Automated extraction of physical parameters from experimentally obtained thermal profiles using a machine learning approach
Explore the source record for details and available documents.
SEARCH · Search NASA
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.
Explore the source record for details and available documents.
This report documents the benchmarks developed and performed for the tools and codes developed under the Nuclear Energy Advanced Modeling and Simulation (NEAMS) program utilizing the MSRE experiment data. There are three focus areas explored in this work: fuel evolution, CFD, and system transient analyses. Fuel depletion is important for MSR reactor physics and fuel cycle analysis. Fuel depletion capability is currently being developed in the Griffin reactor physics code. To support this, a benchmark problem is developed using the MSRE data and an existing depletion code with flowing fuel analysis functionality (ADDER) to validate this new capability in Griffin. The ADDER calculation is compared with the TRITON-MSR results as well as the MSRE experimental data. This benchmark can be directly used to verify and validate the Griffin fuel depletion capability in future years. For CFD simulation, the MSRE CAD model is obtained from the Copenhagen Atomics and a mesh is generated for the MSRE inlet plenum. The Nek5000 code will be employed to find the flow distribution in the MSRE inlet plenum based on the mesh generated. For system analysis, a simple benchmark problem for three types of MSR system transients is developed. The SAM and SPECTRA codes are used to solve the simple problem, and the results from the two codes agree well. The SAM code is also employed to obtain the frequency response of the MSRE system at the 5 and 8 MWt power levels. The results agree well with the experiment measurements. These verification and validation efforts gives SAM users more confidence in MSR system transient simulations using SAM. Lastly, the SAM-Griffin Multiphysics coupling work is initiated. A simple one-way coupling of evaluating the precursor drift effect on the MSRE k-eff is demonstrated. This preliminary coupling work serves as the basis for future more complex Multiphysics simulation and modelling.
Abstract In multi‐mineral fractured rocks, the altered porous layer on the fracture surface resulting from preferential dissolution of the fast‐reacting minerals can have profound impacts on subsequent chemical‐physical alteration of the fractures. This study adopts the micro‐continuum approach to provide further understanding of reactive transport processes in the altered layer (AL), and mass exchanges with the bordering matrix and fracture. The modeling framework couples the Darcy‐Brinkman‐Stokes (DBS) solver in COMSOL Multiphysics and the geochemical modeling capability of CrunchFlow. Three‐dimensional steady state simulations with systematically varied chemical‐physical parameters of the AL were performed to examine the impacts of individual factors and processes. Our simulation results confirm previous observations that dissolution of the fast‐reacting mineral (i.e., calcite) is largely controlled by diffusion across the AL. We also show that dissolution of the slow‐reacting mineral (i.e., dolomite), which controls AL development and fracture enlargement, increases with surface area and has a complex dependence on different local rate‐limiting processes. In particular, advection can result in evident spatial variations in the local dissolution rates of dolomite, although it does not affect the bulk chemistry significantly. The difference in the spatial patterns between simulations with and without advection in the AL is more noticeable in the locations with smaller apertures, with up to 20% difference in local reaction rates. Therefore, it is important to include a full depiction of advection, diffusion, and reactions for accurately capturing local dynamics that control long‐term fracture evolution.
Vibrations generated by machine driveline components can cause excessive noise and structural damage. Magnetostrictive materials, including Galfenol (iron-gallium alloys) and Terfenol-D (terbium-iron-dysprosium alloys), are able to convert mechanical energy to magnetic energy. A magnetostrictive vibration ring is proposed, which generates electrical energy and dampens vibration, when installed in a machine driveline. A 2D axisymmetric finite element (FE) model incorporating magnetic, mechanical, and electrical dynamics is constructed in COMSOL Multiphysics. Based on the model, a parametric study considering magnetostrictive material geometry, pickup coil size, bias magnet strength, flux path design, and electrical load is conducted to maximize loss factor and average electrical output power. By connecting various resistive loads to the pickup coil, the maximum loss factors for Galfenol and Terfenol-D due to electrical energy loss are identified as 0.14 and 0.34, respectively. The maximum average electrical output power for Galfenol and Terfenol-D is 0.21 W and 0.58 W, respectively. The loss factors for Galfenol and Terfenol-D are increased to 0.59 and 1.83, respectively, by using an L-C resonant circuit.
