Search NASA⌕ Search

SEARCH · Search NASA

Results for “Latin hypercube sampling”

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 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Estimating epistemic uncertainty in soil parameters for nonlinear site response analyses: Introducing the Latin Hypercube Sampling technique

This study quantifies the effects of epistemic uncertainty in soil parameters on nonlinear (NL) site response analysis (SRA) results, validated against the data recorded at a well-instrumented geotechnical downhole array located in Japan. To this end, a one-dimensional soil column model of the Service Hall Array (SHA) near the Kashiwazaki-Kariwa Nuclear Power Plant (KKNPP) is developed using the finite element (FE) program LS-DYNA. The dynamic stress–strain relationship is characterized by a modified two-stage hyperbolic (MTH) NL backbone curve formulation capable of capturing soil behavior at both small- and large-shear strains. The model is then validated against the ground motion recordings to capture the model bias. The uncertainties associated with the shear-wave velocity profile (a small-strain soil property) and soil shear strength (a large-strain soil property) are incorporated in NL SRA to quantify their separate and joint randomization effects on the results. This study proposes using the Latin Hypercube Sampling (LHS) method as an efficient alternative to commonly used methods, such as Standard Monte Carlo (SMC), to account for uncertainty propagation in such reliability analysis. Both low-intensity and design-level records from the recordings at the SHA are applied to study the contribution of the small- and large-strain NL dynamic soil properties. Results from 46,200 NL FE analyses (23,100 per input motion) are presented. Measured and predicted site response, using recorded ground motions at this downhole array, is compared to assess the significance of soil parameter uncertainty on the observed ground motion dispersions. It is demonstrated that increasing the number of soft realizations and implementing higher level earthquake intensity lead to higher ground motion dispersion. Unlike past studies in randomization of Vs profiles with the SMC method, the LHS method is shown to have no significant effect on the predicted median surface response spectra and amplification factors (AFs) for this case study.

Engineering↗
DOE OSTI2022

A comprehensive study of non-adaptive and residual-based adaptive sampling for physics-informed neural networks

Physics-informed neural networks (PINNs) have shown to be effective tools for solving both forward and inverse problems of partial differential equations (PDEs). PINNs embed the PDEs into the loss of the neural network using automatic differentiation, and this PDE loss is evaluated at a set of scattered spatio-temporal points (called residual points). The location and distribution of these residual points are highly important to the performance of PINNs. However, in the existing studies on PINNs, only a few simple residual point sampling methods have mainly been used. Here, we present a comprehensive study of two categories of sampling for PINNs: non-adaptive uniform sampling and adaptive nonuniform sampling. We consider six uniform sampling methods, including (1) equispaced uniform grid, (2) uniformly random sampling, (3) Latin hypercube sampling, (4) Halton sequence, (5) Hammersley sequence, and (6) Sobol sequence. We also consider a resampling strategy for uniform sampling. To improve the sampling efficiency and the accuracy of PINNs, we propose two new residual-based adaptive sampling methods: residual-based adaptive distribution (RAD) and residual-based adaptive refinement with distribution (RAR-D), which dynamically improve the distribution of residual points based on the PDE residuals during training. Hence, we have considered a total of 10 different sampling methods, including six non-adaptive uniform sampling, uniform sampling with resampling, two proposed adaptive sampling, and an existing adaptive sampling. We extensively tested the performance of these sampling methods for four forward problems and two inverse problems in many setups. Our numerical results presented in this study are summarized from more than 6000 simulations of PINNs. Here, we show that the proposed adaptive sampling methods of RAD and RAR-D significantly improve the accuracy of PINNs with fewer residual points for both forward and inverse problems. Furthermore, the results obtained in this study can also be used as a practical guideline in choosing sampling methods.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Comparison of multiobjective optimization methods for the $\mathrm{LCLS-II}$ photoinjector

