SEARCH · Engineering Papers

Results for “range prediction”

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.

SourceResearch connectionFrom yearThrough year

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

At least 847 records · Page 47

All sources NASA NTRS DOE OSTI

DOE OSTI2020

Exploiting Machine Learning to Efficiently Predict Multidimensional Optical Spectra in Complex Environments

The excited-state dynamics of chromophores in complex environments determine a range of vital biological and energy capture processes. Time-resolved, multidimensional optical spectroscopies provide a key tool to investigate these processes. Although theory has the potential to decode these spectra in terms of the electronic and atomistic dynamics, the need for large numbers of excited-state electronic structure calculations severely limits first-principles predictions of multidimensional optical spectra for chromophores in the condensed phase. In this work, we leverage the locality of chromophore excitations to develop machine learning models to predict the excited-state energy gap of chromophores in complex environments for efficiently constructing linear and multidimensional optical spectra. By analyzing the performance of these models, which span a hierarchy of physical approximations, across a range of chromophore–environment interaction strengths, we provide strategies for the construction of machine learning models that greatly accelerate the calculation of multidimensional optical spectra from first principles.

36 MATERIALS SCIENCE↗

DOE OSTI2024

Influence of alloy solidification path on melt pool behavior in additive manufacturing

Numerical models used to study transport phenomena in laser-powder bed fusion processes often rely on assumptions and simplifications to reduce their computational expense. One common simplification is in the description of latent heat evolution during the solid-liquid phase change (i.e., the solidification pathway), justified by the fact that the mushy zone thickness is similar to the numerical grid spacing used for continuum transport models. The lack of resolution of transport phenomena in the mushy zone motivates the use of computationally convenient solidification paths such as linear or sigmoidal relationships over pathways derived from fundamental solidification theory such as equilibrium or Scheil models. In the present work, an uncertainty quantification (UQ) framework is used to analyze the influence of solidification pathway selection on the solidification dynamics and melt pool geometries in laser based additive manufacturing (AM) of IN625. Results show the solidification pathway has a quantifiable influence on the cooling rate at the liquidus isotherm, mushy zone thickness, and solidification time. Due to similarities in the latent heat evolution at the beginning of solidification, the equilibrium and Scheil models predict similar cooling rates near the liquidus isotherm, however the wider freezing range of Scheil leads to a wider mushy zone compared to equilibrium. The non-physical latent heat release profiles of sigmoidal and linear paths lead to significant overpredictions of cooling rates at the liquidus isotherm compared to equilibrium and Scheil. Finally, these results indicate that careful consideration should be given to the choice of solidification pathway to ensure reliable model predictions.

36 MATERIALS SCIENCE↗

DOE OSTI2025

Permeate fluxes from desalination of brines and produced waters: A reactive transport modeling study

The increasing interest in the use of membrane systems to desalinate inland brackish water, agricultural drainage, and industrially produced wastewater demands improved means of predicting desalination system performance under variable feedwater compositions. The interaction among water flow, solute transport, and chemical composition in these systems impacts permeate flux evolution. Here, an established multicomponent reactive transport simulator that accounts for these coupled processes is applied to compute osmotic pressure and permeate fluxes in reverse osmosis (RO) systems. The model is first validated by predicting permeate fluxes for a set of benchtop crossflow experiments subject to a range of feed flow rates and compositions, under fouling and non-fouling conditions. Results compare favorably with measured data that show that solutions with similar total dissolved solids concentrations but different compositions result in different permeate fluxes. The model is then applied to predict permeate fluxes from the desalination of produced waters using a commercial spiral wound RO module. For NaCl-dominant brines, at total dissolved salt concentrations (TDS) below about 70 g/L, permeate fluxes are inversely proportional to water mole fraction as the latter is a reasonable approximation of water activity (i.e. ideal mixing). In the case of Ca–Cl-, Na–CO3- and Na–SO4-dominant brines below about 70 g/L TDS, this relationship does not hold as well and tends to overpredict osmotic pressure and thus underpredict permeate fluxes. However, the opposite becomes true at higher TDS values for typical produced waters. The scaling potential of these waters is also computed by allowing the precipitation of minerals above their saturation limit on the RO membrane. This work demonstrates how reactive transport models developed for the analysis of waters from geological systems can be extended to improve process design, optimization, and control in desalination systems from produced waters and beyond.

