Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Deep Operator Networks”

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 1,279 records · Page 71
All sourcesNASA NTRSDOE OSTI
NASA NTRS2018

Superconducting Sensors for Microwave and Optical Photon-Starved Communications (Plenary)

NASA mission modeling indicates a desire for approximately a tenfold data improvement per decade from the Deep Space Network through the year 2040. Some of this improvement will come from aperture enhancement and antenna arraying, increased use of Ka-band over X-band, and high performance optical terminals (e.g. 12 m ground telescope). The remainder will depend on "disruptive" technologies. Deep-space communications differs from near earth communications. One way light travel times are measured in minutes rather than seconds and distances involved are large enough such that optical signals are photon limited and microwave signals have extremely low flux density. High frequency SQIF receivers and single photon detectors for optical communications will be described. Specifically, progress towards a "noiseless" receiver based on arrays of incommensurate area SQUIDs, and single photon counting detectors based on superconducting nanowires and kinetic inductance effects, will be discussed. Finally, the design of an optimal array of optical telescopes to emulate performance of a monolithic 12 m telescope will be outlined. In this case, optimal means minimizing the initial capital investment and operational cost while maintaining performance requirements of the deep-space link.

Romanofsky, Bob↗
DOE OSTI2022

An Autonomous Robot for Shell and Tube Heat Exchanger Inspection

Shell and tube heat exchangers (STHEs) are critical to energy conversion efficiency of power plants. Eddy current examination is a way to evaluate working conditions of these tubes. However, the current testing apparatus requires human to manually insert an eddy current testing (ECT) probe into and extract it out of individual tubes, and meanwhile monitor measurement results for diagnosis. It is a time-consuming and labor-intensive procedure even for an experienced technician. To tackle this challenge, in this work, we developed a robot enabled ECT system for autonomous inspection of STHEs. The robotic platform employs Mecanum wheeled chassis for high mobility, machine vision to locate tube bundle and tube inlets, a rotational Cartesian mechanism to operate at planes with all possible inclinations, and a task-specific mechanism for ECT probe delivery. Machine vision locates tube bundle and tube inlets by an April tag detection algorithm and a Circle Hough Transform (CHT) algorithm, respectively. Assisted by a guidance cone, the ECT probe is continuously fed into the tubes with a fill factor of 0.819. During this process, the eddy current data are automatically collected and real-time analyzed by convolutional neural networks (CNN), showing accuracy of nearly 100% for identifying defective and non-defective tubes and 85% for four types of defective tubes and non-defective tubes.

autonomy, deep learning, eddy current testing, hea↗
DOE OSTI2023

DeePMD-kit v2: A software package for deep potential models

DeePMD-kit is a powerful open-source software package that facilitates molecular dynamics simulations using machine learning potentials known as Deep Potential (DP) models. This package, which was released in 2017, has been widely used in the fields of physics, chemistry, biology, and material science for studying atomistic systems. The current version of DeePMD-kit offers numerous advanced features, such as DeepPot-SE, attention-based and hybrid descriptors, the ability to fit tensile properties, type embedding, model deviation, DP-range correction, DP long range, graphics processing unit support for customized operators, model compression, non-von Neumann molecular dynamics, and improved usability, including documentation, compiled binary packages, graphical user interfaces, and application programming interfaces. This article presents an overview of the current major version of the DeePMD-kit package, highlighting its features and technical details. Additionally, this article presents a comprehensive procedure for conducting molecular dynamics as a representative application, benchmarks the accuracy and efficiency of different models, and discusses ongoing developments.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1987

Ka-band (32-GHz) performance of 70-meter antennas in the Deep Space Network

Two models are provided of the Deep Space Network (DSN) 70 m antenna performance at Ka-band (32 GHz) and, for comparison purposes, one at X-band (8.4 GHz). The baseline 70 m model represents expected X-band and Ka-band performance at the end of the currently ongoing 64 m to 70 m mechanical upgrade. The improved 70 m model represents two sets of Ka-band performance estimates (the X-band performance will not change) based on two separately developed improvement schemes: the first scheme, a mechanical approach, reduces tolerances of the panels and their settings, the reflector structure and subreflector, and the pointing and tracking system. The second, an electronic/mechanical approach, uses an array feed scheme to compensate fo lack of antenna stiffness, and improves panel settings using microwave holographic measuring techniques. Results are preliminary, due to remaining technical and cost uncertainties. However, there do not appear to be any serious difficulties in upgrading the baseline DSN 70 m antenna network to operate efficiently in an improved configuration at 32 GHz (Ka-band). This upgrade can be achieved by a conventional mechanical upgrade or by a mechanical/electronic combination. An electronically compensated array feed system is technically feasible, although it needs to be modeled and demonstrated. Similarly, the mechanical upgrade requires the development and demonstration of panel actuators, sensors, and an optical surveying system.

