Toward Multi-Modal Foundation Models for Inertial Confinement Fusion: Combining Structured Features and Diagnostic Images
Explore the source record for details and available documents.
Computer science↗
SEARCH · Engineering Papers
Results for “feature fusion”
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.
A two-stage optical fusion framework for wildfire severity mapping across the conterminous United States
Accurate wildfire severity mapping (WSM) is essential for post-fire recovery planning, erosion risk assessment, ecosystem monitoring, and disaster risk reduction. Although Landsat and Sentinel optical imagery have been widely used for burn severity assessment, the added value of fusing multiple optical sensors has not been sufficiently quantified across diverse fire events, particularly since the launch of Landsat-9. This study evaluates whether multisensor optical fusion improves wildfire severity mapping relative to single-sensor baselines using Sentinel-2, Landsat-8, and Landsat-9 imagery across 40 wildfire events in the conterminous United States. We tested a two-stage fusion framework that combines feature-level fusion with pixel-level dimensionality reduction. First, feature-level fused datasets were created through early fusion by combining standardized post-fire bands from each sensor into a single predictor stack. Both raw reflectance bands and pairwise spectral transforms were retained to capture within- and cross-sensor spectral interactions. Second, Linear Discriminant Analysis was applied to both single-sensor and fused datasets to produce comparable low-dimensional feature spaces. Six machine-learning classifiers were then used to benchmark model performance with repeated spatially buffered train–test splits. Results show that Landsat-9 was the strongest single-sensor baseline. Among the fusion strategies, Sentinel-2 + Landsat-9 produced the most consistent improvement and reduced performance variability. Landscape-condition analysis further showed that this fusion was most beneficial in shrubland-dominated and high-terrain fires, where it achieved the highest overall mean accuracy and the fewest failures. In contrast, its benefits were less reliable in evergreen forests, mixed vegetation, and low- to moderate-elevation terrain. In operational settings, the Sentinel-2 + Landsat-9 configuration offers a practical solution for post-fire recovery planning, erosion-risk assessment, watershed management, and ecological monitoring when field observations are available and timely satellite-based information is needed.
Landsat↗
Peak2Patch: High-Fidelity Functional Group Identification through Attention-Based Fusion of Infrared and Mass Spectra
Identifying molecular structure based on spectroscopic readings is a key task in a variety of chemical and biological applications. Common spectroscopy techniques, such as Infrared (IR) Spectroscopy and Mass Spectrometry (MS), provide detailed information on the structure of molecular compounds but nonetheless require expert-level knowledge to decode. Machine learning has emerged as a potential solution for automating structure prediction from chemical spectra; however, current approaches generally focus on single sensor modalities, neglecting to leverage the complementary information contained within differing spectra. In this paper, we introduce Peak2Patch, a novel approach to fusion-enhanced prediction of functional groups from IR and mass spectra. First, we perform a detailed comparison of backbone networks for encoding both sparse mass spectra and dense IR spectra and demonstrate the superior performance of transformer neural networks over current state-of-the-art convolutional neural networks. Second, we evaluate three broad categories of fusion: early (raw feature), middle (deep feature), and late (decision) fusion, demonstrating the potential of a deep feature fusion-based approach. Lastly, we present Peak2Patch, our attention-based fusion scheme, which leverages cross-attention to mix features between encoded tokens of the two modalities. We validate our approach on a publicly available multimodal spectroscopic data set of 790k simulated molecules, demonstrating a large improvement in functional group prediction over both the previous state-of-the-art and our own strong single-modal baselines.