Molins, Sergi↗

DOE OSTI2025

Predicting Multicomponent Aqueous Phase Adsorption Equilibria of Organic Acids Using the Generalized Brunauer–Emmett–Teller Isotherm Model

Here, to support process development of adsorptive separation of organic acids, this work presents a thermodynamic modeling methodology to predict multicomponent aqueous phase organic acid adsorption equilibria from single-component adsorption isotherms using the generalized Brunauer–Emmett–Teller isotherm model (gBET). With the organic acid fugacities rigorously accounting for the aqueous phase organic acid solution chemistry and solution nonideality, gBET precisely represents the single-component adsorption isotherms and accurately predicts the binary and ternary mixed-acid adsorption equilibria for the ranges of initial pH (∼3–7), acid concentration (100–400 mmol/L), and temperature (298.15–328.15 K) with less than 10% average absolute relative deviation. In addition, gBET with pH-independent parameters provides insights into the underlying adsorption phenomena, including the adsorbate loadings and compositions in the monolayer and subsequent layers under varying initial pH, temperature, concentration, and composition. The gBET model predictions outperform the predictions from the overloading model and the Ideal Dilute Solution Theory.

09 BIOMASS FUELS↗

DOE OSTI2022

Holocene paleoclimate change in the western US: The importance of chronology in discerning patterns and drivers

We report sediment in lakes and meadows forms a powerful archive that can be used to reconstruct environmental change through time. Reconstructions of lake level, of chemical, biological, and hydrological conditions, and of surrounding vegetation provide detailed information about past climate conditions, both locally and regionally. Indeed, most of our current knowledge of centennial- to millennial-scale climate variability in the arid western United States, where information about past hydroclimate is particularly important, comes from such sediment-based reconstructions. The pressing need for robust, precise predictions of future conditions is a significant motivation for paleoclimate science, and current research questions frequently require Holocene reconstructions to be resolved at sub-centennial timescales. Increasingly, regional syntheses seek to identify synoptic-scale patterns similar to those defined from modern observations (seasonal, interannual, multi-decadal, etc.) or to compare with the output of climate model simulations. However, the age control on existing records, especially those more than about 20 years old, is often sufficient only for millennial-scale interpretation. Here we assess the age control for 84 published and unpublished records from lakes and meadows in the Great Basin, California, and desert southwest, and use Bayesian modeling to evaluate the 95% uncertainty ranges for the 42 best-dated records. In the Late Holocene, about half of the 42 records have <400-year mean uncertainty ranges; however, high-precision age control is especially critical for young records, used to develop an accurate understanding of a proxy’s response to known climate variations. In the Middle Holocene, records vary from 400 to >800-year mean uncertainty and records of the Early Holocene have 600- to >1400-year mean uncertainty ranges. We find that the largest control on modeled uncertainties is dating density, with at least 2 dates/kyr being optimal and suggest obtaining “range-finder” dates at the onset of a study to better predict the total number of dates needed for an adequate age model. Such a density avoids a commonly observed phenomenon of significant peaks in uncertainty arising in gaps between age control points. Analysis of the uncertainties associated with proxy shifts reveal that more than half are >400 years. Although such large uncertainties currently prevent sub-centennial interpretations in most cases, increased dating density, strategic use of limited funds (including budgeting for a 2 date/kyr minimum at the proposal stage), construction of age-depth models with Bayesian methods, and critical evaluation of chronological uncertainty will shed light on past climate variability at finer timescales, enhancing our understanding of global and regional drivers of western U.S. climate.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2021

Tool-workpiece stick-slip conditions and their effects on torque and heat generation rate in the friction stir welding