Physics-informed neural networks (PINNs) have become a popular choice for solving high-dimensional partial differential equations (PDEs) due to their excellent approximation power and generalization ability. Recently, extended PINNs (XPINNs) based on domain decomposition methods have attracted considerable attention due to their effectiveness in modeling multiscale and multiphysics problems and their parallelization. However, theoretical understanding of their convergence and generalization properties remains unexplored. In this study, we take an initial step towards understanding how and when XPINNs outperform PINNs. Specifically, for general multilayer PINNs and XPINNs, we first provide a prior generalization bound via the complexity of the target functions in the PDE problem and a posterior generalization bound via the posterior matrix norms of the networks after optimization. Moreover, based on our bounds, we analyze the conditions under which XPINNs improve generalization. Concretely, our theory shows that the key building block of XPINN, namely, the domain decomposition, introduces a tradeoff for generalization. On the one hand, XPINNs decompose the complex PDE solution into several simple parts, which decreases the complexity needed to learn each part and boosts generalization. On the other hand, decomposition leads to less training data being available in each subdomain, and hence such a model is typically prone to overfitting and may become less generalizable. Empirically, we choose five PDEs to show when XPINNs perform better than, similar to, or worse than PINNs, hence demonstrating and justifying our new theory.
Many problems in physics, engineering, meteorology, computer science, and other fields involve a certain level of coupling between different physical quantities. With the development of advanced computational modeling, simulation, and hardware, simulations for many important multiphysics/multiscale coupling effects are now more tractable. In this project, we have developed robust, efficient, and practical numerical algorithms for such simulation software that are capable of tackling these complex mathematical models in many fundamental ways such as stability, wellposedness, and consistency. We expect that such an integrated theoretical and algorithmic framework will have a substantial impact on numerous multiscale and multiphysics modeling and simulations.
Abstract Synthetic self‐propelled nano and microparticles have a growing appeal for targeted drug delivery, collective functionality, and manipulation at the nanoscale. However, it is challenging to control their positions and orientations under confinement, e.g., in microchannels, nozzles, and microcapillaries. This study reports on the synergistic effect of acoustic and flow‐induced focusing in microfluidic nozzles. In a microchannel with a nozzle, the balance between the acoustophoretic forces and the fluid drag due to streaming flows generated by the acoustic field controls the microparticle's dynamics. This study manipulates the positions and orientations of dispersed particles and dense clusters inside the channel at a fixed frequency by tuning the acoustic intensity. The main findings are: first, this study successfully manipulates the positions and orientations of individual particles and dense clusters inside the channel at a fixed frequency by tuning the acoustic intensity. Second, when an external flow is applied, the acoustic field separates and selectively extrudes shape‐anisotropic passive particles and self‐propelled active nanorods. Finally, the observed phenomena are explained by multiphysics finite‐element modeling. The results shed light on the control and extrusion of active particles in confined geometries and enable applications for acoustic cargo (e.g., drug) delivery, particle injection, and additive manufacturing via printed self‐propelled active particles.
The interaction between the high-temperature superconducting (HTS) bulk and permanent magnet guideway (PMG) has been a topic among the existing HTS magnetic levitation systems. Up to now, the theories about the electromagnetic and force characteristics have been relatively consummated. Here, using the finite element model by COMSOL Multiphysics 5.3a, we investigated the impact of the intrinsic and external factors on the levitation characteristics of HTS maglev systems, the application laws are concluded. Factors including the critical current density and geometric dimensions of the bulk superconductor, as well as the moving velocity, are discussed. The effects of the multiple back and forth movements are also analyzed. By comparing calculation results of the distributions of the induced current and the levitation force of the bulk superconductor, we can clearly confirm the effect laws of various factors on the electromagnetic characteristics and levitation force. These observations support the physical parameters that are difficult to be tested by experiments. And they also have a significant contribution to the experimental condition selections and optimization of the high-temperature superconducting magnetic levitation.