Particle accelerators are among some of the largest science experiments in the world and can consist of thousands of components with a wide variety of input ranges. These systems can easily become unwieldy optimization problems during design and operations studies. Starting in the early 2000s, searching for better beam dynamics configurations became synonymous with heuristic optimization methods in the accelerator physics community. Genetic algorithms and particle swarm optimization are currently the most widely used. These algorithms can take thousands of simulation evaluations to find optimal solutions for one machine prototype. For large facilities such as the Linac Coherent Light Source (LCLS) and others, this equates to a limited exploration of many possible design configurations. In this paper, the LCLS-II photoinjector is optimized with three optimization algorithms. All optimizations were started from both a uniform random and Latin hypercube sample. In all cases, the optimizations started from Latin hypercube samples outperformed optimizations started from uniform samples. All three algorithms were able to optimize the photoinjector, with the model-based methods approximating the Pareto front in fewer simulation evaluations. This work, in combination with previous optimization observations, indicates objective penalties have a strong impact on the efficiency of such methods. In general, we recommend heuristic methods for initial optimizations and model-based methods when information about the objective space is available.

43 PARTICLE ACCELERATORS↗
DOE OSTI2022

Identification of Distribution of Elastic Constants of SiC/SiC Braided Tubes

There has been considerable interest in chemical vapor infiltration (CVI) manufactured silicon carbide fiber and silicon carbide matrix (SiCf/SiCm) composite tubes due to their superior mechanical properties. There are many parameters in the manufacturing of SiCf/SiCm composite tubes such as elastic constants of constituent SiC fiber and SiC matrix, braiding angle, porosity, and others whose values vary over a range. A sensitivity analysis will be carried out to determine the dominant parameters among the abovementioned ones. We intend to determine the variability in elastic constants of SiCf/SiCm composite tubes considering the variability in dominant parameters. A chosen number of combinations of dominant variables will be sampled through Latin hypercube sampling (LHS) method. The elastic constants of composite tube namely Young's moduli in circumferential and longitudinal direction, Poisson's ratio and shear modulus will be calculated at sampled points through finite element (FE) analysis. A polynomial response surface (PRS) will be built for each of the elastic constants to be used a surrogate to FE analysis. once the predictive accuracy of PRS is verified at the validation set of data points, the PRS will be invoked 105 times in Monte Carlo simulations (MCS). The variability can be calculated based on MCS-generated data which will be presented at the conference.

42 ENGINEERING↗
DOE OSTI2022

Mach Conference, 2022.

There has been considerable interest in chemical vapor infiltration (CVI) manufactured silicon carbide fiber and silicon carbide matrix (SiC/SiC) composite tubes due to their superior mechanical properties. The SiC/SiC tubes are known to remain stable even after prolonged exposure to radiation. However, there are many variables in the manufacturing of SiC/SiC composite tubes such as elastic constants of constituent SiC fiber and SiC matrix, braiding angle, porosity, and others whose values vary over a range. The intention is to identify the distribution of elastic constants of SiC/SiC tubes which would help model the performance of the novel cladding material. To do so, a sensitivity analysis will be carried out to determine the dominant variables that influence the elastic constants of SiC/SiC tubes. A chosen number of combinations of dominant variables will be sampled through the Latin hypercube sampling (LHS) method. The elastic constants of composite tube namely Young's moduli in the circumferential and longitudinal direction, Poisson's ratio, and shear modulus will be calculated at sampled points through finite element (FE) analysis. A polynomial response surface (PRS) will be built for each of the elastic constants to be used as a surrogate to FE analysis. Once the predictive accuracy of PRS is verified at the validation set of data points, the PRS will be invoked 105 times in Monte Carlo simulations (MCS). The uncertainty can be quantified by calculating the coefficient of variation (CV) based on MCS-generated data which will be presented at the conference.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2024