Friction stir welding (FSW) has found increased applications in automotive and aerospace industries due to its advantages of solid-state bonding, no fusion and melting, and versatility in various working conditions and material combinations. However, the relationship among processing parameters, material properties, and bonding extent and fidelity remains largely empirical, primarily because of the lack of the mechanistic understanding of the tool-workpiece frictional behavior that affects our subsequent understanding of microstructural evolution and interface bonding formation. While the tool-workpiece stick-slip condition is believed to dictate the resulting torque and heat generation rate during the welding process, it remains rare and elusive to conduct a quantitative experimental measurement of such interfacial field. On the other hand, numerical simulations based on Computational Fluid Dynamics (CFD) rely on ad hoc assumptions of interfacial pressure and shear-stress conditions, but predictions can only be validated via the medium- and far-range temperature field which is known to be insensitive to the interfacial frictional behavior. This work first presents a comparison among two CFD-based simulation methodologies and the Coupled Eulerian Lagrangian (CEL) model in finite element method, the last of which uses the Coulomb friction so that the stick-slip is naturally developed. Based on the Hill-Bower similarity relationship in the contact analysis, an analytical model is developed here to prove why a constant stick-slip fraction will be developed in the steady state, to correlate the stick-slip fraction to processing parameters such as the tool spin rate, and further to derive dimensionless functions for torque and heat-generation-rate predictions. Pros and cons of various numerical approaches in predicting stick-slip are discussed, and our analytical model has been found to agree well with our numerical simulation and literature experimental results. These analyses provide the critical strain-rate and temperature fields that are needed for the bonding analysis in our future work.

36 MATERIALS SCIENCE↗

DOE OSTI2025

Physics-informed latent neural operator for real-time predictions of time-dependent parametric PDEs

Deep operator network (DeepONet) has shown significant promise as surrogate models for systems governed by partial differential equations (PDEs), enabling accurate mappings between infinite-dimensional function spaces. However, when applied to systems with high-dimensional input-output mappings arising from large numbers of spatial and temporal collocation points, these models often require heavily overparameterized networks, leading to long training times. Latent DeepONet addresses some of these challenges by introducing a two-step approach: first learning a reduced latent space using a separate model, followed by operator learning within this latent space. While efficient, this method is inherently data-driven and lacks mechanisms for incorporating physical laws, limiting its robustness and generalizability in data-scarce settings. Here, in this work, we propose PI-Latent-NO, a physics-informed latent neural operator framework that integrates governing physics directly into the learning process. Our architecture features two coupled DeepONets trained end-to-end: a Latent-DeepONet that learns a low-dimensional representation of the solution, and a Reconstruction-DeepONet that maps this latent representation back to the physical space. By embedding PDE constraints into the training via automatic differentiation, our method eliminates the need for labeled training data and ensures physics-consistent predictions. The proposed framework is both memory and compute-efficient, exhibiting near-constant scaling with problem size and demonstrating significant speedups over traditional physics-informed operator models. We validate our approach on a range of parametric PDEs, showcasing its accuracy, scalability, and suitability for real-time prediction in complex physical systems.

Latent representations↗

DOE OSTI2023

Dimensionally reduced machine learning model for predicting single component octanol–water partition coefficients

Abstract MF-LOGP, a new method for determining a single component octanol–water partition coefficients ( $$LogP$$ LogP ) is presented which uses molecular formula as the only input. Octanol–water partition coefficients are useful in many applications, ranging from environmental fate and drug delivery. Currently, partition coefficients are either experimentally measured or predicted as a function of structural fragments, topological descriptors, or thermodynamic properties known or calculated from precise molecular structures. The MF-LOGP method presented here differs from classical methods as it does not require any structural information and uses molecular formula as the sole model input. MF-LOGP is therefore useful for situations in which the structure is unknown or where the use of a low dimensional, easily automatable, and computationally inexpensive calculations is required. MF-LOGP is a random forest algorithm that is trained and tested on 15,377 data points, using 10 features derived from the molecular formula to make $$LogP$$ LogP predictions. Using an independent validation set of 2713 data points, MF-LOGP was found to have an average $$RMSE$$ RMSE = 0.77 ± 0.007, $$MAE$$ MAE = 0.52 ± 0.003, and $${R}^{2}$$ R 2 = 0.83 ± 0.003. This performance fell within the spectrum of performances reported in the published literature for conventional higher dimensional models ( $$RMSE$$ RMSE = 0.42–1.54, $$MAE$$ MAE = 0.09–1.07, and $${R}^{2}$$ R 2 = 0.32–0.95). Compared with existing models, MF-LOGP requires a maximum of ten features and no structural information, thereby providing a practical and yet predictive tool. The development of MF-LOGP provides the groundwork for development of more physical prediction models leveraging big data analytical methods or complex multicomponent mixtures. Graphical Abstract

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2020

Benchmarking materials property prediction methods: the Matbench test set and Automatminer reference algorithm