The most important component in the aluminum industry is the aluminum reduction cell; it has received considerable interests and resources to conduct research to improve its productivity and energy efficiency. Currently, we focus on the integration of numerical simulation data and virtual reality technology to create a scientifically and practically realistic virtual aluminum reduction cell by presenting complex cell structures and physical–chemical phenomena. The multiphysical field simulation models were first built and solved in ANSYS software (ANSYS Inc., Canonsburg, PA, USA). Then, the methodology of combining the simulation results with virtual reality was introduced, and a virtual aluminum reduction cell was created. Our demonstration showed that a computer-based world could be created in which people who are not analysis experts can see the detailed cell structure in a context that they can understand easily. With the application of the virtual aluminum reduction cell, even people who are familiar with aluminum reduction cell operations can gain insights that make it possible to understand the root causes of observed problems and plan design changes in much less time.
In this work, to accelerate the establishment of fundamental understanding of the additive manufacturing (AM) process and its influence on microstructural evolution and related properties, we develop a multiphysics and multiscale modeling framework that integrates: (1) a high-fidelity powder-scale three-dimensional simulation of transient heat transfer and melt flow dynamics, (2) cellular automaton simulation of solidification grain structure and texture, (3) phase-field modeling of precipitation and dissolution of second-phase precipitate during repeated thermal cycles, and (4) microstructure-based micro- and mesoscopic elastic response calculation. Using Ti-6Al-4V as a model system, we demonstrate the application of the integrated framework to simulate complex microstructure evolution during a single-track laser powder bed fusion process and the associated mechanical response. Our modeling framework successfully captures the solidification β grain structure as a function of laser power and scanning speed, α precipitation upon subsequent cooling with different rates, and elastic response of the resulting (α + β) two-phase microstructure. The key features of solidification and second-phase precipitate microstructures, and their dependence on processing parameters, agree well with existing experimental observations. Finally, the established modeling framework is generally applicable to other metallic materials fabricated by AM.
Electroconvection is a multiphysics problem involving coupling of the flow field with the electric field as well as the cation and anion concentration fields. Here, we use electroconvection as a benchmark problem to put forward a new data assimilation framework, the DeepM&Mnet, for simulating multiphysics and multiscale problems at speeds much faster than standard numerical methods using pre-trained neural networks. We first pre-train DeepONets that can predict independently each field, given general inputs from the rest of the fields of the coupled system. DeepONets can approximate nonlinear operators and are composed of two sub-networks, a branch net for the input fields and a trunk net for the locations of the output field. DeepONets, which are extremely fast, are used as building blocks in the DeepM&Mnet and form constraints for the multiphysics solution along with some sparse available measurements of any of the fields. We demonstrate the new methodology and document the accuracy of each individual DeepONet, and subsequently we present two different DeepM&Mnet architectures that infer accurately and efficiently 2D electroconvection fields for unseen electric potentials. Furthermore, the DeepM&Mnet framework is general and can be applied for building any complex multiphysics and multiscale models based on very few measurements using pre-trained DeepONets in a “plug-and-play” mode.
Seismic analysis and risk assessment of safety-critical infrastructure like hospitals, nuclear power plants, dams, and facilities handling radioactive materials involve computationally intensive numerical models and coupled multiphysics scenarios. They are also performed in a strict regulatory environment that requires high software quality assurance standards, and in the case of safety-related nuclear facilities, a conformance to the American Society of Mechanical Engineers Nuclear Quality Assurance (NQA-1) standard. This paper introduces the open-source finite-element software, MASTODON (Multi-hazard Analysis of Stochastic Time-Domain Phenomena), which implements state-of-the-art seismic analysis and risk assessment tools in a quality-controlled environment. MASTODON is built on MOOSE (Multi-physics Object-Oriented Simulation Environment), which is a highly parallelizable, NQA-1 conforming, coupled multiphysics, finite-element framework developed at Idaho National Laboratory. MASTODON is capable of fault rupture and source-to-site wave propagation using the domain reduction method, nonlinear site response, and soil-structure interaction analysis, implicit and explicit time integration, automated stochastic simulations, and seismic probabilistic risk assessment. When coupled with other MOOSE applications, MASTODON can also solve strongly and weakly coupled multiphysics problems. This paper presents a summary of the capabilities of MASTODON and some demonstrative examples.
