Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “MODEL SIMULATION DATA”

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 487 records · Page 27

Machine Learning-Based Regression Models for Ironmaking Blast Furnace Automation

Computational fluid dynamics (CFD)-based simulation has been the traditional way to model complex industrial systems and processes. One very large and complex industrial system that has benefited from CFD-based simulations is the steel blast furnace system. The problem with the CFD-based simulation approach is that it tends to be very slow for generating data. The CFD-only approach may not be fast enough for use in real-time decisionmaking. To address this issue, in this work, the authors propose the use of machine learning techniques to train and test models based on data generated via CFD simulation. Regression models based on neural networks are compared with tree-boosting models. In particular, several areas (tuyere, raceway, and shaft) of the blast furnace are modeled using these approaches. The results of the model training and testing are presented and discussed. The obtained R 2 metrics are, in general, very high. The results appear promising and may help to improve the efficiency of operator and process engineer decisionmaking when running a blast furnace.

97 MATHEMATICS AND COMPUTING↗

Modeling the Radiative Effect on Microphysics in Cirrus Clouds Against Satellite Observations

The radiative effect on microphysics (REM) plays an important role in the dew/frost formation near the surface. How REM impacts cirrus clouds is investigated in this study, using bin microphysical model simulations and coincident data of the CloudSat and Global Precipitation Measurement (GPM) satellites. REM affects ice crystal spectrum with two types: radiative cooling and warming. Radiative cooling, as predicted by the bin-model simulations, favors the formation of horizontally oriented ice crystals (HOICs), but radiative warming does not. Hence, a test of REM can be transformed to a test of HOICs, because HOICs can be measured by the microwave polarization observations of the GPM Microwave Imager (GMI) at 166 GHz. To analyze the GMI data for their HOIC distribution, clouds are sorted into four groups with different optical depth and altitude, based on the radiative cooling/warming ratio (or eta) computed with satellite-retrieved ice water content. Their HOIC distributions (e.g., the midlevel thick clouds have more HOICs than the high-level ones) agree well with those predicted by the bin-model simulations. The general agreement between the GMI observations and bin-model simulations suggests that REM is common in cirrus clouds and impacts cirrus clouds significantly.

clouds↗

Impact Of Melter Internal Design On Off-Gas Flammability

The purpose of this study was to: (1) identify the more dominant design parameters that can serve as the quantitative measure of how prototypic a given melter is, (2) run the existing DWPF models to simulate the data collected using both DWPF and non-DWPF melter configurations, (3) confirm the validity of the selected design parameters by determining if the agreement between the model predictions and data is reasonably good in light of the design and operating conditions employed in each data set, and (4) run Computational Fluid Dynamics (CFD) simulations to gain new insights into how fluid mixing is affected by the configuration of melter internals and to further apply the new insights to explaining, for example, why the agreement is not good.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

On the Sensitivity of Atmospheric Ensembles to Cloud Microphysics in Long-Term Cloud-Resolving Model Simulations

Month-long large-scale forcing data from two field campaigns are used to drive a cloud-resolving model (CRM) and produce ensemble simulations of clouds and precipitation. Observational data are then used to evaluate the model results. To improve the model results, a new parameterization of the Bergeron process is proposed that incorporates the number concentration of ice nuclei (IN). Numerical simulations reveal that atmospheric ensembles are sensitive to IN concentration and ice crystal multiplication. Two- (2D) and three-dimensional (3D) simulations are carried out to address the sensitivity of atmospheric ensembles to model dimensionality. It is found that the ensembles with high IN concentration are more sensitive to dimensionality than those with low IN concentration. Both the analytic solutions of linear dry models and the CRM output show that there are more convective cores with stronger updrafts in 3D simulations than in 2D, which explains the differing sensitivity of the ensembles to dimensionality at different IN concentrations.

