process modeling, to learn the parameter-space-latent-space mapping in addition to the epistemic and aleatoric uncertainty in the data. Finally, to cope with training large datasets, we demonstrate the use of variational Gaussian process models that provide a tractable alternative to the conventional Gaussian process models for large datasets. Furthermore, we introduce the use of active learning to adaptively build and improve a conventional Gaussian process model predictive capability. Overall, we find that our approach provides accurate approximations of the wind-turbine wake flow field that can be queried at an orders-of-magnitude cheaper cost than those generated with high-fidelity physics-based simulations.

Deep neural networks↗
DOE OSTI2020

Calibration of hyperelastic constitutive models: the role of boundary conditions, search algorithms, and experimental variability

The calibration of hyperelastic constitutive models of soft tissue and tissue surrogates is often treated as an exercise in curve-fitting to the average experimental response, and many of the complicating factors such as experimental boundary conditions and data variability are ignored. Here, we focus on three questions that arise in this area: the ramifications of ignoring the experimental boundary conditions, the use of local optimizers, and the role of data variability. Using data from a uniaxial extension experiment on a tissue surrogate, we study how these three factors affect the calibration of isotropic hyperelastic constitutive models. Our results show that even with the simplest of constitutive models, it is necessary to look beyond a "good fit" to the average.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2020

Effect and measurement of residual water in CaCl 2 intended for use as electrolyte in molten salt electrochemical processing

We report CaCl 2 has applications for electrochemical processing of nuclear materials. Thermal dehydration leads to formation of oxide ions, which are shown to react and cause precipitation of dissolved CeCl 3 that was selected as a surrogate for actinide chlorides. Thus, measurement of residual water in CaCl 2 is an essential capability. Thermogravimetric analysis (TGA) was shown to underestimate starting water concentration. Subsequent analysis of solid samples via acid–base titration and cyclic voltammetry (CV) of the molten salt yielded consistent values within 5% for residual water. Hydroxides were shown to be unstable, thus oxygen is retained as oxide ions. Thus, total water in a sample of CaCl 2 can be quantified by combining TGA with with either CV of molten salt or titration of salt samples. The importance of quantifying residual water in the salt was demonstrated by showing that cerium chloride (surrogate for actinide chlorides) will react with oxide ions in CaCl 2 to form insoluble oxides and oxychlorides.

38 RADIATION CHEMISTRY, RADIOCHEMISTRY, AND NUCLEA↗
DOE OSTI2021

Multifractal detrended fluctuation analysis of soil radon ( 222 Rn) and thoron ( 220 Rn) time series

In this paper, we present results of Multifractal Detrended Fluctuation Analysis (MF-DFA) of soil Radon ( 222 Rn) and Thoron ( 220 Rn) time series to examine the scaling and multifractal features. Data consists of 15,692 measurements taken each over forty-minute interval. RTM 1688-2, 222 Rn/ 220 Rn electronic device has been used for measurement purposes. Seasonal periodicities have been removed from the data and MF-DFA was employed on de-seasonalized data. Original 222 Rn and 220 Rn time series have been converted into surrogate and shuffled time series. MFDFA has also been applied to the surrogate and shuffled series to examine the multifractality nature. Results of the study show that $^{222}_{Orig}Rn$, $^{222}_{Surr}Rn$ and $^{220}_{Orig}Rn$, $^{220}_{Surr}Rn$ time series are longrange positively correlated and an increase or decrease in either of radionuclide concentrations will be followed by another increase or decrease of concentrations in future measurements. The $^{220}_{Shuff}Rn$ time series has an independent or short-range dependent structure called white noise and the shuffled time series exhibit a very weak multifractality.

38 RADIATION CHEMISTRY, RADIOCHEMISTRY, AND NUCLEA↗
DOE OSTI2024

Investigation of candidates for reactor produced radioactive materials in support of radiological training exercises

Bromine-82, Potassium-42 and Copper-64 have been successfully adopted as radioactive surrogates for outdoor large area contamination training. The goal of this project was to discover new materials that could supplement potassium bromide (KBr) and copper pellets in radiological dispersal device (RDD) training events to reduce the down time of the training fields and to broaden the toolbox of the RDD surrogate training event program at Idaho National Laboratory. Of the ten different materials investigated, sodium nitrite, gallium metal, and gallium oxide presented the greatest promise as potential materials to replace potassium bromide in RDD training events.