Abstract We present a benchmark test suite and an automated machine learning procedure for evaluating supervised machine learning (ML) models for predicting properties of inorganic bulk materials. The test suite, Matbench, is a set of 13 ML tasks that range in size from 312 to 132k samples and contain data from 10 density functional theory-derived and experimental sources. Tasks include predicting optical, thermal, electronic, thermodynamic, tensile, and elastic properties given a material’s composition and/or crystal structure. The reference algorithm, Automatminer, is a highly-extensible, fully automated ML pipeline for predicting materials properties from materials primitives (such as composition and crystal structure) without user intervention or hyperparameter tuning. We test Automatminer on the Matbench test suite and compare its predictive power with state-of-the-art crystal graph neural networks and a traditional descriptor-based Random Forest model. We find Automatminer achieves the best performance on 8 of 13 tasks in the benchmark. We also show our test suite is capable of exposing predictive advantages of each algorithm—namely, that crystal graph methods appear to outperform traditional machine learning methods given ~10 4 or greater data points. We encourage evaluating materials ML algorithms on the Matbench benchmark and comparing them against the latest version of Automatminer.

36 MATERIALS SCIENCE↗

DOE OSTI2021

Choice of 16S ribosomal RNA primers affects the microbiome analysis in chicken ceca

We evaluated the effect of applying different sets of 16S rRNA primers on bacterial composition, diversity, and predicted function in chicken ceca. Cecal contents from Ross 708 birds at 1, 3, and 5 weeks of age were collected for DNA isolation. Eight different primer pairs targeting different variable regions of the 16S rRNA gene were employed. DNA sequences were analyzed using open-source platform QIIME2 and the Greengenes database. PICRUSt2 was used to determine the predicted function of bacterial communities. Changes in bacterial relative abundance due to 16S primers were determined by GLMs. The average PCR amplicon size ranged from 315 bp (V3) to 769 bp (V4–V6). Alpha- and beta-diversity, taxonomic composition, and predicted functions were significantly affected by the primer choice. Beta diversity analysis based on Unweighted UniFrac distance matrix showed separation of microbiota with four different clusters of bacterial communities. Based on the alpha- and beta-diversity and taxonomic composition, variable regions V1–V3(1) and (2), and V3–V4 and V3–V5 were in most consensus. Our data strongly suggest that selection of particular sets of the 16S rRNA primers can impact microbiota analysis and interpretation of results in chicken as was shown previously for humans and other animal species.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2023

Patterns and rates of soil movement and shallow failures across several small watersheds on the Seward Peninsula, Alaska

Abstract. Thawing permafrost can alter topography, ecosystems, and sediment and carbon fluxes, but predicting landscape evolution of permafrost-influenced watersheds in response to warming and/or hydrological changes remains an unsolved challenge. Sediment flux and slope instability in sloping saturated soils have been commonly predicted from topographic metrics (e.g., slope, drainage area). In addition to topographic factors, cohesion imparted by soil and vegetation and melting ground ice may also control spatial trends in slope stability, but the distribution of ground ice is poorly constrained and hard to predict. To address whether slope stability and surface displacements follow topographic-based predictions, we document recent drivers of permafrost sediment flux present on a landscape in western Alaska that include creep, solifluction, gullying, and catastrophic hillslope failures ranging in size from a few meters to tens of meters, and we find evidence of rapid and substantial landscape change on an annual timescale. We quantify the timing and rate of surface movements using a multi-pronged, multi-scalar dataset including aerial surveys, interannual GPS surveys, synthetic aperture radar interferometry (InSAR), and climate data. Despite clear visual evidence of downslope soil transport of solifluction lobes, we find that the interannual downslope surface displacement of these features does not outpace downslope displacement of soil in locations where lobes are absent (downslope movement means: 7 cm yr−1 for lobes over 2 years vs. 10 cm yr−1 in landscape positions without lobes over 1 year). Annual displacements do not appear related to slope, drainage area, or modeled total solar radiation but are likely related to soil thickness, and volumetric sediment fluxes are high compared to temperate landscapes of comparable bedrock lithology. Time series of InSAR displacements show accelerated movement in late summer, associated with intense rainfall and/or deep thaw. While mapped slope failures do cluster at slope–area thresholds, a simple slope stability model driven with hydraulic conductivities representative of throughflow in mineral and organic soil drastically overpredicts the occurrence of slope failures. This mismatch implies permafrost hillslopes have unaccounted-for cohesion and/or throughflow pathways, perhaps modulated by vegetation, which stabilize slopes against high rainfall. Our results highlight the breadth and complexity of soil transport processes in Arctic landscapes and demonstrate the utility of using a range of synergistic data collection methods to observe multiple scales of landscape change, which can aid in predicting periglacial landscape evolution.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2023

