Search NASA⌕ Search

SEARCH · Search NASA

Results for “Distributed 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.

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

At least 1,459 records · Page 81

Microgrid Procurement Checklist Training: Part 1

Learn how to navigate the Microgrid Procurement Checklist, a key resource for federal agencies developing microgrid projects. This first session will provide a step-by-step walkthrough of the checklist, helping participants understand its structure, key tasks, and how to apply it in project planning.

25 ENERGY STORAGE↗

Generation of random geological models using multi-randomization for machine learning

Generating high-fidelity geological models is essential for advancing machine learning (ML) methods in automated seismic interpretation. For instance, seismic images paired with corresponding fault labels are foundational for ML-based fault detection from seismic migration sections. While several open-access datasets of random geological models exist, open-source tools specifically designed to produce large volumes of such models for ML applications remain scarce. To address this gap, we present RGM (Random Geological Model), an open-source software package for efficiently generating 2D and 3D synthetic geological models tailored for ML workflows. RGM supports the creation of diverse model components, including medium property distributions (P-/S-wave velocities and density), seismic reflectivity images (i.e., synthetic migration sections), relative geological time, and discrete fault attributes such as probability, dip, strike, rake, and displacement. It also accommodates the creation of complex geological features such as salt bodies and unconformities. The model generation algorithm employs a multi-randomization strategy, yielding an effectively infinite-dimensional model space that encompasses a wide range of geological scenarios and associated seismic features. Furthermore, RGM incorporates a method to generate synthetic elastic migration images using analytical elastic reflection coefficients combined with frequency-dependent scaling. This functionality enables the creation of training datasets for ML models that leverage elastic seismic images. RGM is implemented in modern object-oriented Fortran, allowing users to flexibly control statistical parameters governing model variability. We demonstrate the capability, performance, and geological realism of the package through comprehensive 2D and 3D examples.

58 GEOSCIENCES↗

CMU DeepLens: deep learning for automatic image-based galaxy–galaxy strong lens finding

Galaxy-scale strong gravitational lensing can not only provide a valuable probe of the dark matter distribution of massive galaxies, but also provide valuable cosmological constraints, either by studying the population of strong lenses or by measuring time delays in lensed quasars. Due to the rarity of galaxy-scale strongly lensed systems, fast and reliable automated lens finding methods will be essential in the era of large surveys such as Large Synoptic Survey Telescope, Euclid and Wide-Field Infrared Survey Telescope. To tackle this challenge, we introduce CMU DeepLens, a new fully automated galaxy-galaxy lens finding method based on deep learning. This supervised machine learning approach does not require any tuning after the training step which only requires realistic image simulations of strongly lensed systems. We train and validate our model on a set of 20 000 LSST-like mock observations including a range of lensed systems of various sizes and signal-to-noise ratios (S/N). We find on our simulated data set that for a rejection rate of non-lenses of 99 per cent, a completeness of 90 per cent can be achieved for lenses with Einstein radii larger than 1.4 arcsec and S/N larger than 20 on individual g-band LSST exposures. Finally, we emphasize the importance of realistically complex simulations for training such machine learning methods by demonstrating that the performance of models of significantly different complexities cannot be distinguished on simpler simulations. We make our code publicly available at https://github.com/McWilliamsCenter/CMUDeepLens.

gravitational lensing: strong↗

THE SCUBA-2 COSMOLOGY LEGACY SURVEY: MULTIWAVELENGTH COUNTERPARTS TO 10{sup 3} SUBMILLIMETER GALAXIES IN THE UKIDSS-UDS FIELD