38 RADIATION CHEMISTRY, RADIOCHEMISTRY, AND NUCLEA↗
DOE OSTI2024

Modeling delayed thermal runaway in nitric acid-soaked cat litter mixed with radioactive waste

Thermal ignition of radioactive waste within a 55-gallon drum was simulated by using a pressure-dependent waste decomposition model (Hobbs et al. in Process Saf Environ Prot https://doi.org/10.1016/j.psep.2022.09.047, 2022) calibrated with data from full-scale drum experiments (Parker et al. in The thermolytic response of a surrogate RNS waste mixture at the drum scale. Los Alamos National Laboratory Report LA-UR-16-21760, 2016) and validated with experiments from multiple laboratories (Hobbs et al. in Thermal analysis of aged nitric acid-soaked kitty litter in TRU waste drums-23370.WM2023 Conference, Phoenix, AZ, 2023). The acceleration of nitric acid chemistry reacting with an organic cat litter leading to thermal ignition was likely triggered by a restricted vent in the drum. Here, we address whether the form of the rate equation in (Hobbs et al. in Process Saf Environ Prot https://doi.org/10.1016/j.psep.2022.09.047, 2022) is sufficient to extrapolate thermal ignition within aged drums of similar content that have been stored in Texas for over nine years by investigating four different reaction rate forms for waste decomposition. A critical reaction rate reduction analysis is performed on each of these models to determine if delayed thermal runaway within vented aged waste is possible after nine years. We found that a pressure-dependent first-order rate expression not only predicted the accidental ignition of the waste drum, but the form also matches multiple experiments from different laboratories. Even though the waste composition decreases over time, the model predicts that acceleration leading to thermal runaway is possible if the waste is confined, even after 9 years. In conclusion, waste containing oxidizers such as nitric acid should not be mixed with organic adsorbents, especially if the waste is confined.

cookoff↗
DOE OSTI2021

A systematic analysis of phase stability in refractory high entropy alloys utilizing linear and non-linear cluster expansion models

We report that obtaining surrogate models for various types of materials has long been a goal of computational material scientists and physicists, as they allow the calculation of thermodynamic quantities and properties of interest far more efficiently than DFT. Many surrogate models, including interatomic potentials generated by machine learning algorthims and cluster expansions, rely on incorporating many-body interactions such as three-body or four-body interactions. When utilizing machine learning algorithms for interatomic potentials, it is common to start with two-body interactions and systematically include three- and four-body clusters to improve the fit, stopping when including the additional clusters does not measurably change the predictive power of the model. In cluster expansions for FCC and related lattices, it is often enough to include mostly two-body interactions and a handful of three-body interactions. However, it remains to be seen whether those will be enough in the case of lattices with a lower packing fraction, such as BCC.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Incorporation of market signals for the optimal design of post combustion carbon capture systems

Recent studies have shown that fossil generators equipped with post-combustion carbon capture (PCC) systems are needed to reduce the cost of deep decarbonization. Such generators need to be flexible and responsive to grid conditions, particularly in a high variable renewable energy (VRE) environment. In this work, we evaluate the net present value (NPV) of retrofitting an existing natural gas combined cycle (NGCC) unit with a flexible PCC system while incorporating market signals from a high VRE grid. We use our industrial partner’s NGCC configuration as representative of existing NGCC units and Svante’s rapid-temperature swing adsorption (TSA) for PCC. Because of its ability to rapidly startup/shutdown and ramp-up/ramp-down, the chosen capture technology is very attractive for load-following operations. For a given set of market signals, we formulate a two-stage stochastic multi-period optimization problem, under the price-taker assumption, to simultaneously optimize the design of the capture system and operation of the entire plant. Rigorous models for the NGCC unit, PCC system, and compression system are developed using commercial process simulators and validated with either plant or vendor data. For computational tractability, we develop surrogate/reduced-order models for use in the optimization problem. The surrogate model for the NGCC plant is constructed by linearizing the rigorous dynamic model at 75% load, while data-driven nonlinear surrogate models for the capture and compression systems are constructed using simulation data from the rigorous models. The optimization problem, formulated as a mixed integer bilinear program, is implemented in the IDAES® integrated platform and solved to global optimality using Gurobi 9.5. Using this formulation, we determine the profitability of retrofitting an existing NGCC unit with the chosen capture system for multiple regions in the U.S. under two scenarios with different carbon prices. Importantly, the results show that the optimal decision strongly depends on the region and on the carbon price, thereby demonstrating the importance of the inclusion of market signals in the design process.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2021

