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.
Project Overview and Goal: Create a Python Toolkit for UQ (PyTUQ) to facilitate the interoperability of FASTMath UQ tools (Dakota, UQTk, KLPC), as well as other ASCR funded capabilities for UQ workflows.
Large Language Models (LLMs) struggle out of the box when answering factually about detailed questions, especially in domains that are sparsely represented in their training data. This causes hallucinations and reduces reliability making it difficult for them to be used in practice. This work shows that using RAG techniques can improve factual accuracy and reliability, allowing for the application of LLMs in specialized areas, even when those areas that aren’t extensively covered in their initial training.
Type-5 wind turbines are unique in their use of a permanent magnet synchronous generator, as well as their use of a hydraulic torque converter. This architecture presents an opportunity to provide steady and grid-ready energy without the need for a power converter. With infrastructure continuity and reliability being an important topic amongst renewable energies, researchers have been prompted to further investigate the benefits of type-5 turbines’ unique electromechanical configuration on stable electricity generation. Researchers involved in the WindSG project, SG standing for synchronous generator, are aiming to model a type-5 turbine using Real Time Digital Simulation (RTDS) to evaluate its efficacy in the grid. RSCAD, the software run on the RTDS, comes pre-loaded with electrical and electromechanical components to help simulate electrical generation and grid conditions. However, within this repertoire there is a lack of a component to represent a gearbox with high-fidelity. Within RSCAD’s case studies, the gearbox is often represented simply by a gear ratio value. This presented the task of developing a high-fidelity gearbox model in RSCAD for use in the larger RTDS type-5 wind turbine model. This presentation describes a method of developing a lumped parameter mathematical model to represent a planetary-parallel-parallel gearbox in RSCAD for use in RTDS.
Previous work on exhaustive search methodologies for extracting best-match parameters pertaining to dynamic surface quantities from PDV was done by cross-correlating synthetically generated PDV waveforms with observed counterparts using the circular-convolution theorem. This work was further developed into an open-source PDV analysis toolkit called CCPDVANALYSIS which expands upon and enhances the previously tested methods by parallelizing serial algorithmic components and incorporating a comprehensive script library for different flavors of instantaneous frequency functions utilized in generating synthetic PDV waveforms. Results of these enhancements have been shown to markedly decrease execution times of exhaustive search and extraction algorithms and produce improved velocity recoveries for low-velocity and dynamically varying velocity signals. The CCPDVANALYSIS script library demonstrates an advanced method for extracting velocities from low-velocity and non-constant velocity signals further extending and improving the methods beyond capabilities of traditional frequency domain tools.
As part of a U.S. DOE Geothermal Technologies Office funding opportunity, Geysers Power Company, LLC, an indirect subsidiary of Calpine Corporation, partnered with Sandia National Labs, EGI at the University of Utah, and Texas A&M University to demonstrate increased drilling performance at The Geysers Geothermal Field. The performance target in the drilling demonstrations is at least a 25% improvement in rates of penetration with increased time on bottom for each bit. The planned drilling demonstrations are being conducted as part of an existing drilling campaign intended to enhance reservoir utilization. A major goal of the project is to assess the effectiveness of implementing mechanical specific energy (MSE) and drilling dysfunction diagnosis and remediation in these challenging environments, as well as alternate bit technologies. The first demonstration well has been completed, with a total of 15 PDC bit runs in the 17.5”, 12.25” and 8.5” sections. Initial analysis shows ROP gains in all three hole sections, especially in the 17.5” and 12.25” sections, compared with conventional roller cone bit runs in the demonstration well and offset wells. However in the 8.5” hole, wear and damage to the PDC bits resulted in relatively short bit runs. Analysis is underway to take advantage of the positive results and remediate the challenges.
Liquid Metal Jetting (LMJ) is a metal additive manufacturing technique that involves jetting molten metal droplets at high frequencies to build solid metal parts. As an alternative to industry standard Lazer Powder Bed Fusion (LPBF) and Direct Metal Writing (DMW) techniques, LMJ poses significant advantages including no powder feed stock, no post sintering process, very high deposition rates, and the capability to print various metals. This summer, I was tasked with improving my old system to process at much higher resolution, Increasing the capture rate, and implementing this diagnostic on various LMJ setups. This system provides a vital longitudinal study to understand individual droplets in the LMJ process.
This project conducts a comparative analysis of DNA LLM classification techniques using Evo2, Grover, and UTRML, focusing on intra-layer feature extraction in Evo2. By extracting features from multiple layers of Evo2 and integrating them into an autoencoder stack with a binary classification head, we evaluate its effectiveness in classifying genomic sequences compared to smaller DNA language models. My findings demonstrate that Evo2 outperforms Grover and UTRML in classification accuracy on a dataset provided by department 08625, CAO2021, while UTRML offers competitive performance with lower computational costs. This study highlights the potential of advanced embedding techniques in enhancing genomic data analysis and informs future research in bioinformatics.
The application of increased pressure during the curing process of B-stage epoxy has been hypothesized to enhance the adhesive properties and bond strength in printed wiring board (PWB) assembly. This study aimed to investigate the correlation between pressure application and the performance of B-stage epoxy.
At mm-wave frequencies, signals need to be supported by specialized, low-loss transmission lines such as waveguides. Substrate-integrated waveguides (SIWs) combine the good electrical properties of rectangular waveguides with the compactness of planar transmission lines. Additionally, SIWs can be fabricated using standard PCB manufacturing techniques and can be easily integrated with other mm-wave components. This study explores SIWs in the 50-75 GHz range, with designs optimized for simple fabrication.
Explore the source record for details and available documents.
At Sandia National Laboratories, there is a high level of importance placed on identifying and developing solutions to the nation’s current and future security problems. One of these problem would be hardware electronics validation and verification.
Cement production involves the decomposition of limestone (calcium carbonate) at high temperatures (~900°C) to produce CaO, a major constituent in Portland cement (60-70%). Our unique approach will eliminate the high temperature process and introduce a low-temperature electrochemical process to produce Ca(OH) 2 which can be converted into CaO through dehydration process. This program will allow SRNL to become a leading organization in an open and unexplored field that addresses many of the technical challenges.
Abstract not provided.
Fuel debris removal operations at the Fukushima-Daiichi Nuclear Power Station (1F) present significant complexities in many engineering disciplines this presentation will cover this issue among others.
Abstract not provided.
Nuclear data (ND) is vital to predictive simulations such as those completed with MCNP®. ND sensitivities are used to optimize the design of benchmark experiments. Diverse benchmarks (including neutron noise) are required to further our understanding of ND. Measured and simulated data of a 4.5-kg Pu sphere was used to calculate neutron noise parameters and associated ND Sensitivities
Nuclear data (ND) underpins predictive neutron transport simulations like MCNP. The current nuclear data pipeline is time intensive and iterative. PARADIGM aims to reduce uncertainties by restructuring the ND Pipeline.
he project PARADIGM (PARallel Approach of Differential and InteGral Measurements) answers this question by selecting via machine learning (ML) an optimal combination of differential and integral experiments to reduce 239 Pu nuclear data uncertainties from 1-600 keV by 50%..