Search NASA⌕ Search

SEARCH · Search NASA

Results for “Transfer Learning”

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
All sourcesNASA NTRSDOE OSTI
NASA NTRS2009

Low Impact Docking System (LIDS)

Since 1996, NASA has been developing a docking system that will simplify operations and reduce risks associated with mating spacecraft. This effort has focused on developing and testing an original, reconfigurable, active, closed-loop, force-feedback controlled docking system using modern technologies. The primary objective of this effort has been to design a docking interface that is tunable to the unique performance requirements for all types of mating operations (i.e. docking and berthing, autonomous and piloted rendezvous, and in-space assembly of vehicles, modules and structures). The docking system must also support the transfer of crew, cargo, power, fluid, and data. As a result of the past 10 years of docking system advancement, the Low Impact Docking System or LIDS was developed. The current LIDS design incorporates the lessons learned and development experiences from both previous and existing docking systems. LIDS feasibility was established through multiple iterations of prototype hardware development and testing. Benefits of LIDS include safe, low impact mating operations, more effective and flexible mission implementation with an anytime/anywhere mating capability, system level redundancy, and a more affordable and sustainable mission architecture with reduced mission and life cycle costs. In 1996 the LIDS project, then known as the Advanced Docking Berthing System (ADBS) project, launched a four year developmental period. At the end of the four years, the team had built a prototype of the soft-capture hardware and verified the control system that will be used to control the soft-capture system. In 2001, the LIDS team was tasked to work with the X- 38 Crew Return Vehicle (CRV) project and build its first Engineering Development Unit (EDU).

LaBauve, Tobie E.↗
DOE OSTI2024

Carbon Capture through Membranes - Leveraging Multiphysics Modeling, Dimensional Analysis and Machine Learning to Scale up and Optimize Devices and Processes for Decarbonization

We study the separation performance using membrane modules through dimensional analysis (DA). We formulate the main process equations to identify relevant dimensionless numbers inherent in the physics. In particular, we identify that the critical step in the separation process is mass transfer through the selective layer. Remarkably, the dimensionless feed flow (DFfeed) emerges as a crucial factor in describing this process. Not only does DFfeed directly appear in the governing equations, but it also holds a physical significance associated with the time scales for the mass transfer across the feed side and through the selective layer. Regarding the output performance variables, we consider the recovery, stage cut, productivity and purity. In this context, we profit from experimental data and CFD simulations to evaluate the separation performance of the modules when varying the input flowrate, the scale of the module, and the CO2 permeance. These datasets enable us to establish correlations between performance metrics and the dimensionless feed flow (DFfeed). Using simple power functions of DFfeed, we obtain R2 coefficients exceeding 0.99, indicating the accuracy of the correlations built in the present work. In the future, we wish to use DA to understand key transport mechanisms, predict and control module performance, and challenge the universality of these findings by testing various gas separations across different membrane modules beyond our case study.

Pedrozo, Hector A.↗
DOE OSTI2026

Optimization of the FRIB beam dump: a hybrid genetic algorithm and reinforcement learning approach

The operational envelope of high-power-density systems, such as particle accelerators and advanced nuclear energy systems, is critically constrained by the need to manage extreme thermal loads. To address this, we present a novel hybrid optimization framework combining a genetic algorithm (GA) with a soft actor-critic (SAC) deep reinforcement learning agent. This framework was applied to a practical high-heat-flux problem: redesigning the beam dump at the Facility for Rare Isotope Beams (FRIB) for a power upgrade from 20 kW to 50 kW. The resulting design, validated by three-dimensional conjugate heat transfer simulations, suppresses hazardous hot spots and yields a markedly more uniform temperature distribution. This provides a robust operating margin, increasing the average power-handling capability by 72% relative to the current design, demonstrating the framework’s potential to solve complex thermal management challenges in both accelerator technology and advanced nuclear systems.

