Search NASA⌕ Search

SEARCH · Search NASA

Results for “inverse design”

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 OSTI2021

Surface Charge Distribution Mapping Using an Electrostatic Probe

Undesired accumulation of charge on dielectric materials causing electrostatic discharges can be an issue in pulsed power systems and electrical systems in general. As such, an understanding of surface charge distribution, charge accumulation, and decay is required. An in-house no-contact electrostatic probe designed with a 2 mm resolution was fabricated to measure and map surface charge distribution. Designed as a contactless instrument, the probe consisting of a metal rod and low leakage amplifier circuit probes the electric potential near the surface. A scan of the surface then provides the raw data, a potential distribution. The actual probe response - i.e., the surface charge to voltage transfer function, is accounted for in post-processing. This is accomplished via an Inverse Wien Filter - a technique often applied in image processing - to deconvolve the probe response from the measured data. A commercially available electrostatic probe, the Trek 341B meter with a 3455ET probe, capable of measuring +/- 20 kV was compared to the in- house probe that is designed to cover a wider range from +/- 35 kV. A resolution better than the simple probe resolution is achieved through the distribution’s scanning voltage method and deconvolution. Applying repeated scans, surface charge decay was tracked on various polymer materials to determine the material and environment dependence; materials included were PA6, PTFE, and others. As an example of material dependence, samples charged to 20 kV at 65% humidity experience full charge decay in approximately 45 seconds for PA5 (152 mm dia.), while it took about 100 times longer for PTFE (51 mm dia.) to observe the same decay/redistribution of charge.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Reflections on the anomalous ANITA events: the Antarctic subsurface as a possible explanation

The Antarctic Impulsive Transient Antenna (ANITA) balloon experiment was designed to detect radio signals initiated by high-energy neutrinos and cosmic ray (CR) air showers. These signals are typically discriminated by the polarization and phase inversions of the radio signal. The reflected signal from CRs suffer phase inversion compared to a direct ‘tau neutrino’ event. In this paper, we study subsurface reflection, which can occur without phase inversion, in the context of the two anomalous up-going events reported by ANITA. It is found that subsurface layers and firn density inversions may plausibly account for the events, while ice fabric layers and wind ablation crusts could also play a role. This hypothesis can be tested with radar surveying of the Antarctic region in the vicinity of the anomalous ANITA events. Future experiments should not use phase inversion as a sole criterion to discriminate between down-going and up-going events, unless the subsurface reflection properties are well understood.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

A New Galerkin Quadrature Method Not Requiring a Matrix Inverse

We derive a new Galerkin quadrature (GQ) method for S 𝑛 calculations that differs from the two methods preceding it in that a matrix inverse for an 𝑁 𝑑 × 𝑁 𝑑 matrix, where 𝑁𝑑 is the number of directions in the quadrature set, is no longer required. Galerkin quadrature methods are designed for calculations with highly anisotropic scattering. Such methods are not simply special angular quadratures but also are methods for representing the S 𝑛 scattering source that offers several advantages relative to the standard scattering source representation when highly truncated Legendre cross-section expansions must be used. Galerkin quadrature methods are also useful when the scattering is moderately anisotropic, but the quadrature being used is not sufficiently accurate for the order of the scattering source expansion that is required. Furthermore, we derive the new method and present computational results showing that its performance for two challenging problems is comparable to those of the two GQ methods that preceded it.

Galerkin quadrature↗
DOE OSTI2020

Iterative Learning Control for Video-rate Atomic Force Microscopy

We present a control scheme for video-rate atomic force microscopy with rosette pattern. The controller structure involves a feedback internal-model-based controller and a feedforward iterative learning controller. The iterative learning controller is designed to improve tracking performance of the feedback-controlled scanner by rejecting the repetitive disturbances arising from the system nonlinearities. We investigate the performance of two inversion techniques for constructing the learning filter. We conduct tracking experiments using a two-degree-of-freedom microelectromechanical system (MEMS) nanopositioner at frame rates ranging from 5 to 20 frames per second. Furthermore, the results reveal that the algorithm converges rapidly and the iterative learning controller significantly reduces both the transient and steady-state tracking errors. We acquire and report a series of high-resolution time-lapsed video-rate AFM images with the rosette pattern.