Imbriale, W. A.↗
NASA NTRS1987

Models of weather effects on noise temperature and attenuation for Ka- and X-band telemetry performance analysis

Models that show the effects of weather on noise temperature and attenuation of deep space telemetry signals received by the Deep Space Network (DSN) at Ka- and X-band (32 and 8.5 GHz) are developed. These models were used to compare the performance of telemetry links at these two frequencies. The models build on an earlier 1982 model that used three months of water vapor radiometer measurements (31.4 GHz) at Goldstone, augmented with one year of radiosonde measurements made at Edwards Air Force Base. This 1986 model accounts for annual variations of rainfall and extends to a model for Canberra, Australia, and Madrid, Spain. The results show, for example, that at Ka-band, 30 degrees elevation angle, Goldstone weather adds less than 23 + or - 2 K to the system temperature 80% of the time, while Canberra or Madrid weather adds less than 32 + or - 5 K 80% of the time. At X-band, the comparable numbers are 5.1 + or - 0.2 K and 5.7 + or - 0.4 K. A simple analysis shows a substantial telemetry system signal-to-noise ratio advantage when operating at Ka-band compared to X-band.

Slobin, S. D.↗
NASA NTRS1998

RIACS

The Research Institute for Advanced Computer Science (RIACS) was established by the Universities Space Research Association (USRA) at the NASA Ames Research Center (ARC) on June 6, 1983. RIACS is privately operated by USRA, a consortium of universities that serves as a bridge between NASA and the academic community. Under a five-year co-operative agreement with NASA, research at RIACS is focused on areas that are strategically enabling to the Ames Research Center's role as NASA's Center of Excellence for Information Technology. The primary mission of RIACS is charted to carry out research and development in computer science. This work is devoted in the main to tasks that are strategically enabling with respect to NASA's bold mission in space exploration and aeronautics. There are three foci for this work: (1) Automated Reasoning. (2) Human-Centered Computing. and (3) High Performance Computing and Networking. RIACS has the additional goal of broadening the base of researcher in these areas of importance to the nation's space and aeronautics enterprises. Through its visiting scientist program, RIACS facilitates the participation of university-based researchers, including both faculty and students, in the research activities of NASA and RIACS. RIACS researchers work in close collaboration with NASA computer scientists on projects such as the Remote Agent Experiment on Deep Space One mission, and Super-Resolution Surface Modeling.

Moore, Robert C.↗
DOE OSTI2020

Neural networks for modeling and control of particle accelerators

Charged particle accelerators support a wide variety of scientific, industrial, and medical applications. They range in scale and complexity from systems with just a few components for beam acceleration and manipulation, to large scientific user facilities that span many kilometers and have hundreds-to-thousands of individually-controllable components. Specific operational requirements must be met by adjusting the many controllable variables of the accelerator. Meeting these requirements can be challenging, both in terms of the ability to achieve specific beam quality metrics in a reliable fashion and in terms of the time needed to set up and maintain the optimal operating conditions. One avenue toward addressing this challenge is to incorporate techniques from the fields of machine learning (ML) and artificial intelligence (AI) into the way particle accelerators are modeled and controlled. While many promising approaches within AI/ML could be used for particle accelerators, this dissertation focuses on approaches based on neural networks. Neural networks are particularly well-suited to modeling, control, and diagnostic analysis of nonlinear systems, as well as systems with large parameter spaces. They are also very appealing for their ability to process high-dimensional data types, such as images and time series (both of which are ubiquitous in particle accelerators). In this work, key studies that demonstrated the potential utility of modern neural network-based approaches to modeling and control of particle accelerators are presented. The context for this work is important: at the start of this work in 2012, there was little interest in AI/ML in the particle accelerator community, and many of the advances in neural networks and deep learning that enabled its present success had not yet been made at that time. As such, this work was both an exploration of possible application areas and a generator of proof-of-concept demonstrations in these areas.

