SEARCH · Search NASA

Results for “compressed training”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 289 records · Page 16

All sources NASA NTRS DOE OSTI

DOE OSTI2025

FPGA-accelerated SpeckleNN with SNL for real-time X-ray single-particle imaging

We present the implementation of a specialized version of our previously published unified embedding model, SpeckleNN, for real-time speckle pattern classification in X-ray Single-Particle Imaging (SPI), using the SLAC Neural Network Library (SNL) on an FPGA platform. This hardware realization transitions SpeckleNN from a prototypic model into a practical edge solution, optimized for running inference near the detector in high-throughput X-ray free-electron laser (XFEL) facilities, such as those found at the Linac Coherent Light Source (LCLS). To address the resource constraints inherent in FPGAs, we developed a more specialized version of SpeckleNN. The original model, which was designed for broader classification across multiple biological samples, comprised ~5.6 million parameters. The new implementation, while reducing the parameter count to 64.6K (a 98.8% reduction), focuses on maintaining the model's essential functionality for real-time operation, achieving an accuracy of 90%. Furthermore, we compressed the latent space from 128 to 50 dimensions. This implementation was demonstrated on the KCU1500 FPGA board, utilizing 71% of available DSPs, 75% of LUTs, and 48% of FFs, with an average power consumption of 9.4W according to the Vivado post-implementation report. The FPGA performed inference on a single image with a latency of 45.015 microseconds at a 200 MHz clock rate. In comparison, running the same inference on an NVIDIA A100 GPU resulted in an average power consumption of ~73W and an image processing latency of around 400 microseconds. Our FPGA-accelerated version of SpeckleNN demonstrated significant improvements, achieving an 8.9 × speedup and a 7.8 × reduction in power consumption compared to the GPU implementation. Key advancements include model specialization and dynamic weight loading through SNL, which eliminates the need for time-consuming FPGA design re-synthesis, allowing fast and continuous deployment of models (re)trained online. These innovations enable real-time adaptive classification and efficient vetoing of speckle patterns, making SpeckleNN more suited for deployment in XFEL facilities. This implementation has the potential to significantly accelerate SPI experiments and enhance adaptability to evolving experimental conditions.

47 OTHER INSTRUMENTATION↗

DOE OSTI2023

Stable Solid Molecular Hydrogen above 900 K from a Machine-Learned Potential Trained with Diffusion Quantum Monte Carlo

Here, we survey the phase diagram of high-pressure molecular hydrogen with path integral molecular dynamics using a machine-learned interatomic potential trained with quantum Monte Carlo forces and energies. Besides the HCP and C2/c–24 phases, we find two new stable phases both with molecular centers in the Fmmm–4 structure, separated by a molecular orientation transition with temperature. The high temperature isotropic Fmmm–4 phase has a reentrant melting line with a maximum at higher temperature (1450 K at 150 GPa) than previously estimated and crosses the liquid-liquid transition line around 1200 K and 200 GPa.

08 HYDROGEN↗

NASA NTRS2025

Shock-Stationary Application of Pseudoshock Models During High-Amplitude Combustion-Driven Unsteadiness

The isolator pseudo-shock provides necessary compression within a dual-mode scramjet engine and buffers the engine system against unstart. Quasi-1D flux-conserved models are the state-of-the-art reduced-order model for optimization and online control of dual-mode scramjet engines. The stability and efficacy of this modeling approach is evaluated against data from a combustor-driven direct-connect experiment. The experiment exhibited strong combustor-driven unsteadiness that produced upstream propagating weak shocks into the isolator, interacting with the pseudo-shock. While this configuration resulted in unsteadiness that is atypical of standard operation, the experiment provided an opportunity to evaluate the modeling techniques in highly transient states. Such transients could occur during maneuvering or result from unexpected combustor events. A flexible quasi-1D formulation, the Fievet flux-conserved model is fit using Bayesian inference in the laboratory and shock-stationary reference frames. Model performance is analyzed using the Bayesian posteriors and model evaluations over the measured shock train speed range. It is concluded that to produce consistent isolator pressure profile estimates in this unsteady environment, the model must be implemented in a shock-stationary reference frame. Implementing this conclusion in model-based engine controllers may reduce needed unstart safety margins and increase maximum performance.

