Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “error quantification”

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 109 records · Page 6
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Iterative Workflow for Quantification and Minimization of Reduced Chemistry-Induced Uncertainties in Reacting Flow Simulation

Reduced chemistry models mitigate computational cost but introduce two sources of uncertainties in reacting flow simulation, including chemical information loss due to model reduction, and approximation errors due to non-optimal projection. We present an iterative workflow for quantification and minimization of reduced chemistry-induced uncertainties in reacting flow simulations.

chemistry↗
DOE OSTI2024

Evidential Deep Learning for Probabilistic Modelling of Extreme Storm Events

Uncertainty quantification (UQ) methods play an important role in reducing errors in weather forecasting. Conventional approaches in UQ for weather forecasting rely on generating an ensemble of forecasts from physics-based simulations to estimate the uncertainty. However, it is computationally expensive to generate many forecasts to predict real-time extreme weather events. Evidential Deep Learning (EDL) is an uncertainty-aware deep learning approach designed to provide confidence about its predictions using only one forecast. It treats learning as an evidence acquisition process where more evidence is interpreted as increased predictive confidence. We apply EDL to storm forecasting using real-world weather datasets and compare its performance with traditional methods. Our findings indicate that EDL not only reduces computational overhead but also enhances predictive uncertainty. This method opens up novel opportunities in research areas such as climate risk assessment, where quantifying the uncertainty about future climate is crucial.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

An Experimental Investigation of Students’ Learning Effects When Using a Simplified Nuclear Simulator

Securing enough data has been a main challenge in human reliability analysis (HRA). Many researchers and institutes have made a lot of efforts for collecting HRA data to produce reasonable human error probabilities (HEPs) as well as reduce the uncertainty of HRA quantification. Representatively, U.S. Nuclear Regulatory Commission (U.S. NRC), Korea Atomic Energy Research Institute (KAERI) and Idaho National Laboratory (INL) have led lots of empirical research regarding the HRA data collection. The U.S. NRC and KAERI have mainly carried out full-scope simulator research collecting HRA data through experiments using full-scope simulators with actual operators. In contrast, INL has experimentally collected the data using simplified simulators and student operators. INL has proposed the Simplified Human Error Experimental Program (SHEEP) framework to complement full-scope data collection efforts by suggesting a way to infer full-scope data based on experimental data collected from students operating simplified simulators, specifically the Rancor Microworld Simulator (Rancor) and the Compact Nuclear Simulator (CNS). The aim of the SHEEP framework is to lower the entry point for collecting useful HRA data by securing large sample sizes at a reasonable amount of cost and labor while also guaranteeing a high degree of freedom when designing experiments. The authors’ previous research investigated whether data collected from the SHEEP framework could support a representative full-scope study. Besides, human performance differences between professional and student operators when using Rancor and CNS have been analyzed to understand the lack of fidelity of the simplified simulators and student operators within the SHEEP study. As a follow up research, this study experimentally investigates students’ learning effects and the performance trends over a certain period when using Rancor. This study aims to find out 1) how much training or education is required to collect HRA data from non-experts (i.e., students) when using Rancor and 2) how much differences there are in human performance measures between students and professional operators. In this study, a longitudinal experiment is developed. The four experiment trials with two weeks interval are carried out for sixteen undergraduate students majoring nuclear engineering at Chosun University. Totally four scenarios randomly selected from ten Rancor scenarios are used in each experiment trial. Four human performance measurements (i.e., workload, situation awareness, time and error) are considered in the experiment. Lastly, the trend of students’ performance is compared with operator data having been collected from the previous experiment.

99 GENERAL AND MISCELLANEOUS↗
DOE OSTI2023

Uncertainty quantification in scientific machine learning: Methods, metrics, and comparisons

