Search NASA⌕ Search

SEARCH · Search NASA

Results for “Conditional”

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 289 records · Page 16
All sourcesNASA NTRSDOE OSTI
DOE OSTI2026

Magnus method for electronic structure calculations at extreme conditions

We present the application of Magnus based methods to the solution of first order coupled ordinary differential equations in High Energy Density (HED) physics applications. Our focus is on the application to quantum mechanical methods, specifically on the solution of the radial Dirac equation for real and complex energies. HED applications require accurate solutions across a wide range of spatial and energy domains, including regimes where the solutions exhibit pronounced oscillatory behavior. Such cases pose significant computational challenges. We demonstrate that Magnus-based integrators can efficiently and accurately address these challenges. We discuss the implementation of the Magnus method for the solution of the radial Dirac equation, including practical considerations such as the evaluation of matrix exponentials, numerical integration, error estimation, and adaptive step size control. We also discuss the application of these methods to complex energy Green’s function techniques and the efficient approximation of integrals of the solutions relevant to HED electronic structure calculations. Here, we demonstrate the accuracy and robustness of the resulting method in applications to the free-particle case, for which analytic solutions are available for comparison, as well as the challenging case of gold at HED conditions.

general physics↗
DOE OSTI2025

Enhancing approximate modular Bayesian inference by emulating the conditional posterior

In modular Bayesian analyses, complex models are composed of distinct modules, each representing different aspects of the data or prior information. In this context, fully Bayesian approaches can sometimes lead to undesirable feedback between modules, compromising the integrity of the inference. The “cut-distribution” prevents unwanted influence between modules by “cutting” feedback. The direct sampling (DS) algorithm is standard practice for approximating the cut-distribution, but it can be computationally intensive, especially when the number of imputations required is large. An enhanced method is proposed, the Emulating the Conditional Posterior (ECP) algorithm, which leverages emulation to increase the number of imputations. Through numerical experiment it is demonstrated that the ECP algorithm outperforms the traditional DS approach in terms of accuracy and computational efficiency, particularly when resources are constrained. Here, it is also shown how the DS algorithm can be improved using ideas from design of experiments. Some practical recommendations are given for algorithm choice in modular Bayesian analyses.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Coupling thermal energy storage with a thermally anisotropic building envelope for building demand-side management across various US climate conditions

Here, the thermally anisotropic building envelope (TABE) is a novel active building envelope that enhances energy efficiency and thermal comfort in buildings by transferring heat and cold between building envelopes and hydronic loops. When coupled with thermal energy storage (TES) units, the TABE + TES enables the storage of both heat and cold energy captured by the TABE roof or exterior walls. This stored energy can be later released by the TABE floor for indoor heating and cooling, benefiting both the grid and the end user. This paper evaluates the merits of TABE + TES for building demand-side management across various US climate conditions, focusing on peak load shaving, annual energy savings, and cost savings under time-of-use (TOU) electric rate schedules. Simulations were conducted by integrating time-of-day–informed, rule-based control strategies in MATLAB, TABE components and TES units in COMSOL Multiphysics, and whole-building energy analysis in EnergyPlus. A case study using the US Department of Energy’s prototype single-family detached house model in Birmingham, Alabama; Los Angeles, California; Oak Ridge, Tennessee; and Denver, Colorado, showed that the TABE + TES system achieved (1) 70 % peak load shaving in Los Angeles and Denver and 20 % in Birmingham and Oak Ridge; (2) significant peak electricity savings of 351–497 kWh, reducing peak energy consumption by 38 %–78 %; and (3) annual heating cost savings of 0.79 $\$$/m2–1.17 $\$$/m 2 and cooling cost savings of 0.60 $\$$/m 2 –1.17 $\$$/m 2 using a normal utility rate or low-TOU rate. The benefits of employing the TABE + TES system are even more significant under high TOU rates.