Behavioral Ensemble CLM5 Hydrological Parameter Sets

This repository contains hydrological parameter sets derived using the hybrid regionalization method for three distinct streamflow signatures: Streamflow Signatures: Q10: Represents low flow, indicating the nonexceedance probability of 0.1 for daily streamflow. Q90: Represents high flow, with a nonexceedance probability of 0.9 for daily streamflow. Qmean: Indicates the mean annual flow. Parameters for 464 CAMELS Basins: CAMELS_1000_parameters.csv: Contains 1,000 ensemble parameter sets generated using the Latin hypercube sampling method for CLM5, encompassing 15 hydrological parameters. CAMELS_q10_behavioral_parameter_num.csv: Provides the behavioral ensemble parameter sets for the Q10 streamflow signature for each basin. The associated ID number refers to entries in the CAMELS_1000_parameters.csv file. A minimum of 10 ensemble parameter sets are available for each basin. CAMELS_q90_behavioral_parameter_num.csv: Similar to the above file but for the Q90 streamflow signature. CAMELS_qmean_behavioral_parameter_num.csv: Corresponds to the Qmean streamflow signature, similar to the previous files. Parameters for 50,629 1/8° CONUS Land Grid Cells: CONUS_350_parameters.csv: Contains 350 ensemble parameter sets derived using the Latin hypercube sampling method for CLM5's 15 hydrological parameters within 1/8° CONUS land grid cells. CONUS_q10_behavioral_parameter_num.csv: Holds the behavioral ensemble parameter sets for the Q10 streamflow signature, organized for each grid cell. The ID number relates to entries in CONUS_350_parameters.csv. A minimum of 10 ensemble parameter sets are provided for each grid cell. CONUS_q90_behavioral_parameter_num.csv: Similar to the above file but focusing on the Q90 streamflow signature. CONUS_qmean_behavioral_parameter_num.csv: Corresponds to the Qmean streamflow signature, following a similar structure to the previous files.

Yan, Hongxiang↗
DOE OSTI2021

Cosmic Inference: Constraining Parameters with Observations and a Highly Limited Number of Simulations

We look at cosmological probes that pose an inverse problem where the measurement result is obtained through observations, and the objective is to infer values of model parameters that characterize the underlying physical system-our universe, from these observations and theoretical forward-modeling. The only way to accurately forward-model physical behavior on small scales is via expensive numerical simulations, which are further “emulated” due to their high cost. Emulators are commonly built with a set of simulations covering the parameter space with Latin hypercube sampling and an interpolation procedure; the aim is to establish an approximately constant prediction error across the hypercube. In this paper, we provide a description of a novel statistical framework for obtaining accurate parameter constraints. The proposed framework uses multi-output Gaussian process emulators that are adaptively constructed using Bayesian optimization methods with the goal of maintaining a low emulation error in the region of the hypercube preferred by the observational data. In this paper, we compare several approaches for constructing multi-output emulators that enable us to take possible inter-output correlations into account while maintaining the efficiency needed for inference. Using a Lyα forest flux power spectrum, we demonstrate that our adaptive approach requires considerably fewer-by a factor of a few in the Lyα P(k) case considered here-simulations compared to the emulation based on Latin hypercube sampling, and that the method is more robust in reconstructing parameters and their Bayesian credible intervals.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2026

RxnRover/amlro

AMLRO (Active Machine Learning Reaction Optimizer) is an open-source framework designed to accelerate chemical reaction optimization using active learning with classical machine learning regression models. AMLRO integrates space-filling sampling strategies (e.g., Sobol and Latin Hypercube sampling) with iterative model training, prediction, and experiment selection to efficiently navigate complex reaction spaces. The platform supports multiple regression models, flexible multi-objective definitions, and user-defined parameter bounds, enabling data-efficient optimization from small initial datasets. AMLRO is designed for ease of use by experimentalists and can operate as a standalone decision-support tool or be integrated into closed-loop automated experimentation workflows.