Jacobson, Philip [Sandia National Laboratories (SN↗
Data-driven prediction of geometry- and toolpath sequence-dependent intra-layer process conditions variations in laser powder bed fusion
Geometrical features and toolpath sequence are two important factors that cause process condition variations, such as variations in the meltpool temperature or meltpool size, that might lead to undesired material properties in the laser powder bed fusion (LPBF) process. Due to the high dynamics and complex physics of the LPBF process, it is difficult to predict variations in process conditions with simulations alone. Advances in measurement technology and computational technologies open up new possibilities for smart manufacturing. In this paper, a data-driven method to predict intra-layer variations in the processing conditions that source from the toolpath sequence and part geometry is presented. The approach is demonstrated using two-color on-axis pyrometer measurements. Three demonstration cases are presented in which it is demonstrated (1) how the trained predictive model can be used as a filter to ease the interpretation of process variations and discover patterns related to toolpath and part geometry, and (2) how to generate predictions that can be used for feedforward control, i.e., for adjusting laser power or scanning speed along the toolpath using a meltpool temperature prediction model generated based on on-axis measurements. Results show that the developed prediction model is able to meaningfully predict process variations resulted from toolpath sequence and geometry. Predictions are aligned with the results from the related work of others and for the case of 180° laser path turnarounds in our high-speed X-ray imaging experiments. In conclusion, the potential issues related to the current maturity status of the process and measuring equipment that could in practice affect the performance of the proposed solutions are also discussed.
42 ENGINEERING↗
A Comparison Study of Feature Extraction and Data Fusion Techniques for Improving Cyber-Physical Situational Awareness
Explore the source record for details and available documents.
Fragkos, Georgios↗
Magnetization around mix jets entering inertial confinement fusion fuel
Engineering features are known to cause jets of ablator materials to enter the fuel hot-spot in inertial confinement fusion implosions. The Biermann battery mechanism wraps them in a self-generated magnetic field. Here we show that higher-Z jets have an additional thermoelectric magnetic source term that is not present for hydrogen jets, verified here through a kinetic simulation. It has similar magnitude to the Biermann term. We then include this in an extended magneto-hydrodynamic approach to post process an xRAGE radiation-hydrodynamic implosion simulation. The simulation includes an accurate model for the capsule fill tube, producing a dense carbon jet that becomes wrapped in a 4000 T magnetic field. A simple spherical carbon mix model shows that this insulates the electron heat conduction enough to cause contraction of the jet to an optically thick equilibrium. The denser magnetized jet hydrodynamics could change its core penetration and therefore the final mix mass, which is known to be well correlated with fusion yield degradation. Fully exploring this will require self-consistent magneto-hydrodynamic simulations. Experimental signatures of this self-magnetization may emerge in the high energy neutron spectrum.
70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
Hybridized Discontinuous Galerkin Methods for Computational Fluid Dynamics
Hybridizable Discontinuous Galerkin (HDG) methods hold promise for any applications with significant advection character, including thermal hydraulics in light water reactors and advanced reactor concepts and fluid models of plasmas in magnetic confinement fusion. Its features include natural upwinding, local element conservation, and extensibility to arbitrarily high order accuracy. In the last fiscal year we have implemented HDG in the Multiphysics Object-Oriented Simulation Environment (MOOSE). We developed a first-of-its-kind automatic static condensation system in MOOSE’s underlying finite element library libMesh which can condense out arbitrarily many internal variables. Finally, we developed the first preconditioner for HDG discretizations of the Navier-Stokes equations which shows robust performance across a wide range of problem sizes and Reynolds numbers. This preconditioner yields solution times that are equivalent to the fastest developed for industry standard finite volume methods. Moreover, the arbitrarily high-order nature of HDG makes it a prime candidate for acceleration via graphical processing units (GPUs). We believe these developments will hold significant importance for future DOE Nuclear Energy (NE) and Fusion Energy Science (FES) programs.
97 MATHEMATICS AND COMPUTING↗
Decision-Based Fusion for Vehicle Matching
In this work, a framework is proposed for decision fusion utilizing features extracted from vehicle images and their detected wheels. Siamese networks are exploited to extract key signatures from pairs of vehicle images. Our approach then examines the extent of reliance between signatures generated from vehicle images to robustly integrate different similarity scores and provide a more informed decision for vehicle matching. To that end, a dataset was collected that contains hundreds of thousands of side-view vehicle images under different illumination conditions and elevation angles. Experiments show that our approach could achieve better matching accuracy by taking into account the decisions made by a whole-vehicle or wheels-only matching network.