Bayesian↗

DOE OSTI2025

A Linear-Complexity Tensor Butterfly Algorithm for Compressing High-Dimensional Oscillatory Integral Operators

This paper presents a multilevel tensor compression algorithm called tensor butterfly algorithm for efficiently representing large-scale and high-dimensional oscillatory integral operators, including Green's functions for wave equations and integral transforms such as Radon transforms and Fourier transforms. The proposed algorithm leverages a tensor extension of the so-called complementary low-rank property of existing matrix butterfly algorithms. The algorithm partitions the discretized integral operator tensor into subtensors of multiple levels and factorizes each subtensor at the middle level as a Tucker-type interpolative decomposition, whose factor matrices are formed in a multilevel fashion. For a d-dimensional (d > 1) integral operator discretized into a 2d-mode tensor with n2d entries, the overall CPU time and memory requirement scale as O(nd), in stark contrast to the O(nd log n) complexity of existing matrix algorithms such as matrix butterfly algorithms and fast Fourier transforms (FFTs), where n is the number of points per direction. When comparing with other tensor algorithms such as quantized tensor train (QTT), the proposed algorithm also shows superior CPU and memory performance for tensor contraction. Remarkably, the tensor butterfly algorithm can efficiently model high-frequency Green's function interactions between two unit cubes, each spanning 512 wavelengths per direction, which represents problems of scale over 512× larger than that existing butterfly algorithms can handle, with the same amount of computation resources. On the other hand, for a problem representing 64 wavelengths per direction, which is the largest size existing algebraic matrix algorithms can handle, our tensor butterfly algorithm exhibits 200x speedups and 30× memory reduction compared with existing ones. Moreover, the tensor butterfly algorithm also permits O(nd)-complexity FFTs and Radon transforms up to d = 6 dimensions.

Kielstra, P Michael↗

NASA NTRS2017

Zoledronate Prevents Simulated Weightlessness-Induced Bone Loss in the Cancellous Compartment While Blunting the Efficacy of a Mechanical Loading Countermeasure

Astronauts using high-force resistance training while weightless show a high-turnover remodeling state within the skeletal system, with resorption and formation biomarkers being elevated. One countermeasure for the skeletal health of astronauts includes an antiresorptive of the bisphosphonate (BP) drug class. We asked, does the combination of an anti-resorptive and high-force exercise during weightlessness have negative effects on bone remodeling and strength? In this study, we developed an integrated model to mimic the mechanical strain of exercise via cyclical loading (CL) in mice treated with the BP Zoledronate (ZOL) combined with hind limb unloading (HU) to simulate weightlessness. We hypothesized that ZOL prevents structural degradation from simulated weightlessness and that CL and ZOL interact to render CL less effective. Thirty-two C57BL/6 mice (male, 16 weeks old, n=8/group) were exposed to 3 weeks of either HU or normal ambulation (NA). Cohorts of mice received one subcutaneous injection of ZOL (45μg/kg), or saline vehicle (VEH), prior to the start of HU. The right tibia was axially compressed in vivo 60x/day to 9N (+1200μstrain on the periosteal surface) and repeated 3x/week during HU. Left tibiae served as a within subject, non-compressed control. Ex vivo μCT was performed on all subjects to determine cancellous and cortical architectural parameters. Static and dynamic histomorphometry were carried out for the left and right tibiae to determine osteoclast- and osteoblast relevant surfaces. Further, micro damage was assessed in select groups by basic-fuchsin staining to test whether CL had an effect. For all assays, a multivariate (2x2x2) ANCOVA model was used to account for body weight changes. Additionally, for the tibiae, we incorporated a random effect for the subject (hence, a mixed model) to account for observations of both left and right tibiae within each subject. P < 0.05 was considered significant. In the cancellous compartment of the proximal tibial metaphysis, we observed a main effect from each independent variable, as determined by structural and histomorphometric assays. Specifically, as expected, ZOL showed an increase in the cancellous bone volume to total volume fraction (BV/TV, +32%) and trabecular number (+18%) compared to the VEH. As expected, ZOL decreased osteoclast surface (OC/BS) by -45% compared to VEH. Surprisingly, ZOL reduced mineralizing surface (MS/BS) and bone formation rate (BFR), indicators of osteoblast activity, by -40% and -54%, respectively, compared to VEH. Altogether, ZOL-treated mice displayed a low turnover state of remodeling in the metaphysis. In the context of skeletal aging, we speculate that ZOL prevented age-related cancellous strut loss during the experiment. As a main effect, as expected, HU decreased BV/TV by - 31% via reductions in both trabecular thickness (-11%) and number (-22%) compared to NA controls. Additionally, HU decreased MS/BS by -38% and bone formation rate (BFR) by -50% compared to NA controls. Altogether, these data are consistent with structural degradation resulting from imbalanced remodeling that favors resorption. As a main effect, CL increased BV/TV by +15% via increased trabecular thickness (+12%) compared to the noncompressed limb. As expected, CL increased MS/BS (+20%) and BFR (+24%), indicating osteoblast mineralization contributed to bone gains. These data show that CL provided an anabolic stimulus to the cancellous tissue. We observed unique interactions in ZOL*CL and HU*CL. First, ZOL prevented CL-induced increases in BV/TV and trabecular number, as compared to VEH. In the context of skeletal aging, these data suggest no added benefit from CL in the ZOL-treated mice. Interestingly, no microdamage was observed in mice that were unloaded and treated with ZOL (independent of CL). Secondly, HU prevented CL-induced increases in BFR, as compared to NA controls. These data suggest that either exercise is less effective or the kinetics of formation are slower during simulated weightlessness. Osteoclast surface was unchanged by either treatment. Thus, in contrast to exercising astronauts, these data do not suggest a high-turnover state in the metaphysis. To assess mechanical properties as a function of HU or ZOL, we tested the left femur in three-point bending ex vivo. As expected, HU decreased stiffness (-30%) compared to NA, and ZOL increased stiffness compared to VEH (+28%). Interestingly, HU increased the post-yield displacement, related to collagenous tensile loading, compared to NA (+20%). ZOL increased yield force (+11%) and ultimate force (+17%), which seems to explain the significant effect of ZOL increasing total energy (work-to-fracture, +15%), while not affecting the post yield displacement. Taken together, ZOL did not have detrimental affect on mechanical properties. Our integrated model simulates the combination of weightlessness, exercise-induced mechanical strain, and anti-resorptive treatment that astronauts experience during space missions. We conclude that Zoledronate was an effective countermeasure against weightlessness-induced bone loss, though zoledronate, as well as weightlessness, rendered exercise-related mechanical loading less effective.