25 ENERGY STORAGE↗
DOE OSTI2025

Analysis of Line Distance Elements for Various Ibr Controllers and System Conditions

The large-scale penetration of inverter-based resources in power systems has challenged protection engineers because of the different fault behaviors these sources provide compared to conventional generation systems. The main challenges include a low level of fault current magnitude, unpredictable angles of sequence currents, and lack of inertia that can lead to maloperation of conventional phasor-based protection elements. This paper presents a sensitivity analysis of transmission line distance protection elements during phase-to-ground and phase-to-phase faults for different inverter controllers and power system conditions. It also summarizes which line protection elements remain secure near IBR terminations and identifies the ones affected by the inverter-based response. The paper concludes by highlighting that regulating negative-sequence current injection during the fault aids correct protection decisions, but does not address the entire challenge. Finally, alternative protection elements to those affected by the inverter fault response are discussed.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2024

Comprehensive evaluation of commercially scalable atomic-layer-deposited alumina coating impact on full cell battery performance across varied test conditions

Atomic Layer Deposition (ALD) has emerged as a strategic enhancement method for lithium-ion battery (LIB) materials offering potential benefits and durability benefits for industrial battery production. However, the translation from laboratory achievements to commercial-scale applications has been limited. Here, this study aims to bridge this gap by comprehensively evaluating the effects of commercially scalable Al 2 O 3 ALD coatings using full pouch cell performance as a means to assess the ALD impact. We utilized large-scale slot-die coating techniques to ensure consistent electrode quality and tested four configurations of pouch cells to analyze the individual effects of ALD coating on anode and cathode electroactive materials. Our extensive testing matrix included long-term cycling, fast discharge, fast charge, leakage current, and high voltage tests. While at lower C-rates (<~1C), the influence of Al 2 O 3 coatings on cell performance is not significant. Fast charging conditions reveal that the anode ALD coating significantly enhances performance via a passivating effect, while on the cathode, it is detrimental, potentially due to increased resistance of the thin interfacial layer formed during the ALD processing. Leakage current and high-voltage tests show that the application of ALD coatings on either anode or cathode effectively minimizes side reactions at the electrode-electrolyte interface. Additionally, ALD coatings significantly mitigate concentrated and localized lithium plating on the anodes. These insights provide a valuable understanding of the potential of ALD technologies in LIB manufacturing to tailor cell performance, paving the way for safer, more efficient, and cost-effective battery solutions.

25 ENERGY STORAGE↗
DOE OSTI2025

Using X-ray radiography to study oxygen flow in a proton exchange membrane electrolyzer operating under balanced pressure conditions

Of the various water electrolyzer technologies, the proton exchange membrane electrolyzer (PEMWE) is one of the best solutions for producing clean hydrogen without releasing CO 2 . In order to allow for widespread use of clean hydrogen, it is necessary to decrease its cost, which is intrinsically related to system operation. Current PEMWE plants operate in differential mode, directly pressurizing hydrogen and benefiting from thermodynamic compression, which increases overall system efficiency. However, high differential pressure above 30 bar can cause membrane stress, resulting in membrane creeping and failure. Pressurizing the water and operating at balanced pressure allows hydrogen to be produced at higher pressures while preserving the integrity of the membrane and porous layers. Nevertheless, the impact of pressurizing water on PEMWE performance must be better understood to maximize performance under balanced pressure conditions. Here, this study examined the impact of water pressure on two-phase flow. A high-pressure electrolyzer setup was developed to perform operando X-ray radiography and examine oxygen transport with high temporal resolution. The imaging segmentation process, developed to capture bubble properties in the channel, was applied to a specific experiment. The results clearly showed that as pressure increased up to 30 bars, the initial bubbly flow transitioned to slug flow, which led to channel saturation with oxygen. This work demonstrates that two-phase flow in an electrolyzer can be studied using X-ray radiography, which has the advantages of fast measurements and the ability to probe dense materials, such as those required for pressurized electrolyzers.

Balanced high pressure operation↗
DOE OSTI2025

Phosphate-modulated transformation of Sb(V)-bearing ferrihydrite under microbial iron- and sulfate-reducing conditions

