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 1,603 records · Page 89

A two-phase three-field modeling framework for heat pipe application in nuclear reactors

Heat pipes and two-phase thermosyphons are highly efficient heat transfer devices utilizing continuous evaporation and condensation of working fluid for two-phase heat transport in closed systems. Because of the nearly isothermal and fully passive phase-change heat transfer mechanism, heat pipes and thermosyphons have found many applications in nuclear engineering, space technologies, and other energy systems. High-temperature heat pipes are used in nuclear microreactors to remove fission power from the primary system and are coupled with power conversion systems or process heat applications. Modeling of the two-phase flow phenomena inside a heat pipe is essential to its design and safety analysis. In this study, a comprehensive one-dimensional two-phase three-field flow model has been developed for the analysis of heat pipes in normal operation conditions and transients. The conservation or field equations of mass, momentum, and energy were developed for the liquid film, vapor, and droplet. In addition, constitutive models or correlations were reviewed thoroughly and provided for the closure of the three-field equations. Specific constitutive equations regarding interfacial mass and heat transfer at two interfaces, namely film-gas interface and gas-droplet interface, were reviewed for droplet entrainment and deposition rates as well as film and droplet evaporation rates. Furthermore, mechanistic correlations of annular flow film thickness were recommended for the modeling of the thermosyphons without a wick as a critical constitutive correlation. Furthermore, experimental data needs from new experiments using a prototype working fluid or surrogate fluids for the model validation of high-temperature heat pipes in microreactors were recommended for future research.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Differentiable Multiphysics Codes: A Breakthrough Technology for Simulation and Computing

This document summarizes the findings of a strategic planning exercise commissioned by the Weapons Simulation and Computing, Computational Physics (WSC/CP) program at the Lawrence Livermore National Laboratory (LLNL) in FY24. During the year, the committee met with multiple stakeholder communities to gather input, opinions, suggestions and concerns which have been incorporated throughout this document. The key findings from this exercise are summarized: • The development of multiphysics modelling and simulation (mod/sim) codes and software technologies, their deployment on exascale compute platforms, and their broad adoption across the NNSA is a major success of the Advanced Simulation and Computing (ASC) program and the Exascale Computing Project (ECP). Sustained investment in these core technologies is essential. • Today’s state of the art involves running ensembles of O(100K) simulations to perform uncertainty quantification (UQ) and design studies using multiple statistical methods such as Bayesian optimization to understand sensitivities of our models and explore parameterized design spaces. Even with exascale computing, we are practically limited to O(10) parameters in these studies since the number of simulations required to sample the space scales exponentially with the number of design parameters. • The data from these simulation ensembles is increasingly being used to train machine learned (ML) surrogates (or reduced order models, ROMs) which can then be used for optimization or real time design exploration. However, the trained surrogates are still limited in the number of parameters they can represent due to the sampling limitations previously noted. • Augmenting our suite of integrated multiphysics simulation codes, both current and emerging, with the ability to compute gradients (solution derivatives) of arbitrary simulation outputs with respect to (some or all) simulation inputs would be a breakthrough technology, opening the door to a new era of efficient and automated inverse design based on verified and validated mod/sim capabilities. • This capability, which we refer to as differentiable multiphysics codes (DMCs), would revolutionize both UQ and optimization studies by breaking the curse of dimensionality that presently limits our “gradient-free” ensemble based computing approach. A similar breakthrough occurred in the AI/ML community once the ability to compute gradients of arbitrary loss functions using back-propagation became commonplace. Gradient information from the multiphysics codes can also be used to dramatically improve the efficiency and scale of training of ML/ROM surrogates for rapid assessments. • Achieving this in our suite of codes will be a grand challenge, similar to the amount of effort that was required to transition from CPU to GPU computing. It will require buy-in from the entire WSC/CP program and beyond, including all integrated codes, physics and engineering models, third-party library dependencies and performance portability abstractions. It will also require investment in research and development of numerical methods for computing adjoints of coupled physics across multiple adaptively refined moving meshes and of stochastic (Monte Carlo) and mesh free (SPH) methods. • New software and numerical techniques, largely pioneered by the AI/ML community, make this feasible. Chief among these is automatic differentiation (AD), the ability to employ AD at point-wise locations in a physics calculation (instead of traditional black-box approaches) and the ability to perform “back-propagation in time” (or reverse mode AD) for non-linear partial differential equations (PDEs). Fundamentally, the conclusion of this strategic planning exercise is that the time is right to undertake a large scale effort in WSC, centered on the existing integrated codes, to continue the natural evolution of mod/sim in the age of AI/ML. Instead of attempting to replace mod/sim with purely data driven AI/ML models, we believe the key to success is to integrate AI/ML by building on top of the decades of hard-won knowledge and the verified/validated multiphysics modelling capability that is the hallmark of the ASC program.

