Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Compact model”

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 361 records · Page 20
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Acoustic and Thermoacoustic Jet Propulsion

This manuscript bridges the fields of jet propulsion, synthetic jets and thermoacoustic engines. We introduce the case that synthetic jets in air can be considered a consistent approach for propulsion systems. To date, studies on synthetic jet propulsion have been scattered and rather subsidiary. Furthermore, investigating the perspective of propulsion can provide new insights into the field of synthetic jets and the physics of propulsion. In this work, synthetic jet propulsion in air is demonstrated and studied for acoustic and thermoacoustic cases. We developed synthetic jet systems and characterized the resulting propulsion regarding several relevant parameters such as geometric factors and the frequency and power of acoustic jets electromechanically and thermally produced. We demonstrated the 10.4-mN and 2.7-mN thrust in the acoustic and thermoacoustic modes, respectively, using compact tabletop assemblies. The physical mechanism of the jets has been modeled, simulated, verified with Schlieren imaging and laid in the perspective of adherent literature. Similarities and differences regarding traditional jet propulsion systems are discussed. Synthetic jet propulsion in air using a thermal cycle is experimentally demonstrated for the first time.

33 ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2022

Long-term 3D MHD simulations of black hole accretion discs formed in neutron star mergers

ABSTRACT We examine the long-term evolution of accretion tori around black hole (BH) remnants of compact object mergers involving at least one neutron star, to better understand their contribution to kilonovae and the synthesis of r-process elements. To this end, we modify the unsplit magnetohydrodynamic (MHD) solver in flash 4.5 to work in non-uniform three-dimensional spherical coordinates, enabling more efficient coverage of a large dynamic range in length scales while exploiting symmetries in the system. This modified code is used to perform BH accretion disc simulations that vary the initial magnetic field geometry and disc compactness, utilizing a physical equation of state, a neutrino leakage scheme for emission and absorption, and modelling the BH’s gravity with a pseudo-Newtonian potential. Simulations run for long enough to achieve a radiatively inefficient state in the disc. We find robust mass ejection with both poloidal and toroidal initial field geometries, and suppressed outflow at high disc compactness. With the included physics, we obtain bimodal velocity distributions that trace back to mass ejection by magnetic stresses at early times, and to thermal processes in the radiatively inefficient state at late times. The electron fraction distribution of the disc outflow is broad in all models, and the ejecta geometry follows a characteristic hourglass shape. We test the effect of removing neutrino absorption or nuclear recombination with axisymmetric models, finding ${\sim} 50{{\ \rm per\ cent}}$ less mass ejection and more neutron-rich composition without neutrino absorption, and a subdominant contribution from nuclear recombination. Tests of the MHD and neutrino leakage implementations are included.

Fahlman, Steven (ORCID:0000000348759940)↗
DOE OSTI2026

Dynamic crushing of metal lattice metamaterials: Shock mode diagrams and transition to topology-independent compaction regime

Additively manufactured lattice metamaterials offer design versatility in strength and energy absorption and provide an additional degree of freedom through the selection of the lattice topology. Under quasistatic loading, the unit cell structure can strongly affect the stiffness, yield, and post-yield behavior, but whether and to what degree the effect of lattice topology persists into dynamic loading scenarios, up to the compaction shock regime, has not been established. LLNL’ s ALE3D hydrocode was used to perform a computational investigation of dynamic loading in multiple lattice types, including the gyroid, octet, Schwarz D, and rhombic dodecahedron, under impact velocities from 0.25 to 2.25 km/s. Shock Hugoniots for each lattice topology are generated and compared, suggesting that above a critical velocity, distinctions between architectures may not persevere and compacted lattices behave similarly. Here, to investigate the transition between topology-dependent quasistatic compression and the topology-independent regime above the critical velocity, a one-dimensional elastic-linear hardening plasticity-densified solid (E-LHP-DS) shock model for lattice materials was developed that relies upon confined compression to link the quasistatic and shock mechanics. Unlike similar works, the model does not assume rigid behavior prior to yield or locking behavior at densification, allowing a richer exploration of lattice mechanics. With only six parameters, the analytical model simultaneously fit quasistatic confined compression simulations for relative densities 0.1 $≤ \bar{ρ} ≤$ 0.9 and predicted dynamic compaction behavior to traverse several distinct shock modes, each defined by a critical impact speed (equivalently, critical stresses). Comparing the numerical results to the one-dimensional E-LHP-DS shock model predictions suggests that the topology-independence under strong shocks is linked to the onset of densification, which can be predicted based on quasistatic confined compression results.

