Search NASA⌕ Search

SEARCH · Search NASA

Results for “Generalized additive models”

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 127 records · Page 7
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

LLMs for Mfg.—On the State of Large Language Models and Applications to Manufacturing

Additive Manufacturing (AM), referred to as 3D printing, has emerged as a key pillar of Industry 4.0 enabling layer-by-layer fabrication of intricate geometries from CAD models. In parallel, Large Language Models (LLMs), deep learning models for natural language generation trained on vast text corpora, have demonstrated unprecedented capabilities in understanding and generating human-like text. The convergence of these trends opens new opportunities at the intersection of AM and AI/ML, where LLMs can assist engineers and researchers in design, manufacture planning, and knowledge discovery. Recent academic work has begun to explore LLM applications in AM and adjacent fields, such as material science, mechanical engineering, and design for additive manufacturing. This exploration ranges from intelligent process planning to domain-specific knowledge retrieval. This survey provides a comprehensive review of current developments, focusing on peer-reviewed literature contributions that apply, adapt, and advance LLMs in general and domain-specific domains. We analyze state-of-the-art (SOTA) techniques, such as fine-tuning foundational models for specific domains, retrieval-augmented generation (RAG) pipelines, knowledge graph integration, and delve into the architectures and evaluation methods employed. The goal of this survey is to inform researchers and practitioners of the current capabilities and limitations of LLMs in general and in domain-specific applications, and to outline how these models are being tailored to meet the requirements of these applications.

36 MATERIALS SCIENCE↗
DOE OSTI2026

MTL_TX: A Multi-Task Transformer Model for Improved Radiation Time-Series Estimation

Controlling radiation doses at potential radioactive facilities is critical to ensuring the safety of both personnel and the public. At the Thomas Jefferson National Accelerator Facility (JLab), multiple sensors are deployed around the three experimental halls to monitor key parameters, including single-beam current, energy levels, current leakage, and radiation values during accelerator operations. In this study, we developed a Multi-task Transformer model, MTL_TX, to accurately estimate radiation doses at sensor locations based on historical data, with the aim of enhancing safety in accelerator facilities and surrounding public areas. To improve estimation accuracy, we integrated two innovative components into the proposed model: hierarchical feature embedding (HFE) and multi-level decomposition attention (MDA). Additionally, the multi-task learning (MTL) framework effectively leverages correlations among multiple sensors, enabling individual estimations for each sensor. MTL_TX achieved outstanding results on data collected in 2018, with an MSE of 0.1464, an RMSE of 0.2353, and an R 2 score of 0.8584. Furthermore, when trained on 2018 data, MTL_TX exhibited excellent generalization capability to unseen datasets from 2016 to 2019, achieving an MSE of 0.1407, an RMSE of 0.2263, and an R 2 score of 0.8831. These results demonstrate a significant improvement over existing state-of-the-art models.

Transformer↗
DOE OSTI2025

Benchmark microgravity experiments and computations for 3D dendritic-array stability in directional solidification

In this study, we present a comprehensive quantitative analysis of stability bands for dendritic arrays during directional solidification of a transparent succinonitrile-0.46 wt % camphor alloy, spanning a broad range of pulling velocities. Taking advantage of the microgravity environment aboard the International Space Station where most convection effects are suppressed, we obtain unique measurements that quantify the stable primary spacing range of spatially extended three-dimensional dendritic array structures under purely diffusive growth conditions. Through carefully designed velocity jump experiments and detailed examination of sub-grain boundary dynamics, we characterize key instabilities, including elimination and tertiary branching, shedding new light on the mechanisms governing dynamic dendritic spacing selection in extended 3D arrays. Phase field simulations are performed to characterize the stability limits of dendritic array structures for quantitative comparison with the flight experiments. Although the simulations capture general trends, significant deviations are noted at the upper stability boundary, indicating the influence of additional, unexplored factors. These findings contribute to a deeper understanding of dendritic growth dynamics and offer valuable benchmark data that could aid in refining predictive models and improving control of dendritic microstructures in metallurgical applications.