43 PARTICLE ACCELERATORS↗
NASA NTRSDate not supplied

Navigation Performance of the BioSentinel Deep Space CubeSat Mission

The BioSentinel mission was recently launched aboard the SLS launch vehicle (LV) as part of the Artemis- 1 campaign. The BioSentinel navigation team successfully tracked and guided the spacecraft through a lunar gravity assist to its destination Earth-trailing heliocentric orbit. This 6U CubeSat carries live yeast cells to analyze the effects of radiation at large distances from Earth, becoming the first biological payload in Deep Space. Prelaunch activities included mission design updates, orbit determination rehearsals and the development of a tracking schedule in coordination with the Artemis-1 payload office and the Deep Space Network (DSN). An important influence on the trajectories of Artemis I secondaries was the uncertainty associated with deployment from the Interim Cryogenic Propulsion System (ICPS), the upper stage of the SLS LV. The ICPS was rotating at a rate of 1 rpm; there was also an uncertainty in the spin axis attitude, which translated into an unknown clock angle of deployment. The variability in this angle and magnitude of deployment implied the existence of a non-negligible risk of a lunar impact, which was evaluated for various potential launch dates. We present the results of Monte Carlo analyses and compute the pertinent maneuvers to avoid it. In addition, we present a comparison with the actual deployment once the mission launched by reconstructing our trajectory with tracking data. On November 16th 2022 BioSentinel successfully deployed from ICPS and the navigation team started to receive 2-way Doppler and Sequential Ranging data from the DSN. We processed early data to try to obtain a first ephemeris using Initial Orbit Determination (IOD) methods such as the least squares. Soon after deployment, the spacecraft was tumbling and entered safe mode, creating a period where the tracking data were sparse. The mission team recovered the spacecraft and after four tracking passes, we solved for a first ephemeris that was sent to the DSN for better tracking of the spacecraft. After propagating this first ephemeris solution, we determined that we avoided impact with a margin of a few hundred km from the lunar surface. More tracking data over the next few days (from DSN as well as ESA antennas) allowed for a more refined orbit solution predicting a periselene altitude of 406 km and a lunar eclipse lasting 36.5 minutes. Therefore, BioSentinel operators aborted any correction maneuvers. This periselene altitude also gave us the necessary energy to achieve a heliocentric orbit. The next challenge was due to the necessary adjustments in our orbit determination method due to the large energy boost resulting from the lunar flyby. After a series of tracking passes we were able to get a nominal solution that resulted into a stable trajectory. This paper discusses in detail the navigation performance using the X-band IRIS transponder, as well as the challenges and lessons learned prior to and during this deep space, CubeSat mission.

Andres Dono Perez↗
DOE OSTI2022

Distributed Real-time Plume Monitoring for Deep Sea Mineral Extraction​

In the emerging industry of deep-sea mining for minerals and deposits (e.g. polymetallic nodules for nickel, cobalt, copper, and manganese), more data is required to understand the effects of sediment plume generation and predict the distribution of disturbed sediment. There are two main sources of plume generation, the first being at the active mining site where the “collector” directly removes the top layer of the sea floor. The other is the “midwater plume” consisting of unwanted sediment that was collected during extraction that is pumped back into the aphotic zone. The vast majority of plume generation is caused by the collector, causing detrimental and long-lasting impacts on seafloor ecosystems due to the lack of wave activity or strong currents at the sea floor. Therefore, it is crucial to invest in the infrastructure to support the study and constant monitoring over a large area of the sea floor where plume generation is present. Due to the limited number of usable channels and power requirements, current subsea wireless communications technologies are not well suited to instrumenting the large areas of the sea floor needed to monitor plume migration. The scope of this effort is to transition experimental demonstrations of high-bandwidth, full-duplex scalable underwater laser communications to the seafloor in an open ocean environment. Specifically tackling challenges associated with the dynamic nature of the subsea world, including but not limited to, deployment logistics, sustainability, and range. The goal is to enable the internet of underwater things for deep sea industries by broadening the capabilities of subsea communications. By using high-precision laser transmitters, many of the challenges current subsea optical systems face can be circumvented, such as power consumption, interference, and bandwidth limitations. This approach lends itself to wireless interlinking multi-node networks, in series or parallel, facilitating the implementation of a wide array of sensor types. This interlinking allows all the data gathered from the network to be processed through a single hardline uplink to the surface, lowering the complexity required for near real-time data processing. Additionally, the laser control systems produce metadata that can be used to help characterize the water column between the nodes. Combining data from various sensors such as turbidity, temperature, current velocity with metadata such as beam attenuation and deflection can produce a high-resolution model of sea floor conditions around an active mining zone. The resulting near real-time model can be used to optimize location and flow rate of the mining operation to minimize and quantify the environmental impact.