Cellular material↗
DOE OSTI2025

SymbolNet: neural symbolic regression with adaptive dynamic pruning for compression

Abstract Compact symbolic expressions have been shown to be more efficient than neural network (NN) models in terms of resource consumption and inference speed when implemented on custom hardware such as field-programmable gate arrays (FPGAs), while maintaining comparable accuracy (Tsoi et al 2024 EPJ Web Conf. 295 09036). These capabilities are highly valuable in environments with stringent computational resource constraints, such as high-energy physics experiments at the CERN Large Hadron Collider. However, finding compact expressions for high-dimensional datasets remains challenging due to the inherent limitations of genetic programming (GP), the search algorithm of most symbolic regression (SR) methods. Contrary to GP, the NN approach to SR offers scalability to high-dimensional inputs and leverages gradient methods for faster equation searching. Common ways of constraining expression complexity often involve multistage pruning with fine-tuning, which can result in significant performance loss. In this work, we propose S y m b o l N e t , a NN approach to SR specifically designed as a model compression technique, aimed at enabling low-latency inference for high-dimensional inputs on custom hardware such as FPGAs. This framework allows dynamic pruning of model weights, input features, and mathematical operators in a single training process, where both training loss and expression complexity are optimized simultaneously. We introduce a sparsity regularization term for each pruning type, which can adaptively adjust its strength, leading to convergence at a target sparsity ratio. Unlike most existing SR methods that struggle with datasets containing more than O ( 10 ) inputs, we demonstrate the effectiveness of our model on the LHC jet tagging task (16 inputs), MNIST (784 inputs), and SVHN (3072 inputs).

Tsoi, Ho Fung (ORCID:0000000225502184)↗
DOE OSTI2026

Probing jet base emission of M87* with the 2021 Event Horizon Telescope observations

We investigate the presence and spatial characteristics of the jet base emission in M87* at 230 GHz, enabled by the significantly enhanced (u,v) coverage in the 2021 Event Horizon Telescope (EHT) observations. The integration of the 12−m Kitt Peak Telescope (USA) and NOEMA (France) stations into the array introduces two critical intermediate-length baselines to SMT (USA) and IRAM 30−m (Spain), providing sensitivity to emission structures at spatial scales of ∼250 μas and ∼2500 μas (∼ 0.02 pc and ∼ 0.02 pc). Without these new baselines, previous EHT observations of the source in 2017 and 2018 lacked the capability to constrain emission on large scales, where a “missing flux” of order ∼1 Jy is expected to reside. To probe these scales, we analyzed closure phases–robust against station-based gain calibration errors–and model the jet base emission using a simple Gaussian component offset from the compact ring emission at spatial separations > 100 μas. Our analysis revealed a Gaussian feature centered at (ΔRA ≈ 320 μ as, ΔDec. ≈ 60 μ as), projected separation of ≈ 5500 AU, with an estimated flux density of only ∼60 mJy, implying that most of the missing flux identified in previous EHT studies had to originate from different, larger scales. Brighter emission at the relevant spatial scales is firmly ruled out, and the data do not favor more complex models. This component aligns with the inferred position of the large-scale jet and is therefore physically consistent with the emission of the jet base. While our findings point to detectable jet base emission at 230 GHz, the limited coverage provided by only two intermediate baselines limits our ability to robustly reconstruct its morphology. Consequently, we treated the recovered Gaussian as an upper limit on the jet base flux density. Future EHT observations with expanded intermediate baseline coverage will be essential to constrain the structure and nature of this component with higher precision.