bone↗

DOE OSTI2023

Grid-Integrated Production of Fischer-Tropsch Synfuels from Nuclear Power

Idaho National Laboratory (INL) investigates the relative economic profitability of an integrated energy system (IES) coupling an NPP with a synfuel production process at selected case study locations across the United States. In the synfuel IES, a high-temperature steam electrolysis (HTSE) plant is thermally and electrically coupled with an NPP to produce zero-carbon hydrogen. The synthetic fuel is produced from this H 2 combined with a CO 2 supply using the reverse water gas shift process followed by the Fischer-Tropsch (FT) reaction. This analysis considers a system in which the CO 2 is sourced from regional CO 2 emitters via the construction and operation of pipeline-based CO 2 supply networks. Locating the FT plant at the same site as the NPP and HTSE plants enables the NPP to provide zero-carbon heat and power to the HTSE plant and zero-carbon power to the FT plant as well as avoid the requirement for long-distance H 2 product transport from the HTSE plant to the FT plant. Hydrogen storage is used to enable the NPP to dispatch power to the electrical grid (instead of the HTSE plant) when grid demand increases, thus enabling the FT plant to continue to operate in a steady-state production mode. The ability to cease hydrogen production for several hours within each day enables the NPP to provide power to the grid to balance the electricity market during peak periods and maximize revenues for the nuclear synfuel IES. The FT process design considered has a 99% carbon conversion efficiency. The use of nuclear energy and nuclear energy-derived hydrogen enables synfuel production to achieve this high level of carbon utilization. Additionally, the life-cycle carbon emissions of the nuclear-based synfuel production process are very low, with WTW emissions of approximately 25 gCO 2 e/MJ, including steam credits (generated from FT process excess heat), and approximately 7 gCO 2 e/MJ, if steam credits are excluded. This compares favorably with the WTW emissions of 90.5 gCO 2 e/MJ for a compression-ignition, direct injection (CIDI) vehicle with a fuel economy of 31.6 miles per gallon gasoline equivalent (MPGGE), using low-sulfur diesel produced using conventional petroleum production and refining processes. Several NPPs in various regions of the U.S. are considered as case study analyses. Supply locations and transportation via pipeline of the CO 2 feedstock to the NPP site are analyzed through the National Energy Technology Laboratory (NETL) CO 2 Transport Cost model. The team finds that the amount of CO 2 generated by different sectors is sufficient for the synfuel production process at all locations considered. The CO 2 transportation costs are functions of the distance of the source to the NPP location, the CO 2 capture cost at the source, and the quantity of CO 2 transported. Historical electricity prices for the NPP case study locations are collected and analyzed. Monthly average prices, price range, and duration of negative-price periods vary among these locations. For each location, an auto-regressive moving average (ARMA) model is trained on historical electricity price data. ARMA validation is done to ensure the synthetic price distributions represent one of historical prices with high fidelity. Synthetic time series from these ARMA models are used in a coupled dispatch and system optimization in the Holistic Energy Resource Optimization Network (HERON) to compute the differential net present value (NPV) of the IES. The team finds that this econometric is positive, ranging from $14M–1.3bn (2020) depending on the location. The optimal synfuel IES configuration to obtain this increase in NPV often maximizes the size of the synfuel production process with regards to the size of the NPP. However, the team shows that the NPP still plays a stabilizing role for the grid: In periods of high prices and high loads, more electricity from the NPP is sent to the grid. A high variability of electricity prices and extreme maximum prices tend to drive up electricity production. While it requires significant investment, the synfuel IES could increase the economic profitability for the existing fleet of LWRs across the country while still maintaining the grid stabilizer role of NPPs. During its lifetime, the main costs for the nuclear synfuel IES are the carbon feedstock transportation costs, followed by the capital expenses (CAPEX) and operation and maintenance (O&M) costs while the revenue comes first from the IRA H 2 production tax credit (PTC) and then from the sales of synfuel products. The profitability of the synfuel IES is most sensitive to the value of the hydrogen PTC and the synfuel products as well as the cost of the carbon feedstock, highlighting the importance of governmental incentives regarding hydrogen, carbon emissions, and synfuel in driving the deployment of future nuclear synfuel IESs.