We present multiwavelength identifications for the counterparts of 1088 submillimeter sources detected at 850 μm in the SCUBA-2 Cosmology Legacy Survey study of the UKIRT Infrared Deep Sky Survey-Ultra-Deep Survey (UDS) field. By utilizing an Atacama Large Millimeter Array (ALMA) pilot study on a subset of our bright SCUBA-2 sample as a training set, along with the deep optical–near-infrared (OIR) data available in this field, we develop a novel technique, Optical–IR Triple Color (OIRTC), using z − K, K − [3.6], [3.6] − [4.5] colors to select the candidate submillimeter galaxy (SMG) counterparts. By combining radio identification and the OIRTC technique, we find counterpart candidates for 80% of the Class = 1 ≥ 4σ SCUBA-2 sample, defined as those that are covered by both radio and OIR imaging and the base sample for our scientific analyses. Based on the ALMA training set, we expect the accuracy of these identifications to be 82% ± 20%, with a completeness of 69% ± 16%, essentially as accurate as the traditional p-value technique but with higher completeness. We find that the fraction of SCUBA-2 sources having candidate counterparts is lower for fainter 850 μm sources, and we argue that for follow-up observations sensitive to SMGs with S{sub 850} ≳ 1 mJy across the whole ALMA beam, the fraction with multiple counterparts is likely to be >40% for SCUBA-2 sources at S{sub 850} ≳ 4 mJy. We find that the photometric redshift distribution for the SMGs is well fit by a lognormal distribution, with a median redshift of z = 2.3 ± 0.1. After accounting for the sources without any radio and/or OIRTC counterpart, we estimate the median redshift to be z = 2.6 ± 0.1 for SMGs with S{sub 850} > 1 mJy. We also use this new large sample to study the clustering of SMGs and the far-infrared properties of the unidentified submillimeter sources by stacking their Herschel SPIRE far-infrared emission.

79 ASTRONOMY AND ASTROPHYSICS↗

Joint Modeling of Quasar Variability and Accretion Disk Reprocessing Using Latent Stochastic Differential Equations

Quasars are bright active galactic nuclei powered by the accretion of matter around supermassive black holes at the center of galaxies. Their stochastic brightness variability depends on the physical properties of the accretion disk and black hole. The upcoming Rubin Observatory Legacy Survey of Space and Time (LSST) is expected to observe tens of millions of quasars, so there is a need for efficient techniques like machine learning that can handle the large volume of data. Quasar variability is believed to be driven by an X-ray corona, which is reprocessed by the accretion disk and emitted as UV/optical variability. We are the first to introduce an auto-differentiable simulation of the accretion disk and reprocessing. We use the simulation as a direct component of our neural network to jointly model the driving variability and reprocessing, trained with supervised learning on simulated LSST-like 10 yr quasar light curves. We encode the light curves using a transformer encoder, and the driving variability is reconstructed using latent stochastic differential equations, a physically motivated generative deep learning method that can model continuous-time stochastic dynamics. By embedding the physical processes of the driving signal and reprocessing into our network, we achieve a model that is more robust and interpretable. We demonstrate that our model outperforms a Gaussian process regression baseline and can infer accretion disk parameters and time delays between wave bands, even for out-of-distribution driving signals. Our approach provides a powerful framework that can be adapted to solve other inverse problems in multivariate time series.

Fagin, Joshua [City Univ. of New York (CUNY), NY (↗

Assessment of Geothermal Resource Potential at a High-Priority Area on the Utah Testing and Training Range–South (UTTR–S)

Field investigations conducted during 2011 support and expand the conclusion of the original Preliminary Report that discovery of a viable geothermal system is possible in the northwestern part of the Utah Testing and Training Range-South (UTTR-S), referred to henceforth as Focus Area 1. The investigations defined the southward extent of the Wendover graben into and near Focus Area 1, enhanced the understanding of subsurface conditions, and focused further geothermal exploration efforts towards the northwestern-most part of Focus Area 1. Specifically, the detailed gravity survey shows that the Wendover graben, first defined by Cook et al. (1964) for areas north of Interstate Highway 80, extends and deepens southwest-ward to the northwest corner of Focus Area 1. At its deepest point, the intersection with a northwest-trending graben there is favorable for enhanced permeability associated with intersecting faults. Processing and modeling of the gravity data collected during 2011 provide a good understanding of graben depth and distribution of faults bounding the graben and has focused the interest area of the study. Down-hole logging of temperatures in wells made available near the Intrepid, Inc., evaporation ponds, just north of Focus Area 1, provide a good understanding of the variability of thermal gradients in that area and corroborate the more extensive temperature data reported by Turk (1973) for the depth range of 300-500 m. Moderate temperature gradients in the northern part of the Intrepid area increase to much higher gradients and bottom-hole temperatures southeastward, towards graben-bounding faults, suggesting upwelling geothermal waters along those faults. Water sampling, analysis, and temperature measurements of Blue Lakes and Mosquito Willey's springs, on the western boundary of Focus Area 1, also show elevated temperatures along the graben-bounding fault system. In addition, water chemistry suggests origin of those waters in limestone rocks beneath the graben in areas with temperatures as high as 140 C (284 F). In conclusion, all of the field data collected during 2011 and documented in the Appendices of this report indicate that there is reasonable potential for a viable geothermal resource along faults that bound the Wendover graben. Prospects for a system capable of binary electrical generation are especially good, and the possibility of a flash steam system is also within reason. The next steps should focus on securing the necessary funding for detailed geophysical surveys and for drilling a set of temperature gradient wells to further evaluate the resource, and to focus deep exploration efforts in the most promising areas.