accretion, accretion disks↗
DOE OSTI2022

Kinetic frustration by limited bond availability controls the LAT protein condensation phase transition on membranes

LAT is a membrane-linked scaffold protein that undergoes a phase transition to form a two-dimensional protein condensate on the membrane during T cell activation. Governed by tyrosine phosphorylation, LAT recruits various proteins that ultimately enable condensation through a percolation network of discrete and selective protein-protein interactions. Here, we describe detailed kinetic measurements of the phase transition, along with coarse-grained model simulations, that reveal that LAT condensation is kinetically frustrated by the availability of bonds to form the network. Unlike typical miscibility transitions in which compact domains may coexist at equilibrium, the LAT condensates are dynamically arrested in extended states, kinetically trapped out of equilibrium. Modeling identifies the structural basis for this kinetic arrest as the formation of spindle arrangements, favored by limited multivalent binding interactions along the flexible, intrinsically disordered LAT protein. These results reveal how local factors controlling the kinetics of LAT condensation enable formation of different, stable condensates, which may ultimately coexist within the cell.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

NREL’s Wind Turbine Drivetrain Condition Monitoring and Wind Plant Operation and Maintenance Research During the 2010s: A US Land-Based Perspective

The wind industry has seen tremendous growth during the past two decades, with the global cumulative installation capacity reaching more than 650 gigawatts by the end of 2019. Despite performance and reliability improvements of utility-scale wind turbines over the years, the industry still experiences premature component failures, leading to increased operation and maintenance (O&M) costs. Among various turbine components, gearboxes—and, more broadly, drivetrains—have shown to be costly to maintain throughout the design life of a wind turbine. The problem of premature component failure is industry wide. As early as 2007, the US Department of Energy (DOE) started to address this challenge through the National Renewable Energy Laboratory’s (NREL’s) reliability initiative that first focused on gearboxes and more recently expanded to entire drivetrains. The wind turbine drivetrain condition monitoring and wind plant O&M research that is the subject of this paper is part of the NREL initiative and includes a few research and development (R&D) activities conducted during the 2010s. These activities included technology evaluation during the first few years; novel monitoring technique investigation (specifically, compact filter analysis) during the middle years; and data and physics domain modeling for fault detection and prediction in recent years. A high-level summary of these activities is provided in this paper along with some key observations from each activity. Most of the work discussed has been published and can be referred to for more information. They reflect the expected evolution of wind turbine condition monitoring and O&M in the US market—primarily, a land-based perspective. In addition, we have identified several R&D opportunities that can be picked up by the research community to help industry advance in related areas, making wind power more cost competitive in the future.

17 WIND ENERGY↗
DOE OSTI2020

A numerical study of the meso-structure variability in the compaction process of prepreg platelet molded composites

Compaction deformations in a thermoplastic composite system molded from chopped prepreg platelets were analyzed to obtain statistical distributions of local distortions that define the final consolidated meso-structure. This study provides an insight into and allows quantification of the intrinsic geometric variability of the composite meso-structure that controls the variability of the effective composite mechanical properties. A two-phase viscoelastic finite element model was utilized to treat the behavior of molten prepreg platelets, which were assumed to be incompressible and inextensible in the fiber direction during processing. A two-step simulation procedure with discretely modeled platelets and platelet-to-platelet contacts is proposed. The analysis begins with a drop simulation of the uncontrolled material deposition followed by a 3D finite element analysis of the preform compaction and bulk material flow. The proposed model provides a method for generating a composite meso-structure ab initio starting from a statistical sampling of the initial conditions for platelet deposition. The predicted platelet thickness distribution, intra-platelet fiber waviness and out-of-plane platelet undulations were found to be in good agreement with experimental measurements. An estimate of the volume of resin pockets was inferred from the compaction simulation.