Antimony (Sb) is a toxic metalloid that poses environmental risks in terrestrial and aquatic systems. The fate of the Sb(V) oxyanion, Sb(OH) 6 − , is governed by complex biogeochemical processes, including immobilization by ferric (Fe(III)) oxides, reduction by sulfide, and the less-explored competitive adsorption with other anions such as phosphate (PO 4 3− ). Here, this study investigates the interplay between such mechanisms in controlling the behavior of Sb(V) associated with ferrihydrite (Fh) under Fe(III)- and sulfate-reducing conditions, with a particular emphasis on the role of phosphate in influencing Sb(V) mobility and transformation. Anoxic reactors that contained Sb(V)-coprecipitated Fh, varying PO 4 3− concentrations (0, 0.2, and 2 mM), and sulfate, were inoculated with a microbial community sourced from Sb-contaminated soil. Additionally, abiotic reactors with either Sb(V)-adsorbed or Sb(V)-coprecipitated Fh and different PO 4 3− loadings (0–100 mM) were created to investigate the competitive adsorption mechanisms in the absence of microbial activity. Results from the abiotic reactors suggest that Sb(V) is likely incorporated into the Fh structure, with only minor amounts remaining surface-bound and extractable by PO 4 3− . In the biotic reactors, microbial Fe(III) and sulfate reduction were more extensive in the presence of PO 4 3− . At 0.2 mM PO 4 3− , microbial activity transformed Fh into siderite and led to the complete reduction of Sb(V) to Sb(III) as stibnite (Sb 2 S 3 ). At 2 mM PO 4 3− , the greater coverage by PO 4 3− stabilized Fh and decreased the extent of both Fe(III) and Sb(V) reduction, and shifted the reduced products to mackinawite and Sb(III) adsorbed onto Fh, in addition to stibnite formation. This study demonstrates that while PO 4 3− may not directly compete with Sb(V) for sorption sites, it can influence Sb mobility in Fe(III)- and sulfate-reducing environments by enhancing microbial activity and altering the mineralization pathways.

Microbial Fe(III) and sulfate reduction↗
DOE OSTI2024

Analysis of heat and mass transfer potential of a dew-point cooling tower in different climatic conditions

In this study, the performance of the Dew-Point Cooling Tower (DPCT) was analyzed for different factors in a variety of climate conditions. For this purpose, a dedicated numerical model describing heat and mass transfer processes was developed and validated. The results of the numerical simulations allowed to analyze the potential of utilizing the heat and mass transfer process with the dew-point phenomenon for water cooling. It was established that the operational parameters that have a high impact on the performance of the DPCT are: inlet water temperature and inlet air humidity ratio. It was also established that DPCT achieves the highest COP and Specific Cooling Capacity for cold subtropical highland climates and that it achieves highest Wet-bulb Effectiveness for monsoon-influenced humid subtropical climates. In conclusion, the regions where all the efficiency factors achieved above-average values included warm, arid, and desert climates.

42 ENGINEERING↗
DOE OSTI2025

2D CFD of lean premixed hydrogen–air flame quenching under locomotive engine conditions

Hydrogen (H 2 ) is a promising fuel for reducing emissions in heavy-duty internal combustion engines (ICEs), but its low quenching distance increases the risk of flame propagation into narrow crevice regions, such as piston-liner gaps. This work uses detailed CFD simulations with finite-rate chemistry to investigate premixed H 2 –air flame quenching in a two-dimensional (2D) region consistent with the piston-liner gap of a diesel ICE. Model accuracy was assessed by comparison with experimentally measured quenching distances in an annular stepwise diverging tube (ASDT). A parametric study was conducted to assess the influence of crevice width (0.05–1.18 mm), pressure (50–150 bar), unburned gas temperature (431–573 K), and equivalence ratio ( Φ= 0.3–0.6). Results show that the critical Péclet number for flame survival is no greater than 3.25, consistent with prior literature, even under ultra-lean and high-pressure conditions (Φ ≤ 0.3, > 50 bar). Additionally, flames with Péclet numbers exceeding 6.15 exhibited front wrinkling, suggesting the onset of velocity-driven instabilities and enhanced flame robustness. These findings help define thresholds for flame quenching in confined geometries and support the safe design of H 2 fueled ICEs.