A quantitative method to determine the region not influenced by injected interstitial and surface effects during void swelling in ion-irradiated metals

We propose and demonstrate a microstructurally-based experimental method to quantitively determine the depth regions in self-ion-irradiated metals that are affected by the injected interstitial effect and various surface effects, focusing on the choice of safe analysis zones to minimize the impact of these phenomena. The goal is to define the depth ranges where extracted data can be confidently applied to ion-neutron correlations for reactor application. Since ion energies in the range of 1–5 MeV are most frequently employed by the radiation effects community, irradiations were conducted at four energies in this range, all proceeding at 475 °C. The experiment was conducted on relatively pure single crystal iron to focus only on physical phenomena, avoiding the influence of possibly confounding chemical or segregation processes. Care was also taken to minimize the influence of other physical factors such as crystalline orientation. It was shown that, at 475 °C, ion energies of ≤1 MeV were too shallow in penetration and did not yield a safe depth range, but irradiations at 2.5 MeV and above yielded useful safe zones with predicted swelling behavior becoming independent of ion energy. The surface-affected zone width was found to be roughly twice that of the void-denuded zone width and to be independent of accumulated displacement dose. The largest injected-interstitial effect arises from the injected interstitial depression of void nucleation and growth. The interstitial-affected region starts at about one half of the projected range and does not show any “spreading” of its influence in depth as the peak damage level increases from 50 to 100 dpa. This study provides some confidence that enhances the credibility of ion simulation when applied to prediction of void swelling in neutron environments.

36 MATERIALS SCIENCE↗

DOE OSTI2023

Description of quantum interference using mixed quantum/classical theory of inelastic scattering

Manifestation of the quantum interference effect in the oscillation of scattering cross section is explored using the N 2 + O system as a case study. Calculations are carried out for two electronic PESs of the system, for various initial rotational states of N 2 , in a broad range of N 2 + O collision energies and using three theoretical methods: two versions of the approximate mixed quantum/classical theory (MQCT and AT-MQCT) and the accurate full-quantum coupled-channel method (implemented in MOLSCAT). A good agreement between different methods is observed, especially at high energies. Elastic scattering cross-sections oscillate as a function of collision energy, which is the result of quantum interference. The effects of initial rotational excitation and of the PES properties are studied in detail. For the final (thermally averaged) cross sections, both MOLSCAT and MQCT calculations predict a rather regular pattern of quantum oscillations that persist through a broad range of collision energies and expand into the low-energy regime where quantum scattering resonances are common. The difference between cross sections predicted by MQCT and MOLSCAT decreases from ~8% at low energies to ~2% at high energies. Finally, experimental data available at high collision energies are well reproduced.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2024

Flat Spectra of Energetic Particles in Interplanetary Shock Precursors

The observed energy spectra of accelerated particles at interplanetary shocks often do not match the diffusive shock acceleration (DSA) theory predictions. In some cases, the particle flux forms a plateau over a wide range of energies, extending upstream of the shock for up to seven flux e-folds before submerging into the background spectrum. Remarkably, at and downstream of the shock we have studied in detail, the flux falls off in energy as ϵ -1 , consistent with the DSA prediction for a strong shock. The upstream plateau suggests a particle transport mechanism different from those traditionally employed in DSA models. We show that a standard (linear) DSA solution based on a widely accepted diffusive particle transport with an underlying resonant wave–particle interaction is inconsistent with the plateau in the particle flux. To resolve this contradiction, we modify the DSA theory in two ways. First, we include a dependence of the particle diffusivity κ on the particle flux F (nonlinear particle transport). Second, we invoke short-scale magnetic perturbations that are self-consistently generated by, but not resonant with, accelerated particles. They lead to the particle diffusivity increasing with the particle energy as ∝ϵ 3/2 that simultaneously decreases with the particle flux as 1/F. The combination of these two trends results in the flat spectrum upstream. We speculate that nonmonotonic spatial variations of the upstream spectrum, apart from being time-dependent, may also result from non-DSA acceleration mechanisms at work upstream, such as stochastic Fermi or magnetic pumping acceleration.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2024