Neural networks (NNs) are currently changing the computational paradigm on how to combine data with mathematical laws in physics and engineering in a profound way, tackling challenging inverse and ill-posed problems not solvable with traditional methods. However, quantifying errors and uncertainties in NN-based inference is more complicated than in traditional methods. This is because in addition to aleatoric uncertainty associated with noisy data, there is also uncertainty due to limited data, but also due to NN hyperparameters, overparametrization, optimization and sampling errors as well as model misspecification. Although there are some recent works on uncertainty quantification (UQ) in NNs, there is no systematic investigation of suitable methods towards quantifying the total uncertainty effectively and efficiently even for function approximation, and there is even less work on solving partial differential equations and learning operator mappings between infinite-dimensional function spaces using NNs. In this work, we present a comprehensive framework that includes uncertainty modeling, new and existing solution methods, as well as evaluation metrics and post-hoc improvement approaches. Further, to demonstrate the applicability and reliability of our framework, we present an extensive comparative study in which various methods are tested on prototype problems, including problems with mixed input-output data, and stochastic problems in high dimensions. In the Appendix, we include a comprehensive description of all the UQ methods employed. Further, to help facilitate the deployment of UQ in Scientific Machine Learning research and practice, we present and develop in [1] an open-source Python library (github.com/Crunch-UQ4MI/neuraluq), termed NeuralUQ, that is accompanied by an educational tutorial and additional computational experiments.

11 physics-informed neural networks↗
DOE OSTI2024

Conservative discontinuous Galerkin interpolation: Sheared boundary conditions

Local studies of accretion disks and laboratory magnetized plasmas employ analytical coordinate mappings that introduce sheared boundary conditions (BCs). Here, we present a discontinuous Galerkin (DG) algorithm to apply such BCs based on projections and quadrature-free integration. The procedure is high-order accurate, preserves moments exactly and works in multiple dimensions. Tests of increasing complexity are provided, beginning with translations of one and two dimensional fields, followed by 3D and 5D simulations with sheared (twist-shift) BCs. Results show that the algorithm is (p + 1)-order accurate in the DG representation and (p + 2) -order accurate in the cell averages, with p being the order of the polynomial basis. Quantification of the algorithm's hyperdiffusion and discussion of aliasing errors are given. This technique enables conservative local simulations of plasma turbulence with DG, not possible until now.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2020

Resolving atomic transport through iron oxide under irradiation using isotopic tracers

Structural materials in nuclear environments are exposed to complex and harsh conditions, including high temperatures, corrosive solutions, and irradiation. Material stability and likewise degradation are fundamentally controlled by atomic transport mechanisms. Here, we investigate the influence of radiation-induced, non-equilibrium point defect populations on mass transport within model oxide systems using isotopic tracers and atom probe tomography. An isotopically labelled hematite (a-Fe2O3) film was synthesized with 16O- and 18O-rich layers (>30nm and 10 nm, respectively) sequentially deposited via molecular beam epitaxy. One region of the film was proton irradiated to ~0.1 dpa at 450°C, and regions outside the beam spot provided a non-irradiated control of thermal annealing only. 3D APT was used to directly visualize the elemental and O-isotopic redistribution following both thermal and combined thermal and irradiation conditions. Complementary scanning transmission electron microscopy (STEM) was also employed to study the microstructural evolution. Error function fits of the 18O-concentration profiles enabled quantification of diffusion coefficients.

isotopic tracers, diffusion, proton irradiation, h↗
DOE OSTI2023

Chiral uncertainties in ab initio elastic nucleon-nucleus scattering

The effective interaction between a nucleon and a nucleus is one of the most important ingredients for reaction theories. Theoretical formulations were introduced early by Feshbach and Watson, and efforts of deriving and computing those ‘optical potentials’ in a microscopic fashion have a long tradition. However, only recently the leading order term in the Watson multiple scattering approach could be calculated fully ab initio, meaning that the same nucleon-nucleon (NN) interaction enters both the structure as well as the reaction pieces on equal footing. This allows the uncertainties from the underlying chiral effective NN interaction to be systematically explored in nucleon-nucleus elastic scattering observables. In this contribution the main ingredients for arriving at the ab initio leading order of the effective nucleon-nucleus interaction in the Watson approach will be reviewed. Concentrating on one specific chiral NN interaction from the LENPIC collaboration and light nuclei with a 0 + ground state, the leading order nucleon-nucleus interaction is calculated using up to the third chiral order (N2LO) in the nucleon-nucleon potential, and elastic scattering observables are extracted. Then pointwise as well as correlated uncertainty quantification is used for the estimation of the chiral truncation error. Elastic scattering observables for 4 He, 12 C, and 16 O for between 65 and 200 MeV projectile energy will be analyzed.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
NASA NTRS2011

Supersonic Retro-Propulsion Experimental Design for Computational Fluid Dynamics Model Validation