42 ENGINEERING↗
DOE OSTI2025

First-Principles Studies on Sc 2 RuZ (Z = Si, Ge, Sn) Inverse Heusler Alloys: Structural, Electronic, and Transport Properties

The continuous demand for efficient, nontoxic, and thermally stable materials for room-temperature energy conversion motivates the exploration of novel thermoelectric systems beyond the traditional magnetic Heusler alloys. While full and half-Heusler compounds, especially Co-, Ni-, and Mn-based systems, have demonstrated promising thermoelectric properties, their typically high operating temperatures and magnetic complexities limit their applicability in ambient thermal management. In this context, we investigate whether Sc-based inverse Heusler alloys can offer a viable nonmagnetic alternative with competitive thermoelectric performance. In this work, we perform a systematic first-principles study of the inverse Heusler compounds Sc 2 RuZ (Z = Si, Ge, Sn), focusing on their structural, electronic, mechanical, and thermodynamic-thermoelectric properties. Density Functional Theory (DFT) was employed to compute optimized lattice structures and band dispersion, while dynamical stability was assessed via phonon calculations. Thermoelectric transport coefficients, including Seebeck coefficient, electrical conductivity, and thermal conductivity, were estimated using the semiclassical Boltzmann transport theory within the constant relaxation time approximation. Our results show that all Sc 2 RuZ compounds are thermodynamically stable semiconductors with indirect band gaps of 0.12–0.16 eV and exhibit high elastic moduli, especially Sc 2 RuSn, which demonstrates superior stiffness and incompressibility. Importantly, all compounds display promising room-temperature thermoelectric characteristics, including high Seebeck coefficients and power factors. These findings reveal that Sc 2 RuZ alloys represent a rare class of stable, nonmagnetic inverse Heusler semiconductors with intrinsic thermoelectric potential at room temperature, unlike many existing Heusler systems optimized for spintronics or high-temperature operation. This work expands the known design space for Heusler-based thermoelectrics and offers a theoretical basis for experimental realization of efficient, low-temperature, nonmagnetic thermoelectric materials.

alloys↗
DOE OSTI2024

Programming liquid crystal elastomers for multistep ambidirectional deformability

Ambidirectionality, which is the ability of structural elements to move beyond a reference state in two opposite directions, is common in nature. However, conventional soft materials are typically limited to a single, unidirectional deformation unless complex hybrid constructs are used. We exploited the combination of mesogen self-assembly, polymer chain elasticity, and polymerization-induced stress to design liquid crystalline elastomers that exhibit two mesophases: chevron smectic C (cSmC) and smectic A (SmA). Inducing the cSmC-SmA–isotropic phase transition led to an unusual inversion of the strain field in the microstructure, resulting in opposite deformation modes (e.g., consecutive shrinkage or expansion and right-handed or left-handed twisting and tilting in opposite directions) and high-frequency nonmonotonic oscillations. Further, this ambidirectional movement is scalable and can be used to generate Gaussian transformations at the macroscale.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Progress and R/D challenges for FCC-ee SRF

The FCC-ee machines present a huge challenge for the RF systems, which need to be adapted to very diverse beam conditions going from moderate energy and high current for the Z machine to high energy and low beam current for the ttbar. This inverse scaling results naturally from a fixed budget for the synchrotron radiation, which the SRF cavities need to compensate. A global solution was elaborated for the FCC Conceptual Design Report (Abada in Eur Phys J Spec Top 228):261–623, 2019), and is referred here as the baseline. Recently, further studies have led to a new optimized baseline, still based on traditional elliptical cavities. In parallel, a novel concept, named the Slotted Waveguide ELLiptical (SWELL), was proposed with the potential of greatly simplified logistics and reduced costs. Under several aspects, all these changes call for enhanced performance of the RF systems. A vigorous R&D program has therefore continued since the publication of the CDR, with the aim of pushing the performance and demonstrating the feasibility of a more advanced baseline and, more recently, of the SWELL option. The progress and challenges of this ambitious program were presented in the dedicated SRF sessions at FCC week 2022 and are summarized in this paper.