Kulathunga, Dulitha Prasanna [Iowa State Universit↗
DOE OSTI2021

Quantifying the Uncertainty of the Future Hydrological Impacts of Climate Change: Comparative Analysis of an Advanced Hierarchical Sensitivity in Humid and Semiarid Basins

Comparison and quantification of different uncertainties of future climate change involved in the modeling of a hydrological system are highly important for both hydrological modelers and policy-makers. However, few studies have accurately estimated the relative importance of different sources of uncertainty at different spatiotemporal scales. Here, a hierarchical sensitivity analysis framework (HSAF) incorporated with a variance-based global sensitivity analysis is developed to quantify the spatiotemporal contributions of different uncertainties in hydrological impacts of climate change in two different climatic (humid and semiarid) basins in China. The uncertainty sources include three emission scenarios (ESs), 20 global climate models (GCs), three hydrological models (HMs), and the associated sensitive hydrological parameters (PAs) screened and sampled by the Morris and Latin hypercube sampling methods, respectively. Further, the results indicate that the overall trend of uncertainty is PA > HM > GC > ES, but their uncertainties have discrepancies in projections of different hydrological variables. The HM uncertainty in annual and monthly discharge projections is generally larger than the PA uncertainty in the humid basin than semiarid basin. The PA has greater uncertainty in extreme hydrological event (annual peak discharge) projections than in annual discharge projections for both basins (particularly for the humid basin), but contributes larger uncertainty to annual and monthly discharge projections in the semiarid basin than humid basin. The GC contributes larger uncertainty in all the hydrological variables projections in the humid basin than semiarid basin, while the ES uncertainty is rather limited in both basins. Overall, our results suggest there is greater spatiotemporal variability of hydrological uncertainty in more arid regions.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Topology optimization of an airfoil fin microchannel heat exchanger using artificial intelligence

High-performance microchannel heat exchangers are needed to supply heat for power conversion for nuclear microreactors. An airfoil fin microchannel design, constructed of Alloy 617 with helium as the working fluid, is analyzed and optimized using a design of experiments with artificial intelligence techniques. The use of airfoil fins offers the potential to reduce pressure drop across the heat exchanger, as compared to other types of channel configurations. A framework for topology optimization of airfoil fin printed circuit heat exchangers (PCHEs) has been developed that can be readily extended to different fin sizes and shapes, as well as different inlet and operating conditions, materials of construction, and working fluids. An optimization procedure is developed that employs computational fluid dynamics for a set of design points identified using Latin hypercube sampling. Computational fluid dynamics is used to analyze a simplified two-channel configuration where five design parameters are varied – inlet angle, fin scale, extent of staggering, transverse and longitudinal pitches. Two methods (a 5D polynomial and a regression neural network) are compared for generating surrogate models and the resulting response surface approximation is input to a genetic algorithm that is used to identify a set of optimal parameters. The optimal geometries are found across six channel Reynolds numbers ranging from 1000 to 5000, since inlet conditions affect flow through the heat exchanger. Additionally, a set of optimal designs that maximizes heat transfer and minimizes pressure drop is identified, and a thermal stress analysis is performed on the optimal design. Correlations for the Nusselt number and Darcy friction factor are developed that can be useful for thermal hydraulic analyses using system codes. Thermal stresses are analyzed and a brief discussion of the status of code cases of PCHEs for nuclear applications is given. Testing and thermomechanical modeling is needed to facilitate future code compliance of PCHEs for high pressure and high temperature applications.

42 ENGINEERING↗
DOE OSTI2025

Selection of Sampling and Surrogate Modeling Methods for State-Point Evaluations of an AGN-201M Reactor

Nuclear reactor digital twins (DTs) have been proposed for use as a safeguards technology to efficiently monitor new and novel reactors as they come online. A safeguards DT needs to be capable of detecting misuse and diversion as they occur, requiring physics models to be accurate and efficient. Mathematical surrogate models are capable of achieving the necessary efficiency and can largely maintain the accuracy of higher-order models given a quality training sample. The Multiphysics Object-Oriented Simulation Environment (MOOSE) code framework is specifically equipped to generate training samples and create surrogate models using full-order reactor physics models. Utilizing an operational AGN-201M reactor’s specifications, two surrogate types were trained on samples of variable size, and using Cartesian products, Latin hypercube sampling, and quadrature sampling, each was compared and evaluated on accuracy when compared to a full-order Monte Carlo model. Both surrogate types were able to capture reactivity changes within 0.05 $ of the Monte Carlo model while reducing the computation costs by eight orders of magnitude.