97 MATHEMATICS AND COMPUTING↗

PyApprox: Enabling efficient model analysis

PyApprox is a Python-based one-stop-shop for probabilistic analysis of scientific numerical models. Easy to use and extendable tools are provided for constructing surrogates, sensitivity analysis, Bayesian inference, experimental design, and forward uncertainty quantification. The algorithms implemented represent the most popular methods for model analysis developed over the past two decades, including recent advances in multi-fidelity approaches that use multiple model discretizations and/or simplified physics to significantly reduce the computational cost of various types of analyses. Simple interfaces are provided for the most commonly-used algorithms to limit a user’s need to tune the various hyper-parameters of each algorithm. However, more advanced work flows that require customization of hyper-parameters is also supported. An extensive set of Benchmarks from the literature is also provided to facilitate the easy comparison of different algorithms for a wide range of model analyses. This paper introduces PyApprox and its various features, and presents results demonstrating the utility of PyApprox on a benchmark problem modeling the advection of a tracer in ground water.

97 MATHEMATICS AND COMPUTING↗

Optimization with Neural Network Feasibility Surrogates: Formulations and Application to Security-Constrained Optimal Power Flow

In many areas of constrained optimization, representing all possible constraints that give rise to an accurate feasible region can be difficult and computationally prohibitive for online use. Satisfying feasibility constraints becomes more challenging in high-dimensional, non-convex regimes which are common in engineering applications. A prominent example that is explored in the manuscript is the security-constrained optimal power flow (SCOPF) problem, which minimizes power generation costs, while enforcing system feasibility under contingency failures in the transmission network. In its full form, this problem has been modeled as a nonlinear two-stage stochastic programming problem. In this work, we propose a hybrid structure that incorporates and takes advantage of both a high-fidelity physical model and fast machine learning surrogates. Neural network (NN) models have been shown to classify highly non-linear functions and can be trained offline but require large training sets. In this work, we present how model-guided sampling can efficiently create datasets that are highly informative to a NN classifier for non-convex functions. We show how the resultant NN surrogates can be integrated into a non-linear program as smooth, continuous functions to simultaneously optimize the objective function and enforce feasibility using existing non-linear solvers. Overall, this allows us to optimize instances of the SCOPF problem with an order of magnitude CPU improvement over existing methods.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Weighted greedy-optimal design of computer experiments for kernel-based and Gaussian process model emulation and calibration

This article is concerned with the approximation of high-dimensional functions by kernel-based methods. Motivated by uncertainty quantification, which often necessitates the construction of approximations that are accurate with respect to a probability density function of random variables, we aim at minimizing the approximation error with respect to a weighted $L^p$-norm. We present a greedy procedure for designing computer experiments based upon a weighted modification of the pivoted Cholesky factorization. The method successively generates nested samples with the goal of minimizing error in regions of high probability. Numerical experiments validate that this new importance sampling strategy is superior to other sampling approaches, especially when used with non-product probability density functions. We also show how to use the proposed algorithm to efficiently generate surrogates for inferring unknown model parameters from data.