The development of supersonic retro-propulsion, an enabling technology for heavy payload exploration missions to Mars, is the primary focus for the present paper. A new experimental model, intended to provide computational fluid dynamics model validation data, was recently designed for the Langley Research Center Unitary Plan Wind Tunnel Test Section 2. Pre-test computations were instrumental for sizing and refining the model, over the Mach number range of 2.4 to 4.6, such that tunnel blockage and internal flow separation issues would be minimized. A 5-in diameter 70-deg sphere-cone forebody, which accommodates up to four 4:1 area ratio nozzles, followed by a 10-in long cylindrical aftbody was developed for this study based on the computational results. The model was designed to allow for a large number of surface pressure measurements on the forebody and aftbody. Supplemental data included high-speed Schlieren video and internal pressures and temperatures. The run matrix was developed to allow for the quantification of various sources of experimental uncertainty, such as random errors due to run-to-run variations and bias errors due to flow field or model misalignments. Some preliminary results and observations from the test are presented, although detailed analyses of the data and uncertainties are still on going.

Berry, Scott A.↗
NASA NTRS2011

Supersonic Retropropulsion Experimental Results from the NASA Langley Unitary Plan Wind Tunnel

A new supersonic retropropulsion experimental effort, intended to provide code validation data, was recently completed in the Langley Research Center Unitary Plan Wind Tunnel Test Section 2 over the Mach number range from 2.4 to 4.6. The experimental model was designed using insights gained from pre-test computations, which were instrumental for sizing and refining the model to minimize tunnel wall interference and internal flow separation concerns. A 5-in diameter 70-deg sphere-cone forebody with a roughly 10-in long cylindrical aftbody was the baseline configuration selected for this study. The forebody was designed to accommodate up to four 4:1 area ratio supersonic nozzles. Primary measurements for this model were a large number of surface pressures on the forebody and aftbody. Supplemental data included high-speed Schlieren video and internal pressures and temperatures. The run matrix was developed to allow for the quantification of various sources of experimental uncertainty, such as random errors due to run-to-run variations and bias errors due to flow field or model misalignments. Preliminary results and observations from the test are presented, while detailed data and uncertainty analyses are ongoing.

Berry, Scott A.↗
NASA NTRS2014

Supersonic Retropropulsion Test 1853 in NASA LaRC Unitary Plan Wind Tunnel Test Section 2

A supersonic retropropulsion experiment was conducted in the Langley Research Center Unitary Plan Wind Tunnel Test Section 2 at Mach numbers of 2.4, 3.5, and 4.6. Intended as a code validation effort, this study used pretest computations to size and refine the model such that tunnel blockage and internal flow separations were minimized. A 5-in diameter 70 degree sphere-cone forebody, which can accommodate up to four 4:1 area ratio nozzles, followed by a 9.55 inches long cylindrical aft body was selected for this test after computational maturation. The primary measurements for this experiment were high spatial-density surface pressures. In addition, high speed schlieren video and internal pressures and temperatures were acquired. The test included parametric variations in the number of nozzles utilized, thrust coefficients (roughly 0 to 4), and angles of attack (-8 to 20 degrees). The run matrix was developed to also allow quantification of various sources of experimental uncertainty, such as random errors due to run-to-run variations and systematic errors due to flowfield or model misalignments. To accommodate the uncertainty assessment, many runs and replicates were conducted with the model at various locations within the tunnel and with model roll angles of 0, 60, 120, and 180 degrees. This test report provides operational details of the experiment, contains a review of trends, and provides all schlieren and pressure results within appendices.

Berry, Scott A.↗
DOE OSTI2024

Uncertainty quantification of graph convolution neural network models of evolving processes

The application of neural network models to scientific machine learning tasks has proliferated in recent years. In particular, neural networks have proved to be adept at modeling processes with spatial–temporal complexity. Nevertheless, these highly parameterized models have garnered skepticism in their ability to produce outputs with quantified error bounds over the regimes of interest. Hence there is a need to find uncertainty quantification methods that are suitable for neural networks. In this work we present comparisons of the parametric uncertainty quantification of neural networks modeling complex spatial–temporal processes with Hamiltonian Monte Carlo and Stein variational gradient descent and its projected variant. Specifically we apply these methods to graph convolutional neural network models of evolving systems modeled with recurrent neural network and neural ordinary differential equations architectures. We show that Stein variational inference is a viable alternative to Monte Carlo methods with some clear advantages for complex neural network models. For our exemplars, Stein variational interference gave similar pushed forward uncertainty profiles through time compared to Hamiltonian Monte Carlo, albeit with generally more generous variance. As a result, projected Stein variational gradient descent also produced similar uncertainty profiles to the non-projected counterpart, but large reductions in the active weight space were confounded by the stability of the neural network predictions and the convoluted likelihood landscape.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Greedy emulators for nuclear two-body scattering