Accelerator↗
DOE OSTI2026

NeuroCoreX: An Open-Source FPGA-Based Spiking Neural Network Emulator with On-Chip Learning

Spiking Neural Networks (SNNs) are computational models inspired by the event-driven communication and connectivity patterns of biological neural circuits. They enable high energy efficiency and natural support for diverse architectures ranging from layered networks to small-world and graphstructured topologies. In this work, we introduce NeuroCoreX, an open-source, FPGA-based spiking neural network emulator that provides real-time, on-chip learning and flexible network organization. NeuroCoreX supports both feedforward sensory inputs streamed directly from sensors or PCs via UART and recurrent on-chip connectivity, enabling simultaneous processing and learning from external stimuli and internal network dynamics-capabilities rarely available in existing FPGA SNN platforms. The system implements a Leaky Integrate-and-Fire (LIF) neuron model with current-based synapses and supports pair-based STDP learning on both feedforward and recurrent synapses. A lightweight Python interface enables interactive configuration, live monitoring, weight read-back, and experiment control. Importantly, NeuroCoreX is tightly integrated with the SuperNeuroMAT simulator, allowing SNN models to be transferred seamlessly from software to hardware for hardware-in-the-loop development. By combining real-time plasticity, flexible connectivity, and an open-source VHDL implementation, NeuroCoreX provides an extensible and accessible platform for neuromorphic research, algorithm-hardware co-design, and energy-efficient edge intelligence.

Gautam, Ashish [ORNL]↗
DOE OSTI2026

Development of a High-Efficiency Hybrid Dry Cooler System for sCO 2 Power Cycles in CSP Applications

This project addressed a major gap in supercritical CO 2 (sCO 2 ) power cycle research by focusing on the pre-cooler, a component that had received little attention despite its significant impact on cycle efficiency and plant economics. The team developed a compact dry cooler using brazed/diffusion-bonded microchannel passages paired with formed air-side fins, advancing the technology from TRL-2 toward commercial readiness. Compared to conventional fin-tube coolers, the design cuts installation footprint by roughly half for 10+ MWth systems while achieving better heat transfer and lower approach temperatures, translating into a projected LCOE reduction from 6.04 ¢/kWh to between 5.85 and 5.94 ¢/kWh. While fabrication of an aluminum MW-scale prototype revealed brazing and sealing challenges at larger scales, lessons learned informed a subsequent 1 MWth unit that was successfully built and delivered for integration into Sandia's Gen3 Particle Pilot Plant, advancing the technology to TRL-7 with a path toward TRL-8 pending successful testing.

30 DIRECT ENERGY CONVERSION↗
NASA NTRS1971

Investigation of the neurological correlates of information reception

Animals trained to respond to a given pattern of electrical stimuli applied to pathways or centers of the auditory nervous system respond also to certain patterns of acoustic stimuli without additional training. Likewise, only certain electrical stimuli elicit responses after training to a given acoustic signal. In most instances, if a response has been learned to a given electrical stimulus applied to one center of the auditory nervous system, the same stimulus applied to another auditory center at either a higher or lower level will also elicit the response. This kind of transfer of response does not take place when a stimulus is applied through electrodes implanted in neural tissue outside of the auditory system.

Source record↗
NASA NTRS1997

(abstract) Mars Pathfinder Active Thermal Control System: Ground and Flight Performance of a Mechanically Pumped Cooling Loop

A key element of the Mars Pathfinder thermal control system is the Heat Rejection System (HRS). The HRS of Mars Pathfinder is designed to actively control the temperatures of the various parts of the spacecraft.This is achieved by mechanically circulating single-phase Freon 11 liquid through the lander and cruise electronics box heat exchangers and transferring the heat to an external radiator on the cruise stage. This is the first time in spacecraft history that a mechanically pumped cooling loop has been used on a long duration spacecraft mission. Many lessons have been learned during the testing and ground and flight operation of the HRS. This paper will present the performance of the mechanically pumped cooling loop during the ground and flight operations. Based on the lessons learned from this experience, recommendations on the design and operation of the pumped cooling loops for future space missions will be made.