97 MATHEMATICS AND COMPUTING↗

Bayesian calibration of bubble size dynamics applied to CO 2 gas fermenters

To accelerate the scale-up of gaseous CO 2 fermentation reactors, computational models need to predict gas-to-liquid mass transfer which requires capturing the bubble size dynamics, i.e. bubble breakup and coalescence. However, the applicability of existing models beyond air–water mixtures remains to be established. Here, an inverse modeling approach, accelerated with a neural network surrogate, calibrates the breakup and coalescence closure models, that are used in class methods for population balance modeling (PBM). The calibration is performed based on experimental results obtained in a CO 2 -air–water-coflowing bubble column reactor. Bayesian inference is used to account for noise in the experimental dataset and bias in the simulation results. To accurately capture gas holdup and interphase mass transfer, the results show that the breakage rate needs to be increased by one order of magnitude. In conclusion, the inferred model parameters are then used on a separate configuration and shown to also improve bubble size distribution predictions.

09 BIOMASS FUELS↗

Modeling of CBM production, CO2 injection, and tracer movement at a field CO2 sequestration site

Sequestration of carbon dioxide in unmineable coal seams is a potential technology mainly because of the potential for simultaneous enhanced coalbed methane production (ECBM). Several pilot tests have been performed around the globe leading to mixed results. Numerous modeling efforts have been carried out successfully to model methane production and carbon dioxide (CO{sub 2}) injection. Sensitivity analyses and history matching along with several optimization tools were used to estimate reservoir properties and to investigate reservoir performance. Geological and geophysical techniques have also been used to characterize field sequestration sites and to inspect reservoir heterogeneity. The fate and movement of injected CO{sub 2} can be determined by using several monitoring techniques. Monitoring of perfluorocarbon (PFC) tracers is one of these monitoring technologies. As a part of this monitoring technique, a small fraction of a traceable fluid is added to the injection wellhead along with the CO{sub 2} stream at different times to monitor the timing and location of the breakthrough in nearby monitoring wells or offset production wells. A reservoir modeling study was performed to simulate a pilot sequestration site located in the San Juan coal basin of northern New Mexico. Several unknown reservoir properties at the field site were estimated by modeling the coal seam as a dual porosity formation and by history matching the methane production and CO{sub 2} injection. In addition to reservoir modeling of methane production and CO{sub 2} injection, tracer injection was modeled. Tracers serve as a surrogate for determining potential leakage of CO{sub 2}. The tracer was modeled as a non-reactive gas and was injected into the reservoir as a mixture along with CO{sub 2}. Geologic and geometric details of the field site, numerical modeling details of methane production, CO{sub 2} injection, and tracer injection are presented in this paper. Moreover, the numerical predictions of the tracer arrival times were compared with the measured field data. Results show that tracer modeling is useful in investigating movement of injected CO{sub 2} into the coal seam at the field site. Also, such new modeling techniques can be utilized to determine potential leakage pathways, and to investigate reservoir anisotropy and heterogeneity.

Reservoir modeling↗

A unified non-equilibrium phase change model for injection flow modeling

The homogenous relaxation model (HRM) is one of the most widely used models to describe the liquid- gas phase transition. However, in its original formulation, it is unable to handle multispecies vapor-liquid equilibrium (VLE), which limits its applicability to single-component fluids. In this work, a unified non-equilibrium phase change model that considers the VLE of multicomponent mixtures is proposed building upon the HRM's structure. A time factor is introduced to mimic the effect of different phase change timescales due to different mechanisms, e.g., cavitation, flash-boiling, and evaporation. Here to assess the model's performance, computational fluid dynamics simulations of the internal and near-nozzle injection flow of the Engine Combustion Network's Spray G injector were performed using the nine-component PACE-20 fuel with both the unified model and the original HRM. The predicted fuel density in the near-nozzle region matched well with X-ray tomography measurements. The simulation results indicated that, whereas the HRM failed to capture the vaporization due to convective mixing between the fuel and ambient gas, the unified model performed well in predicting the mixing-driven vaporization and the corresponding evaporative cooling. Further comparisons using the nine-component fuel formula and a single-component fuel surrogate demonstrated the unified model's ability to predict preferential vaporization, which affects the predictions of local mixture composition and rate of vaporization. Finally, it is shown that the unified model is capable of representing multiple phase change mechanisms, and the relaxation time factor plays an important role in determining the degree of phase change due to the different mechanisms.