08 HYDROGEN↗
DOE OSTI2025

Statistical evaluation of microscale stress conditions leading to void nucleation in the weak shock regime

Here, we investigate the heterogeneity of the stress state driven by anisotropic deformation response at the single crystal level through five statistical volume element (SVE) calculations of polycrystalline BCC tantalum. This work focuses on grain boundaries as a prominent material defect type prone to void nucleation based upon experimental observations of predominantly intergranular void nucleation in this material. The SVEs are constructed to be statistically representative of larger volumes of material and are meshed such that mean and standard deviation of grain size and orientation information is reconstructed. The computational meshes feature hexahedral (brick) elements and smooth conformal grain boundaries where significant stress concentration is known to occur, a tail effect of interest in the extreme events process of dynamic ductile damage. An existing micromechanical crystallographic plasticity model shown to capture the single crystal behavior of BCC tantalum well is used to perform the polycrystal calculations. The model includes representation of the non-Schmid effect of non-planar screw dislocation kinetics in tantalum. A three-dimensional stress state time profile predicted by damage modeling of a flyer plate impact experiment is applied as boundary conditions to each SVE. Resulting grain boundary stress state statistics are strongly non-Gaussian. Significant structural evolution is observed within the compressive hold before unloading into tension in the stress profile. Strong angular dependence of grain boundary traction magnitude with shock direction is observed. Non-Schmid effects continue to suggest their influence on propensity of microstructural defect types to nucleate voids. A general void nucleation criterion is proposed using probability theory. The general framework is specified to polycrystalline BCC tantalum in the weak shock regime to include the SVE calculations and literature molecular dynamics calculations of grain boundary void nucleation strength. Probability density functions (PDFs) are used to describe the interaction between the local stress state heterogeneity and the distributed grain boundary void nucleation strength state. A causation entropy maximization procedure removes the requirement for ad hoc selection of a PDF functional form and provides a rigorous procedure for data-based PDF determination. The resulting physically informed PDF describes the spatial appearance frequency of nucleated voids as a function of applied macroscale pressure. Lower length scale physics are thus packaged in a precise and computationally efficient way to provide computational plasticity insight to macroscale dynamic ductile damage models.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Influence of initial conditions on data-driven model identification and information entropy for ideal mhd problems

Data-driven methods of model identification are able to discern governing dynamics of a system from data. Such methods are well suited to help us learn about systems with unpredictable evolution or systems with ambiguous governing dynamics given our current understanding. Many plasma problems of interest fall into these categories as there are a wide range of models that exist, however each model is only useful in a certain regime and often limited by computational complexity. To ensure data-driven methods align with theory, they must be consistent and predictable when acting on data whose governing dynamics are known. Weak Sparse Identification of Nonlinear Dynamics (WSINDy) is a recently developed data-driven method that has shown promise in learning governing dynamics from data with high noise levels [1]. This work examines how WSINDy acts on ideal MHD test problems as the initial conditions are varied and specifies limiting requirements for successful equation identification. Furthermore, it is hard to recover the governing dynamics from data that emphasize a single dominant behavior. In these low information cases, Shannon information entropy is able to pick up on the redundancies in the data that affect recoverability.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

The impact of chromium coatings on Zircaloy cladding deformation behavior under reactivity-initiated accident-like mechanical loading conditions

A reactivity-initiated accident (RIA) occurs when a control rod ejection or control blade drop causes an increase in the fission rate. The injection of energy results in an increase in fuel temperature which in turn causes rapid thermal expansion of the fuel pellet. This thermal expansion may result in pellet-cladding mechanical interaction (PCMI) in which the fuel imparts a mechanical strain to the cladding. PCMI may cause the cladding to fail, and thus, the mechanical response of cladding due to PCMI must be investigated when characterizing new cladding materials. Chromium-coated Zircaloy-4 is a near-term accident-tolerant fuel cladding that exhibits improved high-temperature oxidation resistance. Modified burst testing was utilized to experimentally simulate the effects of PCMI on both uncoated and chromium-coated Zircaloy cladding samples at hot zero power conditions. Samples were coated using either cold spraying or physical vapor deposition to understand the differences in behavior that the coating application method may cause. Digital image correlation was used to analyze images of the deforming specimens to extract the in-situ strain behavior of the cladding. The uncoated specimens burst at hoop strains ranging from 8.8 % to 17.2 %. The cold-spray chromium-coated Zircaloy specimens burst at hoop strains of 7.0 % to 11.0 %. The physical vapor deposition coated tubes burst at hoop strains of 9.1 % to 11.5 %. These results indicate that the chromium coating causes a loss in the ductility of the cladding. Furthermore, the higher burst hoop strains of the physical vapor deposition-coated samples relative to the cold-spray samples indicate that the cold-spraying technique causes more of a loss in ductility than physical vapor deposition. All samples burst at higher hoop strains than those expected to occur in an RIA for fresh cladding.