Zeng, Xiping↗

The Importance of Being Adaptable: An Exploration of the Power and Limitations of Domain Adaptation for Simulation-Based Inference with Galaxy Clusters

The application of deep machine learning methods in astronomy has exploded in the last decade, with new models showing remarkably improved performance on benchmark tasks. Not nearly enough attention is given to understanding the models' robustness, especially when the test data are systematically different from the training data, or "out of domain." Domain shift poses a significant challenge for simulation-based inference, where models are trained on simulated data but applied to real observational data. In this paper, we explore domain shift and test domain adaptation methods for a specific scientific case: simulation-based inference for estimating galaxy cluster masses from X-ray profiles. We build datasets to mimic simulation-based inference: a training set from the Magneticum simulation, a scatter-augmented training set to capture uncertainties in scaling relations, and a test set derived from the IllustrisTNG simulation. We demonstrate that the Test Set is out of domain in subtle ways that would be difficult to detect without careful analysis. We apply three deep learning methods: a standard neural network (NN), a neural network trained on the scatter-augmented input catalogs, and a Deep Reconstruction-Regression Network (DRRN), a semi-supervised deep model engineered to address domain shift. Although the NN improves results by 17% in the Training Data, it performs 40% worse on the out-of-domain Test Set. Surprisingly, the Scatter-Augmented Neural Network (SANN) performs similarly. While the DRRN is successful in mapping the training and Test Data onto the same latent space, it consistently underperforms compared to a straightforward Yx scaling relation. These results serve as a warning that simulation-based inference must be handled with extreme care, as subtle differences between training simulations and observational data can lead to unforeseen biases creeping into the results.

Ntampaka, Michelle [Baltimore, Space Telescope Sci↗

Chapter 14: Machine Learning of Combustion LES Models from Reacting Direct Numerical Simulation

In this chapter we demonstrate how supervised deep learning techniques can be used to construct models for the filtered progress variable source term necessary for large eddy simulation (LES). The source data for the model is a direct numerical simulation (DNS) of a reacting flow in a low swirl burner configuration. Filtered quantities taken from the DNS data are used to train a deep neural network (DNN)-based model. An efficient data sampling strategy was devised to ensure that a uniform representation of all the states observed in the filtered DNS data are equally present in the training dataset. A-priori testing of the DNN-based model highlights the representative power of DNN to accurately reproduce the filtered reaction progress variable source term over a range of scales and various flame regimes as seen in an industrial burner.

combustion LES models↗

Transferability of data-driven, many-body models for CO 2 simulations in the vapor and liquid phases

Here, extending on the previous work by Riera et al. [J. Chem. Theory Comput. 16, 2246–2257 (2020)], we introduce a second generation family of data-driven many-body MB-nrg models for CO 2 and systematically assess how the strength and anisotropy of the CO 2 –CO 2 interactions affect the models’ ability to predict vapor, liquid, and vapor–liquid equilibrium properties. Building upon the many-body expansion formalism, we construct a series of MB-nrg models by fitting one-body and two-body reference energies calculated at the coupled cluster level of theory for large monomer and dimer training sets. Advancing from the first generation models, we employ the charge model 5 scheme to determine the atomic charges and systematically scale the two-body energies to obtain more accurate descriptions of vapor, liquid, and vapor–liquid equilibrium properties. Challenges in model construction arise due to the anisotropic nature and small magnitude of the interaction energies in CO 2 , calling for the necessity of highly accurate descriptions of the multidimensional energy landscape of liquid CO 2 . These findings emphasize the key role played by the training set quality in the development of transferable, data-driven models, which, accurately representing high-dimensional many-body effects, can enable predictive computer simulations of molecular fluids across the entire phase diagram.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Nonperturbative Guiding Center Model for Magnetized Plasmas

Perturbative guiding center theory adequately describes the slow drift motion of charged particles in the strongly magnetized regime characteristic of thermal particle populations in various magnetic fusion devices. However, it breaks down for particles with large-enough energy. Here, we report on a data-driven method for learning a nonperturbative guiding center model from full-orbit particle simulation data. We show the data-driven model significantly outperforms traditional asymptotic theory in magnetization regimes appropriate for fusion-born α particles in stellarators, thus opening the door to nonperturbative guiding center calculations.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

ASC FY2023 Implementation Plan Revision 0

The DOE National Nuclear Security Administration (NNSA) Stockpile Stewardship Program (SSP) is an integrated technical program for maintaining the safety, security, and reliability of the U.S. nuclear stockpile. The SSP incorporates nuclear test data, computational modeling and simulation, and experimental facilities to advance understanding of nuclear weapons. The suite of data analyzed comes from activities including previous nuclear tests, stockpile surveillance, experimental research, and development and engineering programs. This integrated national program requires the continued use of experimental facilities and the computational capabilities to support the SSP missions. These component parts, in addition to an appropriately scaled production capability, enable NNSA to support stockpile requirements. The ultimate goal of the SSP, and thus of the Advanced Simulation and Computing (ASC) Program, is to ensure that the U.S. maintains a safe, secure, and effective strategic deterrent.