15 GEOTHERMAL ENERGY↗

Deep Neural Network Algorithm for CMC Microstructure Characterization and Variability Quantification

Microstructure characterization and variability quantification are crucial for understanding ceramic matrix composites (CMCs) mechanical behavior and deformation mechanisms across length scales. Traditionally, analyses of the micrographs obtained from microscopy are labor-intensive. However, with the vast improvement in computer vision (CV) and deep learning (DL), an automated algorithm can be designed to extract essential microstructure variability from micrographs which can then be used to construct a statistically representative volume element (SRVE). The DL-based algorithm spans the taxonomy of microstructure analyses, including semantic segmentation of microstructure constituents, secondary phases, matrix/fiber interface, and defects, and quantifying the microstructure variability in terms of probability distributions. In this work, C/SiNC and SiC/SiNC CMCs microstructures are semantically segmented through a deep convolutional neural network, followed by variability quantification through the implementation of a fully connected regression layer, hence forming a deep regression network. The deep regression network operates in a feedforward regime, in which the neuron output signal traverses through the network in a unidirectional manner. The weight tensor associated with each layer is updated through a backpropagation stochastic gradient descent approach. The input gray-scale image obtained through in-house scanning electron microscope and confocal microscope micrographs is augmented through affine transformations to increase the training set size, which is then processed through four strided convolutional layers. This compresses the image resolution by half at each layer while increasing the image depth by applying different filters (image encoding). The class activation maps (CAMs) corresponding to the applied filters highlight the key architectural features and assist with the semantic segmentation of the microstructure.

Hamza, Mohamed H.↗

Restoring Critical Loads in Resilient Distribution Systems Using a Curriculum Learned Controller: Preprint

In this paper, we propose a curriculum learned reinforcement learning (RL) controller to facilitate distribution system critical load restoration (CLR), leveraging RL's fast online response and its outstanding optimal sequential control capability. Like many grid control problems, CLR is complicated due to the large control action space and renewable uncertainty in a heavily constrained non-linear environment with strong intertemporal dependency. The nature of the problem oftentimes causes the RL policy to converge to a poor-performing local optimum if learned directly. To overcome this, we design a two-stage curriculum in which the RL agent will learn generation control and load restoration decision under different scenarios progressively. Via curriculum learning, the trained RL controller is expected to achieve a better control performance, with critical loads restored as rapidly and reliably as possible. Using the IEEE 13-bus test system, we illustrate the performance of the RL controller trained by the proposed curriculum-based method.

61 RADIATION PROTECTION AND DOSIMETRY↗

Challenges for unsupervised anomaly detection in particle physics

Anomaly detection relies on designing a score to determine whether a particular event is uncharacteristic of a given background distribution. One way to define a score is to use autoencoders, which rely on the ability to reconstruct certain types of data (background) but not others (signals). In this paper, we study some challenges associated with variational autoencoders, such as the dependence on hyperparameters and the metric used, in the context of anomalous signal (top and W) jets in a QCD background. We find that the hyperparameter choices strongly affect the network performance and that the optimal parameters for one signal are non-optimal for another. In exploring the networks, we uncover a connection between the latent space of a variational autoencoder trained using mean-squared-error and the optimal transport distances within the dataset. We then show that optimal transport distances to representative events in the background dataset can be used directly for anomaly detection, with performance comparable to the autoencoders. Whether using autoencoders or optimal transport distances for anomaly detection, we find that the choices that best represent the background are not necessarily best for signal identification. These challenges with unsupervised anomaly detection bolster the case for additional exploration of semi-supervised or alternative approaches.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Development of Ensemble Neural Network Convection Parametrizations for Climate Models Using ARM Data