47 OTHER INSTRUMENTATION↗
A Comparison of Infectious Disease Forecasting Methods across Locations, Diseases, and Time
Accurate infectious disease forecasting can inform efforts to prevent outbreaks and mitigate adverse impacts. This study compares the performance of statistical, machine learning (ML), and deep learning (DL) approaches in forecasting infectious disease incidences across different countries and time intervals. We forecasted three diverse diseases: campylobacteriosis, typhoid, and Q-fever, using a wide variety of features (n = 46) from public datasets, e.g., landscape, climate, and socioeconomic factors. We compared autoregressive statistical models to two tree-based ML models (extreme gradient boosted trees [XGB] and random forest [RF]) and two DL models (multi-layer perceptron and encoder–decoder model). The disease models were trained on data from seven different countries at the region-level between 2009–2017. Forecasting performance of all models was assessed using mean absolute error, root mean square error, and Poisson deviance across Australia, Israel, and the United States for the months of January through August of 2018. The overall model results were compared across diseases as well as various data splits, including country, regions with highest and lowest cases, and the forecasted months out (i.e., nowcasting, short-term, and long-term forecasting). Overall, the XGB models performed the best for all diseases and, in general, tree-based ML models performed the best when looking at data splits. There were a few instances where the statistical or DL models had minutely smaller error metrics for specific subsets of typhoid, which is a disease with very low case counts. Feature importance per disease was measured by using four tree-based ML models (i.e., XGB and RF with and without region name as a feature). The most important feature groups included previous case counts, region name, population counts and density, mortality causes of neonatal to under 5 years of age, sanitation factors, and elevation. This study demonstrates the power of ML approaches to incorporate a wide range of factors to forecast various diseases, regardless of location, more accurately than traditional statistical approaches.
59 BASIC BIOLOGICAL SCIENCES↗
Optimization of stochastic feature properties in laser powder bed fusion
Process parameter selection in laser powder bed fusion (LPBF) controls the as-printed dimensional tolerances, pore formation, surface quality and microstructure of printed metallic structures. Measuring the stochastic mechanical performance for a wide range of process parameters is cumbersome both in time and cost. As such, in this study, we overcome these hurdles by using high-throughput tensile (HTT) testing of over 250 dogbone samples to examine process-driven performance of strut-like small features, ~1 mm 2 in austenitic stainless steel (316 L). The output mechanical properties, porosity, surface roughness and dimensional accuracy were mapped across the printable range of laser powers and scan speeds using a continuous wave laser LPBF machine. Tradeoffs between ductility and strength are shown across the process space and their implications are discussed. While volumetric energy density deposited onto a substrate to create a melt-pool can be a useful metric for determining bulk properties, it was not found to directly correlate with output small feature performance.
316 L stainless steel↗
Physics and technology considerations for the deuterium–tritium fuel cycle and conditions for tritium fuel self sufficiency
The tritium aspects of the DT fuel cycle embody some of the most challenging feasibility and attractiveness issues in the development of fusion systems. The review and analyses in this paper provide important information to understand and quantify these challenges and to define the phase space of plasma physics and fusion technology parameters and features that must guide a serious R&D in the world fusion program. We focus in particular on components, issues and R&D necessary to satisfy three 'principal requirements': (1) achieving tritium self-sufficiency within the fusion system, (2) providing a tritium inventory for the initial start-up of a fusion facility, and (3) managing the safety and biological hazards of tritium. A primary conclusion is that the physics and technology state-of-the-art will not enable DEMO and future power plants to satisfy these principal requirements. We quantify goals and define specific areas and ideas for physics and technology R&D to meet these requirements. A powerful fuel cycle dynamics model was developed to calculate time-dependent tritium inventories and flow rates in all parts and components of the fuel cycle for different ranges of parameters and physics and technology conditions. Dynamics modeling analyses show that the key parameters affecting tritium inventories, tritium start-up inventory, and tritium self-sufficiency are the tritium burn fraction in the plasma (f b ), fueling efficiency (η f ), processing time of plasma exhaust in the inner fuel cycle (t p ), reactor availability factor (AF), reserve time (tr) which determines the reserve tritium inventory needed in the storage system in order to keep the plant operational for time t r in case of any malfunction of any part of the tritium processing system, and the doubling time (t d ). Results show that η f f b > 2% and processing time of 1–4 h are required to achieve tritium self-sufficiency with reasonable confidence. For η f f b = 2% and processing time of 4 h, the tritium start-up inventory required for a 3 GW fusion reactor is ~11 kg, while it is <5 kg if η f f b = 5% and the processing time is 1 h. To achieve these stringent requirements, a serious R&D program in physics and technology is necessary. The EU-DEMO direct internal recycling concept that carries fuel directly from the plasma exhaust gas to the fueling systems without going through the isotope separation system reduces the overall processing time and tritium inventories and has positive effects on the required tritium breeding ratio (TBR R ). A significant finding is the strong dependence of tritium self-sufficiency on the reactor availability factor. Simulations show that tritium self-sufficiency is: impossible if AF < 10% for any η f f b , possible if AF > 30% and 1% ≤ η f f b ≤ 2%, and achievable with reasonable confidence if AF > 50% and η f f b > 2%. These results are of particular concern in light of the low availability factor predicted for the near-term plasma-based experimental facilities (e.g. FNSF, VNS, CTF), and can have repercussions on tritium economy in DEMO reactors as well, unless significant advancements in RAMI are made. There is a linear dependency between the tritium start-up inventory and the fusion power. The required tritium start-up inventory for a fusion facility of 100 MW fusion power is as small as 1 kg. Since fusion power plants will have large powers for better economics, it is important to maintain a 'reserve' tritium inventory in the tritium storage system to continue to fuel the plasma and avoid plant shutdown in case of malfunctions of some parts of the tritium processing lines. But our results show that a reserve time as short as 24 h leads to unacceptable reserve and start-up inventory requirements. Therefore, high reliability and fast maintainability of all components in the fuel cycle are necessary in order to avoid the need for storing reserve tritium inventory sufficient for continued fusion facility operation for more than a few hours. The physics aspects of plasma fueling, tritium burn fraction, and particle and power exhaust are highly interrelated and complex, and predictions for DEMO and power reactors are highly uncertain because of lack of experiments with burning plasma. Fueling by pellet injection on the high field side of tokamak has evolved to be the preferred method to fuel a burning plasma. Extrapolation from the DIII-D penetration scaling shows fueling efficiency expected in DEMO to be <25%, but such extrapolations are highly uncertain. The fueling efficiency of gas in a reactor relevant regime is expected to be extremely poor and not very useful for getting tritium into the core plasma efficiently. Gas fueling will nonetheless be useful for feedback control of the divertor operating parameters. Extensive modeling has been carried out to predict burn fraction, fueling requirements, and fueling efficiency for ITER, DEMO, and beyond. The fueling rate required to operate Q = 10 ITER plasmas in order to provide the required core fueling, helium exhaust and radiative divertor plasma conditions for acceptable divertor power loads was calculated. If this fueling is performed with a 50–50 DT mix, the tritium burn fraction in ITER would be ~0.36%, which is too low to satisfy the self-sufficiency conditions derived from the dynamics modeling for fusion reactors. Extrapolation to DEMO using this approach would also yield similarly low burn fraction. Extensive analysis presented shows that specific features of edge neutral dynamics in ITER and fusion reactors, which are different from present experiments, open possibilities for optimization of tritium fueling and thus to improve the burn fraction. Using only tritium in pellet fueling of the plasma core, and only deuterium for edge density, divertor power load and ELM control results in significant increase of the burn fraction to 1.8–3.6%. These estimates are performed with physics models whose results cannot be fully validated for ITER and DEMO plasma conditions since these cannot be achieved in present tokamak experiments. Thus, several uncertainties remain regarding particle transport and scenario requirements in ITER and DEMO. The safety standard requirements for protection of the public and release guidelines for tritium have been reviewed. General safety approaches including minimizing tritium inventories, reducing tritium permeation through materials, and decontaminating material for waste disposal have been suggested.