97 MATHEMATICS AND COMPUTING↗

Carbon Capture Simulation Initiative: A Case Study in Multi-Scale Modeling and New Challenges

Abstract: Advanced multi-scale modeling and simulation has the potential to dramatically reduce development time, resulting in considerable cost savings. The Carbon Capture Simulation Initiative is a partnership among national laboratories, industry and universities that is developing and deploying a suite of multi-scale modeling and simulation tools including basic data submodels, steady-state and dynamic process models, process optimization and uncertainty quantification tools, an advanced dynamic process control framework, high-resolution filtered computational-fluid-dynamic (CFD) submodels, validated high-fidelity device-scale CFD models with quantified uncertainty, and a risk analysis framework. These tools and models enable basic data submodels, including thermodynamics and kinetics, to be used within detailed process models to synthesize and optimize a process. The resulting process informs the development of process control systems and more detailed simulations of potential equipment to improve the design and reduce scale-up risk. Quantification and propagation of uncertainty across scales is an essential part of these tools and models.

42 ENGINEERING↗

Turbulent transport and mixing in transitional Rayleigh-Taylor unstable flow: A priori assessment of gradient-diffusion and similarity modeling

Data from a 1152×760×1280 direct numerical simulation of a Rayleigh-Taylor mixing layer modeled after a small-Atwood-number water-channel experiment is used to investigate the validity of gradient diffusion and similarity closures a priori. The budgets of the mean flow, turbulent kinetic energy, turbulent kinetic energy dissipation rate, heavy-fluid mass fraction variance, and heavy-fluid mass fraction variance dissipation rate transport equations across the mixing layer were previously analyzed at different evolution times to identify the most important transport and mixing mechanisms. Here a methodology is introduced to systematically estimate model coefficients as a function of time in the closures of the dynamically significant terms in the transport equations by minimizing the L2 norm of the difference between the model and correlations constructed using the simulation data. It is shown that gradient-diffusion and similarity closures used for the turbulent kinetic energy K, turbulent kinetic energy dissipation rate ε, heavy-fluid mass fraction variance S, and heavy-fluid mass fraction variance dissipation rate χ equations capture the shape of the exact, unclosed profiles well over the nonlinear and turbulent evolution regimes. Using order-of-magnitude estimates for the terms in the exact transport equations and their closure models, it is shown that several of the standard closures for the turbulent production and dissipation (destruction) must be modified to include Reynolds-number scalings appropriate for Rayleigh-Taylor flow at small to intermediate Reynolds numbers. The late-time, large Reynolds number coefficients are determined to be different from those used in shear flow applications and largely adopted in two-equation Reynolds-averaged Navier-Stokes (RANS) models of Rayleigh-Taylor turbulent mixing. In addition, it is shown that the predictions of the Boussinesq model for the Reynolds stress agree better with the data when additional buoyancy-related terms are included. It is shown that an unsteady RANS paradigm is needed to predict the transitional flow dynamics from early evolution times, analogous to the small Reynolds number modifications in RANS models of wall-bounded flows in which the production-to-dissipation ratio is far from equilibrium. Although the present study is specific to one particular flow and one set of initial conditions, the methodology could be applied to calibrations of other Rayleigh-Taylor flows with different initial conditions (which may give different results during the early-time, transitional flow stages, and perhaps asymptotic stage). Finally, the implications of these findings for developing high-fidelity eddy viscosity-based turbulent transport and mixing models of Rayleigh-Taylor turbulence are discussed.