36 MATERIALS SCIENCE↗
DOE OSTI2026

Simulation of Channel Flow with Square Ribs for Blanket First-Wall Cooling: Geometry-Specific Tuning of k-ω Model Using Adjoint Method

Cooling of the plasma-facing first wall is challenging in the design of blanket components because of the high heat flux (on the order of 𝑀𝑊/𝑚2) from the plasma, especially when a low thermal mass medium like helium is chosen as the coolant. Therefore, heat transfer enhancement in which the convective heat transfer rate is augmented by the addition of turbulence-promoting structures becomes a key initiative for providing sufficient cooling capability with helium. Previously, computational fluid dynamics simulations had been performed on pipe flows with different transverse and longitudinal ribbed geometries at Oak Ridge National Laboratory to compare the enhancement performance among different ribbed geometries. Rib shape morphing had been conducted to obtain an optimized rib profile. In the work presented here, the adjoint method is adopted in the ANSYS Fluent solver for turbulence model augmentation, and the Generalized k-ω (GEKO) turbulence model is employed because of its ability of tuning the turbulence model. The Nusselt number and pressure drop obtained from the channel flow with bottom ribbed wall experiments are used as the targets. Sensitivity analysis provides information as guidance to improve the turbulence model accuracy. The augmented GEKO model is tuned for the studied ribbed channel geometry and flow conditions, providing improved predictive accuracy within this context. Extension to other configurations offers potential but may require additional tuning and validation.

Xu, Tracy [ORNL] (ORCID:0009000193700887)↗
DOE OSTI2025

Challenges in predicting protein-protein interactions of understudied viruses: Arenavirus-human interactions

Understanding protein-protein interactions (PPIs) between viruses and host organisms is crucial for uncovering infection mechanisms and identifying potential therapeutic targets. The ability to generalize PPI predictive models across understudied viruses presents a significant challenge. In this work, we use arenavirus-human PPIs to illustrate the difficulties associated with model generalization, which are compounded by a lack of both positive and negative data. We employ a Transfer Learning approach to investigate arenavirus-human PPIs by utilizing models trained on better-studied virus-human and human-human PPIs. Additionally, we curate and assess four types of negative sampling datasets to evaluate their impact on model performance. Despite the overall high accuracies (93–99 %) and AUPRC scores (0.8–0.9) appearing promising, further analysis indicates that these performance metrics can be misleading due to data leakage, data bias, and overfitting, especially concerning under-represented viral proteins. We reveal these gaps and assess the impact of data imbalance using standard k-fold cross-validation and Independent Blind Testing with a Balanced Dataset, resulting in a drop in accuracy below 50 %. We propose a viral protein-specific evaluation framework that categorizes viral proteins into majority and minority classes based on their representation in the dataset, enabling comparison of model performance across these groups using balanced accuracies. This framework offers a more robust evaluation of model generalizability, addressing biases inherent in standard evaluation techniques and paving the way for more reliable PPI prediction models for understudied viruses.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

Thermal Adaptation of Enzyme‐Mediated Processes Reduces Simulated Soil CO2 Fluxes Upon Soil Warming