36 MATERIALS SCIENCE↗
DOE OSTI2026

Agentic framework for programmatic crystal structure generation using a fine-tuned worker–supervisor large language model

Platinum group metals (PGMs) underpin many catalytic technologies but face severe supply constraints, motivating the search for alternative materials and computational methods to accelerate discovery. While atomistic simulation tools such as Pymatgen and ASE have streamlined structure manipulation, they require detailed inputs, limiting accessibility for experimentalists and slowing early-stage exploration. Here, in this study, we present an AI-driven agentic framework that orchestrates worker–supervisor large language models (LLMs). The worker translates natural-language prompts of varying abstraction into valid crystallographic structures using a compact LLM fine-tuned with low-rank adaptation on a curated text–code–CIF dataset, emphasizing energy-efficient training. Benchmarking against the baseline CodeGen-350M-mono model shows that fine-tuning reduces hallucination rates from 100% to as low as 5% and improves structural match accuracy to up to 82% for fully specified inputs. Accuracy declines with decreasing prompt detail but remains nontrivial even when only stoichiometry and space group are provided, underscoring the LLM’s capacity for crystallographic inference. The supervisor Claude LLM evaluates the outputs and triggers iterative refinement through the worker’s built-in structure manipulation capabilities (e.g., supercell scaling, strain, vacancy, and substitution operations). We further demonstrate use cases for technologically relevant catalysts, including IrO 2 , pyrochlore Pb 2 Ir 2 O 7 , Ni 2 FeO 4 , and Ni 3 Mo, where the framework generates physically consistent structures that can be refined via geometry optimization. This work introduces a low-energy, language-driven pathway for integrating human and machine intelligence in materials design, paving the way for AI-assisted synthesis planning and high-throughput screening of complex oxides.

AI agent↗
DOE OSTI2023

Collaboration for Advanced Modeling of Particle Accelerators

The pverarching purpose is to accelerate and expand the scope of discoveries from high energy physics (HEP) particle accelerators by enabling the design of accelerators that are significantly more compact and cheaper to build and run. This will be realized through (i) developing high-performance computing (HPC) accelerator and beam modeling capabilities to design the full range of systems required (ii) developing community simulation ecosystems that seamlessly integrate accelerator elements to facilitate the design and control of next-generation accelerators.

43 PARTICLE ACCELERATORS↗
DOE OSTI2024

A soft departure from jamming: the compaction of deformable granular matter under high pressures

Here, the high-pressure compaction of three-dimensional granular packings is simulated using a bonded particle model (BPM) to capture linear elastic deformation. In the model, grains are represented by a collection of point particles connected by bonds. A simple multibody interaction is introduced to control Poisson's ratio and the arrangement of particles on the surface of a grain is varied to model both high- and low-frictional grains. At low pressures, the growth in packing fraction and coordination number follows the expected behavior near jamming and exhibit friction dependence. As the pressure increases, deviations from the low-pressure power-law scaling emerge after the packing fraction grows by approximately 0.1 and results from simulations with different friction coefficients converge. These results are compared to predictions from traditional discrete element method simulations which, depending on the definition of packing fraction and coordination number, may only differ by a factor of two. As grains deform under compaction, the average volumetric strain and asphericity, a measure of the change in the shape of grains, are found to grow as power laws and depend heavily on the Poisson's ratio of the constituent solid. Larger Poisson's ratios are associated with less volumetric strain and more asphericity and the apparent power-law exponent of the asphericity may vary. The elastic properties of the packed grains are also calculated as a function of packing fraction. In particular, we find the Poisson's ratio near jamming is 1/2 but decreases to around 1/4 before rising again as systems densify.