33 ADVANCED PROPULSION SYSTEMS↗

Survey of Multifidelity Methods in Uncertainty Propagation, Inference, and Optimization

In many situations across computational science and engineering, multiple computational models are available that describe a system of interest. These different models have varying evaluation costs and varying fidelities. Typically, a computationally expensive high-fidelity model describes the system with the accuracy required by the current application at hand, while lower-fidelity models are less accurate but computationally cheaper than the high-fidelity model. Outer-loop applications, such as optimization, inference, and uncertainty quantification, require multiple model evaluations at many different inputs, which often leads to computational demands that exceed available resources if only the high-fidelity model is used. This work surveys multifidelity methods that accelerate the solution of outer-loop applications by combining high-fidelity and low-fidelity model evaluations, where the low-fidelity evaluations arise from an explicit low-fidelity model (e.g., a simplified physics approximation, a reduced model, a data-fit surrogate) that approximates the same output quantity as the high-fidelity model. The overall premise of these multifidelity methods is that low-fidelity models are leveraged for speedup while the high-fidelity model is kept in the loop to establish accuracy and/or convergence guarantees. We categorize multifidelity methods according to three classes of strategies: adaptation, fusion, and filtering. The paper reviews multifidelity methods in the outer-loop contexts of uncertainty propagation, inference, and optimization.

97 MATHEMATICS AND COMPUTING↗

Inferring Positions of Tumor and Nodes in Stage III Lung Cancer From Multiple Anatomical Surrogates Using Four-Dimensional Computed Tomography

Purpose: To investigate the feasibility of modeling Stage III lung cancer tumor and node positions from anatomical surrogates. Methods and Materials: To localize their centroids, the primary tumor and lymph nodes from 16 Stage III lung cancer patients were contoured in 10 equal-phase planning four-dimensional (4D) computed tomography (CT) image sets. The centroids of anatomical respiratory surrogates (carina, xyphoid, nipples, mid-sternum) in each image set were also localized. The correlations between target and surrogate positions were determined, and ordinary least-squares (OLS) and partial least-squares (PLS) regression models based on a subset of respiratory phases (three to eight randomly selected) were created to predict the target positions in the remaining images. The three-phase image sets that provided the best predictive information were used to create models based on either the carina alone or all surrogates. Results: The surrogate most correlated with target motion varied widely. Depending on the number of phases used to build the models, mean OLS and PLS errors were 1.0 to 1.4 mm and 0.8 to 1.0 mm, respectively. Models trained on the 0%, 40%, and 80% respiration phases had mean ({+-} standard deviation) PLS errors of 0.8 {+-} 0.5 mm and 1.1 {+-} 1.1 mm for models based on all surrogates and carina alone, respectively. For target coordinates with motion >5 mm, the mean three-phase PLS error based on all surrogates was 1.1 mm. Conclusions: Our results establish the feasibility of inferring primary tumor and nodal motion from anatomical surrogates in 4D CT scans of Stage III lung cancer. Using inferential modeling to decrease the processing time of 4D CT scans may facilitate incorporation of patient-specific treatment margins.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Unified Language Frontend for Physic-Informed AI/ML

Artificial intelligence and machine learning (AI/ML) are becoming important tools for scientific modeling and simulation as in several other fields such as image analysis and natural language processing. ML techniques can leverage the computing power available in modern systems and reduce the human effort needed to configure experiments, interpret and visualize results, draw conclusions from huge quantities of raw data, and build surrogates for physics based models. Domain scientists in fields like fluid dynamics, microelectronics and chemistry can automate many of their most difficult and repetitive tasks or improve the design times by use of the faster ML-surrogates. However, modern ML and traditional scientific highperformance computing (HPC) tend to use completely different software ecosystems. While ML frameworks like PyTorch and TensorFlow provide Python APIs, most HPC applications and libraries are written in C++. Direct interoperability between the two languages is possible but is tedious and error-prone. In this work, we show that a compiler-based approach can bridge the gap between ML frameworks and scientific software with less developer effort and better efficiency. We use the MLIR (multi-level intermediate representation) ecosystem to compile a pre-trained convolutional neural network (CNN) in PyTorch to freestanding C++ source code in the Kokkos programming model. Kokkos is a programming model widely used in HPC to write portable, shared-memory parallel code that can natively target a variety of CPU and GPU architectures. Our compiler-generated source code can be directly integrated into any Kokkosbased application with no dependencies on Python or cross-language interfaces.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Final Report, DOE Early Career Award: Predictive modeling of complex physical systems: new tools for statistical inference, uncertainty quantification, and experimental design