Mars↗
NASA NTRS2015

SAGE III on ISS Lessons Learned on Thermal Interface Design

The Stratospheric Aerosol and Gas Experiment III (SAGE III) instrument - the fifth in a series of instruments developed for monitoring vertical distribution of aerosols, ozone, and other trace gases in the Earth's stratosphere and troposphere - is currently scheduled for delivery to the International Space Station (ISS) via the SpaceX Dragon vehicle in 2016. The Instrument Adapter Module (IAM), one of many SAGE III subsystems, continuously dissipates a considerable amount of thermal energy during mission operations. Although a portion of this energy is transferred via its large radiator surface area, the majority must be conductively transferred to the ExPRESS Payload Adapter (ExPA) to satisfy thermal mitigation requirements. The baseline IAM-ExPA mechanical interface did not afford the thermal conductance necessary to prevent the IAM from overheating in hot on-orbit cases, and high interfacial conductance was difficult to achieve given the large span between mechanical fasteners, less than stringent flatness specifications, and material usage constraints due to strict contamination requirements. This paper will examine the evolution of the IAM-ExPA thermal interface over the course of three design iterations and will include discussion on design challenges, material selection, testing successes and failures, and lessons learned.

Davis, Warren↗
DOE OSTI2026

Multi-task Parallelism for Robust Pre-training of Graph Foundation Models on Multi-source, Multi-fidelity Atomistic Modeling Data

Graph foundation models using graph neural networks promise sustainable, efficient atomistic modeling. To tackle challenges of processing multi-source, multi-fidelity data during pre-training, recent studies employ multi-task learning, in which shared message passing layers initially process input atomistic structures regardless of source, then route them to multiple decoding heads that predict data-specific outputs. This approach stabilizes pre-training and enhances a model’s transferability to unexplored chemical regions. Preliminary results on approximately four million structures are encouraging, yet questions remain about generalizability to larger, more diverse datasets and scalability on supercomputers. We propose a multi-task parallelism method that distributes each head across computing resources with GPU acceleration. Implemented in the open-source HydraGNN architecture, our method was trained on over 24 million structures from five datasets and tested on the Perlmutter, Aurora, and Frontier supercomputers, demonstrating efficient scaling on all three highly heterogeneous super-computing architectures.

Lupo Pasini, Massimiliano [ORNL] (ORCID:0000000249↗
DOE OSTI2025

A machine-learning-driven data labeling pipeline for scientific analysis in MLExchange

This study introduces a novel labeling pipeline to accelerate the labeling process of scientific data sets by using artificial intelligence (AI)-guided tagging techniques. This pipeline includes a set of interconnected web-based graphical user interfaces (GUIs), where Data Clinic and MLCoach enable the preparation of machine learning (ML) models for data reduction and classification, respectively, while Label Maker is used for label assignment. Throughout this pipeline, data can be accessed through a direct connection to a file system or through Tiled for access through Hypertext Transfer Protocol (HTTP). Our experimental results present three use cases where this labeling pipeline has been instrumental for the study of large X-ray scattering data sets in the area of pattern recognition, the remote analysis of resonant soft X-ray scattering data and the fine-tuning process of foundation models. These use cases highlight the labeling capabilities of this pipeline, including the ability to label large data sets in a short period of time, to perform remote data analysis while minimizing data movement and to enhance the fine-tuning process of complex ML models with human involvement.

Chavez, Tanny (ORCID:0000000193172896)↗
NASA NTRSDate not supplied

Gateway Power Quality Lessons Learned

Power Quality is a physical description of the electrical characteristics that allow the system to function properly without significant loss of performance or life. This physical description includes steady state voltage limits, transient voltage limits in normal/abnormal conditions, ripple voltage, stability, fault conditions, and more. All which are vital for improving reliability, ensuring stable operation, defining proper fault recovery, and ensuring a ‘plug and play’ approach to design and integration. Typically, a specification for Power Quality is created based on expected system performance or an existing standard, such as the International Space Power System Interoperability Standards (ISPSIS). One example of such a specification is the Gateway Power Quality Specification. It defines the requirements and characteristics of the 120 Volt direct current electrical power system for the Gateway Electrical Power System (EPS) and the Gateway Electrical Power Consuming Equipment (EPCE). This specification also maintains a separate requirement verification section that defines test methods for requirement verification. The test methods include analysis, test, inspection, and demonstration. These test methodologies and requirements are used to ensure that the loads operate when connected to the specified power quality and performance as defined by this specification. The challenge with developing a specification is that desired system characteristics are not always fully matured before the specification is needed and many performance requirements may be application specific. This drives the need to utilize lessons learned through extensive analysis and testing as well as historical knowledge to finalize requirements. Some example requirements where this is important are Small and Large Signal Stability, Ripple Voltage, Inrush/Surge Currents, and Fault Containment. Lessons learned are also important in the testing, analysis, and verification to ensure consistent and accurate results to verify performance. This presentation will cover the lessons learned for power quality relative to ripple, inrush/surge, fault containment, testing/verification, and more.