47 OTHER INSTRUMENTATION↗
DOE OSTI2024

Fiber Sorbents – A Versatile Platform for Sorption-Based Gas Separations

Increasing demand for high-purity fine chemicals and a drive for process intensification of large-scale separations have driven significant work on the development of highly engineered porous materials with promise for sorption-based separations. While sorptive separations in porous materials offer energy-efficient alternatives to longstanding thermal-based methods, the particulate nature of many of these sorbents has sometimes limited their large-scale deployment in high-throughput applications such as gas separations, for which the necessary high feed flow rates and gas velocities accrue prohibitive operational costs. These processability limitations have been historically addressed through powder shaping methods aimed at the fabrication of structured sorbent contactors based on pellets, beads or monoliths, commonly obtained as extrudates. These structures overcome limitations such as elevated pressure drops commonly recorded across powder adsorption beds but often accrue thermal limitations arising from elevated particle density and aggregation, which ultimately cap their maximum separation performance. Furthermore, the harsh mechanical strain to which powder particles are subjected during contactor fabrication, in the form of extrusion/compression forces, can result in partial pore occlusion and framework degradation, further limiting their performance. Here, we present the development of porous fiber sorbents as an alternative sorbent contactor design capable of addressing sorbent processability limitations while enabling an array of performance-maximizing heat integration capabilities. This new sorbent form factor leverages pre-existing know-how from hollow fiber spinning to produce fiber-shaped sorbent contactors through the phase inversion of known polymers in a process known as dry-jet/wet quenching. The process of phase inversion allows microporous sorbent particles to be latched onto a macroporous polymer matrix under mild processing conditions, thus making it compatible with soft porous materials prone to amorphization under traditional pelletization conditions. Sorbent fibers can be created with different geometries through control of the spinning apparatus and process, offering the possibility to produce monolithic and hollow fibers alike, the latter of which can be integrated with thermalization fluid flows. In this Account, we summarize our progress in the field of fiber sorbents from both design and application standpoints. We further guide the reader through the evolution of this field from the early inceptive work on zeolite hollow fibers to recent developments on MOF fibers. We highlight the versatile nature of fiber sorbents, both from the composition, fabrication and structure points of view, and further demonstrate how fiber sorbents offer alternative paths in tackling new and challenging chemical separation challenges like direct air capture (DAC), with a final perspective on the future of the field.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Learning and meta-learning of stochastic advection–diffusion–reaction systems from sparse measurements

Physics-informed neural networks (PINNs) were recently proposed in [18] as an alternative way to solve partial differential equations (PDEs). A neural network (NN) represents the solution, while a PDE-induced NN is coupled to the solution NN, and all differential operators are treated using automatic differentiation. Here, we first employ the standard PINN and a stochastic version, sPINN, to solve forward and inverse problems governed by a non-linear advection–diffusion–reaction (ADR) equation, assuming we have some sparse measurements of the concentration field at random or pre-selected locations. Subsequently, we attempt to optimise the hyper-parameters of sPINN by using the Bayesian optimisation method (meta-learning) and compare the results with the empirically selected hyper-parameters of sPINN. In particular, for the first part in solving the inverse deterministic ADR, we assume that we only have a few high-fidelity measurements, whereas the rest of the data is of lower fidelity. Hence, the PINN is trained using a composite multi-fidelity network, first introduced in [12], that learns the correlations between the multi-fidelity data and predicts the unknown values of diffusivity, transport velocity and two reaction constants as well as the concentration field. For the stochastic ADR, we employ a Karhunen–Loève (KL) expansion to represent the stochastic diffusivity, and arbitrary polynomial chaos (aPC) to represent the stochastic solution. Correspondingly, we design multiple NNs to represent the mean of the solution and learn each aPC mode separately, whereas we employ a separate NN to represent the mean of diffusivity and another NN to learn all modes of the KL expansion. For the inverse problem, in addition to stochastic diffusivity and concentration fields, we also aim to obtain the (unknown) deterministic values of transport velocity and reaction constants. The available data correspond to 7spatial points for the diffusivity and 20 space–time points for the solution, both sampled 2000 times. We obtain good accuracy for the deterministic parameters of the order of 1–2% and excellent accuracy for the mean and variance of the stochastic fields, better than three digits of accuracy. In the second part, we consider the previous stochastic inverse problem, and we use Bayesian optimisation to find five hyper-parameters of sPINN, namely the width, depth and learning rate of two NNs for learning the modes. Here, we obtain much deeper and wider optimal NNs compared to the manual tuning, leading to even better accuracy, i.e., errors less than 1% for the deterministic values, and about an order of magnitude less for the stochastic fields.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Bimetallic-Derived Catalytic Structures for CO 2 -Assisted Ethane Activation