Abstract Understanding factors influencing carbon effluxes from soils to the atmosphere is important in a world experiencing climatic change. Two important uncertainties related to soil organic carbon (SOC) stock responses to a changing climate are (a) whether soil microbial communities acclimate or adapt to changes in soil temperature and (b) how to represent this process in SOC models. To further explore these issues, we included thermal adaptation of enzyme‐mediated processes in a mechanistic SOC model (ReSOM) using the macromolecular rate theory. Thermal adaptation is defined here to encompass all potential responses of soil microbes and microbial communities following a change in temperature. To assess the effects of thermal adaptation of enzyme‐mediated processes on simulated SOC losses, ReSOM was applied to data collected from a 13‐year soil warming experiment. Results show that a model omitting thermal adaptation of enzyme‐mediated processes substantially overestimates observed CO 2 effluxes during the initial years of soil warming. The bias against observed CO 2 effluxes was lower for models including thermal adaptation of enzyme‐mediated processes. In addition, for a simulated linear 3°C soil warming over 100 years, models including thermal adaptation of enzyme‐mediated processes simulated SOC losses of a factor of three smaller than models omitting this process. As thermal adaptation of microbial community characteristics is generally not included in models simulating feedback between the soil, biosphere and atmosphere, we encourage future studies to assess the potential impact that microbial adaptation has on soil carbon – climate feedback representations in models. Plain Language Summary A major uncertainty in projecting how much soil organic carbon (SOC) will be converted to CO 2 as a consequence of climate change is related to how soil microbes may adapt to increasing soil temperatures. While this “microbial thermal adaptation” has been shown to occur in short‐term lab incubation experiments, its effect on SOC cycling on a decadal timescale is not clear. To address this knowledge gap, a mechanistic SOC model was used to simulate data collected from a 13‐year soil warming experiment, to assess how microbial thermal adaptation affects predicted SOC losses upon soil warming. The model results show that incorporating microbial thermal adaptation into the model led to reduced CO 2 effluxes from the soil to the atmosphere compared to the common approach of omitting this mechanism. Our results imply that projected SOC losses for the decades to come may be reduced when this mechanism is incorporated in land models. We therefore advocate for more research on the mechanisms controlling microbial thermal adaptation, and how to implement this mechanism in SOC models. Key Points A crucial aspect of soil organic carbon (SOC) models is the representation of soil microbes Predicted soil CO 2 fluxes upon soil warming are reduced when accounting for microbial thermal adaptation On a centennial time scale, this thermal adaptation results in up to a factor of three lower predicted SOC loss

Van de Broek, Marijn↗
DOE OSTI2026

Taylor approximation variance reduction for approximation errors in PDE-constrained Bayesian inverse problems

In numerous applications, surrogate models are used as a replacement for accurate parameter-to-observable mappings when solving large-scale inverse problems governed by partial differential equations (PDEs). The surrogate model may be a computationally cheaper alternative to the accurate parameter-to-observable mappings and/or may ignore additional unknowns or sources of uncertainty. The Bayesian approximation error (BAE) approach provides a means to account for the induced uncertainties and approximation errors, i.e. the errors between the accurate parameter-to-observable mapping and the surrogate. The statistics of these errors are, however, in general unknown a priori, and are thus calculated using Monte Carlo sampling. Although the sampling is typically carried out offline, i.e. before considering the data, the process can still represent a computational bottleneck. In this work, we develop a scalable computational approach for reducing the costs associated with the sampling stage of the BAE approach. Specifically, we consider the Taylor expansion of the accurate and surrogate forward models with respect to the uncertain parameter fields either as a control variate for variance reduction or as a means to directly and efficiently approximate the mean and covariance of the approximation errors. We propose efficient methods for evaluating the expressions for the mean and covariance of the Taylor approximations based on linear(-ized) PDE solves. Furthermore, the proposed approach is independent of the dimension of the uncertain parameter, depending instead on the intrinsic dimension of the data, ensuring scalability to high-dimensional problems. The potential benefits of the proposed approach are demonstrated for two high-dimensional inverse problems governed by PDE examples, namely for the estimation of a distributed Robin boundary coefficient in a linear diffusion problem, and for a coefficient estimation problem governed by a nonlinear diffusion problem.