Numerically Testing Conceptual Models of the Utah FORGE Reservoir Using July 2023 Circulation Test Data

Over the past several years, many new data sets have become available regarding the characterization of the Utah FORGE reservoir. These include, but are not limited to, the stimulation of Well 16A, the drilling and completion of Well 16B, and interwell circulation confirmatory testing. As part of the characterization efforts, conceptual models of the reservoir are re-examined as new data become available. As part of the planning for FORGE activities, numerical models are often used to predict the reservoir response to the planned testing. Stochastic methods are often employed to bound uncertainty and allow for evaluation of comprehensive ranges of key reservoir parameters. For the most recent interwell circulation confirmatory testing (July 2023), a priori numerical model predictions did bound the observed behavior (Xinj et al., 2023), but key deviations from expected behavior prompted the FORGE team to reevaluate our conceptual model of the reservoir. In early October 2023, key members of the development, testing, and monitoring teams met for 2 days to review newly collected data and discuss ‘interesting’ or ‘key’ observations. From these discussions, 15 Key Observations were documented, with several significant ones being that the discrete fracture network developed from the 16A stimulation data may not be appropriate and that the early time pressure data obtained during the summer 2023 reservoir testing were best described using radial solutions. In July 2023, two campaigns of interwell confirmatory testing were conducted, the first set of tests occurred on July 4-5 and the second set on July 18-19. The second set of circulation tests conducted at the Utah FORGE site between the injection well 16A(78)-32 and production well 16B(78)-32 on July 18 and 19, 2023 are used to calibrate material properties in a thermal-hydraulic-mechanical (THM) simulation of the discrete fracture network connecting the wells. The spatially and temporally varying reservoir properties are calibrated to match the time dependent pressure and production profiles from the circulation tests. In future work, this calibrated model will be coupled to the native state THM model of the FORGE reservoir to predict surface deformation and strains resulting from pumping schedules.

58 GEOSCIENCES↗

DOE OSTI2024

Numerically Testing Conceptual Models of the Utah FORGE Reservoir Using July 2024 Circulation Test Data

15 GEOTHERMAL ENERGY↗

DOE OSTI2022

EDGE2D-EIRENE and ERO2.0 predictions of nitrogen molecular break-up and transport in the divertor of JET low-confinement mode plasmas

EDGE2D-EIRENE simulations of nitrogen-seeded partially detached JET L-mode plasmas show that the divertor N I to N V radiation distributions are highly sensitive to the upstream electron density, and less sensitive to changes in the assumed cross-field particle diffusivity for nitrogen ions. The EDGE2D-EIRENE simulations reproduce the peak intensities of N I to N V as measured by vertically viewing divertor filterscopes to within 50 % for a narrow range of the upstream electron density within the experimental uncertainties, while the predicted profiles are narrower than the measured ones. Including nitrogen atoms only in the ERO2.0 simulations implies lower N III and N IV peak intensities by one and two thirds, respectively, compared to EDGE2D-EIRENE. If nitrogen is assumed to recycle exclusively as molecules instead of atoms, ERO2.0 predicts that the N III and N IV intensities in the divertor increase by up to a factor of two and that the time-averaged, volume-integrated number of N 5+ to N 7+ ions in the plasma also increases by approximately a factor of 2.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2026

Transformer-based operator learning framework for self-energy in strongly correlated systems

We introduce Σ-Attention, a transformer-based operator-learning framework for approximating the self-energy operator of strongly correlated electronic systems. By creating a batched dataset that combines results from three complementary approaches, i.e., many-body perturbation theory, strong-coupling expansion, and exact diagonalization, each effective in specific parameter regimes, Σ-Attention is applied to learn an accurate approximation for the self-energy operator that is valid across a wide range of parameter regimes. This hybrid strategy leverages the strengths of existing methods while relying on the transformer's ability to generalize beyond individual limitations. More importantly, the scalability of the transformer architecture allows the learned self-energy to be extended to systems with larger sizes, leading to much improved computational scaling. Using the one-dimensional Hubbard model, we demonstrate that Σ-Attention can accurately predict the Matsubara Green's function of large systems with a wide range of coupling strength. Our framework offers a promising and scalable pathway for studying strongly correlated systems with many possible generalizations.

Zhu, Yuanran↗