SEARCH · Engineering Papers
Results for “Model uncertainty”
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.
Bayesian learning of gas transport in three-dimensional fracture networks
Not Available
Uncertainty propagation in reduced order models based on crystal plasticity
Not Available
A compute-bound formulation of Galerkin model reduction for linear time-invariant dynamical systems
This work aims to advance computational methods for projection-based reduced-order models (ROMs) of linear time-invariant (LTI) dynamical systems. For such systems, current practice relies on ROM formulations expressing the state as a rank-1 tensor (i.e., a vector), leading to computational kernels that are memory bandwidth bound and, therefore, ill-suited for scalable performance on modern architectures. This weakness can be particularly limiting when tackling many-query studies, where one needs to run a large number of simulations. This work introduces a reformulation, called rank-2 Galerkin, of the Galerkin ROM for LTI dynamical systems which converts the nature of the ROM problem from memory bandwidth to compute bound. We present the details of the formulation and its implementation, and demonstrate its utility through numerical experiments using, as a test case, the simulation of elastic seismic shear waves in an axisymmetric domain. We quantify and analyze performance and scaling results for varying numbers of threads and problem sizes. In conclusion, we present an end-to-end demonstration of using the rank-2 Galerkin ROM for a Monte Carlo sampling study. We show that the rank-2 Galerkin ROM is one order of magnitude more efficient than the rank-1 Galerkin ROM (the current practice) and about 970 times more efficient than the full-order model, while maintaining accuracy in both the mean and statistics of the field.
Calibrating hypersonic turbulence flow models with the HIFiRE-1 experiment using data-driven machine-learned models.
In this paper we study the efficacy of combining machine-learning methods with projection-based model reduction techniques for creating data-driven surrogate models of computationally expensive, high-fidelity physics models. Such surrogate models are essential for many-query applications e.g., engineering design optimization and parameter estimation, where it is necessary to invoke the high-fidelity model sequentially, many times. Surrogate models are usually constructed for individual scalar quantities. However there are scenarios where a spatially varying field needs to be modeled as a function of the model’s input parameters. Here we develop a method to do so, using projections to represent spatial variability while a machine-learned model captures the dependence of the model’s response on the inputs. The method is demonstrated on modeling the heat flux and pressure on the surface of the HIFiRE-1 geometry in a Mach 7.16 turbulent flow. The surrogate model is then used to perform Bayesian estimation of freestream conditions and parameters of the SST (Shear Stress Transport) turbulence model embedded in the high-fidelity (Reynolds-Averaged Navier–Stokes) flow simulator, using shock-tunnel data. The paper provides the first-ever Bayesian calibration of a turbulence model for complex hypersonic turbulent flows. We find that the primary issues in estimating the SST model parameters are the limited information content of the heat flux and pressure measurements and the large model-form error encountered in a certain part of the flow.
Bayesian calibration of strength model parameters from Taylor impact data
Materials strength plays a key role in determining the mechanical response of engineered structures. As such, accurate strength models are crucial in simulations involving complex loading conditions, particularly when deformation in the plastic regime is deemed important. In this work, a Gaussian process based surrogate for a finite element simulation of a Taylor impact test is developed and used for Bayesian calibration of the Preston–Tonks–Wallace strength model. Here, the surrogate model is shown to closely approximate the salient features of the Taylor cylinder deformation and is validated against simulation output before being used in the strength model calibration routine. The results show that Taylor impact test data can be used in the calibration of constitutive equations through the use of a combination of data science techniques, namely Gaussian processes and Bayesian inference.
Structural model choices regularly overshadow parametric uncertainty in agent-based simulations of household flood risk outcomes
Not Available
Value of abstraction in performance assessment – When is a higher level of detail necessary?
