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.
OJINN Is an easy-to-use deep neural network model that requires fewer user-specified hyper-parameters than traditional neural networks. The algorithm leverages decision trees trained on the data to determine an appropriate deep neural network architectures and weight Initializations. Optional functions also select the learning rate, batch size. and number of training Iterations necessary to create an accurate model.
Existing interactive visualization tools for deep learning are mostly applied to the training, debugging, and refinement of neural network models working on natural images. However, visual analytics tools are lacking for the specific application of x-ray image classification with multiple structural attributes. In this paper, we present an interactive system for domain scientists to visually study the multiple attributes learning models applied to x-ray scattering images. It allows domain scientists to interactively explore this important type of scientific images in embedded spaces that are defined on the model prediction output, the actual labels, and the discovered feature space of neural networks. Users are allowed to flexibly select instance images, their clusters, and compare them regarding the specified visual representation of attributes. The exploration is guided by the manifestation of model performance related to mutual relationships among attributes, which often affect the learning accuracy and effectiveness. The system thus supports domain scientists to improve the training dataset and model, find questionable attributes labels, and identify outlier images or spurious data clusters. Case studies and scientists feedback demonstrate its functionalities and usefulness.
4 page paper submitted to ANS Student conference. Summary of paper similar to the following abstract: The ability to predict how a reactor will operate, understand when anomalous conditions arise, and ensure a reactor is being operated as expected is crucial for deploying new nuclear facilities. Digital twins serve as a unique solution to recognizing reactor behavior; however, they require data to be useful. For next-generation reactors, this data may not currently be available. To explore how synthetic physics-informed reactor data can be used to predict reactor operations, a recurrent neural network was implemented for the Idaho State University AGN-201 digital twin. The goal of this work is to determine how synthetic data can be used to train a recurrent neural network model for predicting the reactor power of the AGN-201. The recurrent neural network was validated using both synthetic and real operational data. We envision this approach will help bridge the gap between the virtual and physical sides of a digital twin, where reactor physics models based on as-built data can be corrected for actual operating parameters to ensure the virtual model mirrors reality.
Abstract With rapid progress across platforms for quantum systems, the problem of many-body quantum state reconstruction for noisy quantum states becomes an important challenge. There has been a growing interest in approaching the problem of quantum state reconstruction using generative neural network models. Here we propose the ‘attention-based quantum tomography’ (AQT), a quantum state reconstruction using an attention mechanism-based generative network that learns the mixed state density matrix of a noisy quantum state. AQT is based on the model proposed in ‘Attention is all you need’ by Vaswani et al (2017 NIPS ) that is designed to learn long-range correlations in natural language sentences and thereby outperform previous natural language processing (NLP) models. We demonstrate not only that AQT outperforms earlier neural-network-based quantum state reconstruction on identical tasks but that AQT can accurately reconstruct the density matrix associated with a noisy quantum state experimentally realized in an IBMQ quantum computer. We speculate the success of the AQT stems from its ability to model quantum entanglement across the entire quantum system much as the attention model for NLP captures the correlations among words in a sentence.
When evaluated at the scale of individual processes, next-generation technologies may be more energy and emissions intensive than current technology. Furthermore, many advanced technologies have the potential to reduce material and energy consumption in upstream or downstream processing stages. In order to fully understand the benefits and consequences of technology deployment, next-generation technologies should be evaluated in context, as part of a supply chain. This work presents the Materials Flow through Industry (MFI) supply chain modeling tool. The MFI tool is a cradle-to-gate linear network model of the US industrial sector that can model a wide range of manufacturing scenarios, including changes in production technology and increases in industrial energy efficiency. The MFI tool was developed to perform supply chain scale analyses in order to quantify the impacts and benefits of next-generation technologies and materials at that scale. For the analysis presented in this paper, the MFI tool is utilized to explore a case study comparing three lightweight vehicle supply chains to the supply chain of a conventional, standard weight vehicle. Several of the lightweight vehicle supply chains are evaluated under manufacturing scenarios that include next-generation production technologies and next-generation materials. Results indicate that producing lightweight vehicles is more energy and emission intensive than producing the non-lightweight vehicle, but the fuel saved during vehicle use offsets this increase. In this case study, greater reductions in supply chain energy and emissions were achieved through the application of the next-generation technologies than from application of energy efficiency increases.
