Search NASA⌕ Search

SEARCH · Search NASA

Results for “SURROGATE 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.

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

At least 1,315 records · Page 73

Application of Machine Learning to Predict the Response of the Liquid Mercury Target at the Spallation Neutron Source

The Spallation Neutron Source (SNS) at Oak Ridge National Laboratory is currently the most powerful accelerator-driven neutron source in the world. The intense proton pulses strike on SNS’s mercury target to provide bright neutron beams, which also leads to severe fluid-structure interactions inside the target. Prediction of resultant loading on the target is difficult particularly when helium gas is intentionally injected into mercury to reduce the loading and mitigate the pitting damage on the target’s internal walls. Leveraging the power of machine learning and the measured target strain, we have developed machine learning surrogates for modeling the discrepancy between simulations and experimental strain data. We then employ these surrogates to guide the refinement of the high-fidelity mercury/helium mixture model to predict a better match of target strain response.

Lin, Lianshan↗

Coupled Lake‐Atmosphere‐Land Physics Uncertainties in a Great Lakes Regional Climate Model

Abstract This study develops a surrogate‐based method to assess the uncertainty within a convective permitting integrated modeling system of the Great Lakes region, arising from interacting physics parameterizations across the lake, atmosphere, and land surface. Perturbed physics ensembles of the model during the 2018 summer are used to train a neural network surrogate model to predict lake surface temperature (LST) and near‐surface air temperature (T2m). Average physics uncertainties are determined to be 1.5C for LST and T2m over land, and 1.9C for T2m over lake, but these have significant spatiotemporal variations. We find that atmospheric physics parameterizations alone are the dominant sources of uncertainty (45%–53%), while lake and land parameterizations account for 33% and 38% of the uncertainty of LST and T2m over land respectively. Interactions of atmosphere physics parameterizations with those of the land and lake contribute to an additional 13%–17% of the total variance. LST and T2m over the lake are more uncertain in the deeper northern lakes, particularly during the rapid warming phase that occurs in late spring/early summer. The LST uncertainty increases with sensitivity to the lake model's surface wind stress scheme. T2m over land is more uncertain over forested areas in the north, where it is most sensitive to the land surface model, than the more agricultural land in the south, where it is most sensitive to the atmospheric planetary boundary and surface layer scheme. Uncertainty also increases in the southwest during multiday temperature declines with higher sensitivity to the land surface model.

54 ENVIRONMENTAL SCIENCES↗

Embedding physics domain knowledge into a Bayesian network enables layer-by-layer process innovation for photovoltaics

Process optimization of photovoltaic devices is a time-intensive, trial-and-error endeavor, which lacks full transparency of the underlying physics and relies on user-imposed constraints that may or may not lead to a global optimum. Herein, we demonstrate that embedding physics domain knowledge into a Bayesian network enables an optimization approach for gallium arsenide (GaAs) solar cells that identifies the root cause(s) of underperformance with layer-by-layer resolution and reveals alternative optimal process windows beyond traditional black-box optimization. Our Bayesian network approach links a key GaAs process variable (growth temperature) to material descriptors (bulk and interface properties, e.g., bulk lifetime, doping, and surface recombination) and device performance parameters (e.g., cell efficiency). For this purpose, we combine a Bayesian inference framework with a neural network surrogate device-physics model that is 100× faster than numerical solvers. With the trained surrogate model and only a small number of experimental samples, our approach reduces significantly the time-consuming intervention and characterization required by the experimentalist. As a demonstration of our method, in only five metal organic chemical vapor depositions, we identify a superior growth temperature profile for the window, bulk, and back surface field layer of a GaAs solar cell, without any secondary measurements, and demonstrate a 6.5% relative AM1.5G efficiency improvement above traditional grid search methods.