Bayesian approximation error↗
DOE OSTI2026

Tula: Optimizing Time, Cost, and Generalization in Distributed Large-Batch Training

Distributed training increases the number of batches processed per iteration either by scaling-out (adding more nodes) or scaling-up (increasing the batch-size). However, the largest configuration does not necessarily yield the best performance. Horizontal scaling introduces additional communication overhead, while vertical scaling is constrained by computation cost and device memory limits. Thus, simply increasing the batch-size leads to diminishing returns: training time and cost decrease initially but eventually plateaus, creating a knee-point in the time/cost vs. batch-size pareto curve. The optimal batch-size therefore depends on the underlying model, data and available compute resources. Large batches also suffer from worse model quality due to the well-known “generalization gap”. In this paper, we present Tula, an online service that automatically optimizes time, cost, and convergence quality for large-batch training of convolutional models. It combines parallel-systems modeling with statistical performance prediction to identify the optimal batchsize. Tula predicts training time and cost within 7.5−14% error across multiple models, and achieves up to 20× overall speedup and improves test accuracy by ≈9% on average over standard large-batch training on various vision tasks, thus successfully mitigating the generalization gap and accelerating training at the same time.

Tyagi, Sahil [ORNL] (ORCID:0009000783144745)↗
DOE OSTI2024

Bridging Hydrological Ensemble Simulation and Learning Using Deep Neural Operators

Ensemble-based simulation and learning (ESnL) has long been used in hydrology for parameter inference, but computational demands of process-based ESnL can be quite high. To address this issue, we propose a deep neural operator learning approach. Neural operators are generic machine learning algorithms that can learn functional mappings between infinite-dimensional spaces, providing a highly flexible tool for scientific machine learning. Our approach is built upon DeepONet, a specific deep neural operator, and is designed to address several common problems in hydrology, namely, model parameter estimation, prediction at ungaged locations, and uncertainty quantification. Here we demonstrate the effectiveness of our DeepONet-based workflow using an existing large model ensemble created for an eastern U.S. watershed that is instrumented with 10 streamflow gages. Results suggest DeepONet achieves high efficiency in learning an ML surrogate model from the model ensemble, with the modified Kling-Gupta Efficiency exceeding 0.9 on holdout test sets. Parameter inference, carried out using the trained DeepONet surrogate model and genetic algorithm, also yields robust results. Additionally, we formulate and train a separate DeepONet model for physics-informed, seq-to-seq streamflow forecasting, which further reduces biases in the pre-trained DeepONet surrogate model. While this study focuses primarily on a single watershed, our approach is general and may be extended to enable learning from model ensembles across multiple basins or models. Thus, this research represents a significant contribution to the application of hybrid machine learning in hydrology.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Minimal implicit-solvent coarse-grained simulation of Pluronic block copolymers with ionic liquids

Pluronic block copolymers, composed of poly(ethylene oxide) (PEO) and poly(propylene oxide) (PPO) in a triblock structure (PEO–PPO–PEO), are well known for their amphiphilic character and ability to self‐assemble into micelles in aqueous solution. The addition of ionic liquids (ILs) can further modulate the core–shell structures of these copolymers, influencing their stability, critical micellization temperature, and size. However, fully atomistic simulations often become prohibitively expensive due to the size and complexity of these systems. In this work, coarse‐grained simulations using a minimal implicit‐solvent model were performed to examine how two classes of ILs, namely, 1‐alkyl‐3‐methylimidazolium ([C n C 1 im]) and 1‐alkyl‐3‐methylpyrrolidinium ([C n C 1 pyrr]), change the micellization of Pluronic block copolymers in aqueous solution. The effects of IL concentration and alkyl group length were investigated, and the model greatly improved the efficiency of simulating large‐scale micelle systems. Furthermore, the numerical simulations are qualitatively compared with experimental investigations. Our results show that adding ILs expands the micelle core by embedding IL tails among the PPO blocks, thereby increasing overall micelle size. Less polar ILs generally induce more pronounced micellar growth. However, the effect of IL tail length on conformation and micellar packing is non‐monotonic. Up to moderate chain lengths (around C8–C10), the IL tails can extend sufficiently to increase local separation within the micelle; at longer tail lengths, enhanced hydrophobic clustering and steric hindrance cause the tails to bend or fold, capping further expansion. In addition, although block copolymer chains tend to pack more closely in the presence of longer‐tailed ILs, the random coil size of an individual polymer chain does not necessarily shrink. Meanwhile, these insights provide a deeper understanding of how Pluronic/IL systems interact, informing applications in drug delivery, cosmetics, food, and environmental engineering. Finally, our minimal implicit‐solvent model can be applied to larger systems and longer timescales, substantially reducing computational cost while reproducing key structural trends observed experimentally.