Mons, Ishan↗
DOE OSTI2020

AmeriFlux US-xSB NEON Ordway-Swisher Biological Station (OSBS)

This is the AmeriFlux version of the carbon flux data for the site US-xSB NEON Ordway-Swisher Biological Station (OSBS). Site Description - The Ordway-Swisher Biological Station (OSBS) is operated by the University of Florida and comprises over 9,300 acres. It is a year-round field station established for the long-term study and conservation of unique ecosystems through management, research and education. The Station is located approximately 20 miles east of Gainesville in Melrose (Putnam County, Florida). There are two aquatic arrays at Ordway-Swisher, representing the two dominant aquatic features on the landscape: 1) Suggs lake, a shallow surface water lake that is rich in taxa and biologically active in structure and function; and 2) Barco lake, a deep lake connected to ground water. The forest is maintained by fire and has a relatively open structure: it is managed with prescribed burns at a frequency of 3-4 years.

Network), NEON (National Ecological Observatory↗
DOE OSTI2023

AmeriFlux FLUXNET-1F US-xSB NEON Ordway-Swisher Biological Station (OSBS)

This is the AmeriFlux Management Project (AMP) created FLUXNET-1F version of the carbon flux data for the site US-xSB NEON Ordway-Swisher Biological Station (OSBS). This is the FLUXNET version of the carbon flux data for the site US-xSB NEON Ordway-Swisher Biological Station (OSBS) produced by applying the standard ONEFlux (1F) software. Site Description - The Ordway-Swisher Biological Station (OSBS) is operated by the University of Florida and comprises over 9,300 acres. It is a year-round field station established for the long-term study and conservation of unique ecosystems through management, research and education. The Station is located approximately 20 miles east of Gainesville in Melrose (Putnam County, Florida). There are two aquatic arrays at Ordway-Swisher, representing the two dominant aquatic features on the landscape: 1) Suggs lake, a shallow surface water lake that is rich in taxa and biologically active in structure and function; and 2) Barco lake, a deep lake connected to ground water. The forest is maintained by fire and has a relatively open structure: it is managed with prescribed burns at a frequency of 3-4 years.

Network), NEON (National Ecological Observatory↗
DOE OSTI2023

Three-dimensional nanoscale reduced-angle ptycho-tomographic imaging with deep learning (RAPID)

X-ray ptychographic tomography is a nondestructive method for three dimensional (3D) imaging with nanometer-sized resolvable features. The size of the volume that can be imaged is almost arbitrary, limited only by the penetration depth and the available scanning time. Here we present a method that rapidly accelerates the imaging operation over a given volume through acquiring a limited set of data via large angular reduction and compensating for the resulting ill-posedness through deeply learned priors. The proposed 3D reconstruction method “RAPID” relies initially on a subset of the object measured with the nominal number of required illumination angles and treats the reconstructions from the conventional two-step approach as ground truth. It is then trained to reproduce equal fidelity from much fewer angles. After training, it performs with similar fidelity on the hitherto unexamined portions of the object, previously not shown during training, with a limited set of acquisitions. In our experimental demonstration, the nominal number of angles was 349 and the reduced number of angles was 21, resulting in a x140 aggregate speedup over a volume of 4.48 x 93.18 x 3.92 μm 3 and with (14nm) 3 feature size, i.e. ~ 10 8 voxels. RAPID’s key distinguishing feature over earlier attempts is the incorporation of atrous spatial pyramid pooling modules into the deep neural network framework in an anisotropic way. We found that adjusting the atrous rate improves reconstruction fidelity because it expands the convolutional kernels’ range to match the physics of multi-slice ptychography without significantly increasing the number of parameters.