In recent years, the simultaneous upgrading of CO 2 and ethane has emerged as a promising approach for generating valuable gaseous (CO, H 2 , and ethylene) and liquid (aromatics and C 3 oxygenates) chemicals from the greenhouse gas CO 2 and large-reserved shale gas. The key challenges for controlling product selectivity lie in the selective C-H and C-C bond cleavage of ethane with the assistance of CO 2 . Bimetallic-derived catalysts likely undergo alloying or oxygen-induced segregation under reaction conditions, thus providing diverse types of interfacial sites, e.g., metal/support (M/M'O x ) interface and metal oxide/metal (M'O x /M) inverse interface, that are beneficial for selective CO 2 -assisted ethane upgrading. The alloying extent can be initially predicted by cohesive energy and atomic radius (or Wigner-Seitz radius), while the preference for segregation to form the on-top suboxide can be approximated using the work function, electronegativity, and binding strength of adsorbed oxygen. Furthermore, bimetallic-derived catalysts are typically supported on high surface area oxides. Modifying the reducibility and acidity/basicity of the oxide supports and introducing surface defects facilitate CO 2 activation and oxygen supplies for ethane activation. Using in situ synchrotron characterization and density functional theory (DFT) calculations, we found that the electronic properties of oxygen species influence the selective cleavage of C-H/C-C bonds in ethane, with electron-deficient oxygen over the metal (or alloy) surface promoting nonselective bond scission to produce syngas and electron-enriched oxygen over the metal oxide/metal interface enhancing selective C-H scission to yield ethylene. Here, we further demonstrate that the preferred structures of the catalyst surfaces, either alloy surfaces or metal oxide/metal inverse interfaces, can be controlled through the appropriate choice of metal combinations and their atomic ratios. Through a comprehensive comparison of experimental results and DFT calculations, the selectivity of C-C/C-H bond scission is correlated with the thermodynamically favorable bimetallic-derived structures (i.e., alloy surfaces or metal oxide/metal inverse interfaces) under reaction conditions over a wide range of bimetallic catalysts. These findings not only offer structural and mechanistic insights into bimetallic-derived catalysts but also provide design principles for selective catalysts for CO 2 -assisted activation of ethane and other light alkanes. This Account concludes by discussing challenges and opportunities in designing advanced bimetallic-derived catalysts, incorporating new reaction chemistries for other products, employing precise synthesis strategies for well-defined structures with optimized site densities, and leveraging time/spatial/energy-resolved in situ spectroscopy/scattering/microscopy techniques for comprehensive structural analysis. The research methodologies established here are helpful for the investigation of dynamic alloy and interfacial structures and should inspire more efforts toward the simultaneous upgrading of CO 2 and shale gas.