Enhanced Data Efficiency Using Deep Neural Networks and Gaussian Processes for Aerodynamic Design Optimization

Adjoint-based optimization methods are attractive for aerodynamic shape design primarily due to their computational costs being independent of the dimensionality of the input space and their ability to generate high-fidelity gradients that can then be used in a gradient-based optimizer. This makes them very well suited for high-fidelity simulation based aerodynamic shape optimization of highly parametrized geometries such as aircraft wings. However, the development of adjoint-based solvers involve careful mathematical treatment and their implementation require detailed software development. Furthermore, they can become prohibitively expensive when multiple optimization problems are being solved, each requiring multiple restarts to circumvent local optima. In this work, we propose a machine learning enabled, surrogate-based framework that replaces the expensive adjoint solver, without compromising on predicting predictive accuracy. Specifically, we first train a deep neural network (DNN) from training data generated from evaluating the high-fidelity simulation model on a model-agnostic design of experiments on the geometry shape parameters. The optimum shape may then be computed by using a gradient-based optimizer coupled with the trained DNN. Subsequently, we also perform a gradient-free Bayesian optimization, where the trained DNN is used as the prior mean. We observe that the latter framework (DNN-BO) improves upon the DNN-only based optimization strategy for the same computational cost. Overall, this framework predicts the true optimum with very high accuracy, while requiring far fewer high-fidelity function calls compared to the adjoint-based method. Furthermore, we show that multiple optimization problems can be solved with the same machine learning model with high accuracy, to amortize the offline costs associated with constructing our models. Our methodology finds applications in the early stages of aerospace design. (C) 2021 Published by Elsevier Masson SAS.

Renganathan, S. Ashwin↗
DOE OSTI2012

Stepwise renal lineage differentiation of mouse embryonic stem cells tracing in vivo development

Highlights: Black-Right-Pointing-Pointer We induced renal lineages from mESCs by following the in vivo developmental cues. Black-Right-Pointing-Pointer We induced nephrogenic intermediate mesoderm by stepwise addition of factors. Black-Right-Pointing-Pointer We induced two types of renal progenitor cells by reciprocal conditioned media. Black-Right-Pointing-Pointer We propose the potential role of CD24 for the enrichment of renal lineage cells. -- Abstract: The in vitro derivation of renal lineage progenitor cells is essential for renal cell therapy and regeneration. Despite extensive studies in the past, a protocol for renal lineage induction from embryonic stem cells remains unestablished. In this study, we aimed to induce renal lineages from mouse embryonic stem cells (mESC) by following in vivo developmental stages, i.e., the induction of mesoderm (Stage I), intermediate mesoderm (Stage II) and renal lineages (Stage III). For stage I induction, in accordance with known signaling pathways involved in mesoderm development in vivo, i.e., Nodal, bone morphogenic proteins (BMPs) and Wnt, we found that the sequential addition of three factors, i.e., Activin-A (A), a surrogate for Nodal signaling, during days 0-2, A plus BMP-4 (4) during days 2-4, and A4 plus lithium (L), a surrogate for Wnt signaling, during days 4-6, was most effective to induce the mesodermal marker, Brachyury. For stage II induction, the addition of retinoic acid (R) in the continuous presence of A4L during days 6-8 was most effective to induce nephrogenic intermediate mesodermal markers, such as Pax2 and Lim1. Under this condition, more than 30% of cells were stained positive for Pax2, and there was a concomitant decrease in the expression of non-mesodermal markers. For stage III induction, in resemblance to the reciprocal induction between ureteric bud (UB) and metanephric mesenchyme (MM) during kidney development, we found that the exposure to conditioned media derived from UB and MM cells was effective in inducing MM and UB markers, respectively. We also observed the emergence and gradual increase of cell populations expressing progenitor cell marker CD24 from Stage I to Stage III. These CD24{sup +} cells correlated with higher levels of expression of Brachyury at stage I, Pax2 and Lim1 at stage II and MM markers, such as WT1 and Cadherin 11, after exposure to UB-conditioned media at stage III. In conclusion, our results show that stepwise induction by tracing in vivo developmental stages was effective to generate renal lineage progenitor cells from mESC, and CD24 may serve as a useful surface marker for renal lineage cells at stage II and MM cells at stage III.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2018

