Search NASA⌕ Search

SEARCH · Search NASA

Results for “ml”

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 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
NASA NTRS2021

Simons Observatory HoloSim-ML: Machine Learning Applied to the Efficient Analysis of Radio Holography Measurements of Complex Optical Systems

Near-field radio holography is a common method for measuring and aligning mirror surfaces for millimeter andsub-millimeter telescopes. In instruments with more than a single mirror, degeneracies arise in the holographymeasurement, requiring multiple measurements and new fitting methods. We present HoloSim-ML, a Pythoncode for beam simulation and analysis of radio holography data from complex optical systems. This code usesmachine learning to efficiently determine the position of hundreds of mirror adjusters on multiple mirrors with fewmicrometer accuracy. We apply this approach to the example of the Simons Observatory 6 m telescope.

Grace E Chesmore↗
NASA NTRSDate not supplied

A New ML-Based Adaptive Thinning Methodology to Improve the Impact of AIRS and CrIS Assimilation on Global Tropical Cyclone Forecasts

This work builds on previous research performed by this team to improve the forecast of Tropical Cyclones (TCs) by assimilating AIRS and CrIS radiances into the NASA Global Earth Observing System (GEOS). Past published work demonstrated that the assimilation of radiances with variable density was beneficial to TC forecasting in the GEOS. In the previous setup, a fixed-size moving square named 'TC domain' was activated by the so-called TC-vitals, an international real-time message accessible to all NWP forecasting centers, that documents the existence of a TC, its estimated position, and its size. The information from TC-vitals activated a switch in the GEOS, which allowed to reduce the distance used for thinning AIRS and CrIS data inside a 15 degrees by 15 degrees moving TC domain centered on the storm, so that more data were assimilated in the vicinity of the TC during its lifetime. The methodology produced improved TC analyses and led to better forecasts, particularly related to intensity, without damaging the global forecast skill. In the new version, the adaptive thinning methodology is based on a machine-learning technique. The technique searches for TCs and creates TC masks by using cloud-top temperatures from all geostationary satellites without the need for additional information. It is being trained against the International Best Track Archive for Climate Stewardship (IBTrACS) data base. Once a TC mask is created, a switch identical to the one used in the previous adaptive thinning method is activated, allowing the GEOS to ingest more data in the TC-shaped size-changing domain that follows the storm. As of today, the team has been able to successfully assimilate data inside the ML-detected TC domains. Future work includes an improved capability of reducing false alarm rates (i.e., cloud systems that are erroneously labeled as TCs).

Oreste Reale↗
NASA NTRSDate not supplied

Surface Hardness Testing of Mobile Launcher 1 (ML-1) Vehicle Support Posts (VSP) Following Artemis I Launch Exposure

Following the successful launch of Artemis I on November 16th, 2022 (Fig 1), a surface hardness assessment was performed on the eight primary VSPs (Figs. 2, 3) located on the Zero Deck of the ML. This assessment was performed as part of an instrument investigation for strain gauges installed on VSP interior surfaces that had a “strange reaction”. This report does not address the VSP strain gauge investigation, but rather VSP structural integrity for future service (Artemis II, III, …). Surface hardness measurements were taken to assess the strength of the ASTM A148 steel castings to confirm thermal exposure from launch had not reannealed the VSPs. However unlikely, it is possible high temperature may have heated and tempered (reannealed) the VSPs, resulting in a change of built-in stresses. This loss in built-in stress may explain the change in off-set measured by the strain gauges. Initial hardness measurements were made at two interior surfaces of VSPs near the strain gauges and the corresponding exterior surfaces (engine hole-facing). Field hardness measurements on the exterior, engine hole surface-facing surfaces for all eight VSPs ranged from 175 to 226 HB/P, well below the minimum hardness of 302 Brinell (HB). Opposing interior surface hardness measurements near strain gauges ranged from 264 to 358 HB/P. In all cases for flight VSPs, a drop in interior-to-exterior hardness ranging from 70 to 132 HB/P was measured. Surface hardness measurements were then made on the two spare VSPs not exposed to launch temperatures for baseline comparison. These two spare VSPs had a drop in Brinell hardness from interior to exterior surfaces ranging from 3 to 17 HB/P. The large interior-to-exterior delta hardness for eight flight VSPs but not for the two spare VSPs suggests launch exposure temperatures softened the exterior surfaces, raising two questions: 1) Are the VSPs compromised for future use, and 2) How deep is the exterior surface softening effect? To address concerns of low surface hardness values at non-critical exterior surfaces, additional field surface hardness measurements were taken at high-stress locations on VSP3: Both door jams, back radius, and internal left and right radius near the base. These values came in well above the required minimum hardness (317 to 365 HB/P). For a more representative pre-launch versus post-launch surface hardness comparison, VSP Acceptance Data Packages (ADPs) were reviewed, and hardness measurements were made at the same eight original locations on VSP3 where hardness measurements were made at the foundry. To determine the degree of softening, 1/8” of material was machined away at three exterior surfaces of engine hole-facing side (Fig. 4). A significant increase in surface hardness was measured ranging from 50 to 86 HB/P. Values at these three locations were still slightly below the required minimum hardness, so an additional 1/8” material was machined away, and hardness values were retaken. At the ¼” depth, hardness values increased to 321 to 324 HB/P. These values fall above the required minimum, concluding surface hardness testing. Rationale for VSP continued use was reached with no restrictions during a Chief Engineering Review based on hardness values exceeding minimum at critical locations and subsurface locations for VSP #3.