14 SOLAR ENERGY↗

Material Model Parameters Optimization in Liquid Mercury Target Dynamics Simulation With Machine Learning Surrogates

A pulsed spallation target is subjected to very short (∼0.7μs) but intense loads (23.3 kJ) from repeated proton pulses, which knock away neutrons from the mercury atoms’ nuclei for a wide range application in physics, engineering, medicine, petroleum exploration, biology, chemistry, etc. The effect of this pulsed loading on the stainless-steel target module which contains the flowing mercury target material is difficult to predict not only due to its short but intense explosive-like physical reaction, but also the nonlinear material behavior of the liquid mercury in the structure. Injecting small helium bubbles in the mercury has been an efficient method of mitigating the pressure wave at high power level stage. However, prediction of the resultant loading on the target is more difficult when helium gas is intentionally injected into the mercury. A 2-phase material model that incorporates the Rayleigh-Plesset (R-P) model is expected to address this complex multi-physics dynamics problem by including the bubble dynamics in the liquid mercury. A parameter sensitivity study was firstly employed to understand their impact on the simulation strains. The investigated parameters included E, μ, γ, σ, n, VFgas, and gas cumulative volume curve control parameters a and b. Verification and validation results from sparse polynomial expansions (SPE) method and directional Gaussian smoothing (DGS) optimization show that the surrogate model had training error of ∼7% and validation error of ∼15%, indicating that machine learning methods and surrogate models can help optimize the uncertain parameters in the complex 2-phase material model. This approach is expected to fill the knowledge gap between unknown liquid-gas mixture material model and measured vessel strain responses.

Lin, Lianshan↗

SU-E-J-73: Generation of Volumetric Images with a Respiratory Motion Model Based On An External Surrogate Signal

Purpose: Respiratory motion during radiotherapy treatment can differ significantly from motion observed during imaging for treatment planning. Our goal is to use an initial 4DCT scan and the trace of an external surrogate marker to generate 3D images of patient anatomy during treatment. Methods: Deformable image registration is performed on images from an initial 4DCT scan. The deformation vectors are used to develop a patient-specific linear relationship between the motion of each voxel and the trajectory of an external surrogate signal. Correlations in motion are taken into account with principal component analysis, reducing the number of free parameters. This model is tested with digital phantoms reproducing the breathing patterns of ten measured patient tumor trajectories, using five seconds of data to develop the model and the subsequent thirty seconds to test its predictions. The model is also tested with a breathing physical anthropomorphic phantom programmed to reproduce a patient breathing pattern. Results: The error (mean absolute, 95th percentile) over 30 seconds in the predicted tumor centroid position ranged from (0.8, 1.3) mm to (2.2, 4.3) mm for the ten patient breathing patterns. The model reproduced changes in both phase and amplitude of the breathing pattern. Agreement between prediction and truth over the entire image was confirmed by assessing the global voxel intensity RMS error. In the physical phantom, the error in the tumor centroid position was less than 1 mm for all images. Conclusion: We are able to reconstruct 3D images of patient anatomy with a model correlating internal respiratory motion with motion of an external surrogate marker, reproducing the expected tumor centroid position with an average accuracy of 1.4 mm. The images generated by this model could be used to improve dose calculations for treatment planning and delivered dose estimates. This work was partially funded by a research grant from Varian Medical Systems.

60 APPLIED LIFE SCIENCES↗

Market-Integrated Optimization of Wind-Battery-Hydrogen Hybrids for Peaking Capacity via Storage

As Integrated Energy Systems (IES) combine multiple energy and storage technologies to provide potentially more value and less risk via resource diversification, complementary overbuild, increased flexibility, and revenue-stacking, IES value is dependent on electricity market dispatch and grid interactions should play an important role in IES design and operation. This study hybridizes and retrofits wind and combustion turbine plants to study the impacts of replacing gas generation capacity with wind, battery, PEM electrolysis, hydrogen tanks and hydrogen turbines. The optimized design is co-simulated in a production cost model with different bidding strategies in order to compare performance and highlight the importance of grid-interactions. We analyze the revenue and dispatch changes as well as the price and cost implications of wind-battery-hydrogen IESs.

electrolysis↗

Optimization of the generator coordinate method with machine-learning techniques for nuclear spectra and neutrinoless double- β decay: Ridge regression for nuclei with axial deformation