08 HYDROGEN↗

DOE OSTI2022

HyMARC Seedling: Optimized Hydrogen Adsorbents via Machine Learning and Crystal Engineering

This final report is presented in two parts: In Part 1, the H 2 capacities of a diverse set of 918,734 metal-organic frameworks (MOFs) sourced from 19 databases is predicted via machine learning (ML). Using only 7 structural features as input, ML identifies 8,282 MOFs with the potential to exceed the capacities of state-of-the-art materials. The identified MOFs are predominantly hypothetical compounds having low densities (<0.31 g/cm3) in combination with high surface areas (>5,300 m2/g), void fractions ($0.90), and pore volumes (>3.3 cm3/g). The relative importance of the input features are characterized, and dependencies on the ML algorithm and training set size are quantified. The most important features for predicting H 2 uptake are pore volume (for gravimetric capacity) and void fraction (for volumetric capacity). The ML models are available on the web, allowing for rapid and accurate predictions of the hydrogen capacities of MOFs from limited structural data; the simplest models require only a single crystallographic feature. In part 2, ways to improve the poor powder packing density of MOFs is discussed. More specifically, a strategy that improves packing efficiency and volumetric hydrogen gas storage density dramatically through engineered morphologies and controlled-crystal size distributions is presented that holds promise for maximizing storage capacity for a given MOF. The packing density improvement, demonstrated for the benchmark sorbent MOF-5, leads to a significant enhancement of volumetric hydrogen storage performance relative to commercial MOF-5. System model projections demonstrate that engineering of crystal morphology/size or use of a bimodal distribution of cubic crystal sizes in tandem with system optimization can surpass the 25 g/L volumetric capacity of a typical 700 bar compressed storage system and exceed the DOE targets 2020 volumetric capacity (30 g/L). Finally, a critical link between improved powder packing density and reduced damage upon compaction is revealed leading to sorbents with both high surface area and high density.