Power Quality↗
DOE OSTI2024

Acoustic-based monitoring and machine learning of component status for microreactor applications

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.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2026

Explainable machine learning for incipient anomaly detection in compact molten salt heat exchanger with overlapping feature distributions

High-temperature molten salt-cooled reactors (MSCRs) are a promising next-generation nuclear technology option, offering efficient power conversion and inherent safety features. However, the reliability of these systems depends on the robust operation of heat exchangers (HXs), which are susceptible to failure due to temperature gradients and channel plugging caused by fluid freezing. Conventional monitoring methods, relying on inlet and outlet measurements, lack the spatial resolution needed to detect early-stage faults. We propose a novel design of a compact salt-to-salt matrix-type HX design consisting of interleaved arrays of parallel tubes, with integrated synthetic fiber optic distributed temperature sensing (DTS) to enable localized detection of incipient faults. To evaluate performance of this design, we generate high-fidelity synthetic data using heat transfer computational modeling to simulate channel plugging, and introduce sensor noise for realistic modeling of measurements. The dataset comprises of 97% normal operation and 3% anomaly cases, with each anomaly class representing 1% of the data. These early anomalies result in overlapping temperature profiles between normal and faulty channels, producing a non-separable dataset that challenges traditional classification techniques. We benchmark eight supervised machine learning (ML) models and demonstrate that XGBoost achieves the highest performance. To improve transparency, we develop an explainability framework combining Shapley values and partially ordered sets (POSETs) to quantify and structurally analyze feature importance. This approach identifies both dominant predictors and ambiguous feature relationships, enhancing trust and interpretability. Our results highlight the potential of combining DTS and explainable ML with intelligent feature selection to improve predictive maintenance and ensure operational resilience in advanced nuclear systems.

Prantikos, Konstantinos [Argonne National Laborato↗
NASA NTRS2005

Auxiliary Propulsion Activities in Support of NASA's Exploration Initiative

The Space Launch Initiative (SLI) procurement mechanism NRA8-30 initiated the Auxiliary Propulsion System/Main Propulsion System (APS/MPS) Project in 2001 to address technology gaps and development risks for non-toxic and cryogenic propellants for auxiliary propulsion applications. These applications include reaction control and orbital maneuvering engines, and storage, pressure control, and transfer technologies associated with on-orbit maintenance of cryogens. The project has successfully evolved over several years in response to changing requirements for re-usable launch vehicle technologies, general launch technology improvements, and, most recently, exploration technologies. Lessons learned based on actual hardware performance have also played a part in the project evolution to focus now on those technologies deemed specifically relevant to the Exploration Initiative. Formal relevance reviews held in the spring of 2004 resulted in authority for continuation of the Auxiliary Propulsion Project through Fiscal Year 2005 (FY05), and provided for a direct reporting path to the Exploration Systems Mission Directorate. The tasks determined to be relevant under the project were: continuation of the development, fabrication, and delivery of three 870 lbf thrust prototype LOX/ethanol reaction control engines; the fabrication, assembly, engine integration and testing of the Auxiliary Propulsion Test Bed at White Sands Test Facility; and the completion of FY04 cryogenic fluid management component and subsystem development tasks (mass gauging, pressure control, and liquid acquisition elements). This paper presents an overview of those tasks, their scope, expectations, and results to-date as carried forward into the Exploration Initiative.

Best, Philip J.↗
NASA NTRS2009

Experimental Results of Integrated Refrigeration and Storage System Testing

Launch operations engineers at the Kennedy Space Center have identified an Integrated Refrigeration and Storage system as a promising technology to reduce launch costs and enable advanced cryogenic operations. This system uses a close cycle Brayton refrigerator to remove energy from the stored cryogenic propellant. This allows for the potential of a zero loss storage and transfer system, as well and control of the state of the propellant through densification or re-liquefaction. However, the behavior of the fluid in this type of system is different than typical cryogenic behavior, and there will be a learning curve associated with its use. A 400 liter research cryostat has been designed, fabricated and delivered to KSC to test the thermo fluid behavior of liquid oxygen as energy is removed from the cryogen by a simulated DC cycle cryocooler. Results of the initial testing phase focusing on heat exchanger characterization and zero loss storage operations using liquid oxygen are presented in this paper. Future plans for testing of oxygen densification tests and oxygen liquefaction tests will also be discussed. KEYWORDS: Liquid Oxygen, Refrigeration, Storage