The generator coordinate method (GCM) is an important tool of choice for modeling large-amplitude collective motion in atomic nuclei. The computational complexity of the GCM increases rapidly with the number of collective coordinates. It imposes a strong restriction on the applicability of the method. In this work, we propose a subspace-reduction algorithm that employs optimal statistical ML models as surrogates for exact quantum-number projection calculations for norm and Hamiltonian kernels. The model space of the original GCM is reduced to a subspace relevant for nuclear low energy spectra and the NME of ground state to ground state 0νββ decay based on the orthogonality condition (OC) and the energy-transition-orthogonality procedure (ENTROP), respectively. For simplicity, the polynomial ridge regression (RR) algorithm is used to learn the norm and Hamiltonian kernels of axially deformed configurations. The efficiency and accuracy of this algorithm are illustrated for 76 Ge and 76 Se by comparing results obtained using the optimal RR models to direct GCM calculations. The low-lying energy spectra of 76 Ge and 76 Se, as well as the 0νββ-decay NME between their ground states, are computed. Furthermore, the results show that the performance of the GCM+OC/ENTROP+RR is more robust than that of the GCM+RR alone, and the former can reproduce the results of the original GCM calculation accurately with a significantly reduced computational cost.

59 ≤ A ≤ 89↗

Kinetic Modeling of Toluene Oxidation for Surrogate Fuel Applications

New environmental issues, like the effect of combustion-generated greenhouse gases, provide motivation to better characterize oxidation of hydrocarbons. Transportation, in particular, significantly contributes to energy consumption and CO{sub 2} emissions. Kinetic studies about the combustion of fuels under conditions typical of internal combustion engines provides important support to improve mechanism formulation and to eventually provide better computational tools that can be used to increase the engine performance. It is foreseeable that at least in the next 30 years the main transportation fuels will be either gasoline or diesel. Unfortunately, these fuels are very complex mixtures of many components. Moreover, their specifications and performance requirements significantly change the composition of these fuels: gasoline and diesel mixtures are different if coming from different refineries or they are different from winter to summer. At the same time a fuel with a well defined and reproducible composition is needed for both experimental and modeling work. In response to these issues, surrogate fuels are proposed. Surrogate fuels are defined as mixtures of a small number of hydrocarbons whose relative concentrations is adjusted in order to approximate the chemical and physical properties of a real fuel. Surrogate fuels are then very useful both for the design of reproducible experimental tests and also for the development of reliable kinetic models. The primary reference fuels (PRF) are a typical and old example of surrogate fuel: n-heptane and iso-octane mixtures are used to reproduce antiknock propensity of complex mixtures contained in a gasoline. PRFs are not able to surrogate gasoline in operating conditions different from standard ones and new surrogates have been recently proposed. Toluene is included in all of them as a species able to represent the behavior of aromatic compounds. On the other side, the toluene oxidation chemistry is not so well established and uncertainties still remain in the mechanism. This is especially true in the low temperature regime (< 850K). In these conditions, the toluene reactivity is too low to be conveniently investigated. Nonetheless, gasoline surrogates work in the engine at low temperatures, because of the presence of very reactive alkanes. The effect of these component interactions have to be taken into account. This work's aim is to present the model activity carried out by two different research groups, comparing the main pathways and results, matching data carried out in different devices both for pure toluene and mixtures. This is the starting point for a further activity to improve the two kinetic schemes.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Parametric model-order reduction for radiation transport using multi-resolution proper orthogonal decomposition

For parametric high-fidelity simulations, it is often desirable to utilize a reduced-order model (ROM) to emulate, at a reduced computational cost, parametric solutions of the governing partial differential equations (PDEs) for unseen parameter values. One commonly employed option is to utilize a data-driven, projection-based ROM supplemented with subspace identification via proper orthogonal decomposition (POD). POD discovers the ROM subspace by computing the singular value decomposition (SVD) of a set of training data from the full-order model (FOM). In streaming-dominated radiation transport simulations with localized sources, solutions often greatly vary over the spatial domain by many orders of magnitude. In such cases, machine-precision arithmetic can be insufficient to obtain an accurate SVD, resulting in a poorly performing ROM. We present a method called multiresolution POD (mrPOD) that mitigates these inaccuracies. The mrPOD method works by decomposing the spatial domain into regions and performing proper orthogonal decomposition on the training dataset separately in each region. In conclusion, mrPOD is tested on single energy group and multigroup atmospheric shielding transport problems and is shown to outperform classic POD.