Unsteady aerodynamics will be an important part of the floating wind turbines of the future operating under high shear across the rotor disk coupled with platform motion and atmospheric turbulence. We develop an unsteady aerodynamics and dynamic stall model using a long short-term memory variant of recurrent neural networks. The neural network model is trained using the oscillating airfoil data set from Ohio State University. The predictions from our machine learning (ML)-based model show good agreement with the experimental data and other state-of-the-art dynamic stall models for a wide range of airfoils, Reynolds numbers and reduced frequencies. In some cases the predictions are better than the Beddoes-Leishman model implementation in OpenFAST, when using the default coefficients. The ML-based model is also able to capture the key physics associated with dynamic stall, such as the precedence of moment stall before lift stall and cycle-to-cycle variations in the aerodynamic response. The new unsteady aerodynamics model is expected to improve prediction of fatigue loads for yaw-based wake-steering control scenarios in actuator-line and actuator-disc simulations of wind farms. Our methodology for training the ML-model provides a pathway for improving design level tools using high-fidelity computational fluid dynamics (CFD) simulations in the future.
Neutrino-nucleus scattering cross sections are critical theoretical inputs for long-baseline neutrino oscillation experiments. However, robust modeling of these cross sections remains challenging. For a simple but physically motivated toy model of the DUNE experiment, we demonstrate that an accurate neural-network model of the cross section—leveraging only Standard-Model symmetries—can be learned from near-detector data. We perform a neutrino oscillation analysis with simulated far-detector events, finding that oscillation analysis results enabled by our data-driven cross-section model approach the theoretical limit achievable with perfect prior knowledge of the cross section. We further quantify the effects of flux shape and detector resolution uncertainties as well as systematics from cross-section mismodeling. This proof-of-principle study highlights the potential of future neutrino near-detector datasets and data-driven cross-section models.
1. Segmentation We provide scripts for the model definition of the U-net architecture adapted from: https://github.com/jvanvugt/pytorch-unet/blob/master/unet.py We developed scripts for preparing the lung segmentation data set. We developed scripts for training the U-Net architecture. We developed scripts for applying the trained U-Net model to perform the ROI Hide and Seek protocol on the classification dataset to create the modified dataset. 2. Classification We provide scripts for the training of the COVID-Net models provided by Linda Wang, this code is adapted from her github repository: https://github.com/lindawangg/COVID-Net/tree/d7b36831d854f57de5bc7557217f5439e86e016f. These scripts were modified to save training log information as well as to load the provided models in their github repo. We developed scripts for training standard Neural Network Models (resnet 50, vgg 11, Alexnet) on the COVID datsets along with the ROI Hide and Seek altered datasets.
Neutrino-nucleus scattering cross sections are critical theoretical inputs for long-baseline neutrino oscillation experiments. However, robust modeling of these cross sections remains challenging. For a simple but physically motivated toy model of the DUNE experiment, we demonstrate that an accurate neural-network model of the cross section -- leveraging Standard Model symmetries -- can be learned from near-detector data. We then perform a neutrino oscillation analysis with simulated far-detector events, finding that the modeled cross section achieves results consistent with what could be obtained if the true cross section were known exactly. This proof-of-principle study highlights the potential of future neutrino near-detector datasets and data-driven cross-section models.
Neutrino-nucleus scattering cross sections are critical theoretical inputs for long-baseline neutrino oscillation experiments. However, robust modeling of these cross sections remains challenging. For a simple but physically motivated toy model of the DUNE experiment, we demonstrate that an accurate neural-network model of the cross section—leveraging only Standard-Model symmetries— can be learned from near-detector data. We perform a neutrino oscillation analysis with simulated far-detector events, finding that oscillation analysis results enabled by our data-driven cross-section model approach the theoretical limit achievable with perfect prior knowledge of the cross section. We further quantify the effects of flux shape and detector resolution uncertainties as well as systematics from cross-section mismodeling. This proof-of-principle study highlights the potential of future neutrino near-detector datasets and data-driven cross-section models.