42 ENGINEERING↗

Computationally efficient CFD prediction of bubbly flow using physics-guided deep learning

To realize efficient computational fluid dynamics (CFD) prediction of two-phase flow, a multi-scale framework was proposed in this paper by applying a physics-guided data-driven approach. Instrumental to this framework, Feature Similarity Measurement (FSM) technique was developed for error estimation in two-phase flow simulation using coarse-mesh CFD, to achieve a comparable accuracy as fine-mesh simulations with fast-running feature. In this work, by defining physics-guided parameters and variable gradients as physical features, FSM has the capability to capture the underlying local patterns in the coarse-mesh CFD simulation. Massive low-fidelity data and respective high-fidelity data are used to explore the underlying information relevant to the main simulation errors and the effects of phenomenological scaling. By learning from previous simulation data, a surrogate model using deep feedforward neural network (DFNN) can be developed and trained to estimate the simulation error of coarse-mesh CFD. In a demonstration case of two-phase bubbly flow, the DFNN model well captured and corrected the unphysical “peaks” in the velocity and void fraction profiles near the wall in the coarse-mesh configuration, even for extrapolative predictions. The research documented supports the feasibility of the physics-guided deep learning methods for coarse mesh CFD simulations which has a potential for the efficient industrial design.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

DEMONSTRATION OF A DATA-DRIVEN PHYSICS-BASED APPROACH FOR COMPUTATIONALLY EFFICIENT CFD PREDICTION OF TWO-PHASE BUBBLY FLOW

To realize efficient computational fluid dynamics (CFD) prediction of two-phase flow, a multi-scale physics-guided data-driven approach, Feature Similarity Measurement (FSM) technique was developed for error estimation in two-phase flow simulation using coarse-mesh CFD, to achieve a comparable accuracy as fine-mesh simulations with fast-running feature. By defining physics-guided parameters and variable gradients as physical features, FSM has the capability to capture the underlying local patterns in the coarse-mesh CFD simulation. Massive low-fidelity data and respective high-fidelity data are used to explore the underlying information relevant to the main simulation errors and the effects of phenomenological scaling. By learning from previous simulation data, a surrogate model using deep feedforward neural network (DFNN) can be developed and trained to estimate the simulation error of coarse-mesh CFD. In a demonstration case of two-phase bubbly flow, the DFNN model well captured and corrected the unphysical “peaks” in the velocity and void fraction profiles near the wall in the coarse-mesh configuration, even for extrapolative predictions. The research documented supports the feasibility of the physics-guided deep learning methods for coarse mesh CFD simulations which has a potential for the efficient industrial design.

42 ENGINEERING↗

Transfer learning-based soybean LAI estimations by integrating PROSAIL, UAV, and PlanetScope imagery

Accurate Leaf Area Index (LAI) estimations at the soybean plot scale is achievable using high-resolution Unmanned Aerial Vehicle (UAV) imagery and field measurement samples. However, the limited coverage of UAV flights restricts large-scale remote sensing monitoring in expansive soybean fields. This study leverages the broad coverage and 3-m resolution of PlanetScope satellite imagery to extend LAI prediction from UAV to satellite scales through transfer learning, using UAV-scale LAI estimates as a benchmark to validate cross-scale consistency. To address this challenge, this study proposed the LAI-TransNet, a two-stage transfer learning framework designed for precise and scalable soybean LAI prediction across large areas, demonstrating its effectiveness in cross-scale monitoring. In Stage 1, a UAV-scale benchmark is established using PROSAIL-simulated UAV reflectance data (UAV-Sim) and field-measured soybean LAI. Traditional machine learning, deep learning, and transfer learning models are trained on a hybrid UAV-Sim and field-measured dataset (UAV-Sim_Measured), with the transfer learning model CNN-TL, fine-tuned using pre-trained weights derived from UAV-Sim, achieving the highest accuracy (R 2 = 0.81, RMSE = 0.64 m 2 /m 2 , rRMSE = 11.5 %). In Stage 2, LAI-TransNet is developed by fine-tuning the CNN-TL model on PlanetScope simulated data (PS-Sim), preprocessed via cross-domain mapping to align UAV and satellite spectral features. Real PlanetScope imagery is corrected for reflectance consistency with reference to UAV imagery spectral profiles. LAI-TransNet outperforms other deep learning models trained directly on PS-Sim (R 2 = 0.69 vs. 0.60–0.63), ensuring robust cross-scale consistency. In conclusion, by bridging UAV and satellite scales, LAI-TransNet enables large-scale soybean LAI monitoring, enhancing precision agriculture management through improved monitoring with the PlanetScope imagery.