The novel neural network (NN) approach has been formulated and used for development of a NN ensemble stochastic convection parametrization for climate models. This fast parametrization is built based on data from Cloud Resolving Model (CRM) simulations initialized with and forced by TOGA-COARE data. The SAM (System for Atmospheric Modeling), developed by D. Randall, M. Khairoutdinov, and their collaborators, has been used for CRM simulations. The observational data are also used for validation of model simulations. The SAM-simulated data have been averaged and projected onto the GCM space of atmospheric states to implicitly define a stochastic convection parametrization. This parametrization is emulated using an ensemble of NNs. An ensemble of NNs with different NN parameters has been trained and tested. The inherent uncertainty of the stochastic convection parametrization derived in such a way is estimated. Due to these inherent uncertainties, NN ensemble is used to constitute a stochastic NN convection parametrization. The developed NN convection parametrization have been validated in a diagnostic CAM (CAM-NN) run vs. the control CAM run. Actually, CAM inputs have been used, at every time step of the control/original CAM integration, for parallel calculations of the NN convection parametrization (CAM-NN) to produce its outputs as a diagnostic byproduct. Total precipitation (P) and cloudiness (CLD) time series, diurnal cycles, and P and CLD distributions for the large Tropical Pacific Ocean for the parallel CAM-NN and CAM runs show similarity and consistency with the NCEP reanalysis. The P and CLD distributions for the tropical area for the parallel runs have been analyzed first for the TOGA-COARE boreal winter season (November 1992 through February 1993) and then for the winter seasons of the follow-up parallel decadal simulations. The obtained results are encouraging and practically meaningful. They show the validity of the NN approach. This constitutes an important practical conclusion of the study: the obtained results on NN ensembles as a stochastic physics parametrization show a realistic possibility of development of NN convection parametrization for climate (and NWP) models based on learning cloud physics from CRM/SAM simulated data.

97 MATHEMATICS AND COMPUTING↗

Development and validation of a second-order thermal network model for residential buildings

Heating, Ventilation, and Air Conditioning (HVAC) systems can maintain the space air temperature of residential buildings, either directly by heating/cooling the air, or indirectly via heat transfer to and from the building structure that acts as a thermal mass. Hence, HVAC systems can help achieve load shifting, peak load reduction, and/or energy cost saving, thus enabling grid-interactive HVAC operation. A home thermal model that can accurately reflect the dynamics of the space air and interior wall surface temperatures, is therefore valuable. This paper develops such a model using the standard RC (resistance-capacitance) approach. The model contains a virtual envelope node and an internal space node and is thus second-order. A hybrid parameter identification scheme, made up of the least-squares and optimal search methods, is also developed. The proposed model and scheme were validated using data collected from a test home. It was found that a modest amount of training data was sufficient to yield reliable parameter estimates and accurate prediction. It was also found that when making 24-hour-ahead prediction of the space air temperature, both methods had comparable performances when the training data began in a transition season. However, when they began in an HVAC season, the optimal search method performed better. Furthermore, the least-squares method is recommended during a transition season due to its lower computational burden, while the optimal search method is recommended during an HVAC season due to its better estimation performance.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Reliable extrapolation of deep neural operators informed by physics or sparse observations

Deep neural operators can learn nonlinear mappings between infinite-dimensional function spaces via deep neural networks. As promising surrogate solvers of partial differential equations (PDEs) for real-time prediction, deep neural operators such as deep operator networks (DeepONets) provide a new simulation paradigm in science and engineering. Pure data-driven neural operators and deep learning models, in general, are usually limited to interpolation scenarios, where new predictions utilize inputs within the support of the training set. However, in the inference stage of real-world applications, the input may lie outside the support, i.e., extrapolation is required, which may result to large errors and unavoidable failure of deep learning models. Here, we address this challenge of extrapolation for deep neural operators. First, we systematically investigate the extrapolation behavior of DeepONets by quantifying the extrapolation complexity, via the 2-Wasserstein distance between two function spaces and propose a new strategy of bias–variance trade-off for extrapolation with respect to model capacity. Subsequently, we develop a complete workflow, including extrapolation determination, and we propose five reliable learning methods that guarantee a safe prediction under extrapolation by requiring additional information—the governing PDEs of the system or sparse new observations. The proposed methods are based on either fine-tuning a pre-trained DeepONet or multifidelity learning. We demonstrate the effectiveness of the proposed framework for various types of parametric PDEs. Furthermore, our systematic comparisons provide practical guidelines for selecting a proper extrapolation method depending on the available information, desired accuracy, and required inference speed.