When evaluated at the process level, next-generation technologies may be more energy and emissions intensive than current technology. However, many advanced technologies have the potential to reduce material and energy consumption in upstream or downstream processing stages. In order to fully understand the benefits and consequences of technology deployment, next-generation technologies should be evaluated in context, as part of a supply chain. This work presents the Material Flows through Industry (MFI) scenario modeling tool. The MFI tool is a cradle-to-gate linear network model of the U.S. industrial sector that can model a wide range of manufacturing scenarios, including changes in production technology, increases in industrial energy efficiency, and substitution between functionally equivalent materials. The MFI tool was developed to perform supply chain scale analyses in order to quantify the impacts and benefits of next-generation technologies and materials at that scale. For the analysis presented in this paper, the MFI tool is utilized to explore a case study comparing a steel supply chain to the supply chains of several functionally equivalent materials. Several of the alternatives to the baseline steel supply chain include next-generation production technologies and materials. Results of the case study show that aluminum production scenarios can out-perform the steel supply chain by using either an advanced smelting technology or an increased aluminum recycling rate. The next-generation material supply chains do not perform as well as either aluminum or steel, but may offer additional use phase reductions in energy and emissions that are outside the scope of the MFI tool. Future work will combine results from the MFI tool with a use phase analysis.
Line profiles of optically thin extreme-ultraviolet emission lines observed in a quiet sun region at positions within and above the white-light limb with the NRL slit spectrograph (S082-B) on Skylab are discussed. Absolute line intensities and full widths at half-maximum are presented for lines formed over the temperature range from about 10,000 to 22,000 K. The line intensities are compared with the predictions of simple atmospheric models consisting of a spicule component and a thin spherically symmetric or network models, but can be explained by assuming that the emission arises from spicule-like inhomogeneities. Random mass-motion velocities are calculated. The velocity increases with increasing temperature of line formation. Near the limb and above about 4000 K the calculated velocity is consistent with the predictions of a constant acoustic flux passing through the transition zone. For the ions formed at temperatures not less than about 63,000 K, the velocity is found to increase with increasing height above the white-light limb.
The human innate immune response, particularly the type-I interferon (IFN) response, is highly robust and effective first line of defense against virus invasion. IFN molecules are produced and secreted frominfected cells upon virus infection and recognition. They then act as signaling/communicationmolecules to activate an antiviral response in neighboring cells so that those cells become refractory to infection. Previous experimental studies have identified the detailed molecular mechanisms for the IFN signaling and response. However, the principles underlying how host cells use IFN to communicate with each other to collectively and robustly halt an infection is not understood. Here we take a multiplex network modeling approach to provide a theoretical framework to identify key factors that determine the effectiveness of the IFN response against virus infection of a host. In this approach, we consider the virus spread among host cells and the interferon signaling to protect host cells as a competition process on a two-layer multiplex network. We focused on two types of network topology, i.e., the Erdős-Rényi (ER) network and the Geometric Random (GR) network, which represent the scenarios when infection of cells is mostly well mixed (e.g., in the blood) and when infection is spatially segregated (e.g., in tissues), respectively.We show that in general, the IFN response works effectively to stop viral infection when virus infection spreads spatially (a most likely scenario for initial virus infection of a host at the peripheral tissue). Importantly, we show that the effectiveness of the IFN response is robust against large variations in the distance of IFN diffusion as long as IFNs diffuse faster than viruses and they can effectively induce antiviral responses in susceptible host cells. This suggests that the effectiveness of the IFN response is insensitive to the specific arrangement of host cells in peripheral tissues. Thus, our work provides a quantitative explanation of why the IFN response can serve an effective and robust response in different tissue types to a wide range of viral infections of a host.
An artificial convolutional neural network was created to efficiently mimic a micromechanics model, the High Fidelity Generalized Method of Cells, for use in multi-scale structural finite element analysis. The network was found to quickly and accurately replicate the stiffness predicted by the micromechanics model using a 2D image of an idealized representative volume element of a fiber/matrix microstructure. The long-term goal of this work is to efficiently apply multi-scale methods for predicting the damage progression of a composite structure.