Accident-tolerant fuel↗
DOE OSTI2026

Simulations of grain growth in tungsten armor materials under ARC plasma edge operation conditions using an integrated plasma-edge/materials model

An integrated model of grain growth deuterium-exposed tungsten polycrystals, consisting of a two-dimensional vertex dynamics model fitted to atomistic data, has been developed to assess the grain growth kinetics of deuterium-exposed polycrystalline tungsten (W). The model tracks the motion of grain boundaries under the effect of driving forces stemming from grain boundary curvature and differential deuterium concentration accumulation. Here, we apply the model to experimentally synthesized tungsten polycrystals under deuterium-saturated conditions relevant to the ARC concept design. The results indicate rapid grain growth kinetics in the near-surface region adjacent to the plasma, where the temperature reaches 1400 K, whereas the microstructure remains stable deeper in the material with the lower temperature of 1000 K. The combined modeling and analysis further reveal that monolithic tungsten produced via conventional fabrication routes is highly susceptible to grain coarsening at temperatures exceeding 1000 K, largely independent of the magnitude of the applied driving force. Moreover, the accumulation of deuterium near grain boundaries has a pronounced inhibitory effect on grain boundary migration. High-angle grain boundaries ( > 50°) contribute more significantly to the overall grain growth process.

36 MATERIALS SCIENCE↗
DOE OSTI2024

High protonic resistance of hydrocarbon-based cathodes in PEM fuel cells under low humidity conditions: Origin, implication, and mitigation

Hydrocarbon-based electrodes for proton-exchange membrane fuel cells face challenges in closing the performance gap with electrodes based on perfluorosulfonic acid ionomers, particularly under low humidity conditions. Alongside increased oxygen transport resistance and higher kinetic-induced overpotentials, the protonic resistance of these fluorine-free electrodes is the primary hurdle to improved performance. This study systematically investigates the origin and impact of the cathode protonic resistance on fuel cell performance, utilizing sulfonated phenylated polyphenylenes as hydrocarbon ionomers. Electrochemical characterization at low relative humidity (≤50 %) reveal a high protonic resistance arising from both lower conductivity of the hydrocarbon thin film compared to the bulk membrane and increased cathode tortuosity at a gas transport-optimized ionomer to carbon (I/C) ratio of 0.2. The poor protonic resistance at low relative humidities leads to a non-homogeneous current distribution across the thickness of the cathode electrode, resulting in lower catalyst utilization. To address this issue, reducing the thickness of the cathode CL while maintaining a constant Pt loading (i.e., increasing the Pt on carbon ratio) significantly reduces protonic resistance. This improvement compensates for the kinetic disadvantages of highly loaded carbon particles and results in a considerable performance increase by 40 % at 0.75 V under low relative humidities.

25 ENERGY STORAGE↗
DOE OSTI2025

Improved guided-wave acoustic defect detection and localization in pipes under varying temperature conditions using deep learning

Early defect detection in pipelines is critical across industries, particularly in the oil and gas sector, where failures result in significant maintenance costs and operational disruptions. Acoustic guided-wave techniques are widely used for nondestructive evaluation of pipeline defects due to their long-distance propagation capability. However, environmental variations, sensitivity limitations, and complex signal interpretation challenges limit the effectiveness of traditional signal processing approaches with guided-wave signals. Recent advances in deep learning methods have demonstrated remarkable success in solving complex real-world problems in many fields. In particular, deep-learning-based signal processing holds substantial promise to overcome limitations and challenges of conventional signal processing. This study presents a deep learning framework for pipeline inspection using acoustic guided-wave signals under temperature varying environments. The proposed framework employs a dual-path one-dimensional convolutional autoencoder that combines defect detection, localization, and temperature prediction functions. The proposed system utilizes multi-mode and broadband acoustic waves with an optimized number of sensors that provide high accuracy while retaining practical simplicity. Experimental validation is performed on a carbon steel pipe. The results indicate exceptional defect detection accuracy and precise defect localization with a mean absolute error of 66 mm. The proposed technique also predicts the effective average temperature of the pipe with a mean absolute error of 0.2°C. Comparative analysis shows superior performance of the proposed method over a traditional method previously developed by the authors' team. These results highlight the potential of integrating deep learning methods into guided-wave pipeline inspection systems to improve reliability under varying environmental conditions.