In this study, different approaches in performance assessment (PA) of the long-term safety of a repository for radioactive waste were examined. This investigation was carried out as part of the DECOVALEX-2023 project, an international collaborative effort for research and model comparison. One specific task of the DECOVALEX-2023 project was the Salt Performance Assessment Modelling task (Salt PA), which aimed at comparing various models and methods employed in the performance assessment of deep geological repositories in salt. In the context of the Salt PA task, three distinct teams from SNL (United States), Quintessa Ltd (United Kingdom), and GRS (Germany) examined the consequences of employing different levels of abstractions when modelling the repository's geometry and implementing various features and processes, using the example of a simple hypothetical repository structure in domal salt. Each team applied their own tools: PFLOTRAN (SNL), QPAC (Quintessa) and LOPOS (GRS). These differ essentially regarding numerical concept and degree of detail in the representation of the underlying physical processes. The discussion focused on when simplifications can be appropriately applied and what consequences result from them. Furthermore, it was explored when and if a higher level of fidelity in geometry or physical processes is required.
A comprehensive techno-eco-assessment of CO 2 enhanced oil recovery projects using a machine-learning assisted workflow
Carbon dioxide enhanced oil recovery (CO 2 -EOR) projects not only extract residual oil but also sequestrate CO 2 in the depleted reservoirs. Here, this study develops a machine-learning-based workflow to co-optimize the hydrocarbon recovery, CO 2 sequestration volume and project net present value (NPV) simultaneously. Considering the trade-off relationships among the objective functions, support vector regression with Gaussian kernel (Gaussian- SVR) proxies are coupled with multi-objective particle swarm optimization (PSO) protocol and generate Pareto optimal solutions. Taking advantage of the high computational efficacy of the proxy model, economic uncertainties introduced by tax credits, capital costs and oil price are investigated by this study. The results indicate that the tax incentive policy (Section 45Q) plays a vital role in enhancing the economic returns of CO 2 -EOR projects, especially under the depression of crude oil market. The proposed workflow has been successfully implemented to optimize a water alternative CO 2 (CO 2 -WAG) injection project in a depleted oil sand in the US. The optimization results yield an incremental oil production of 15.8 MM STB and 1.37 MM metric tons of CO 2 storage in a 20-year development strategy, with the highest project NPV to be 205.6 MM US dollars.
Ensemble modeling of global lake evaporation under climate change
Global projections of lake evaporation are typically based on simulations using single mechanistic models. However, because of its complex interactions with various lake physical properties, environmental and anthropogenic drivers, lake evaporation is highly variable and sensitive to the choice of model used. In this study, we present a multi-model analysis to investigate differences across global simulations of lake evaporation during the warm-season using three different lake models driven by outputs from four general circulation models (GCM) (i.e. 12 model combinations in total) for historic and future scenarios. Our results suggest substantial differences among lake-climate model simulations of lake evaporation. These differences varied throughout the 20th and 21st century, with model driver data explaining 74% of the variance in future projections of warm- season lake evaporation. Our projections indicate that, by the end of the 21st century (2070–2099), global annual lake evaporation rates will increase by 10–27% under Representative Concentration Pathways (RCPs) 2.6–8.5. Finally, we highlight the importance of using a multi-model approach for the prediction of future global lake evaporation responses to climate change.
Reconstruction of fast neutron direction in segmented organic detectors using deep learning
A method for reconstructing the direction of a fast neutron source using a segmented organic scintillator-based detector and deep learning model is proposed and analyzed. Here, the model is based on recurrent neural network, which can be trained by a sequence of data obtained from an event recorded in the detector and suitably pre-processed. The performance of deep learning-based model is compared with the conventional double-scatter detection algorithm in reconstructing the direction of a fast neutron source. With the deep learning model, the uncertainty in source direction of 0.301 rad is achieved with 100 neutron detection events in a segmented cubic organic scintillator detector with a side length of 46 mm. To reconstruct the source direction with the same angular resolution as the double-scatter algorithm, the deep learning method requires 75% fewer events. Application of this method could augment the operation of segmented detectors operated in the neutron scatter camera configuration for applications such as special nuclear material detection.