47 OTHER INSTRUMENTATION↗
NASA NTRSDate not supplied

NASA’s Evolving Ka-band Network Capabilities to Meet Mission Demand

Space missions are increasingly demanding higher data rates to support the growth in information-intensive mission operations. This growth is reflected in planned and operational missions from low Earth orbit, such as the upcoming NASA-Indian Space Research Organization (ISRO) Synthetic Aperture Radar (NISAR) and Plankton, Aerosol, Cloud and Ocean Ecosystem (PACE) missions, to the future Artemis lunar campaign, and the recently launched James Webb Space Telescope orbiting at the Sun-Earth L2 Lagrange point. JWST was the first L2 mission to be defined as a high data rate mission transmitting at 8 Mbps, or 270 gigabits of science data per day. ISRO and PACE anticipate achieving data throughputs of 5-40 terabits per day. These data rates exceed the capabilities of S-band and X-band frequency allocations and are a key driver for migrating to the 26 GHz Ka-band frequency allocation. The NASA Space Communications and Navigation (SCaN) program has been preparing the networks to support this demand by pursuing critical Ka-band infrastructure. The status of current and evolving network capability, including the Near Space Network’s Initiative for Ka-band Advancement (NIKA), and the Deep Space Network’s Lunar Exploration Upgrades (DLEU), as well as profiling mission usage of Ka-band services, are discussed in detail. The push toward Ka-band, is not only an opportunity for increased performance, but alleviates current challenges with contentious and cluttered spectrum access in S- and X-band. The paper provides an overview of these advantages and advanced techniques that optimize its use before the transition to optical communications becomes an imperative. The challenges and potential mitigations for missions considering selection of Ka-band network services are also discussed.

space communications↗
DOE OSTI2021

Distance-Weighted Graph Neural Networks on FPGAs for Real-Time Particle Reconstruction in High Energy Physics

Graph neural networks have been shown to achieve excellent performance for several crucial tasks in particle physics, such as charged particle tracking, jet tagging, and clustering. An important domain for the application of these networks is the FGPA-based first layer of real-time data filtering at the CERN Large Hadron Collider, which has strict latency and resource constraints. We discuss how to design distance-weighted graph networks that can be executed with a latency of less than one μs on an FPGA. To do so, we consider a representative task associated to particle reconstruction and identification in a next-generation calorimeter operating at a particle collider. We use a graph network architecture developed for such purposes, and apply additional simplifications to match the computing constraints of Level-1 trigger systems, including weight quantization. Using the hls4ml library, we convert the compressed models into firmware to be implemented on an FPGA. Performance of the synthesized models is presented both in terms of inference accuracy and resource usage.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
NASA NTRS1994

A comparative survey of current and proposed tropospheric refraction-delay models for DSN radio metric data calibration

The standard tropospheric calibration model implemented in the operational Orbit Determination Program is the seasonal model developed by C. C. Chao in the early 1970's. The seasonal model has seen only slight modification since its release, particularly in the format and content of the zenith delay calibrations. Chao's most recent standard mapping tables, which are used to project the zenith delay calibrations along the station-to-spacecraft line of sight, have not been modified since they were first published in late 1972. This report focuses principally on proposed upgrades to the zenith delay mapping process, although modeling improvements to the zenith delay calibration process are also discussed. A number of candidate approximation models for the tropospheric mapping are evaluated, including the semi-analytic mapping function of Lanyi, and the semi-empirical mapping functions of Davis, et. al.('CfA-2.2'), of Ifadis (global solution model), of Herring ('MTT'), and of Niell ('NMF'). All of the candidate mapping functions are superior to the Chao standard mapping tables and approximation formulas when evaluated against the current Deep Space Network Mark 3 intercontinental very long baselines interferometry database.

Estefan, J. A.↗
NASA NTRS2024

Former Central Heat Plant SWMU 045 Year 2 Air Sparge System Performance Monitoring Report

This Air Sparge (AS) Performance Monitoring (PM) Report (PMR) presents Year 2 operation, maintenance, and monitoring (OM&M) activities, PM results, and monitoring well installations supporting the AS Interim Measure (IM) at the Former Central Heat Plant (CHP) at Kennedy Space Center (KSC), Florida. CHP has been designated Solid Waste Management Unit 045 under the KSC Resource Conservation and Recovery Act Corrective Action Program. An AS IM was installed at CHP between 2019 and 2021, which included the installation of an AS system to treat a chlorinated solvent groundwater plume. Contaminants of concern (COCs) identified at CHP for the AS IM include tetrachloroethene (PCE), trichloroethene (TCE), cis-1,2-dichloroethene (cDCE), and vinyl chloride (VC). The completed AS system includes a network of 267 AS wells, which treat approximately 1.3 acres of contaminated groundwater. “Hot” compressor technology is used to treat the source zone, while a “cold” compressor is used to treat two hot spot (HS) areas (HS1 and HS2) and the high concentration plume (HCP). The AS system began operation in June-July 2021 and this document includes Year 2 of operation. The overall runtimes for the AS system for the Year 2 reporting period (October 2022 to September 2023) were approximately 69 percent for the cold trailer and 71 percent for the hot trailer. Air samples and vapor screening results collected during the reporting period showed concentrations less than applicable human health and air emissions permit criteria. Groundwater performance monitoring results show that AS treatment continues to be effective in reducing COC concentrations at CHP. At the shallow interval, COC concentrations were all non-detect, less than, or met their respective State of Florida Groundwater Cleanup Target Levels (GCTLs) at the end of Year 2 in September 2023. In the deep interval, 10 of the 15 PM wells detected COCs greater than their respective GCTLs, with two of these wells also exceeding the Natural Attenuation Default Concentration for VC. Based on Year 2 OM&M and PM results, continued operation of the AS system is required to meet the IM objective. It is therefore recommended to continue with AS IM operations at CHP with the following plan for Year 3.