42 ENGINEERING↗

Computational Bayesian Framework for Quantification and Reduction of Predictive Uncertainty in Subsurface Environmental Modeling

Subsurface environmental systems are open and complex, in which intricate hydrologic, microbiologic, and geochemical processes occur and interact at multiple scales. Understanding and predicting system responses to natural forces (e.g., climate changes) and human activities (e.g., contaminant remediation and CO 2 sequestration) is indispensable for managing water resources, cleaning subsurface contamination at many sites of the Department of Energy (DOE), and providing expert analysis to inform policy-making on long-term stewardship of nuclear waste disposal and CO 2 storage sites. Groundwater reactive transport modeling is necessary for understanding and predicting the complex subsurface environmental system. However, as complex systems are never completely predictable, model predictions of the subsurface system are inherently uncertain, and uncertainty is one of the greatest obstacles in simulating the subsurface environment. Quantifying predictive uncertainty of the complex system is critical to test model predictive power, support decision-making, and guide experimental design and data collection for uncertainty reduction. This project tackled two major obstacles to efficient and effective quantification and reduction of predictive uncertainty in simulating the complex subsurface environment. One obstacle is that existing methods of uncertainty quantification are too fragmented and incomplete for understanding and predicting the complex subsurface environment as a whole. The other obstacle is that existing methods of uncertainty quantification are computationally demanding and may not be applicable to groundwater reactive transport modeling that is always computationally expensive. The above two obstacles to uncertainty quantification directly affect the design of data collection schemes for uncertainty reduction. In this project, we developed a comprehensive Bayesian framework to remove the first obstacle. The second obstacle was tackled by using the sparse grid methods, which were originally developed and analyzed for numerical integration in the field of applied mathematics. We integrated the sparse grid methods into the Bayesian framework seamlessly, which led to a sparse-grid-based method of Bayesian uncertainty analysis. The developed computational Bayesian framework has been applied to a number of subsurface environmental problems through collaborating with scientists at the Pacific Northwest National Laboratory and Oak Ridge National Laboratory. The applications demonstrated that the computational Bayesian framework is general and compatible with any models and numerical codes of subsurface modeling. It is thus can be applied to gain insights for subsurface environmental modeling. This project helped the PI achieve his career goal to advance the state-of-the-art computational methods in uncertainty quantification for subsurface environmental modeling. This project supported two doctoral students, two master students, and one post-doc. The two doctoral students, after their graduation, worked as post-docs at the Pacific Northwest National Laboratory and the Oak Ridge National Laboratory. The project produced a total of 29 journal articles (four under review), one book chapter, and a number of conference proceedings and abstracts.

54 ENVIRONMENTAL SCIENCES↗

Hybrid Reynolds-Averaged/Large Eddy Simulation of the Flow in a Model SCRamjet Cavity Flameholder

Steady-state and scale-resolving simulations have been performed for flow in and around a model scramjet combustor flameholder. Experimental data available for this configuration include velocity statistics obtained from particle image velocimetry. Several turbulence models were used for the steady-state Reynolds-averaged simulations which included both linear and non-linear eddy viscosity models. The scale-resolving simulations used a hybrid Reynolds-averaged/large eddy simulation strategy that is designed to be a large eddy simulation everywhere except in the inner portion (log layer and below) of the boundary layer. Hence, this formulation can be regarded as a wall-modeled large eddy simulation. This e ort was undertaken to not only assess the performance of the hybrid Reynolds-averaged / large eddy simulation modeling approach in a flowfield of interest to the scramjet research community, but to also begin to understand how this capability can best be used to augment standard Reynolds-averaged simulations. The numerical errors were quantified for the steady-state simulations, and at least qualitatively assessed for the scale-resolving simulations prior to making any claims of predictive accuracy relative to the measurements. The steady-state Reynolds-averaged results displayed a high degree of variability when comparing the flameholder fuel distributions obtained from each turbulence model. This prompted the consideration of applying the higher-fidelity scale-resolving simulations as a surrogate "truth" model to calibrate the Reynolds-averaged closures in a non-reacting setting prior to their use for the combusting simulations. In general, the Reynolds-averaged velocity profile predictions at the lowest fueling level matched the particle imaging measurements almost as well as was observed for the non-reacting condition. However, the velocity field predictions proved to be more sensitive to the flameholder fueling rate than was indicated in the measurements.