36 MATERIALS SCIENCE↗
DOE OSTI2021

3D Mechanical Analysis of a Compact Nb 3 Sn IR Quadrupole for EIC

Abstract-The Electron Ion Collider (EIC) will require large aperture quadrupole magnets for the Hadron beam in the insertion region. Key requirements include high field, compact size, and tight control of the fringe fields. A 120 mm aperture, 308.4 mm outer diameter actively shielded Nb 3 Sn quadrupole model is under development to support these goals. This work is being carried out by a collaboration of BNL, JLAB and LBNL. A compact shell-based structure preloaded with a bladder and key system was developed for this project. In this paper, the effect of the compact structure on the mechanical behavior was investigated. In particular, the impact of the assembly tolerances and coil size variations on the actual coil stresses and bladder pressures was computed and compared with results from larger bladder and key structures developed for the LHC IR. The longitudinal preload is provided by stainless steel rods. Differently from other bladder and key magnets, the rods are first preloaded axially before coil axial load is applied. This new design aims to increase the overall stiffness of the system and reduce the longitudinal coil displacement during powering. Based on the results of a 3D mechanical analysis, the preliminary preload targets for the EIC quadrupole assembly will be reviewed and discussed.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2023

Emulation of seismic-phase traveltimes with machine learning

SUMMARY We present a machine learning (ML) method for emulating seismic-phase traveltimes that are computed using a global-scale 3-D earth model and physics-based ray tracing. Accurate traveltime predictions based on 3-D earth models are known to reduce the bias of event location estimates, increase our ability to assign phase labels to seismic detections and associate detections to events. However, practical use of 3-D models is challenged by slow computational speed and the unwieldiness of pre-computed lookup tables that are often large and have prescribed computational grids. In this work, we train a ML emulator using pre-computed traveltimes, resulting in a compact and computationally fast way to approximate traveltimes that are based on a 3-D earth model. Our model is trained using approximately 850 million P-wave traveltimes that are based on the global LLNL-G3D-JPS model, which was developed for more accurate event location. The training-set consists of traveltimes between 10 393 global seismic stations and randomly sampled event locations that provide a prescribed, distance-dependent geographic sample density for each station. Prediction accuracy is dependent on event-station distance and whether the station was included in the training set. For stations included in the training set the mean absolute deviation (MAD) of the difference between traveltimes computed using ray tracing through the 3-D model and the ML emulator for local, regional, and teleseismic distances are 0.090, 0.125 and 0.121 s, respectively. For tested station locations not included in the training set, MAD values for the three distance ranges increase to 0.173, 0.219 and 0.210 s, respectively. Empirical traveltime residuals for a global reference data are indistinguishable when ML emulation or the 3-D model is used to compute traveltimes. This result holds regardless of whether the recording station is used in ML training or not.

58 GEOSCIENCES↗
DOE OSTI2025

Control Mechanisms for Self‐Sealing in Activated Clay‐Rich Faults Through Controlled Hydraulic Injection Experiment

Abstract In a high‐pressure injection fault activation experiment conducted at the Mont Terri underground research laboratory in Switzerland, the transmissivity of the Opalinus Clay fault significantly increased due to opening and shearing. The fluid injection, spanning a few hours, generated a 10 m radius fault activation patch. Subsequent pressure pulse tests conducted bi‐weekly for a year revealed the gradual return of fault transmissivity to its initial state. The study utilized fluid pressure decay analysis, optical fiber monitoring, continuous active source seismic measurements and borehole displacement sensors for measuring fault displacements. The fault zone exhibited a dilation of approximately 1.4 mm, associated with both normal and tangential movements during activation, resulting in a sudden transmissivity increase from 1 × 10 −12 to 3.2 × 10 −7 m 2 /s. Early post‐activation, transient compaction and the subsequent slow compaction were observed, transitioning to an extension regime. The pressure pulse tests demonstrated a rapid transmissivity drop by more than two orders of magnitude within the first 10 days, followed by a gradual and less pronounced decrease. Plastic shear and compaction dominated the transmissivity evolution until 70 days after injection ended, followed by a period where additional factors, such as clay mineral swelling, influenced the behavior. Extrapolation suggested a sealing process taking at least 50 years after the initial activation. Plain Language Summary A field‐scale fault activation experiment offers valuable insights into the elasto‐plastic processes governing the sealing of shale faults. The experiment reveals a rapid increase in the fault's transmissivity by approximately five orders of magnitude during activation. Subsequent observations show a gradual transmissivity decrease by about three orders of magnitude post‐activation, with slow long‐term plastic shear and compaction of the fault competing against secondary processes, notably clay mineral swelling. All conceptual models employed to interpret these field data converge on the estimation that the fault's return to its initial low transmissivity state would require a minimum of 50 years. Key Points High‐pressure injection fault activation experiment at the Mont Terri underground research laboratory Continuous transmissivity measurements record self‐sealing inside a clay‐rich fault zone Transmissivity undergoes a phase of domination by slow plastic compaction and shearing during the initial post‐activation period, with mineral swelling exerting its influence over the long term