Predictive simulation of complex physical systems increasingly rests on the interplay of experimental observations with computational models. Key inputs, parameters, or structural aspects of models may be incomplete or unknown, and must be developed from indirect and limited observations. At the same time, quantified uncertainties are needed to qualify computational predictions in the support of design and decision-making. In this context, Bayesian statistics provides a foundation for inference from noisy and limited data, but at prohibitive computional expense. This project intends to make rigorous predictive modeling *feasible* in complex physical systems, via accelerated and scalable tools for uncertainty quantification, Bayesian inference, and experimental design. Specific objectives are as follows: 1. Develop adaptive posterior approximations and dimensionality reduction approaches for Bayesian inference in high-dimensional nonlinear systems. 2. Extend accelerated Bayesian methodologies to large-scale {\em sequential} data assimilation, fully treating nonlinear models and non-Gaussian state and parameter distributions. 3. Devise efficient surrogate-based methods for Bayesian model selection and the learning of model structure. 4. Develop scalable simulation/optimization approaches to nonlinear Bayesian experimental design, for both parameter inference and model selection. 5. Demonstrate these inferential tools on chemical kinetic models in reacting flow, constructing and refining thermochemical and electrochemical models from limited data. Demonstrate Bayesian filtering on canonical stochastic PDEs and in the dynamic estimation of inhomogeneous subsurface properties and flow fields.

97 MATHEMATICS AND COMPUTING↗

Experimental and kinetic modeling study of tetralin: A naphtheno-aromatic fuel for gasoline, jet and diesel surrogates

Distillate fuels contain significant proportions of naphtheno-aromatic components and tetralin is a suitable surrogate component to represent this molecular moiety. The presence of aromatic and naphthyl rings makes kinetic modeling of tetralin very challenging. Primary radicals formed during the oxidation of tetralin can be aryl, benzylic or paraffinic in nature. Using available information on reaction paths and rate constants of naphthenes and alkyl-aromatics, in this study a kinetic model of tetralin has been developed with emphasis on low-temperature chemistry and high-pressure conditions. Due to the lack of high-level quantum chemical calculations on reaction pathways of tetralin, analogous rates from ab-initio studies on benzylic and paraffinic radicals have been adopted here. Some modifications to the reaction rate rules are incorporated to account for the unique characteristics of tetralin's molecular structure. Important reaction channels have been identified using reaction path and brute force sensitivity analyses. In order to investigate the model performance at low temperatures, new experiments are carried out in a rapid compression machine on blends of tetralin and 3-methylpentane. Blending of low-reactivity tetralin with a high-reactivity alkane allowed the investigation of tetralin ignition at very low temperatures (665 – 856 K). The kinetic model developed in the current study is found to predict the current experiments and literature data adequately. The new model will aid in high-fidelity surrogate predictions at engine-relevant conditions.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

OH time-histories during oxidation of n-heptane and methylcyclohexane at high pressures and temperatures

OH concentration time-histories during n-heptane and methylcyclohexane (MCH) oxidation were measured behind reflected shock waves in a heated, high-pressure shock tube. Experimental conditions covered temperatures of 1121 to 1332 K, pressures near 15 atm, and initial fuel concentrations of 750 and 1000 ppm (by volume), and an equivalence ratio of 0.5 with O{sub 2} as the oxidizer and argon as the bath gas. OH concentrations were measured using narrow-linewidth ring-dye laser absorption near the R-branchhead of the OH A-X(0,0) system at 306.47 nm. These current measurements together with our recent results for n-dodecane oxidation [S.S. Vasu, D.F. Davidson, Z. Hong, V. Vasudevan, R.K. Hanson, Proc. Combust. Inst. 32 (2009), doi:10.1016/j.proci.2008.05.006] provide critically needed validation targets for jet fuel surrogate kinetic mechanisms and further improve understanding of high-pressure, high-temperature oxidation chemistry. Detailed comparisons of these OH time-histories with the predictions of various kinetic mechanisms were made. Sensitivity and pathway analyses for these reference fuel components were performed, leading to reaction rate recommendations with improved model performance. Current results are the first quantitative measurements of OH time-histories during high-pressure oxidation of these fuels, and hence are a critical step toward development of accurate reaction models for jet fuel surrogates. (author)