03 NATURAL GAS↗
DOE OSTI2023

Same-sided successive-shock HED instability experiments

Inertial confinement fusion (ICF) and high-energy density (HED) physics experiments experience complicated forcing for instability growth and mix due to the ubiquitous presence of multiple shocks interacting with perturbations on multiple material interfaces. One common driver of instability growth is successive shocks from the same direction. However, there is a severe lack of analytic work and modeling validation for same-sided successive shocks since they are extremely difficult to achieve with conventional (non-HED) drivers. Successive shocks access a large instability parameter space; idealized fluid theory [K. O. Mikaelian, Phys. Rev. A 31, 410 (1985)] predicts 15 different interface evolution scenarios for a sinusoidal perturbation. Growth becomes more complex for multi-mode, compressible HED systems. The Mshock campaign is the first experiment in any fluid regime to probe a wide portion of successive shock parameter space. This is enabled by our development of a hybrid direct/indirect drive platform capable of creating independently controllable successive shocks on the National Ignition Facility. These experiments have delivered the first data capable of rigorously challenging our models and their ability to accurately capture Richtmyer–Meshkov growth under successive shocks. Single-mode and two-mode experiments have successfully demonstrated the ability to access and control the various growth scenarios of the shocked interface, including re-inversion, freeze out, and continued growth. Simulations and theoretical modeling are shown to accurately capture the experimental observations in the linear growth phase, giving us confidence in our ICF/HED design codes.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2025

The Artificial Scientist: in-Transit Machine Learning of Plasma Simulations

Large-scale simulations or scientific experiments produce petabytes of data per run. This poses massive challenges for I/O and storage when scientific analysis workflows are run manually offline. Unsupervised deep learning-based techniques to extract patterns and non-linear relations from these large amounts of data provide a way to build scientific understanding from raw data, reducing the need for manual pre-selection of analysis steps, but require exascale compute and memory to process the full dataset available. In this paper, we demonstrate a heterogeneous streaming workflow in which plasma simulation data is streamed directly to a Machine Learning (ML) application training a model on the simulation data in-transit, completely circumventing the capacity-constrained filesystem bottleneck. This workflow employs openPMD to provide a high level interface to describe scientific data and also uses ADIOS2, to transfer volumes of data that exceed the capabilities of the filesystem. We employ experience replay to avoid catastrophic forgetting in learning from this non-steady state process in a continual manner and adapt it to improve model convergence while learning in-transit. As a proof-of-concept, we approach the ill-posed inverse problem of predicting particle dynamics from radiation in a particle-incell (PIConGPU) simulation of the Kelvin-Helmholtz instability (KHI). We detail hardware-software co-design challenges as we scale PIConGPU to full Frontier, the Top-1 system as of June 2024 Top500 list.

Kelling, Jeffrey [Helmholtz-Zentrum Dresden Rossen↗
DOE OSTI2020

Beyond the acceptance limit of DRAGON: The case of the 6 Li(α,γ) 10 B reaction

Radiative capture reactions play a pivotal role for our understanding of the origin of the elements in the cosmos. Recoil separators provide an effective way to study these reactions, in inverse kinematics, and take advantage of the use of radioactive ion beams. However, a limiting factor in the study of radiative capture reactions in inverse kinematics is the momentum spread of the product nuclei, which can result in an angular spread larger than the geometric acceptance of the separator. The DRAGON facility at TRIUMF is a versatile recoil separator, designed to study radiative capture reactions relevant to astrophysics in the A~10–30 region. In this work we present the first attempt to study with DRAGON a reaction, 6 Li(α, γ) 10 B, for which the recoil angular spread exceeds DRAGON’s acceptance. Furthermore, our result is in good agreement with the literature value, showing that DRAGON can measure resonance strengths of astrophysically important reactions even when not all the recoils enter the separator.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2024

Benchmarking of Different Inverse Point Kinetics Implementations for an Autocorrected Reactimeter Algorithm