08 HYDROGEN↗

NASA NTRS2010

Analysis of Nozzle Jet Plume Effects on Sonic Boom Signature

An axisymmetric full Navier-Stokes computational fluid dynamics (CFD) study was conducted to examine nozzle exhaust jet plume effects on the sonic boom signature of a supersonic aircraft. A simplified axisymmetric nozzle geometry, representative of the nozzle on the NASA Dryden NF-15B Lift and Nozzle Change Effects on Tail Shock (LaNCETS) research airplane, was considered. The highly underexpanded nozzle flow is found to provide significantly more reduction in the tail shock strength in the sonic boom N-wave pressure signature than perfectly expanded and overexpanded nozzle flows. A tail shock train in the sonic boom signature, similar to what was observed in the LaNCETS flight data, is observed for the highly underexpanded nozzle flow. The CFD results provide a detailed description of the nozzle flow physics involved in the LaNCETS nozzle at different nozzle expansion conditions and help in interpreting LaNCETS flight data as well as in the eventual CFD analysis of a full LaNCETS aircraft. The current study also provided important information on proper modeling of the LaNCETS aircraft nozzle. The primary objective of the current CFD research effort was to support the LaNCETS flight research data analysis effort by studying the detailed nozzle exhaust jet plume s imperfect expansion effects on the sonic boom signature of a supersonic aircraft. Figure 1 illustrates the primary flow physics present in the interaction between the exhaust jet plume shock and the sonic boom coming off of an axisymmetric body in supersonic flight. The steeper tail shock from highly expanded jet plume reduces the dip of the sonic boom N-wave signature. A structured finite-volume compressible full Navier-Stokes CFD code was used in the current study. This approach is not limited by the simplifying assumptions inherent in previous sonic boom analysis efforts. Also, this study was the first known jet plume sonic boom CFD study in which the full viscous nozzle flow field was modeled, without coupling to a sonic boom propagation analysis code, from the stagnation chamber of the nozzle to the far field external flow, taking into account all nonisentropic effects in the shocks, boundary layers, and free shear layers, and their interactions at distances up to 30 times the nozzle exit diameter from the jet centerline. A CFD solution is shown in Figure 2. The flow field is very complicated and multi-dimensional, with shock-shock and shockplume interactions. At the time of this reporting, a full three-dimensional CFD study was being conducted to evaluate the effects of nozzle vectoring on the aircraft tail shock strength.

Bui, Trong↗

DOE OSTI2024

DOC-DICAM: Domain Aware One Class Defect Identification in Composite Aerostructure Material

Fiber-reinforced composites are a common material used in the design of aircraft structures due to their good tensile strength and resistance to compression. During the manufacturing process, these structures are thoroughly inspected for flaws and defects to ensure structural integrity during commercial use. Non-destructive testing (NDT) is a collection of inspection methods that allow inspectors to evaluate material without altering it. Due to the high safety standards in aerospace manufacturing, the NDT process is done manually and can be a significant bottleneck in the development workflow. In this paper, we develop an AI-based assistance tool to drastically reduce inspection time. Typical AI workflows require large amounts of annotated data, but defects rarely occur resulting in strong class imbalance. To overcome this, we formulate the problem of defect identification as an anomaly detection task in which our primary focus is learning non-defect characteristics. To do this, we develop a multi-task self-supervised learning framework that embeds problem specific domain knowledge into the deep learning model. We verify our method using fuselage data generated in a production environment. As a result, we show that our method can effectively identify defects and requires minimal training and inference time.

anomaly detection↗

NASA NTRS1993

A petabyte size electronic library using the N-Gram memory engine

A model library containing petabytes of data is proposed by Triada, Ltd., Ann Arbor, Michigan. The library uses the newly patented N-Gram Memory Engine (Neurex), for storage, compression, and retrieval. Neurex splits data into two parts: a hierarchical network of associative memories that store 'information' from data and a permutation operator that preserves sequence. Neurex is expected to offer four advantages in mass storage systems. Neurex representations are dense, fully reversible, hence less expensive to store. Neurex becomes exponentially more stable with increasing data flow; thus its contents and the inverting algorithm may be mass produced for low cost distribution. Only a small permutation operator would be recalled from the library to recover data. Neurex may be enhanced to recall patterns using a partial pattern. Neurex nodes are measures of their pattern. Researchers might use nodes in statistical models to avoid costly sorting and counting procedures. Neurex subsumes a theory of learning and memory that the author believes extends information theory. Its first axiom is a symmetry principle: learning creates memory and memory evidences learning. The theory treats an information store that evolves from a null state to stationarity. A Neurex extracts information data without a priori knowledge; i.e., unlike neural networks, neither feedback nor training is required. The model consists of an energetically conservative field of uniformly distributed events with variable spatial and temporal scale, and an observer walking randomly through this field. A bank of band limited transducers (an 'eye'), each transducer in a bank being tuned to a sub-band, outputs signals upon registering events. Output signals are 'observed' by another transducer bank (a mid-brain), except the band limit of the second bank is narrower than the band limit of the first bank. The banks are arrayed as n 'levels' or 'time domains, td.' The banks are the hierarchical network (a cortex) and transducers are (associative) memories. A model Neurex was built and studied. Data were 50 MB to 10 GB samples of text, data base, and images: black/white, grey scale, and high resolution in several spectral bands. Memories at td, S(m(sub td)), were plotted against outputs of memories at td-1. S(m(sub td)) was Boltzman distributed, and memory frequencies exhibited self-organized criticality (SOC); i.e., 'l/f(sup beta)' after long exposures to data. Whereas output signals from level n may be encoded with B(sub output) = O(-log(2)f(sup beta)) bits, and input data encoded with B(sub input) = O((S(td)/S(td-1))(sup n)), B(sup output)/B(sub input) is much less than 1 always, the Neurex determines a canonical code for data and it is a lossless data compressor. Further tests are underway to confirm these results with more data types and larger samples.