73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
Types of spatter and their features and formation mechanisms in laser powder bed fusion additive manufacturing process
Spatter causes defect formation, powder redistribution and contamination in laser powder bed fusion (LPBF) additive manufacturing process. It is critical to distinguish different types of spatter and understand their features and formation mechanisms. This work reveals the features and formation mechanisms of five unique types of spatter during the LPBF process by in-situ high-speed, high-energy x-ray imaging. Spatters observed during LPBF testing are quantified by their speed, size, and direction. Distinct quantifiable characteristics for each type of spatter are identified. Finally, effects of the laser power, scan speed, and ambient pressure on spatter formation and features are unraveled. A spatter formation map for AlSi10Mg alloy is constructed.
36 MATERIALS SCIENCE↗
Feature issue introduction: laser driven inertial confinement fusion and bridging the gaps to inertial fusion energy systems
Major fusion research milestones have been achieved using laser driven inertial confinement fusion (ICF) in recent years, and these successes have ignited tremendous enthusiasm for inertial fusion energy (IFE). However, the complexity and difficulty of obtaining fusion ignition with a laser driver in a research setting are often underappreciated, as are the gaps to high driver efficiency, high repetition rates, and laser and target durability requirements needs for IFE. On the academic side, several new research laser systems have been constructed over the past few years, enabling researchers to probe the limits of ICF physics and engineering. This feature issue highlights the challenges and capabilities of laser research and development targeted towards advancing IFE.
Physics - Plasma physics↗
Gas scintillation mitigation in gas Cherenkov detectors for inertial confinement fusion (invited)
Gas Cherenkov detectors provide a time resolved measurement of the fusion burn in inertial confinement fusion experiments. The fusion rate delivers critical benchmark figures, such as burn width and bang time. Recent detector improvements pushed temporal resolution to 10 ps to make burn width measurements on igniting targets possible. First high temporal resolution measurements using CO 2 gas fills had a background signal with a long decay length (tail), which was caused by gas scintillation. This gas scintillation limits the ability of the detector to resolve short burn width and high frequency features in the fusion rate measurements. A thorough investigation of the cause of the tail and mitigation options for gas scintillation is presented here. As a near-term resolution, neon gas is being used to extract fusion burn histories. Paths forward for the next generation of gas Cherenkov detectors are identified including the usage of oxygen as a Cherenkov medium.
47 OTHER INSTRUMENTATION↗
Unraveling electronic correlations in warm dense quantum plasmas
The study of matter at extreme densities and temperatures has emerged as a highly active frontier at the interface of plasma physics, material science and quantum chemistry with relevance for planetary modeling and inertial confinement fusion. A particular feature of such warm dense matter is the complex interplay of Coulomb interactions, quantum effects, and thermal excitations, making its rigorous theoretical description challenging. Here, we demonstrate how ab initio path integral Monte Carlo simulations allow us to unravel this intricate interplay for the example of strongly compressed beryllium, focusing on two X-ray Thomson scattering data sets obtained at the National Ignition Facility. We find excellent agreement between simulation and experiment with a very high level of consistency between independent observations without the need for any empirical input parameters. Our results call into question previously used chemical models, with important implications for the interpretation of scattering experiments and radiation hydrodynamics simulations.
70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
Activation Analysis for the Inboard Region of FNSF Using SERPENT
This study describes an application of the SERPENT 2 code with the TENDL-2017 nuclear data library and the latest available model features of the Fusion Energy System Studies–Fusion Nuclear Science Facility (FNSF), to evaluate the activation of components after shutdown at 1, 10, and 100 years, assuming a plant lifetime of 8.5 full-power years. The primary parameters evaluated include the specific activity, decay heat, and waste disposal rating (WDR). The specific activity and decay heat are calculated with SERPENT 2 using a 360-deg model of the FNSF, while the WDR is calculated and classified based on the waste disposal limits established by the U.S. Nuclear Regulatory Commission under 10 CFR 61.55 as well as by using the Fetter approach. A python-based script developed for a previous high-level waste classification and analysis study was implemented and adapted to this research to calculate the WDR by comparing nuclide concentrations to the values established in 10 CFR 61.55 to generate a waste classification for each component surveyed. As only three short-lived isotopes have limitations for classifications beyond Class A, of which only 63Ni is present in appreciable quantities, there is a limit to the amount that short-lived isotopes contribute to the most significant waste analyzed here. In most cases, a handful of long-lived isotopes can be problematic, such as 59Ni and 94Nb, for example, which are solely responsible for multiple Class C classifications. The results herein reported heavily depend on the specific materials and mass/volume fractions in the specific model used in this study, which has changed and evolved since the inception of the FNSF concept and past studies. Furthermore, the more significant contributions of this study may be the development of a modeling and simulation toolkit and a strategy to perform these calculations, so to help evaluate and optimize future fusion facilities.