In November 2017, the Transient Reactor Test Facility returned to operation. Since that time, many transient test series have been completed; each has provided valuable data for materials performance, reactor safety that can be applied in future designs. During each experimental series, detector count rates provided important information on the core behavior during transients. However, a limitation of these data is that variations in the neutron distribution during experiments cause errors when attempting to infer reactivity evolution from detector signals. Neutron physics codes can be used to compute the flux shape variations. However, this is a poor solution when the experimental data is used to do verification, validation and uncertainty quantification (VVUQ) on codes. Indeed, if the output of the code is used both as a reference and to correct what the reference is compared to, the circular dependency limits the quality of the VVUQ approach. To overcome this problem, an Autocorrected Reactimeter Algorithm (ACRA) has been developed. This approach infers time-dependent reactivity evolution by testing different spatial corrections and selecting the one that minimizes reactivity variations when the core is in a frozen configuration (i.e. when there is no variation in parameters affecting reactivity). However, the scope of this method was limited to transients where there were negligible thermal feedback. Indeed, the core is never in a frozen configuration when the fuel temperature varies during the whole transient. This is our motivation for the development of an improved version of the ACRA which does not require frozen configurations

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2026

Physics model enhancements in MARLEY version 2

Large next-generation neutrino detectors, such as those designed for accelerator-based oscillation experiments, are expected to have impressive sensitivity to MeV astrophysical neutrinos generated by the Sun and by supernovae. There are also numerous related prospects to search for MeV-scale phenomena beyond the Standard Model. In contrast to detectors constructed with water and hydrocarbon targets, which will be primarily sensitive to electron antineutrinos via inverse beta decay, the upcoming Deep Underground Neutrino Experiment (DUNE) will provide unparalleled capabilities to measure low-energy electron neutrinos from cosmic sources. Interpreting such future measurements in DUNE will require a high-quality simulation of MeV interaction physics in the argon nucleus. The MARLEY (Model of Argon Reaction Low Energy Yields) neutrino event generator has been designed with the goal of delivering the needed simulation capabilities for DUNE and other argon-based neutrino detectors. In this poster, I will present recent improvements to MARLEY's suite of interaction models, which will be released for general use by the neutrino community in version 2 of the software.

Gardiner, Steven James [Fermilab] (ORCID:000000028↗
DOE OSTI2026

Physics model enhancements in MARLEY version 2

Large next-generation neutrino detectors, such as those designed for accelerator-based oscillation experiments, are expected to have impressive sensitivity to MeV astrophysical neutrinos generated by the Sun and by supernovae. There are also numerous related prospects to search for MeV-scale phenomena beyond the Standard Model. In contrast to detectors constructed with water and hydrocarbon targets, which will be primarily sensitive to electron antineutrinos via inverse beta decay, the upcoming Deep Underground Neutrino Experiment (DUNE) will provide unparalleled capabilities to measure low-energy electron neutrinos from cosmic sources. Interpreting such future measurements in DUNE will require a high-quality simulation of MeV interaction physics in the argon nucleus. The MARLEY (Model of Argon Reaction Low Energy Yields) neutrino event generator has been designed with the goal of delivering the needed simulation capabilities for DUNE and other argon-based neutrino detectors. In this talk, I will present recent improvements to MARLEY's physics models, which will be released for general use by the neutrino community in version 2 of the software.

Gardiner, Steven James [Fermilab] (ORCID:000000028↗
DOE OSTI2023

Current rectification based on noncentrosymmetric quantum materials

Rectification is a process that converts electromagnetic fields into direct current (DC). Such a process underlies a wide range of technologies, including wireless communication, wireless charging, energy harvesting, and infrared detection. Existing rectifiers are mostly based on semiconductor diodes, with limited applicability to small voltages or high frequency inputs. Here, we present an alternative approach to current rectification that uses the electronic properties of quantum crystals without semiconductor junctions. We identify a new mechanism for rectification from skew scattering due to the chirality of itinerant electrons in time-reversal-invariant but inversion-breaking materials. Our calculations reveal large, tunable rectification effects in graphene multilayers and transition metal dichalcogenides. These effects can be used in high-frequency rectifiers by rational material design and quantum wavefunction engineering.

Isobe, Hiroki↗
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↗