MOOSE↗
DOE OSTI2022

Process window estimation in manufacturing through Entropy-Sigma active learning

In manufacturing, there exist boundary identification problems for defining parameter spaces that meet desired thresholds on outcomes. This paper presents an Entropy-Sigma acquisition function for active learning of the process window/map in manufacturing using a Gaussian Process surrogate. Here, the method is applied to identify the stability boundary for the stability process map in machining using time-domain simulations with a periodic sampling stability metric. Results show that the proposed Entropy-Sigma method significantly outperforms Latin hypercube sampling or grid-based methods. The described method can be applied to identify the process window/map for any manufacturing application using a quantitative process outcome metric.

42 ENGINEERING↗
DOE OSTI2022

Accelerated statistical failure analysis of multifidelity TRISO fuel models

Statistical nuclear fuel failure analysis is critical for the design and development of advanced reactor technologies. Although Monte Carlo Sampling (MCS) is a standard method of statistical failure analysis for fuels, the low failure probabilities of some advanced fuel forms and the correspondingly large number of required model evaluations limit its application to low-fidelity (e.g., 1-D) fuel models. In this paper, we present four other statistical methods for fuel failure analysis in Bison, considering tri-structural isotropic (TRISO)-coated particle fuel as a case study. The statistical methods considered are Latin hypercube sampling (LHS), adaptive importance sampling (AIS), subset simulation (SS), and the Weibull theory. Using these methods, we analyzed both 1-D and 2-D representations of TRISO models to compute failure probabilities and the distributions of fuel properties that result in failures. The results of these methods compare well across all TRISO models considered. Overall, SS and the Weibull theory were deemed the most efficient, and can be applied to both 1-D and 2-D TRISO models to compute failure probabilities. Moreover, since SS also characterizes the distribution of parameters that cause TRISO failures, and can consider failure modes not described by the Weibull criterion, it may be preferred over the other methods. Finally, a discussion on the efficacy of different statistical methods of assessing nuclear fuel safety is provided.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2023

Swelling and permeability effects during propellant cookoff

Large rocket motors may violently explode when exposed to accidental fires. Even hot metal fragments from a nearby accident may penetrate the propellant and ultimately cause thermal ignition. A mechanistic understanding of heated propellants leading to thermal runaway is a major unsolved problem. Here we show that thermal ignition in propellants can be predicted using a universal cookoff model coupled to a micromechanics pressurization model. Our model predicts the time to thermal ignition in cookoff experiments with variable headspace volumes. Furthermore, we found that experiments with headspace volumes are more prone to deformation which distorts pores and causes increased permeability when the propellant expands into this headspace. Delayed ignition with larger headspace volume correlates with lower headspace pressures during decomposition. We found that our predictions matched experimental measurements best when the initial propellant was impermeable to gas flow rather than being permeable. Similar behavior is expected with other energetic materials with rubbery binders. Our model is validated using data from a separate laboratory. We also present an uncertainty analysis using Latin Hypercube Sampling (LHS) of thermal ignition caused by a steel fragment embedded in the propellant.

33 ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2025

A statistical approach to screening isotopic signatures in monitoring for underground nuclear explosions

The ability to differentiate between atmospheric radionuclide signatures from underground nuclear explosions (UNEs) and signals from other sources, such as medical isotope-production facilities and nuclear reactors, can be critical to the detection and monitoring of unannounced, low-yield nuclear events. Signatures having anomalously high amplitudes, compared to background levels, remain the best indicator in screening for a UNE. However, isotopic composition can further validate a suspected UNE signature, but separation from any atmospheric background composition is first necessary. To date, evaluating the challenges of performing this separation has typically involved comparing an observed background with a highly idealized deterministic model of radioxenon signature production by a UNE that does not consider the influence of post-detonation chemical/physical processes in the detonation cavity or the subsequent gas transport mechanisms that can also affect the isotopic composition of the detected gas signature. In addition, purely deterministic models, as previously employed, overlook the uncertainty inherent in estimating critical parameters characterizing the UNE and its detonation environment. In this paper, we create detailed, multi-parameter models of radionuclide evolution using the widely accepted England and Rider post-detonation radionuclide decay-chain network coupled to detailed models simulating physical production and transport processes affecting the gas signature. Because these models are governed by uncertain parameters including barometric fluctuations, realistic ranges of variation for each of the parameters influencing isotopic composition are then defined. A Latin-Hypercube sampling approach is used to obtain a random distribution of isotopic production and gas transport results associated with a given value of each parameter. We apply these results to background histories of two stations, one providing 4-isotope background measurements and the other providing two-isotope measurements associated with the 2013 DPRK announced UNE.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗
DOE OSTI2022