Guglielmi, Yves↗
DOE OSTI2020

A Vertex-Aligned Model for Packing 4-Hexagonal Clusters in a Regular Hexagonal Container

This paper deals with a problem the packing polyhex clusters in a regular hexagonal container. It is a common problem in many applications with various cluster shapes used, but symmetric polyhex is the most useful in engineering due to its geometrical properties. Hence, we concentrate on mathematical modeling in such an application, where using the “bee” tetrahex is chosen for the new Compact Muon Solenoid (CMS) design upgrade, which is one of four detectors used in Large Hadron Collider (LHC) experiment at European Laboratory for Particle Physics (CERN). We start from the existing hexagonal containers with hexagonal cells packed inside, and uniform clustering applied. We compare the center-aligned (CA) and vertex-aligned (VA) models, analyzing cluster rotations providing the increased packing efficiency. We formally describe the geometrical properties of clustering approaches and show that cluster sharing is inevitable at the container border with uniform clustering. In addition, we propose a new vertex-aligned model decreasing the number of shared clusters in the uniform scenario, but with a smaller number of clusters contained inside the container. Also, we describe a non-uniform tetrahex cluster packing scheme in the proposed container model. With the proposed cluster packing solution, it is accomplished that all clusters are contained inside the container region. Since cluster-sharing is completely avoided at the container border, the maximal packing efficiency is obtained compared to the existing models.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Advective gas flow in bentonite: Development and comparison of enhanced multi-phase numerical approaches

Understanding the impact of repository gas, generated from degradation of waste and its interaction with the host rock, is essential when assessing the performance and safety function of long-term disposal systems for radioactive waste. Numerical models based on conventional multi-phase flow theory have historically been applied to predict the outcome and impact of gas flow on different repository components. However, they remain unable to describe the full complexity of the physical processes observed in water-saturated experiments (e.g., creation of dilatant pathways) and thus, the development of novel representations for their description is required when assessing fully saturated clay-based systems. This was the primary focus of Task A within the international cooperative project DECOVALEX-2019 (D-2019) and refinement of these approaches is the primary focus of this study (Task B in the current phase of DECOVALEX-2023). This paper summarises development of enhanced numerical representations of key processes and compares the performance of each model against high-quality laboratory test data. Experimental data reveals that gas percolation in water-saturated compacted bentonite is characterised by four key features: (i) a quiescence phase, followed by (ii) the gas breakthrough, which leads to a (iii) peak value, which is then followed by (iv) a negative decay. Three models based on the multiphase flow theory have been developed. These models can provide good initial values and reasonable responses for gas breakthrough (although some of them still predict a too-smooth response). Peak gas pressure values are in general reasonably well captured, although maximum radial stress differences are observed at 48 mm from the base of the sample. Here, numerical peak values of 12.8 MPa are predicted, whereas experimental values are about 11 MPa. These models are also capable of providing a reasonable representation of the negative pressure decay following peak pressure. However, other key specific features (such as the timing of gas breakthrough) still require a better representation. The model simulations and their comparison with experimental data show that these models need to be further improved with respect to model parameter calibration, the numerical representation of spatial heterogeneities in material properties and flow localisation, and the upscaling of the related physical processes and parameters. To further understand gas flow localisation, a new conceptual model has been developed, which shows that discrete channels can possibly be induced through the instability of gas-bentonite interface during gas injection, thus providing a new perspective for modeling gas percolation in low-permeability deformable media.