Restoring ZSM-5 performance for catalytic fast pyrolysis of biomass: Effect of regeneration temperature

Zeolite catalysts used for biomass catalytic fast pyrolysis (CFP) deactivate rapidly, similar to a fluidized catalytic cracking (FCC) catalyst used in refining. To operate effectively when there is rapid deactivation, biomass CFP can take place in a riser FCC-style reactor in which the catalyst has a short contact time (seconds) with reactants before it is regenerated. The regeneration, therefore, has two major needs for effective operation: 1) heat balance, since the heat required for the CFP reactions is brought into the reactor by the hot catalyst and 2) relatively short (minutes) regeneration to restore the catalyst activity to be near its initial state. In order to understand effective conditions to regenerate zeolites used for CFP, a series of experiments were performed to determine the effect of regeneration temperature on the activity of ZSM-5 (SiO 2 /Al 2 O 3 = 30). After use for pine pyrolysis vapor upgrading, the catalyst was oxidized in 4% O 2 at temperatures between 500-700 °C and reevaluated for the upgrading of pine pyrolysis vapors to assess the extent of regeneration. Additional testing was performed using ethylene aromatization as a surrogate reaction to probe regeneration efficiency. Regeneration experiments were performed for either a fixed length of time (20 min) or until there was no further CO 2 measured in the effluent gas. Results from the ethylene aromatization reactions were shown to serve as an excellent surrogate for CFP reactivity and indicated that the use of model compound studies can effectively be used to understand reaction and regeneration processes from biomass CFP. Both sets of results indicate that a spent ZSM-5 used for biomass CFP could be fully regenerated at 650 degrees C and 700 °C within 20 min, whereas regeneration temperatures of 550 °C and 600 °C required longer regeneration temperatures and in the case of regeneration at 500 °C, there may be coke species that are not removed and the catalyst activity may never be fully restored. Characterization by pyridine diffuse reflectance infrared spectroscopy, thermogravimetric analysis coupled with infrared spectroscopy, and N 2 physisorption showed that higher regeneration temperatures are more effective for restoring Bronsted acid sites and catalyst mesoporosity by rapidly removing aromatic coke deposits. In conclusion, regeneration at 650 °C and 700 °C led to a slightly higher total porosity as compared to the pristine catalyst, which was attributed to the formation of additional mesoporosity from catalyst steaming.

09 BIOMASS FUELS↗
DOE OSTI2015

A thermodynamic and kinetic model for paste–aggregate interactions and the alkali–silica reaction

A new conceptual model is developed for ASR formation based on geochemical principles tied to aqueous speciation, silica solubility, kinetically controlled mineral dissolution, and diffusion. ASR development is driven largely by pH and silica gradients that establish geochemical microenvironments between paste and aggregate, with gradients the strongest within the aggregate adjacent to the paste boundary (i.e., where ASR initially forms). Super-saturation of magadiite and okenite (crystalline ASR surrogates) occurs in the zone defined by gradients in pH, dissolved silica, Na{sup +}, and Ca{sup 2} {sup +}. This model provides a thermodynamic rather than kinetic explanation of why quartz generally behaves differently from amorphous silica: quartz solubility does not produce sufficiently high concentrations of H{sub 4}SiO{sub 4} to super-saturate magadiite, whereas amorphous silica does. The model also explains why pozzolans do not generate ASR: their fine-grained character precludes formation of chemical gradients. Finally, these gradients have interesting implications beyond the development of ASR, creating unique biogeochemical environments.