42 ENGINEERING↗

NEW APPROACHES TO PHOTOMETRIC REDSHIFT PREDICTION VIA GAUSSIAN PROCESS REGRESSION IN THE SLOAN DIGITAL SKY SURVEY

Expanding upon the work of Way and Srivastava we demonstrate how the use of training sets of comparable size continue to make Gaussian process regression (GPR) a competitive approach to that of neural networks and other least-squares fitting methods. This is possible via new large-size matrix inversion techniques developed for Gaussian processes (GPs) that do not require that the kernel matrix be sparse. This development, combined with a neural-network kernel function appears to give superior results for this problem. Our best-fit results for the Sloan Digital Sky Survey (SDSS) Main Galaxy Sample using u, g, r, i, z filters gives an rms error of 0.0201 while our results for the same filters in the luminous red galaxy sample yield 0.0220. We also demonstrate that there appears to be a minimum number of training-set galaxies needed to obtain the optimal fit when using our GPR rank-reduction methods. We find that morphological information included with many photometric surveys appears, for the most part, to make the photometric redshift evaluation slightly worse rather than better. This would indicate that most morphological information simply adds noise from the GP point of view in the data used herein. In addition, we show that cross-match catalog results involving combinations of the Two Micron All Sky Survey, SDSS, and Galaxy Evolution Explorer have to be evaluated in the context of the resulting cross-match magnitude and redshift distribution. Otherwise one may be misled into overly optimistic conclusions.

79 ASTRONOMY AND ASTROPHYSICS↗

Efficient Neural Network Approaches for Conditional Optimal Transport with Applications in Bayesian Inference

In this work, we present two neural network approaches that approximate the solutions of static and dynamic conditional optimal transport (COT) problems. Both approaches enable conditional sampling and conditional density estimation, which are core tasks in Bayesian inference—particularly in the simulation-based (“likelihood-free”) setting. Our methods represent the target conditional distribution as a transformation of a tractable reference distribution. Obtaining such a transformation, chosen here to be an approximation of the COT map, is computationally challenging even in moderate dimensions. To improve scalability, our numerical algorithms use neural networks to parameterize candidate maps and further exploit the structure of the COT problem. Our static approach approximates the map as the gradient of a partially input convex neural network. It uses a novel numerical implementation to increase computational efficiency compared to state-of-the-art alternatives. Our dynamic approach approximates the conditional optimal transport via the flow map of a regularized neural ODE; compared to the static approach, it is slower to train but offers more modeling choices and can lead to faster sampling. We demonstrate both algorithms numerically, comparing them with competing state-of-the-art approaches, using benchmark datasets and simulation-based Bayesian inverse problems.

97 MATHEMATICS AND COMPUTING↗

Smart Preprocessing & Robust Integration Emulator

To achieve the desired particle size of biomass feedstocks during preprocessing for trouble-free handling and conversion to produce biofuels and bioproducts, the raw materials must undergo a crucial milling process. The particle size of biomass plays a critical role in subsequent biofuel manufacturing, where a larger area-to-volume ratio facilitates efficient synthesis while balancing the impact of moisture on biomass storage. To optimize biofuel production efficiency and overcome these challenges, it is imperative to accurately predict the particle size distribution (PSD) of the biomass in the design of efficient preprocessing systems. The population balance model (PBM), upon empirical calibration and validation, can provide rapid prediction of post-milling PSD of granular biomass. However, PSD has limitations related to mass conservation and the absence of moisture considerations. To overcome these drawbacks, a deep learning model called the enhanced deep neural operator (DNO+) is implemented in the code. This model not only retains the capabilities of the PBM in handling complex mapping functions but also incorporates additional factors influencing the system. By considering various experimental conditions such as sieve size and moisture content, the trained DNO+ model can effectively predict the PSD after milling for any given feed PSD. To further reduce the reliance on experimental data, the PBM is integrated into the DNO+ model, resulting in a physics-informed DNO+ (PIDNO+). The PIDNO+ model addresses the non-conservation of quality exhibited by the PBM while inheriting the advantages of the DNO+ model in considering multiple influencing factors. Moreover, the PIDNO+ model significantly reduces the amount of data required for model training. Both deep learning models, i.e., DNO+ and PIDNO+, are excellent in predictive performance, offering swift and accurate machine learning-based predictions. The use of this code that contains these models will assist in guiding the proper milling equipment selection and operational conditions to achieve the desired biomass particle sizes, ensuring the efficiency of subsequent biofuel and bioproduct production processes.