58 GEOSCIENCES↗
DOE OSTI2026

Spherical tokamak physics research in preparation for the operation of NSTX-U

The National Spherical Torus Experiment Upgrade (NSTX-U) is preparing to resume operation, representing a crucial step toward realizing compact, cost-effective fusion pilot plants. In advance of this, extensive modeling and data analysis have been conducted to advance the physics basis for low-aspect-ratio, high-performance plasma regimes, focusing on three core objectives: confinement and stability, power and particle handling, and steady-state operation. Significant progress has been made in understanding the electron temperature flattening in high-β plasmas, which is shown to be driven by a complex interplay of magnetohydrodynamic instabilities (e.g. non-resonant infernal modes), fast-ion-driven Alfvén eigenmodes, and electron and ion-scale micro-instabilities, particularly Kinetic Ballooning Modes (KBMs), whose destabilization is strongly dependent on parallel magnetic field fluctuations (δB ∥ ). Furthermore, a new gyrokinetic critical pedestal model was developed, accurately predicting pedestal structure by identifying KBMs as the primary stability limit, offering a critical constraint for future high-confinement scenarios. To address the challenge of high heat flux, novel liquid lithium plasma-facing components were modeled. The analysis confirmed that lithium vapor shielding is a self-regulating mechanism for heat mitigation, while also emphasizing that strong main ion parallel flow is essential to minimize core lithium contamination. Finally, progress toward steady-state operation was anchored by developing the required physics basis and control tools. This includes predictive modeling for reversed magnetic shear sustainment, demonstrating that magnetic island-induced bootstrap current reduction is negligible in STs, and advancing real-time control and disruption avoidance capabilities. The development of high-speed surrogate models (e.g. MMMNet) provides computationally efficient tools vital for non-inductive scenario optimization and integrated, low-disruptivity operations planned for NSTX-U.

NSTX-U↗
DOE OSTI2024

INSURE: An Information Theory iNspired diSentanglement and pURification modEl for Domain Generalization

Domain Generalization (DG) aims to learn a generalizable model on the unseen target domain by only training on the multiple observed source domains. Although a variety of DG methods have focused on extracting domain-invariant features, the domain-specific class-relevant features have attracted attention and been argued to benefit generalization to the unseen target domain. To take into account the class-relevant domain-specific information, in this paper we propose an Information theory iNspired diSentanglement and pURification modEl (INSURE) to explicitly disentangle the latent features to obtain sufficient and compact (necessary) class-relevant feature for generalization to the unseen domain. Specifically, we first propose an information theory inspired loss function to ensure the disentangled class-relevant features contain sufficient class label information and the other disentangled auxiliary feature has sufficient domain information. Additionally, we further propose a paired purification loss function to let the auxiliary feature discard all the class-relevant information and thus the class-relevant feature will contain sufficient and compact (necessary) class-relevant information. Moreover, instead of using multiple encoders, we propose to use a learnable binary mask as our disentangler to make the disentanglement more efficient and make the disentangled features complementary to each other. We conduct extensive experiments on five widely used DG benchmark datasets including PACS, VLCS, OfficeHome, TerraIncognita, and DomainNet. The proposed INSURE achieves state-of-the-art performance. We also empirically show that domain-specific class-relevant features are beneficial for domain generalization. The code is available at https://github.com/yuxi120407/INSURE .

97 MATHEMATICS AND COMPUTING↗