This report provides a description and assessment of recent efforts to couple acoustic-based experimental measurements and characterization with machine learning models in order to enhance structural health monitoring capabilities for nuclear microreactors. With resilient embedded sensors in development by others supported by programs funded by the US Department of Energy’s Office of Nuclear Energy, the work described herein builds upon ongoing efforts to improve non-destructive testing technology that relates measured acoustic signatures to component stresses and/or structural defects, using a combination of new experimental measurements and machine learning architectures. The experimental procedure remained similar to that developed for the previous year’s demonstration of damage detection by the authors, with the same damaged sample tested under similar applied stress conditions. Notably, a new mounting fixture was designed and implemented to improve measurement consistency and a more sophisticated laser Doppler vibrometer was employed to make high-fidelity vibration measurements. Two nominally identical sets of training data were collected for each experimental setup to better understand the repeatability of the experiment and to better test the generality of trained neural network models. Additionally, we obtained new high-quality 3D mode shapes of the damaged test article at various stress and excitation levels, providing greater insights into the physical response of the sample during testing. Previously, we demonstrated that a machine learning model based on a convolutional neural network can predict structural details of an artificially introduced interface (intact, rough cut, smooth cut), and the applied torque level. In this study, we have transitioned to graph-based neural network architectures to better develop and test a flexible framework that is more suitable to being transferred away from controlled benchtop experiments and into more applied settings where less-structured data inputs may be expected. In general, performance testing of a graph neural network on frequency-domain representations of the data indicates strong and consistent identification of test conditions for datasets recorded on damaged components. With goals of predicting damage location and other changing experimental conditions using limited datasets, predictive models using a graph neural network architecture correctly predicted the applied torque level with an accuracy of 85% using only a single measurement point and predicted within one torque level in 95% of test windows. Predictions of damage location had limited success due to the symmetry and minimal number of the damage scenarios presented during model training. Results were ambiguous as to whether the model could detect the location of the artificial damage, or if it was instead learning the location of a given measurement point on the part and subsequently detecting which points were closest to the location of the damage. This finding will be factored into upcoming planned work on damaged graphite components, where new experimental tests with a larger number and variety of damage scenarios are expected to provide improved validation of recent developments in monitoring methodology.
A content-addressable memory (CAM) stores key-value associations such that the key is recalled by providing its associated value. While CAM recall is traditionally performed using recurrent neural network models, we show how to solve this problem using adiabatic quantum optimization. Our approach maps the recurrent neural network to a commercially available quantum processing unit by taking advantage of the common underlying Ising spin model. We then assess the accuracy of the quantum processor to store key-value associations by quantifying recall performance against an ensemble of problem sets. We observe that different learning rules from the neural network community influence recall accuracy but performance appears to be limited by potential noise in the processor. The strong connection established between quantum processors and neural network problems supports the growing intersection of these two ideas.
The hot compressive tests of 6082 aluminum alloy were conducted on a Gleeble-3500 thermomechanical simulator at temperature ranges of 380-530 °C and strain rate range of 0.01-10 s{sup −1}. The constitutive analysis and microstructural evolution of the alloy were investigated. It was indicated that the peak stress increased with increasing strain rate and decreasing temperature. Dynamic recovery and dynamic recrystallization lead to the softening behavior of the alloy. In order to characterize the flow behavior of this alloy, some models were established based on the experimental data including the phenomenological Arrhenius-type model, the physically based Estrin and Mecking (EM) model for work hardening and dynamic recovery, and the EM model, which was combined with the Avrami equation for dynamic recrystallization. An artificial neural network model was also established to predict the flow stress. The results indicate that the Arrhenius-type model is more simple and more efficient than the EM + Avrami model. Moreover, the well-trained ANN model has the best predicting performance.
Abstract Serpentine channels adjacent to a thin, porous medium are a potentially attractive alternative to a conventional thick flow‐through electrode for redox flow batteries. The hydrodynamics of serpentine flow fields were investigated with computational fluid dynamics, a two‐dimensional model of the porous electrode based on Darcy's law, and a resistance network model at the scale of the active area. Predictions from the three models were used to map the available design space. The optimal electrode thickness, in terms of minimizing nonuniformity, was identified and compared to the result for an interdigitated flow field. Serpentine favors thicker electrodes and higher flows than interdigitated, in qualitative agreement with experimental findings. Furthermore, interdigitated designs deliver more uniform intraelectrode velocities and lower overall pressure drops than serpentine flow fields.