Artemis I↗
NASA NTRSDate not supplied

Parameterization of Vertical Cloud Distribution from C3M and MERRA Data Using ML Method

Clouds play a key role in regulating the hydrological cycle and the Earth's radiative energy budget. However, global climate models (GCMs) with a horizontal grid spacing on the order of 100 km have limitations in representing sub-grid cloud dynamics with spatial scales on the order of 1 km, leading to potential uncertainties in cloud radiative feedback on the global scale. In our research, we will leverage the capabilities of Deep Machine Learning (DML) methods to construct parameterizations of sub-grid volumetric cloud fraction (VCF), which is the frequency of occurrence on a grid volume accumulated in the horizontal and vertical directions. Our investigation delves into the intricate relationship between VCF obtained from the NASA CALIPSO-CloudSat-CERES-MODIS (CCCM) satellite observation data and 3-D MERRA-2 reanalysis meteorological profiling data (e.g., wind, relative humidity, temperature). Through a comprehensive one-year data training utilizing the Sequence to Sequence DML method, we have successfully disentangled the complicated cloud formation dynamics across diverse meteorological conditions through a day-to-day analysis framework. Preliminary findings reveal promising statistical agreements in geographical and vertical distributions and seasonal variations of volumetric cloud fraction between ML prediction and satellite measurements. These results underscore the aptitude of our DML model to discern underlying cloud physical processes and accurately represent sub-grid cloud formation dynamics. Additionally, we have also employed trained neural network to analyze uncertainties arising from errors in meteorological data, further enhancing the robustness of our VCF parameterization.

Shan Zeng↗
NASA NTRSDate not supplied

Investigating the Use of Machine Learning (ML) to Assess Tropospheric Doppler Radar Wind Profiler (TDRWP) Data Quality

Manual Quality Control (MQC) of Tropospheric Doppler Radar Wind Profiler (TDRWP) data is essential for defining an accurate climatology for downstream aerospace vehicle assessments. MQC traditionally takes around 30.5 hours per year of radar data. The Marshall Space Flight Center Natural Environments Branch (MSFC NE) used machine learning (ML) to test the feasibility of automating the MQC process, showing a potential to reduce labor by 300%. However, analysis of the model showed some false positives. We compared a neural network to the model to validate it and develop a process for assessing comparable solutions in the future.

Corey Walker↗
NASA NTRSDate not supplied

Coaxial Rotor CFD Validation and ML Surrogate Model Generation

A fixed-pitch speed-controlled coaxial rotor system was tested in the NASA Langley Transonic Dynamics Tunnel in September 2022. The rotors have a diameter of 1.35 meters and an inter-rotor spacing of 25% of the diameter. Though this test was focused on the NASA New Frontiers Dragonfly mission, the resulting dataset is relevant to a wide array of applications of multirotor vehicles, especially those using fixed-pitch variable-speed rotors. Most notably beyond Dragonfly, perhaps, is the application of coaxial rotor pair configurations for eVTOL and Urban Air Mobility (UAM) aircraft. This work provides a thorough CFD validation study quantifying coaxial rotor performance estimation with accuracy on order 5-10% using an efficient hybrid BEMT-URANS flow solver over a wide range of operating conditions. This accuracy was achieved using novel approaches for the construction of both the C81 airfoil performance lookup tables and the BEMT rotor model. These novel approaches are combined with advanced scripting to further accelerate the commercial off-the-shelf CFD solver on GPU accelerated machines. Finally, the development of machine learning surrogate models is presented to produce highly efficient and accurate rotor performance predictions for fixed-pitch variable-speed multirotor aircraft over the complete flight regime including scout, cruise, climb, descent, as well as limiting cases of vortex-ring and windmill-brake states.