36 MATERIALS SCIENCE↗
DOE OSTI2019

A reduced basis approach for PDEs on parametrized geometries based on the shifted boundary finite element method and application to a Stokes flow

Here we propose a model order reduction technique integrating the Shifted Boundary Method (SBM) with a POD-Galerkin strategy. This approach allows to deal with complex parametrized domains in an efficient and straightforward way. The impact of the proposed approach is threefold. First, problems involving parametrizations of complex geometrical shapes and/or large domain deformations can be efficiently solved at full-order by means of the SBM. This unfitted boundary method permits to avoid remeshing and the tedious handling of cut cells by introducing an approximate surrogate boundary. Second, the computational effort is reduced by the development of a Reduced Order Model (ROM) technique based on a POD-Galerkin approach. Third, the SBM provides a smooth mapping from the true to the surrogate domain, and for this reason, the stability and performance of the reduced order basis are enhanced. This feature is the net result of the combination of the proposed ROM approach and the SBM. Similarly, the combination of the SBM with a projection-based ROM gives the great advantage of an easy and fast to implement algorithm considering geometrical parametrization with large deformations. The transformation of each geometry to a reference geometry (morphing) is in fact not required. These combined advantages will allow the solution of PDE problems more efficiently. We illustrate the performance of this approach on a number of two-dimensional Stokes flow problems.

42 ENGINEERING↗
DOE OSTI2020

Data-driven learning of nonlocal physics from high-fidelity synthetic data

A key challenge to nonlocal models is the analytical complexity of deriving them from first principles, and frequently their use is justified a posteriori. Here, we extract nonlocal models from data, circumventing these challenges and providing data-driven justification for the resulting model form. Extracting data-driven surrogates is a major challenge for machine learning (ML) approaches, due to nonlinearities and lack of convexity — it is particularly challenging to extract surrogates which are provably well-posed and numerically stable. Our scheme not only yields a convex optimization problem, but also allows extraction of nonlocal models whose kernels may be partially negative while maintaining well-posedness even in small-data regimes. To achieve this, based on established nonlocal theory, we embed in our algorithm sufficient conditions on the non-positive part of the kernel that guarantee well-posedness of the learnt operator. These conditions are imposed as inequality constraints to meet the requisite conditions of the nonlocal theory. We demonstrate this workflow for a range of applications, including reproduction of manufactured nonlocal kernels; numerical homogenization of Darcy flow associated with a heterogeneous periodic microstructure; nonlocal approximation to high-order local transport phenomena; and approximation of globally supported fractional diffusion operators by truncated kernels.

42 ENGINEERING↗
DOE OSTI2023

Context-aware learning of hierarchies of low-fidelity models for multi-fidelity uncertainty quantification

Multi-fidelity Monte Carlo methods leverage low-fidelity and surrogate models for variance reduction to make tractable uncertainty quantification even when numerically simulating the physical systems of interest with high-fidelity models is computationally expensive. This work proposes a context-aware multi-fidelity Monte Carlo method that optimally balances the costs of training low-fidelity models with the costs of Monte Carlo sampling. It generalizes the previously developed context-aware bi-fidelity Monte Carlo method to hierarchies of multiple models and to more general types of low-fidelity models. When training low-fidelity models, the proposed approach takes into account the context in which the learned low-fidelity models will be used, namely for variance reduction in Monte Carlo estimation, which allows it to find optimal trade-offs between training and sampling to minimize upper bounds of the mean-squared errors of the estimators for given computational budgets. This is in stark contrast to traditional surrogate modeling and model reduction techniques that construct low-fidelity models with the primary goal of approximating well the high-fidelity model outputs and typically ignore the context in which the learned models will be used in upstream tasks. Further, the proposed context-aware multi-fidelity Monte Carlo method applies to hierarchies of a wide range of types of low-fidelity models such as sparse-grid and deep-network models. Numerical experiments with the gyrokinetic simulation code Gene show speedups of up to two orders of magnitude compared to standard estimators when quantifying uncertainties in small-scale fluctuations in confined plasma in fusion reactors. This corresponds to a runtime reduction from 72 days to four hours on one node of the Lonestar6 supercomputer at the Texas Advanced Computing Center.

42 ENGINEERING↗