70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
Fully Implicit Conjugate Heat Transfer Analysis of the ARC-Class Vacuum Vessel
The coupled simulation of fusion reactor blankets including neutronics, thermal-hydraulics and thermo-mechanics is expected to speed up the design cycle of fusion reactor design concepts. In this work we demonstrate tight implicit coupling of conjugate heat transfer using the open-source Computational Fluid Dynamics software OpenFOAM for thermo-fluid mechanics and Diablo for thermo-solid mechanics. The heat transfer analysis is augmented by volumetric energy deposition from neutronic calculations using the Monte Carlo N-particle code on both solid and fluid parts of the vacuum vessel. An additional heat flux is imposed on the first wall estimated from the design power of the reactor. The tight coupling is realized through the open-source coupling library, preCICE, and tested on the vacuum vessel of the affordable, robust, compact reactor design by Commonwealth Fusion Systems. The features of the coupling and the influence of different coupling parameters such as coupling schemes, acceleration techniques and convergence criterion are discussed. The coupled simulation results are compared to a thermal-hydraulics simulation which includes only the fluid domains (the liquid immersion molten salt blanket and cooling channel) to demonstrate usefulness of a coupled simulation. Further analysis is performed to identify regions of hot spots for subsequent design improvement. This introduces the outline for integrating conjugate electromagnetics and fluid/solid mechanics (e.g., allow for deformation of the cooling channel walls) with our present approach for future analysis.
Sircar, Arpan↗
Project DarkStar: Vision for LLNL in 2030
DarkStar was a Strategic Initiative (FY2021-FY2024) to investigate applications of Artificial Intelligence (AI) and Machine Learning (ML) to scientific problems of complex hydrodynamics, shockwave physics and energetic materials. The research focused on physics and engineering design as a process that can be tremendously accelerated through merging AI with advanced physics simulation on exascale-class platforms, and to experimentally validate this revolutionary new approach through dynamic materials campaigns. A central thread of scientific inquiry was in the application of AI to enable human understanding of how to control hydrodynamic instability (which has impacts to areas such as inertial confinement fusion) via engineering features and time-dependent sources. Motivated by an unfinished line of research started by Dr. Johnny von Neumann, AI-enabled simulation approaches were developed that allowed DarkStar researchers to uncover several ground-breaking discoveries regarding hydrodynamic instability, including how to completely suppress Richtmyer-Meshkov instability (RMI). These S&T discoveries, along with other advances, have shown the way for an entirely new approach to time-dependent problems known as inverse design – the idea that complex systems can be developed directly from a final state that is to be achieved and resolve the initial design via satisfying several constraints simultaneously via AI/ML. Through experimental campaigns conducted across a wide range of facilities in the NNSA complex (the High Explosive Application Facility at LLNL, the Dynamic Compression Sector/Advanced Photon Source at Argonne National Lab, and Special Technologies Laboratory at MSTS) the radical new AI/ML approach to engineering complex material dynamics was verified, establishing a new field of study within the realm of shock physics. As advanced manufacturing capabilities continue to develop, the great importance of inverse design as a means to apply that technology effectively for NNSA missions will feature prominently over this decade. DarkStar has positioned NNSA as a world-leader in this newly emerging cross-disciplinary area of AI methods for advanced physics simulation and pioneered multiple novel approaches that have enabled the broader scientific community. By allowing us to see past the horizon, to 2030 and beyond, DarkStar has illuminated the vast potential of AI/ML to impact a wide range of new national security missions and, consequently, multiple areas of further research have already emerged across the NNSA and DOD complex.
42 ENGINEERING↗