Machine learning with knowledge constraints for process optimization of open-air perovskite solar cell manufacturing

Perovskite photovoltaics (PV) have achieved rapid development in the past decade in terms of power conversion efficiency of small-area lab-scale devices; however, successful commercialization still requires further development of low-cost, scalable, and high-throughput manufacturing techniques. One of the critical challenges of developing a new fabrication technique is the high-dimensional parameter space for optimization, but machine learning (ML) can readily be used to accelerate perovskite PV scaling. Herein, we present an ML-guided framework of sequential learning for manufacturing process optimization. We apply our methodology to the Rapid Spray Plasma Processing (RSPP) technique for perovskite thin films in ambient conditions. With a limited experimental budget of screening 100 process conditions, we demonstrated an efficiency improvement to 18.5% as the best-in-our-lab device fabricated by RSPP, and we also experimentally found 10 unique process conditions to produce the top-performing devices of more than 17% efficiency, which is 5 times higher rate of success than the control experiments with pseudo-random Latin hypercube sampling. Our model is enabled by three innovations: (a) flexible knowledge transfer between experimental processes by incorporating data from prior experimental data as a probabilistic constraint; (b) incorporation of both subjective human observations and ML insights when selecting next experiments; (c) adaptive strategy of locating the region of interest using Bayesian optimization first, and then conducting local exploration for high-efficiency devices. Furthermore, in virtual benchmarking, our framework achieves faster improvements with limited experimental budgets than traditional design-of-experiments methods (e.g., one-variable-at-a-time sampling). This framework shows the capability of incorporating researchers’ domain knowledge into the ML-guided optimization loop; therefore, it has the potential to facilitate the wider adoption of ML in scaling to perovskite PV manufacturing.

14 SOLAR ENERGY↗
DOE OSTI2025

A hybrid numerical and machine learning framework for evaluating the performance of a 780 cm 2 aqueous organic redox flow battery

Aqueous organic redox flow battery (AORFB) is a promising cost-competitive technology for large-scale energy storage. Among existing work, the dihydroxyphenazine (DHP)-based AORFB has demonstrated high energy density and low capacity degradation in 10 cm2 cells during lab tests. However, its commercial-scale performance in more complex environments remains unknown, posing a barrier for commercialization. To address this gap, this work presents a comprehensive performance evaluation of a 780 cm 2 DHP-based AORFB by combining physics-based numerical model, machine learning (ML)-based surrogate models, and ML-derived sensitivity quantification. Specifically, we first select 12 key battery parameters that include 10 physicochemical quantities and 2 operation quantities, then select 6 performance metrics that include energy efficiency (EE), discharging capacity, charging energy, and power losses due to concentration, activation, and ohmic over-potentials. With such selection, 12800 combinations of the 12 parameters are subsequently generated using the Latin Hypercube Sampling method. These combinations, together with 38 pre-defined State of Charge, are then integrated to a validated AORFB model developed in COMSOL to compute the performance metrics. With both input parameters and performance metrics, 60 deep neural network (DNN) surrogate models are then trained to approximate the relationship between the 10 physicochemical quantities and 6 performance metrics at each flow rate and current density. Sensitivity scores are then calculated based on the DNN models. Two additional sensitivity analysis tools, i.e., MARS, and SHAP, are also used to cross-validate the sensitivity scores from the DNN. The results demonstrate that 1) the standard potential ranks the first in controlling EE and charging energy, 2) the membrane conductivity is most critical for power loss and EE, and 3) specific area and reaction rate control activation power loss.

25 ENERGY STORAGE↗