Coaxial↗
NASA NTRS1993

Theoretical Study of the H2-ML(+) Binding Energies

The cooperative ligand effects are studied in MLH2(+) and the results are compared to the recent experiments of Kemper et al. The bonding in these compounds is principally electrostatic in origin; however, ligand to metal and metal to ligand donations are important, especially for H2. We show that differences arise among the vanadium, cobalt, and copper complexes which are due to 3d donation to H2. Electron correlation is required to describe the dative interaction, and we find that second order Moller-Plesset perturbation theory (MP2) yields a good description of these systems compared with higher levels of correlation (such as the modified coupled pair functional and coupled cluster approaches) and experiment. However, obtaining quantitative results requires higher levels of theory than MP2.

Maitre, Philippe↗
NASA NTRS2014

NASA One-Dimensional Combustor Simulation--User Manual for S1D_ML

The work presented in this paper is to promote research leading to a closed-loop control system to actively suppress thermo-acoustic instabilities. To serve as a model for such a closed-loop control system, a one-dimensional combustor simulation composed using MATLAB software tools has been written. This MATLAB based process is similar to a precursor one-dimensional combustor simulation that was formatted as FORTRAN 77 source code. The previous simulation process requires modification to the FORTRAN 77 source code, compiling, and linking when creating a new combustor simulation executable file. The MATLAB based simulation does not require making changes to the source code, recompiling, or linking. Furthermore, the MATLAB based simulation can be run from script files within the MATLAB environment or with a compiled copy of the executable file running in the Command Prompt window without requiring a licensed copy of MATLAB. This report presents a general simulation overview. Details regarding how to setup and initiate a simulation are also presented. Finally, the post-processing section describes the two types of files created while running the simulation and it also includes simulation results for a default simulation included with the source code.

combustion↗
NASA NTRSDate not supplied

GeoNEX-ML: A Machine Learning System for Geostationary Satellite Imagery

Improved capabilities of earth monitoring satellites are enabling a wide range of studies on the environmental effects of climate change, often leveraging the recent advancements in machine learning. At the same time, the new capabilities, including higher spatial resolution and temporal frequency, are expanding the amount of data generated at exponential rates. Further, a large majority of archived datasets generated by scientific processing is never used. This motivates the development of an efficient machine learning system for end-to-end processing of multi-level satellite datasets, from level 1 top of atmosphere observations to user friendly environmental variables of interest. Using current generation geostationary satellites GOES-16/17 (NOAA/NASA), Himawari-8/9 (JAXA), and GK-2A (Korea), we present an interchangeable set of machine models to perform spectral adjustment, physical model emulation, LEO-GEO emulation, and optical flow in a high performance computing environment. We use these tools to generate consistent virtual observations across sensors, perform atmospheric correction and cloud detection, and estimate land surface temperature and atmospheric winds. This approach aims to improve the robustness of remotely sensed data processing by learning from diverse sets of observations while enabling near real-time and on-demand capabilities.

Geostationary satellites↗
NASA NTRSDate not supplied

GeoNEX-ML: A Machine Learning System for Earth Observations

Improved capabilities of earth monitoring satellites are enabling a wide range of studies on the environmental effects of climate change, often leveraging the recent advancements in machine learning. At the same time, the new capabilities, including higher spatial resolution and temporal frequency, are expanding the amount of data generated at exponential rates. At the NASA Earth eXchange (NEX), we build deep learning methods to learn from cross sensor satellite-based Earth observations for generating new datasets with efficient processing techniques. Using current generation geostationary satellites on NEX, we present an interchangeable set of machine models to perform spectral adjustment, physical model emulation, LEO-GEO emulation, and optical flow. These tools are used to generate consistent virtual observations across sensors, perform atmospheric correction and cloud detection, and estimate land surface temperature and atmospheric winds. This approach aims to improve the robustness of remotely sensed data processing by learning from diverse sets of observations while enabling near real-time and on-demand capabilities.

Geostationary↗