Kevin Alex Murphy↗
DOE OSTI2025

Deep Learning–Assisted Multiobjective Optimization of Geological CO 2 Storage Performance under Geomechanical Risks

In geological CO 2 storage, designing the optimal well control strategy for CO 2 injection to maximize CO 2 storage while minimizing the associated geomechanical risks is not trivial. This challenge arises due to pressure buildup, CO 2 plume migration, the highly nonlinear nature of geomechanical responses to rock-fluid interaction, and the high computational cost associated with coupled flow and geomechanics simulations. In this paper, we introduce a novel optimization framework to address these challenges. The optimization problem is formulated as follows: maximize total CO 2 storage while minimizing geomechanical risks by adjusting the injection schedules within bounded constraints. The geomechanical risks are primarily driven by injection-induced pressure build-up, which is characterized by ground displacement and the induced microseismicity. We used the Fourier neural operator (FNO)-based deep learning model to construct surrogate models, replacing the time-consuming coupled flow and geomechanics simulations for evaluating the aforementioned objective functions. The developed surrogate models have been incorporated into a multiobjective optimization framework through a genetic algorithm to reduce the computational burden. The proposed optimization framework reduces the computational cost from approximately 2,400 hours, when using objective function evaluations based on physics-based simulations, to around 20 minutes. A set of Pareto-optimal solutions of the proposed workflow yields nontrivial optimal decisions, reducing the microseismicity potential and the vertical displacement. This Pareto front highlights the optimal trade-offs between CO 2 storage amount, safety, and ground displacement, emphasizing the need for careful optimization and management of injection strategies to achieve a balanced outcome. The novelty of this work is twofold. First, we demonstrate the importance of incorporating the minimization of the geomechanical risks as objective functions into the CO 2 storage optimization workflow to mitigate the potential risk of induced microseismicity and ground displacement. Second, we leverage the FNO-based surrogate models to optimize a real-field CO 2 storage operation.

42 ENGINEERING↗
DOE OSTI2022

Prediction of DIII-D Pedestal Structure from Externally Controllable Parameters

The sharp increase of pressure at the edge of a high confinement mode (H-mode) plasma, the pedestal, strongly impacts overall plasma performance. Predicting the pedestal is a necessity to control and optimize tokamak operations. An experimental data-driven machine learning (ML) approach is presented that predicts the pedestal heights and widths of electron density (ne) and electron temperature (Te) profiles as well as the separatrix ne from externally controllable parameters such as the plasma shape, heating method and power, and gas puff rate and integrated gas puff. The OMFIT framework was used with DIII-D data to efficiently, robustly, and automatically build a database of pedestal parameters to train machine learning models. Database creation was enabled by the search engine tool for DIII-D data, TokSearch, which parallelizes data fetching, enabling fast searches through basic signals of thousands of DIII-D shots and selection of relevant time intervals. Principal Component Analysis (PCA) separated the database into three clusters that represent classes of plasma shapes that are regularly used in DIII-D. The most important parameters for setting the pedestal structure were plasma current (Ip), toroidal magnetic field (Bφ), neutral beam heating power (PNBI) and shaping quantities. The Deep Jointly Informed Neural Networks (DJINN) algorithm was applied to identify suitable neural network (NN) architectures that appropriately capture the features of the pedestal database. Separate NNs were implemented for each pedestal parameter, and ensembling methods were used to improve the prediction accuracy and allowed estimation of the prediction uncertainty. The pedestal predictions of the test dataset lie within the measurement uncertainties of the pedestal parameters. The NN outperformed simple Linear Regression (LR) analysis, indicating non-linear dependencies in the pedestal structure. The presented achievements illustrate a promising path for future research, using feature extraction to infer experimental trends and thereby improve pedestal models as well as deploying NN for a fast pedestal prediction in DIII-D scenario development.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