Notardonato, W. U.↗
NASA NTRS2006

Lessons Learned from Daily Uplink Operations during the Deep Impact Mission

The daily preparation of uplink products (commands and files) for Deep Impact was as problematic as the final encounter images were spectacular. The operations team was faced with many challenges during the six-month mission to comet Tempel One of the biggest difficulties was that the Deep Impact Flyby and Impactor vehicles necessitated a high volume of uplink products while also utilizing a new uplink file transfer capability. The Jet Propulsion Laboratory (JPL) Multi-Mission Ground Systems and Services (MGSS) Mission Planning and Sequence Team (MPST) had the responsibility of preparing the uplink products for use on the two spacecraft. These responsibilities included processing nearly 15,000 flight products, modeling the states of the spacecraft during all activities for subsystem review, and ensuring that the proper commands and files were uplinked to the spacecraft. To guarantee this transpired and the health and safety of the two spacecraft were not jeopardized several new ground scripts and procedures were developed while the Deep Impact Flyby and Impactor spacecraft were en route to their encounter with Tempel-1. These scripts underwent several adaptations throughout the entire mission up until three days before the separation of the Flyby and Impactor vehicles. The problems presented by Deep Impact's daily operations and the development of scripts and procedures to ease those challenges resulted in several valuable lessons learned. These lessons are now being integrated into the design of current and future MGSS missions at JPL.

uplink↗
DOE OSTI2024

Predicting Partial Atomic Charges in Metal–Organic Frameworks: An Extension to Ionic MOFs

Molecular simulation is an invaluable tool to predict and understand the usage of metal–organic frameworks (MOFs) for gas storage and separation applications. Accurate partial atomic charges, commonly obtained from density functional theory (DFT) calculations, are often required to model the electrostatic interactions between the MOF and adsorbates, especially when the adsorbates have dipole or quadrupole moments, such as water and CO 2 . Machine learning (ML) models have been previously employed to predict partial charges and avoid the computational cost associated with DFT calculations. However, previous ML models suffer from small training data sets, which limit their scope of application. In this work, we introduce two novel machine learning models, PACMOF2-neutral and PACMOF2-ionic, aimed at predicting the density-derived electrostatic and chemical (DDEC6) partial atomic charges for both neutral and ionic MOFs. These models not only yield DFT-level accuracy at a fraction of the computational cost but also demonstrate a remarkable improvement in prediction of adsorption, as validated with grand canonical Monte Carlo simulations. Furthermore, the robustness and fast computational time of the PACMOF2 models, along with their transferability to other porous materials such as covalent organic frameworks and zeolites, underscores their potential in high-throughput screening of MOFs for diverse applications.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Scaling Ensembles of Data-Intensive Quantum Chemical Calculations for Millions of Molecules

Deep learning models are efficient computational tools that can accelerate the inverse design of molecules with desired functional properties by generating predictions at a fraction of the time required by traditional quantum chemical approaches. To ensure that a model maintains accuracy and transferability across broad regions of the chemical space explored during the inverse design, it must be trained on massively large volumes of simulation data. This requires running large-scale ensemble quantum chemical calculations on high-performance computing (HPC) systems for data collection. However, the efficient execution of such large ensemble calculations and the management of large volumes of output data require tools that can judiciously utilize computational resources and manage metadata overhead on the file system. Therefore, we present a high-performance, scalable, ensemble management framework for performing data-intensive quantum chemical electronic structure calculations for organic molecules. This framework provides abstractions to plug different ab initio, first principles, and first principles-based semi-empirical methods and executes them efficiently at large scale on HPC systems. It dynamically distributes tasks to resources and uses tiered storage for managing large collections of files. We employed this framework to process over ten million organic molecules and generate open-source datasets that provide UV-vis absorption spectra by running time-dependent density-functional tight-binding calculations. It is the largest database containing molecular optical spectra that were simulated with quantum chemical methods in a consistent manner.

Mehta, Kshitij↗