Baurle, R. A.↗

WE-D-303-02: Applications of Volumetric Images Generated with a Respiratory Motion Model Based On An External Surrogate Signal

Purpose: Respiratory motion can vary significantly over the course of simulation and treatment. Our goal is to use volumetric images generated with a respiratory motion model to improve the definition of the internal target volume (ITV) and the estimate of delivered dose. Methods: Ten irregular patient breathing patterns spanning 35 seconds each were incorporated into a digital phantom. Ten images over the first five seconds of breathing were used to emulate a 4DCT scan, build the ITV, and generate a patient-specific respiratory motion model which correlated the measured trajectories of markers placed on the patients’ chests with the motion of the internal anatomy. This model was used to generate volumetric images over the subsequent thirty seconds of breathing. The increase in the ITV taking into account the full 35 seconds of breathing was assessed with ground-truth and model-generated images. For one patient, a treatment plan based on the initial ITV was created and the delivered dose was estimated using images from the first five seconds as well as ground-truth and model-generated images from the next 30 seconds. Results: The increase in the ITV ranged from 0.2 cc to 6.9 cc for the ten patients based on ground-truth information. The model predicted this increase in the ITV with an average error of 0.8 cc. The delivered dose to the tumor (D95) changed significantly from 57 Gy to 41 Gy when estimated using 5 seconds and 30 seconds, respectively. The model captured this effect, giving an estimated D95 of 44 Gy. Conclusion: A respiratory motion model generating volumetric images of the internal patient anatomy could be useful in estimating the increase in the ITV due to irregular breathing during simulation and in assessing delivered dose during treatment. This project was supported, in part, through a Master Research Agreement with Varian Medical Systems, Inc. and Radiological Society of North America Research Scholar Grant #RSCH1206.

60 APPLIED LIFE SCIENCES↗

An Edge-Cloud Integrated Solution for Buildings Demand Response Using Reinforcement Learning

Buildings, as major energy consumers, can provide great untapped demand response (DR) resources for grid services. However, their participation remains low in real-life. One major impediment for popularizing DR in buildings is the lack of cost-effective automation systems that can be widely adopted. Existing optimization-based smart building control algorithms suffer from high costs on both building-specific modeling and on-demand computing resources. To tackle these issues, this paper proposes a cost-effective edge-cloud integrated solution using reinforcement learning (RL). Beside RL’s ability to solve sequential optimal decision-making problems, its adaptability to easy-to-obtain building models and the off-line learning feature are likely to reduce the controller’s implementation cost. Using a surrogate building model learned automatically from building operation data, an RL agent learns an optimal control policy on cloud infrastructure, and the policy is then distributed to edge devices for execution. Simulation results demonstrate the control efficacy and the learning efficiency in buildings of different sizes. A preliminary cost analysis on a 4-zone commercial building shows the annual cost for optimal policy training is only 2.25% of the DR incentive received. Results of this study show a possible approach with higher return on investment for buildings to participate in DR programs.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Simulating the CMS High Granularity Calorimeter with ML

Detector simulation is a key component of physics analysis and related activities in CMS. In the upcoming High Luminosity LHC era, simulation will be required to use a smaller fraction of computing in order to satisfy resource constraints. At the same time, CMS will be upgraded with the new High Granularity Calorimeter (HGCal), which requires significantly more resources to simulate than the existing CMS calorimeters. This computing challenge motivates the use of generative machine learning models as surrogates to replace full physics-based simulation. We study the application of state-of-the-art diffusion models to simulate particle showers in the CMS HGCal. We will discuss methods to overcome the challenges posed by the high-dimensional, irregular geometry of the HGCal. The quality of the showers produced by the diffusion model will be assessed by comparison to the full GEANT4-based simulation. The increase in simulation throughput will be quantified and methods to accelerate the diffusion model inference will also be discussed.