Bugajski, Joseph M.↗

DOE OSTI2020

Estimating Watershed Subsurface Permeability From Stream Discharge Data using Deep Neural Networks, Frontiers in Earth Science: Dataset

This data package contains watershed modeling inputs and outputs as well as deep neural networks training and testing results used in "Estimating Watershed Subsurface Permeability From Stream Discharge Data using Deep Neural Networks" (Cromwell et al., 2021). We train various deep neural network models with different architectures to predict subsurface permeability from stream discharge hydrograph at the watershed outlet. The training data are obtained from ensemble simulations of hydrographs corresponding to an permeability ensemble using a fully-distributed, integrated surface-subsurface hydrologic model. The trained model is then applied to estimate the permeability of the real watershed using its observed hydrograph at the outlet. Our study demonstrates that the permeabilities of the soil and geologic facies that make significant contributions to the outlet discharge can be more accurately estimated from the discharge data. Their estimations are also more robust with observation errors. Compared to the traditional ensemble smoother method, DNNs show stronger performance in capturing the nonlinear relationship between permeability and stream hydrograph to accurately estimate permeability. Our study sheds new light on the value of the emerging deep learning methods in assisting integrated watershed modeling by improving parameter estimation, which will eventually reduce the uncertainty in predictive watershed models. For detailed information regarding watershed model description and DNNs setup, please refer to Cromwell et al., 2021. The deep_learning.zip file contains the inputs and outputs for the different deep neural networks. The watershed_model.zip file contains the inputs and outputs for watershed simulation using ATS. The figures.zip file contains raw figures and their corresponding scripts. For a more detailed description, please see the README.md file within each compressed file.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2024

Accelerating the Discovery of New, Single Phase High Entropy Ceramics via Active Learning

High-entropy ceramics have garnered interest due to their remarkable hardness, compressive strength, thermal stability, and fracture toughness; yet the discovery of new high-entropy ceramics (out of a tremendous number of possible elemental permutations) still largely requires costly, inefficient, trial-and-error experimental and computational approaches. The entropy forming ability (EFA) factor was recently proposed as a computational descriptor that positively correlates with the likelihood that a 5-metal high-entropy carbide (HECs) will form the desired single phase, homogeneous solid solution; however, discovery of new compositions is computationally expensive. If you consider 8 candidate metals, the HEC EFA approach uses 49 optimizations for each of the 56 unique 5-metal carbides, requiring a total of 2744 costly density functional theory calculations. Here, we describe an orders-of-magnitude more efficient active learning (AL) approach for identifying novel HECs. To begin, we compared numerous methods for generating composition-based feature vectors (e.g., magpie and mat2vec), deployed an ensemble of machine learning (ML) models to generate an average and distribution of predictions, and then utilized the distribution as an uncertainty. Here we then deployed an AL approach to extract new training data points where the ensemble of ML models predicted a high EFA value or was uncertain of the prediction. Our approach has the combined benefit of decreasing the amount of training data required to reach acceptable prediction qualities and biases the predictions toward identifying HECs with the desired high EFA values, which are tentatively correlated with the formation of single phase HECs. Using this approach, we increased the number of 5-metal carbides screened from 56 to 15,504, revealing 4 compositions with record-high EFA values that were previously unreported in the literature. Our AL framework is also generalizable and could be modified to rationally predict optimized candidate materials/combinations with a wide range of desired properties (e.g., mechanical stability, thermal conductivity).

36 MATERIALS SCIENCE↗

NASA NTRS2000

THRSTER: A Three-Stream Ejector Ramjet Analysis and Design Tool

An engineering tool for analyzing ejectors in rocket based combined cycle (RBCC) engines has been developed. A key technology for multi-cycle RBCC propulsion systems is the ejector which functions as the compression stage of the ejector ramjet cycle. The THRee STream Ejector Ramjet analysis tool was developed to analyze the complex aerothermodynamic and combustion processes that occur in this device. The formulated model consists of three quasi-one-dimensional streams, one each for the ejector primary flow, the secondary flow, and the mixed region. The model space marches through the mixer, combustor, and nozzle to evaluate the solution along the engine. In its present form, the model is intended for an analysis mode in which the diffusion rates of the primary and secondary into the mixed stream are stipulated. The model offers the ability to analyze the highly two-dimensional ejector flowfield while still benefits from the simplicity and speed of an engineering tool. To validate the developed code, wall static pressure measurements from the Penn-State and NASA-ART RBCC experiments were used to compare with the results generated by the code. The calculated solutions were generally found to have satisfactory agreement with the pressure measurements along the engines, although further modeling effort may be required when a strong shock train is formed at the rocket exhaust. The range of parameters in which the code would generate valid results are presented and discussed.