Robust importance sampling for error estimation in the context of optimal Bayesian transfer learning
Not Available
Bayesian inverse uncertainty quantification of the physical model parameters for the spallation neutron source first target station
Not Available
Machine-learning-based efficient parameter space exploration for energy storage systems
Not Available
Predicting Large‐Scale Systematic Missing Pipe Attributes in Water Distribution Networks
Water distribution network (WDN) models are an essential tool used by water utilities for hydraulic analysis. Unfortunately, missing data and insufficient resources often make creating and maintaining these models unfeasible. Existing methods to address missing pipe properties, like sequential imputation for missing values and reconstruction using graph metrics, are designed to accommodate random patterns of missing information and require a significant percentage of the system's attributes to be known. However, these data completeness assumptions do not always align with real‐world scenarios where large sections of the WDN model have missing data. To address this challenge, this study proposes a data‐driven approach for estimating pipe diameter when considering different spatial patterns and degrees of data completeness (i.e., 0%–90%). Using data from 16 WDNs in Kentucky, this study compares the use of machine learning (ML) using topological and geospatial features against an existing deterministic approach. Results demonstrate that WDN models with pipe diameters predicted by the proposed ML method had comparable hydraulic performance to the ground truth models. Moreover, results showed that ML method performance varies between WDNs of differing topological classification. Insights from this study help advance the ability to leverage partial data to create and maintain WDN models amid uncertainty and inadequate resources.
Code-to-code comparison and validation of the radiation-hydrodynamics capabilities of the FLASH code using a laboratory astrophysical jet
The potential for laser-produced plasmas to yield fundamental insights into high energy density physics (HEDP) and deliver other useful applications can sometimes be frustrated by uncertainties in modeling the properties and behavior of these plasmas using radiation-hydrodynamics codes. In an effort to overcome this and to corroborate the accuracy of the HEDP capabilities in the publicly available FLASH radiation-hydrodynamics code, we present detailed code-to-code comparisons between FLASH and the HYDRA code developed at Lawrence Livermore National Laboratory using previously published HYDRA simulations from Grava et al. [Phys. Rev. E 78, 016403 (2008)]. That study describes a laser experiment that produced a jet-like feature that the authors compare to astrophysical jets. Importantly, the Grava et al. [Phys. Rev. E 78, 016403 (2008)] experiment included detailed x-ray interferometric measurements of electron number densities and a time-integrated measurement of the soft x-ray spectrum. Despite markedly different methods for treating the computational mesh, and different equations of state and opacity models, the FLASH results resemble the results from HYDRA and, most importantly, the experimental measurements of electron density. Having validated the FLASH code in this way, we use the code to further investigate and understand the formation of the jet seen in the Grava et al. [Phys. Rev. E 78, 016403 (2008)] experiment and discuss its relation to the Wan et al. [Phys. Rev. E 55, 6293 (1997)] experiment at the NOVA laser.
Overview of preparation for the American WAKE ExperimeNt (AWAKEN)
The American WAKE ExperimeNt (AWAKEN) is a multi-institutional field campaign focused on gathering critical observations of wind farm–atmosphere interactions. These interactions are responsible for a large portion of the uncertainty in wind plant modeling tools that are used to represent wind plant performance both prior to construction and during operation and can negatively impact wind energy profitability. The AWAKEN field campaign will provide data for validation, ultimately improving modeling and lowering these uncertainties. The field campaign is designed to address seven testable hypotheses through the analysis of the observations collected by numerous instruments at 13 ground-based locations and on five wind turbines. The location of the field campaign in Northern Oklahoma was chosen to leverage existing observational facilities operated by the U.S. Department of Energy Atmospheric Radiation Measurement program in close proximity to five operating wind plants. The vast majority of the observations from the experiment are publicly available to researchers and industry members worldwide, which the authors hope will advance the state of the science for wind plants and lead to lower cost and increased reliability of wind energy systems.