42 ENGINEERING↗
DOE OSTI2025

Optimizing processing conditions for additively reinforced thermoforming (ART) in convergent manufacturing

This study utilized additively reinforced thermoforming (ART) to enhance the thermomechanical properties of polyethylene terephthalate glycol (PETG) sheet. ART materials were produced by overprinting PETG/carbon fiber filament (PETG/CF) on neat PETG sheets at varying conditions. The mechanical properties of the PETG sheet, PETG/CF, and ART materials were assessed, showing that ART exhibited superior tensile strength and modulus of elasticity. The tensile strength and modulus in the x-direction for ART at 265°C were 57.32 ± 2.9 MPa and 3.41 ± 0.4 GPa, respectively, compared to 49.1 ± 0.5 MPa and 1.92 ± 0.09 GPa for neat PETG. Microstructural analysis revealed strong interfacial adhesion between layers, while thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), and heat deflection temperature analysis provided insights into the ART material's thermoforming behavior, aiding design optimization for enhanced stiffness, reduced necking, and improved customization. In conclusion, this information can be used to design for the thermoforming operation.

Additive reinforcement↗
DOE OSTI2026

Robustness of the smartpixels classifier for different simulated sensor geometries and non-ideal detector conditions

Pixel tracking detectors at upcoming collider experiments will see unprecedented charged-particle densities. Real-time data reduction on the detector will enable higher granularity and faster readout, possibly enabling the use of the pixel detector in high-rate online event selection, such as the ATLAS or CMS first-level trigger systems. This data reduction can be accomplished with a neural network (NN) in the readout chip bonded with the sensor that recognizes and rejects tracks with low transverse momentum (p T ) based on the geometrical shape of the charge deposition (“cluster”). To design viable detectors for deployment, the dependence of the NN as a function of the sensor geometry, external magnetic field, irradiation, and noise must be understood. In this paper, we present first studies of the efficiency and data reduction for planar pixel sensors exploring these parameters. For the CMS HL-LHC sensor geometry, we obtain a signal efficiency of (91.9 ± 0.7)% and a data reduction of (29.7 ± 1.0)%. A smaller sensor pitch in the bending direction improves the p T discrimination, but a larger pitch can be partially compensated with detector thickness. Any accumulated radiation damage also changes the cluster shape, reducing the signal efficiency compared to the baseline by approximately 30–60% in absolute terms, but nearly all of the performance can be recovered through retraining of the network and updating the weights. Finally, the impact of noise was investigated, and retraining the network on noise-injected datasets was found to maintain performance within 6% of the baseline network trained and evaluated on noiseless data. •ASIC-compatible track-momentum classifier is robust in realistic detector conditions.•About 90% signal efficiency and 30% data reduction per layer for CMS HL-LHC geometry.•Single-layer signal efficiency increases for smaller pixel pitch or thicker sensors.•Performance with noise or after radiation damage mostly recovered by retraining.

Shekar, Danush [Illinois U., Chicago] (ORCID:00000↗
DOE OSTI2026

Beyond interpolation: Physics-inspired gating transformers for extrapolating irradiation conditions to novel nuclear fuels

The qualification of advanced nuclear fuels relies on irradiation experiments in test reactors that emulate commercial conditions. Designing these tests requires accurate prediction of key irradiation quantities, particularly heat generation rate and burnup, yet obtaining them typically involves computationally expensive multi-step simulation workflows. We propose a physics-inspired gating transformer (PIGT) that integrates an inverse-square, distance-based attenuation into the encoder representation to bias attention toward physically relevant spatial relationships while retaining data-driven flexibility. Using MiniFuel irradiation data from the High Flux Isotope Reactor at Oak Ridge National Laboratory, we benchmark against ensemble methods, feedforward and recurrent networks, convolutional models, and standard transformers. While baseline models perform well under interpolation, they exhibit a pronounced generalization gap when evaluated on fuels not included in the training set. The proposed model consistently improves extrapolative accuracy and stability, yielding the strongest performance on unseen fuel configurations. These results indicate that a lightweight physics structure embedded within attention mechanisms can substantially improve robustness, enabling more reliable surrogate predictions to accelerate the design of nuclear fuel irradiation experiments.

Fuel qualification↗