Chue, R. S.↗

NASA NTRS2000

THRSTER: A THRee-STream Ejector Ramjet Analysis and Design Tool

Chue, R. S.↗

NASA NTRS1961

Post University On-the-Job Training for Engineers

Our national need for qualified scientists and engineers is greater now than at any other time in our history. Fortunately, we can point with pride to this need as a measure of the impact of science and technology on our way of life. In effect, we have made such rapid strides In advancing established sciences and in opening new technological fields that we have proved the value of the scientist and engineer to society, and, as a-result, have created an expanding demand for their services which we must now attempt to satisfy. This demand we face is also due to the changing skills and high degree of specialization required to perform in these new technological fields. The colleges and universities are doing their part to provide current graduates with a modern technical foundation, but we cannot afford to ignore the thousands of experienced engineers and scientists already employed by private industry and government. As employers, we have an obligation to these men and women to see that they are provided with an understanding of the latest advances that modern technology has to offer; that we develop them in particular specialty areas characteristic of a given field of work; and, equally important, that we assist them in the transition from one field to another as the technological emphasis shifts. Practically all technological industries have experienced and continue to experience rapid changes in their activities. The aerospace business, in particular, has been characterized by extremely rapid, in fact revolutionary, changes during the relatively short period of its existence0 At the National Aeronautics and Space Administration, successor to the National Advisory Committee for Aeronautics, for example, we have encountered the fun impact of a changing science and technology. Indeed, as a research organization, we have undoubtedly contributed, in some measure, to this change. Within the NASAs Lewis Research Center, we have approximately 800 research scientists and engineers who have matured professionally in an environment which is essentially one of continuous learning - an experience which comes close to being a form of post graduate training in itself. This environment, in addition to providing continuous evolutionary changes, has also provided two major revolutions which have made this development picture more complex. We will describe these environmental changes which have occurred at the Lewis Research Center and discuss the various techniques and programs we have employed to provide for the professional development of our staff. The Lewis Research Center has had an Interesting and exciting l8-year history of aerospace propulsion research and development. It began during the early years of World War II as an expansion of the Power Plant Division of the NCA Langley Center with the mission of conducting research required for the development of improved reciprocating engines and to study the associated problems of subsonic propulsion aerodynamics, It was only a few years later, however, that turbojet and ramjet propulsion and supersonic flight research became our main concern. This transition to jet type engines and higher speeds was our first major technological change. The aerodynamics of propellers became the aerodynamics of high speed turbine and compressor blades; the fuel ignition and carbon deposition problems were transferred from a cyclical or Intermittent high compression combustion chamber to a continuous combustion zone within a thin-walled metal shell; aerodynamics problems were thrust into the supersonic range; and high temperature materials began to play an increasingly critical role. Although this transition still required the same basic knowledge and principles as before, the new engine types did involve a different emphasis and variety of consideration not generally familiar to our scientists and engineers.

Manganiello, Eugene J.↗

DOE OSTI2022

Super-resolution and segmentation deep learning for breast cancer histopathology image analysis

Traditionally, a high-performance microscope with a large numerical aperture is required to acquire high-resolution images. However, the images’ size is typically tremendous. Therefore, they are not conveniently managed and transferred across a computer network or stored in a limited computer storage system. As a result, image compression is commonly used to reduce image size resulting in poor image resolution. Here, we demonstrate custom convolution neural networks (CNNs) for both super-resolution image enhancement from low-resolution images and characterization of both cells and nuclei from hematoxylin and eosin (H&E) stained breast cancer histopathological images by using a combination of generator and discriminator networks so-called super-resolution generative adversarial network-based on aggregated residual transformation (SRGAN-ResNeXt) to facilitate cancer diagnosis in low resource settings. The results provide high enhancement in image quality where the peak signal-to-noise ratio and structural similarity of our network results are over 30 dB and 0.93, respectively. The derived performance is superior to the results obtained from both the bicubic interpolation and the well-known SRGAN deep-learning methods. In addition, another custom CNN is used to perform image segmentation from the generated high-resolution breast cancer images derived with our model with an average Intersection over Union of 0.869 and an average dice similarity coefficient of 0.893 for the H&E image segmentation results. Finally, we propose the jointly trained SRGAN-ResNeXt and Inception U-net Models, which applied the weights from the individually trained SRGAN-ResNeXt and inception U-net models as the pre-trained weights for transfer learning. The jointly trained model’s results are progressively improved and promising. We anticipate these custom CNNs can help resolve the inaccessibility of advanced microscopes or whole slide imaging (WSI) systems to acquire high-resolution images from low-performance microscopes located in remote-constraint settings.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2023