Xia, Yidong [Idaho National Laboratory (INL), Idah↗

Controlling distributed energy resources via deep reinforcement learning for load flexibility and energy efficiency

Behind-the-meter distributed energy resources (DERs), including building solar photovoltaic (PV) technology and electric battery storage, are increasingly being considered as solutions to support carbon reduction goals and increase grid reliability and resiliency. However, dynamic control of these resources in concert with traditional building loads, to effect efficiency and demand flexibility, is not yet commonplace in commercial control products. Traditional rule-based control algorithms do not offer integrated closed-loop control to optimize across systems, and most often, PV and battery systems are operated for energy arbitrage and demand charge management, and not for the provision of grid services. More advanced control approaches, such as MPC control have not been widely adopted in industry because they require significant expertise to develop and deploy. Recent advances in deep reinforcement learning (DRL) offer a promising option to optimize the operation of DER systems and building loads with reduced setup effort. However, there are limited studies that evaluate the efficacy of these methods to control multiple building subsystems simultaneously. Additionally, most of the research has been conducted in simulated environments as opposed to real buildings. This paper proposes a DRL approach that uses a deep deterministic policy gradient algorithm for integrated control of HVAC and electric battery storage systems in the presence of on-site PV generation. The DRL algorithm, trained on synthetic data, was deployed in a physical test building and evaluated against a baseline that uses the current best-in-class rule-based control strategies. Performance in delivering energy efficiency, load shift, and load shed was tested using price-based signals. The results showed that the DRL-based controller can produce cost savings of up to 39.6% as compared to the baseline controller, while maintaining similar thermal comfort in the building. The project team has also integrated the simulation components developed during this work as an OpenAIGym environment and made it publicly available so that prospective DRL researchers can leverage this environment to evaluate alternate DRL algorithms.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Bayesian learning with Gaussian processes for low-dimensional representations of time-dependent nonlinear systems

This work presents a data-driven method for learning low-dimensional time-dependent physics-based surrogate models whose predictions are endowed with uncertainty estimates. We use the operator inference approach to model reduction that poses the problem of learning low-dimensional model terms as a regression of state space data and corresponding time derivatives by minimizing the residual of reduced system equations. Standard operator inference models perform well with accurate training data that are dense in time, but producing stable and accurate models when the state data are noisy and/or sparse in time remains a challenge. Another challenge is the lack of uncertainty estimation for the predictions from the operator inference models. Our approach addresses these challenges by incorporating Gaussian process surrogates into the operator inference framework to (1) probabilistically describe uncertainties in the state predictions and (2) procure analytical time derivative estimates with quantified uncertainties. The formulation leads to a generalized least-squares regression and, ultimately, reduced-order models that are described probabilistically with a closed-form expression for the posterior distribution of the operators. The resulting probabilistic surrogate model propagates uncertainties from the observed state data to reduced-order predictions. Furthermore, we demonstrate the method is effective for constructing low-dimensional models of two nonlinear partial differential equations representing a compressible flow and a nonlinear diffusion–reaction process, as well as for estimating the parameters of a low-dimensional system of nonlinear ordinary differential equations representing compartmental models in epidemiology.

Data-driven model reduction↗

Demonstration of reconstruction-free static magnetic control of DIII-D plasma with deep reinforcement learning

This paper presents the development and experimental validation of a reinforcement learning (RL)-based magnetic controller on the DIII-D tokamak. The controller directly maps raw magnetic diagnostic signals to actuator commands, replacing the traditional isoflux control algorithm based on equilibrium reconstruction. Four RL controllers are trained using the Soft Actor–Critic algorithm with an asymmetric Actor–Critic architecture in the NSFsim simulator. All controllers are deployed in the DIII-D Plasma Control System and operated with a 4 kHz feedback loop. Two randomization strategies are evaluated during training: evolving kinetic profiles and fixed kinetic profiles within each episode. The latter approach is found to better capture experimental deviations in the current density profile and to provide overall improved control performance. Robust operation is demonstrated across heating power scans in both L- and H-mode plasmas, as well as during transient events such as L–H transitions and pellet injections. Control errors in plasma shape and radial position remained within 1.5–2.0 cm and 1 cm, respectively. A notable discrepancy was observed in the vertical X-point position, with errors of up to approximately 4 cm, attributed to the current density distribution mismatches between simulations and experiments.

DIII-D↗