As part of the Global Threat Reduction Initiative, the Oak Ridge National Laboratory is evaluating conversion of fuel for the High Flux Isotope Reactor (HFIR) from high-enriched uranium to low-enriched uranium. Currently, multiphysics simulations that model fluid-structure interaction phenomena are being performed to ensure the safety of the reactor with the new fuel type. A monolithic solver that fully couples fluid and structural dynamics is used to model deflections in the new design. A classical experiment is chosen to validate the capabilities of the current solver and the method. Here, a single-plate simulation with various boundary conditions as well as a five-plate simulation are presented. Finally, use of the monolithic solver provides stable solutions for the large deflections and the tight coupling of the fluid and structure and the maximum deflections are captured accurately.
Micro-Gas-Analyzers have many applications in detecting chemical compounds present in the air. MEMS valves are used to perform sampling of gasses, as they enable control of fluid flow at the micro level. Current generation electrostatically actuated MEMS valves were tested to determine their ability to hold off a given gauge pressure with an applied voltage. Current valve designs were able to hold off 98 psi with only 82 V applied to the valves. The valves were determined to be 1.83 times more efficient than older valve designs, due to increasing the electrostatic area of the valve and trapping oxide between polysilicon layers. Newer valve designs were also proposed and modeled using ANSYS multiphysics, which should be able to hold off 100 psi with only 29 V needed. This performance would be 2.82 times more efficient than current designs, or 5.17 times more efficient than older valve designs. This will be accomplished by further increasing the valve radius and decreasing the gap between the valve boss and electrode.
This is an update to the final report filed after the extension of this program to May of 2011. The activities of the present program contributed to the goals and objectives of the Fuel Cell element of the Hydrogen, Fuel Cells and Infrastructure Technologies Program of the Department of Energy through five sub-projects. Three of these projects have focused on PEM cells, addressing the creation of carbon-based metal-free catalysts, the development of durable seals, and an effort to understand contaminant adsorption/reaction/transport/performance relationships at low contaminant levels in PEM cells. Two programs addressed barriers in SOFCs; an effort to create a new symmetrical and direct hydrocarbon fuel SOFC designs with greatly increased durability, efficiency, and ease of manufacturing, and an effort to create a multiphysics engineering durability model based on electrochemical impedance spectroscopy interpretations that associate the micro-details of how a fuel cell is made and their history of (individual) use with specific prognosis for long term performance, resulting in attendant reductions in design, manufacturing, and maintenance costs and increases in reliability and durability.
This is an update to the final report filed after the extension of this program to May of 2011. The activities of the present program contributed to the goals and objectives of the Fuel Cell element of the Hydrogen, Fuel Cells and Infrastructure Technologies Program of the Department of Energy through five sub-projects. Three of these projects have focused on PEM cells, addressing the creation of carbon-based metal-free catalysts, the development of durable seals, and an effort to understand contaminant adsorption/reaction/transport/performance relationships at low contaminant levels in PEM cells. Two programs addressed barriers in SOFCs; an effort to create a new symmetrical and direct hydrocarbon fuel SOFC designs with greatly increased durability, efficiency, and ease of manufacturing, and an effort to create a multiphysics engineering durability model based on electrochemical impedance spectroscopy interpretations that associate the micro-details of how a fuel cell is made and their history of (individual) use with specific prognosis for long term performance, resulting in attendant reductions in design, manufacturing, and maintenance costs and increases in reliability and durability.
This work summarizes the findings of a reduced order model (ROM) study performed using Sierra ROM module Pressio_Aria on Sandia National Laboratories' (SNL) Crash-Burn L2 milestone thermal model with pristine geometry. Comparisons are made to full order model (FOM) results for this same Crash-Burn model using Sierra multiphysics module Aria.