Leaf area index (LAI)↗

Bioenergy crop models: Descriptions, data requirements and future challenges

Field studies that address the production of lignocellulosic biomass as a potential source of renewable energy are making available critical information for the development, validation, and use of bioenergy crop models. A literature survey revealed that 14 models have been developed and validated for herbaceous and woody bioenergy crops, and for Crassulacean acid metabolism (CAM) crops adapted to arid lands. These models simulate field-scale production of biomass for switchgrass (ALMANAC, EPIC, and Agro-BGC), miscanthus (MISCANFOR, MISCANMOD, and WIMOVAC), sugarcane (APSIM, AUSCANE, and CANEGRO), and poplar and willow (SECRETS and 3PG). Two models are adaptations of dynamic global vegetation models and simulate biomass yields of miscanthus and sugarcane as plant function types at regional scales (Agro-IBIS and LPJmL). A model of biomass production in CAM plants has been developed (EPI), but lacks the sophistication of the other models. Except for CAM plants, all the models include representations of leaf area dynamics, radiation interception and utilization, biomass production, and partitioning of biomass to roots and shoots. A few of the models are capable of simulating soil water, nutrient, and carbon cycle processes, making them especially useful for assessing environmental consequences (e.g., erosion and nutrient losses) associated with the field-scale deployment of bioenergy crops. Similar to other process-based models, simulations are challenged by computing and data management issues and an integrated framework for model testing and inter-comparison is needed. Considerable work remains concerning the development of models for unconventional bioenergy crops like CAM plants, generation and distribution of high-quality field data for model development and validation, and development of an integrated framework for efficient execution of large-scale simulations for use in planning regional to global sustainable bioenergy production systems.

Surendran Nair, Sujith↗

Science &Technology Review (September 2022)

Computer modeling is essential to scientific research. Models simulate natural phenomena to aid scientists in understanding their underlying principles. While the most complex models running on supercomputers may contain millions of lines of code and generate billions of data points, models never simulate reality perfectly. Experiments—in contrast—have been fundamental to the study of natural phenomena from science’s earliest days. However, some of today’s complex experiments generate too much data for the human mind to interpret.

42 ENGINEERING↗

Dynamic force matching: A method for constructing dynamical coarse-grained models with realistic time dependence

Coarse-grained (CG) models of molecular systems, with fewer mechanical degrees of freedom than an all-atom model, are used extensively in chemical physics. It is generally accepted that a coarse-grained model that accurately describes equilibrium structural properties (as a result of having a well constructed CG potential energy function) does not necessarily exhibit appropriate dynamical behavior when simulated using conservative Hamiltonian dynamics for the CG degrees of freedom on the CG potential energy surface. Attempts to develop accurate CG dynamic models usually focus on replacing Hamiltonian motion by stochastic but Markovian dynamics on that surface, such as Langevin or Brownian dynamics. However, depending on the nature of the system and the extent of the coarse-graining, a Markovian dynamics for the CG degrees of freedom may not be appropriate. In this paper, we consider the problem of constructing dynamic CG models within the context of the Multi-Scale Coarse-graining (MS-CG) method of Voth and coworkers. We propose a method of converting a MS-CG model into a dynamic CG model by adding degrees of freedom to it in the form of a small number of fictitious particles that interact with the CG degrees of freedom in simple ways and that are subject to Langevin forces. The dynamic models are members of a class of nonlinear systems interacting with special heat baths that were studied by Zwanzig [J. Stat. Phys. 9, 215 (1973)]. The properties of the fictitious particles can be inferred from analysis of the dynamics of all-atom simulations of the system of interest. This is analogous to the fact that the MS-CG method generates the CG potential from analysis of equilibrium structures observed in all-atom simulation data. The dynamic models generate a non-Markovian dynamics for the CG degrees of freedom, but they can be easily simulated using standard molecular dynamics programs. We present tests of this method on a series of simple examples that demonstrate that the method provides realistic dynamical CG models that have non-Markovian or close to Markovian behavior that is consistent with the actual dynamical behavior of the all-atom system used to construct the CG model. Both the construction and the simulation of such a dynamic CG model have computational requirements that are similar to those of the corresponding MS-CG model and are good candidates for CG modeling of very large systems.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