Amram, Oz↗

Autoignition of trans -decalin, a diesel surrogate compound: Rapid compression machine experiments and chemical kinetic modeling

Decahydronaphthalene (decalin), with both cis and trans isomers, is a bicyclic alkane that is found in aviation fuels, diesel fuels, and alternative fuels from tar sands and oil shales. Between the two decalin isomers, trans-decalin has a lower cetane number, is energetically more stable, and has a lower boiling point. Moreover, trans-decalin has often been chosen as a surrogate component to represent two-ring naphthenes in transportation fuels. Recognizing the importance of understanding the chemical kinetics of trans-decalin in the development of surrogate models, an experimental and modeling study has been conducted. Experimentally, the autoignition characteristics of trans-decalin were investigated using a rapid compression machine (RCM) by using trans-decalin/O2/N2 mixtures at compressed pressures of PC=10–25 bar, low-to-intermediate compressed temperatures of TC=620–895 K, and varying equivalence ratios of Φ=0.5, 1.0, and 2.0. These new experimental data demonstrate the effects of pressure, fuel loading, and oxygen concentration on autoignition of trans-decalin. The current RCM data of trans-decalin at lower temperatures were also found to complement well with the literature shock tube data of decalin (mixture of cis+trans) at higher temperatures. Furthermore, a chemical kinetic model for the oxidation of trans-decalin has been developed with new reaction rates and pathways, including, for the first time, a fully-detailed representation of low-temperature chemical kinetics for transdecalin. This model shows good agreement with the overall ignition delay results of the current RCM experiments and the literature shock tube studies. Chemical kinetic analyses of the developed model were further conducted to help identify the fuel decomposition pathways and the reactions controlling the autoignition at varying conditions.

10 SYNTHETIC FUELS↗

Convergence of Probability Densities Using Approximate Models for Forward and Inverse Problems in Uncertainty Quantification

Here, we analyze the convergence of probability density functions utilizing approximate models for both forward and inverse problems. We consider the standard forward uncertainty quantification problem where an assumed probability density on parameters is propagated through the approximate model to produce a probability density, often called a push-forward probability density, on a set of quantities of interest (QoI). The inverse problem considered in this paper seeks to update an initial probability density assumed on model input parameters such that the subsequent push-forward of this updated density through the parameter-to-QoI map matches a given probability density on the QoI. We prove that the densities obtained from solving the forward and inverse problems, using approximate models, converge to the true densities as the approximate models converge to the true models. Numerical results are presented to demonstrate convergence rates of densities for sparse grid approximations of parameter-to-QoI maps and standard spatial and temporal discretizations of PDEs and ODEs.

97 MATHEMATICS AND COMPUTING↗

AI-accelerated protein-ligand docking for SARS-CoV-2 is 100-fold faster with no significant change in detection

Protein-ligand docking is a computational method for identifying drug leads. The method is capable of narrowing a vast library of compounds down to a tractable size for downstream simulation or experimental testing and is widely used in drug discovery. While there has been progress in accelerating scoring of compounds with artificial intelligence, few works have bridged these successes back to the virtual screening community in terms of utility and forward-looking development. We demonstrate the power of high-speed ML models by scoring 1 billion molecules in under a day (50 k predictions per GPU seconds). We showcase a workflow for docking utilizing surrogate AI-based models as a pre-filter to a standard docking workflow. Our workflow is ten times faster at screening a library of compounds than the standard technique, with an error rate less than 0.01% of detecting the underlying best scoring 0.1% of compounds. Our analysis of the speedup explains that another order of magnitude speedup must come from model accuracy rather than computing speed. In order to drive another order of magnitude of acceleration, we share a benchmark dataset consisting of 200 million 3D complex structures and 2D structure scores across a consistent set of 13 million “in-stock” molecules over 15 receptors, or binding sites, across the SARS-CoV-2 proteome. We believe this is strong evidence for the community to begin focusing on improving the accuracy of surrogate models to improve the ability to screen massive compound libraries 100 × or even 1000 × faster than current techniques and reduce missing top hits. The technique outlined aims to be a fast drop-in replacement for docking for screening billion-scale molecular libraries.

59 BASIC BIOLOGICAL SCIENCES↗