Applications of reduced basis method emulators are increasing in low-energy nuclear physics because they enable fast and accurate sampling of high-fidelity calculations, enabling robust uncertainty quantification. Here, in this paper, we develop, implement, and test two model-driven emulators based on the (Petrov-)Galerkin projection using the prototypical test case of two-body scattering with the Minnesota potential and a more realistic local chiral potential. The high-fidelity scattering equations are solved with the matrix Numerov method, a reformulation of the popular Numerov recurrence relation for solving special second-order differential equations as a linear system of coupled equations. A novel error estimator based on reduced-space residuals is applied to an active learning approach (a greedy algorithm) to choosing training samples (“snapshots”) for the emulator and contrasted with a proper orthogonal decomposition (POD) approach. Both approaches allow for computationally efficient offline-online decompositions, but the greedy approach requires many fewer snapshot calculations. These developments set the groundwork for emulating scattering observables based on chiral nucleon-nucleon and three-nucleon interactions and optical models, where computational speed-ups are necessary for Bayesian uncertainty quantification. Our emulators and error estimators are widely applicable to linear systems.

Bayesian methods↗
DOE OSTI2023

Analyzing Hardware and Software Common Cause Failures in Digital Instrumentation and Control Systems using Dual Error Propagation Method

This paper develops a methodology for quantifying software common cause failures (CCFs) in digital instrumentation and control (I&C) systems of nuclear power plants. To support the transition of analog I&C systems to digital in nuclear power plants, probabilistic risk assessment (PRA) techniques are used. The hardware components of the I&C systems have reliability databases that can be used in the PRA studies. However, the failure data for redundant software components of the systems is sparse. Failure of components constitutes a CCF, wherein two or more components or systems fail due to a single shared cause and coupling mechanism. This paper proposes a quantification approach that can simultaneously model hardware and software components, incorporate the CCFs of software systems in the models, and bridge the gap between the failure quantification of models and the development of CCF parametric databases. We demonstrate the dual error propagation method (DEPM) by developing I&C systems failure models for a representative digital reactor trip system. The DEPM models are built to simulate the control and data flows within the systems and can accommodate failure states. By expanding DEPM to software CCFs, we generated alpha factor parameter estimates for each of the modeled error propagation mechanisms.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2021

Formative vs. Summative Dependence in Human Reliability Analysis

Dependence in human reliability analysis (HRA) is the concept that once an initial human error has occurred, subsequent human errors are more likely. The approach almost universally adopted in HRA was first introduced in the Technique for Human Error Rate Prediction (THERP). In THERP, calculated human error probabilities (HEPs) are adjusted for dependence in the final step of quantification. This final adjustment for dependence involves anchoring the HEP to a set of values corresponding to low through complete dependence. The effect is to increase the HEP. In this paper, we propose formative dependence, which occurs at the onset of quantification. We demonstrate that dependence should be considered earlier in the calculation process, because it is not actually necessary to calculate the HEP when it is subsequently overridden by the dependence anchors. By applying dependence first in the calculation, processing steps can be eliminated, making for more efficient analysis.

99 GENERAL AND MISCELLANEOUS↗
DOE OSTI2023

Uncertainty quantification of machine learning models to improve streamflow prediction under changing climate and environmental conditions

Machine learning (ML) models, and Long Short-Term Memory (LSTM) networks in particular, have demonstrated remarkable performance in streamflow prediction and are increasingly being used by the hydrological research community. However, most of these applications do not include uncertainty quantification (UQ). ML models are data driven and can suffer from large extrapolation errors when applied to changing climate/environmental conditions. UQ is required to quantify the influence of data noises on model predictions and avoid overconfident projections in extrapolation. In this work, we integrate a novel UQ method, called PI3NN, with LSTM networks for streamflow prediction. PI3NN calculates Prediction Intervals by training 3 Neural Networks. It can precisely quantify the predictive uncertainty caused by the data noise and identify out-of-distribution (OOD) data in a non-stationary condition to avoid overconfident predictions. We apply the PI3NN-LSTM method in the snow-dominant East River Watershed in the western US and in the rain-driven Walker Branch Watershed in the southeastern US. Results indicate that for the prediction data which have similar features as the training data, PI3NN precisely quantifies the predictive uncertainty with the desired confidence level; and for the OOD data where the LSTM network fails to make accurate predictions, PI3NN produces a reasonably large uncertainty indicating that the results are not trustworthy and should avoid overconfidence. PI3NN is computationally efficient, robust in performance, and generalizable to various network structures and data with no distributional assumptions. It can be broadly applied in ML-based hydrological simulations for credible prediction.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Quantification and visualization of uncertainties in reconstructed penumbral images of implosions at Omega