Study of The Compression of an Imploding Plasma and the Magnetic- Field Flux Using Advanced, Non-Intrusive, Spectroscopic Techniques

The main motivation of the project was to measure the magnetic field in pulsed power driven plasma compression using advanced non-intrusive spectroscopic techniques. The strength of the project is that it encompasses a combination of laboratory experiments with diagnostics for both the magnetic field and plasma properties and computational modeling. The following issues were addressed: (1) Plasma and magnetic field dynamics, (2) Field compression and diffusion, (3) Energy balance in the system, (4) Onset of instabilities, and (5) Correlation of the plasma and field evolutions with the initial conditions. The unique feature of the research was the use of advanced spectroscopy to measure both the magnetic field and various plasma properties in multiple locations using both side-on and end-on views, simultaneously. The experimental system allowed for imploding different gas mixtures (which facilitated local measurements) and for flexibility in modifying the initial conditions. Two current generators were used with significantly different current rise times (200 ns and 1.6 µs) but at the same current level of 400 kA. The slower pulsed power generator (at rise time 1.6 µs) was based at the Weizmann Institute of Science (WIS) and imploded plasma into which an axial magnetic field (up to 1 T) was initially embedded. The faster current rise time generator (at rise time 200 ns) based at University of California San Diego (UCSD) was used for plasma parameter studies initially, with collaborative efforts we successfully implemented the advanced spectroscopy techniques at UC San Diego. One graduate student (Nick Aybar) and a postdoctoral scholar (Maylis Dozeries) were trained. They both are based at the Lawrence Livermore National Laboratory now.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2024

Super Resolving Unrolled Neural Networks for Remote Sensing

In remote sensing systems, the capabilities of the system are constrained by the complex interactions between size, weight, and power (SWAP) of potential designs. In electro-optical (EO) systems, examples of these critical parameters include the system’s sensitivity and resolution. Those parameters can be increased by ever larger optical apertures and focal planes but at the cost of more SWAP. Multi-image super resolution (MISR) techniques allow resolution to be enhanced via computation rather than more sophisticated optical hardware. These algorithms combine multiple images together into a single, higher resolution image, trading temporal resolution and computation for spatial resolution. Fielded MISR techniques, such as Drizzle, can require several hundred images to create a single super resolved image, implying reduced temporal resolution, increased data acquisition load, and limiting mission applications. Iterative techniques, such as model-based image reconstruction and compressive sensing, have been shown to create super resolved images using fewer images than Drizzle. They do this by posing an optimization problem that balances accuracy between a highly accurate physical model and an image model. In the case of super resolution, the physical model is defined by the relation between low resolution input images and the desired high resolution output image. The image model encodes some assumptions about the super resolved image. These assumptions are meant to suppress reconstruction artifacts that arise due to deterministic physical model error, stochastic measurement noise, and potential undersampling. In practice, the performance of iterative methods are limited by imaging models compatible with optimization. Deep learning-based methods can effectively learn image models of arbitrary complexity, but lack the theoretical explainability and robustness of iterative techniques. Consensus equilibrium (CE) generalizes the iterative techniques beyond optimization, enabling blackbox algorithms such as traditional and neural image denoisers to be used as the image model. CE-based approaches retain much of the explainability and robustness of iterative techniques while allowing the expressiveness of machine learning image models to be used. Additionally, by unrolling iterations of CE with an embedded image denoiser, the image denoiser can be further trained and specialized to the specific application with potentially higher quality reconstructions. Under this project, we demonstrated the feasibility of training an unrolled neural network based upon CE. While we didn’t train one, we showed that the CE process is differentiable and its gradient can be tractably computed. We also explored the usage of a variants of CE akin to generative neural works. Most importantly, we applied the CE framework to a number of problems including non-blind deconvolution, upsampling, single-image super resolution, MISR, event-based sensing, and saturated deconvolution. Our MISR prototype creates high quality reconstructions with an order of magnitude fewer images than previous approaches and, critically, produces these reconstructions fast enough for practical usage.

47 OTHER INSTRUMENTATION↗