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Uncertainty Analysis of Decomposing Polyurethane Foam

Sensitivity/uncertainty analyses are necessary to determine where to allocate resources for improved predictions in support of our nation's nuclear safety mission. Yet, sensitivity/uncertainty analyses are not commonly performed on complex combustion models because the calculations are time consuming, CPU intensive, nontrivial exercises that can lead to deceptive results. To illustrate these ideas, a variety of sensitivity/uncertainty analyses were used to determine the uncertainty associated with thermal decomposition of polyurethane foam exposed to high radiative flux boundary conditions. The polyurethane used in this study is a rigid closed-cell foam used as an encapsulant. Related polyurethane binders such as Estane are used in many energetic materials of interest to the JANNAF community. The complex, finite element foam decomposition model used in this study has 25 input parameters that include chemistry, polymer structure, and thermophysical properties. The response variable was selected as the steady-state decomposition front velocity calculated as the derivative of the decomposition front location versus time. An analytical mean value sensitivity/uncertainty (MV) analysis was used to determine the standard deviation by taking numerical derivatives of the response variable with respect to each of the 25 input parameters. Since the response variable is also a derivative, the standard deviation was essentially determined from a second derivative that was extremely sensitive to numerical noise. To minimize the numerical noise, 50-micrometer element dimensions and approximately 1-msec time steps were required to obtain stable uncertainty results. As an alternative method to determine the uncertainty and sensitivity in the decomposition front velocity, surrogate response surfaces were generated for use with a constrained Latin Hypercube Sampling (LHS) technique. Two surrogate response surfaces were investigated: 1) a linear surrogate response surface (LIN) and 2) a quadratic response surface (QUAD). The LHS techniques do not require derivatives of the response variable and are subsequently relatively insensitive to numerical noise. To compare the LIN and QUAD methods to the MV method, a direct LHS analysis (DLHS) was performed using the full grid and timestep resolved finite element model. The surrogate response models (LIN and QUAD) are shown to give acceptable values of the mean and standard deviation when compared to the fully converged DLHS model.

Hobbs, Michael L.↗

Diabatic initialization for a mesoscale model simulation experiment

A mesoscale model simulation of a 'surrogate' atmosphere designed for the evaluation of adiabatic initialization and moisture enhancement techniques is briefly described. The simulation provides mesoscale precipitation rates and associated diabatic heating fields which are known exactly and are internally consistent within the context of the model. The advantage of this approach is that a numerical model can easily serve as a source of mesoscale data in the usual absence of internally consistent mesoscale observed data sets.

Kalb, Michael W.↗

Learning Scene Categories from High Resolution Satellite Image for Aerial Video Analysis

Automatic scene categorization can benefit various aerial video processing applications. This paper addresses the problem of predicting the scene category from aerial video frames using a prior model learned from satellite imagery. We show that local and global features in the form of line statistics and 2-D power spectrum parameters respectively can characterize the aerial scene well. The line feature statistics and spatial frequency parameters are useful cues to distinguish between different urban scene categories. We learn the scene prediction model from highresolution satellite imagery to test the model on the Columbus Surrogate Unmanned Aerial Vehicle (CSUAV) dataset ollected by high-altitude wide area UAV sensor platform. e compare the proposed features with the popular Scale nvariant Feature Transform (SIFT) features. Our experimental results show that proposed approach outperforms te SIFT model when the training and testing are conducted n disparate data sources.

99 GENERAL AND MISCELLANEOUS↗