Penumbral imaging is a technique used in plasma diagnostics in which a radiation source shines through one or more large apertures onto a detector. To interpret a penumbral image, one must reconstruct it to recover the original source. The inferred source always has some error due to noise in the image and uncertainty in the instrument geometry. Interpreting the inferred source thus requires quantification of that inference’s uncertainty. Markov chain Monte Carlo algorithms have been used to quantify uncertainty for similar problems but have never been used for the inference of the shape of an image. Because of this, there are no commonly accepted ways of visualizing uncertainty in two-dimensional data. This paper demonstrates the application of the Hamiltonian Monte Carlo algorithm to the reconstruction of penumbral images of fusion implosions and presents ways to visualize the uncertainty in the reconstructed source. This methodology enables more rigorous analysis of penumbral images than has been done in the past.

Instruments & Instrumentation↗
DOE OSTI2024

Assessing the design of integrated methane sensing networks

Abstract While methane is the second largest contributor to global warming after carbon dioxide, it has a larger warming effect over a much shorter lifetime. Despite accelerated technological efforts to radically reduce global carbon dioxide emissions, rapid reductions in methane emissions are needed to limit near-term warming. Being primarily emitted as a byproduct from agricultural activities and energy extraction, methane is currently monitored via bottom–up (i.e. activity level) or top–down (via airborne or satellite retrievals) approaches. However, significant methane leaks remain undetected and emission rates are challenging to characterize with current monitoring frameworks. In this paper, we study the design of a layered monitoring approach that combines bottom–up and top–down approaches as an integrated sensing network. By recognizing that varying meteorological conditions and emission rates impact the efficacy of bottom–up monitoring, we develop a probabilistic approach to optimal sensor placement in its bottom–up network. Subsequently, we derive an inverse Bayesian framework to quantify the improvement that a design-optimized integrated framework has on emission-rate quantifications and their uncertainties. We find that under realistic meteorological conditions, the overall error in estimating the true emission rates is approximately 1.3 times higher, with their uncertainties being approximately 2.4 times higher, when using a randomized network over an optimized network, highlighting the importance of optimizing the design of integrated methane sensing networks. Further, we find that optimized networks can improve scenario coverage fractions by more than a factor of 2 over experimentally-studied networks, and identify a budget threshold beyond which the rate of optimized-network coverage improvement exhibits diminishing returns, suggesting that strategic sensor placement is also crucial for maximizing network efficiency.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

Sparse Bayesian mass mapping with uncertainties: hypothesis testing of structure

ABSTRACT A crucial aspect of mass mapping, via weak lensing, is quantification of the uncertainty introduced during the reconstruction process. Properly accounting for these errors has been largely ignored to date. We present a new method to reconstruct maximum a posteriori (MAP) convergence maps by formulating an unconstrained Bayesian inference problem with Laplace-type l1-norm sparsity-promoting priors, which we solve via convex optimization. Approaching mass mapping in this manner allows us to exploit recent developments in probability concentration theory to infer theoretically conservative uncertainties for our MAP reconstructions, without relying on assumptions of Gaussianity. For the first time, these methods allow us to perform hypothesis testing of structure, from which it is possible to distinguish between physical objects and artefacts of the reconstruction. Here, we present this new formalism, and demonstrate the method on simulations, before applying the developed formalism to two observational data sets of the Abell 520 cluster. Initial reconstructions of the Abell 520 catalogues reported the detection of an anomalous ‘dark core’ – an overdense region with no optical counterpart – which was taken to be evidence for self-interacting dark matter. In our Bayesian framework, it is found that neither Abell 520 data set can conclusively determine the physicality of such dark cores at $99{{\ \rm per\ cent}}$ confidence. However, in both cases the recovered MAP estimators are consistent with both sets of data.

Price, M. A.↗