Atomistic simulations↗
DOE OSTI2025

Local conservation of energy in fully implicit PIC algorithms

We consider the issue of strict, fully discrete local energy conservation for a whole class of fully implicit local-charge- and global-energy-conserving particle-in-cell (PIC) algorithms. Earlier studies demonstrated these algorithms feature strict global energy conservation. However, whether a local energy conservation theorem exists (in which the local energy update is governed by a flux balance equation at every mesh cell) for these schemes is unclear. In this study, we show that a local energy conservation theorem indeed exists. We begin our analysis with the 1D electrostatic PIC model without orbit-averaging, and then generalize our conclusions to account for orbit averaging, multiple dimensions, and electromagnetic models (Darwin). In all cases, a temporally, spatially, and particle-discrete local energy conservation theorem is shown to exist, proving that these formulations (as originally proposed in the literature), in addition to being locally charge conserving and globally energy conserving, are strictly locally energy conserving as well. In contrast to earlier proofs of local conservation in the literature, which only considered continuum time, our result is valid for the fully implicit time-discrete version of all models considered, including important features such as orbit averaging. We demonstrate the local-energy-conservation property numerically with a paradigmatic numerical example.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Uncertainty quantification for molecular property predictions with graph neural architecture search

Graph Neural Networks (GNNs) have emerged as a prominent class of data-driven methods for molecular property prediction. However, a key limitation of typical GNN models is their inability to quantify uncertainties in the predictions. This capability is crucial for ensuring the trustworthy use and deployment of models in downstream tasks. To that end, we introduce AutoGNNUQ, an automated uncertainty quantification (UQ) approach for molecular property prediction. AutoGNNUQ leverages architecture search to generate an ensemble of high-performing GNNs, enabling the estimation of predictive uncertainties. Our approach employs variance decomposition to separate data (aleatoric) and model (epistemic) uncertainties, providing valuable insights for reducing them. In our computational experiments, we demonstrate that AutoGNNUQ outperforms existing UQ methods in terms of both prediction accuracy and UQ performance on multiple benchmark datasets, and generalizes well to out-of-distribution datasets. Additionally, we utilize t-SNE visualization to explore correlations between molecular features and uncertainty, offering insight for dataset improvement. AutoGNNUQ has broad applicability in domains such as drug discovery and materials science, where accurate uncertainty quantification is crucial for decision-making.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Evaluating the Trustworthiness of Explainable Artificial Intelligence (XAI) Methods Applied to Regression Predictions of Arctic Sea Ice Motion

Abstract Recent advances in explainable artificial intelligence (XAI) methods show promise for understanding predictions made by machine learning (ML) models. XAI explains how the input features are relevant or important for the model predictions. We train linear regression (LR) and convolutional neural network (CNN) models to make 1-day predictions of sea ice velocity in the Arctic from inputs of present-day wind velocity and previous-day ice velocity and concentration. We apply XAI methods to the CNN and compare explanations to variance explained by LR. We confirm the feasibility of using a novel XAI method [i.e., global layerwise relevance propagation (LRP)] to understand ML model predictions of sea ice motion by comparing it to established techniques. We investigate a suite of linear, perturbation-based, and propagation-based XAI methods in both local and global forms. Outputs from different explainability methods are generally consistent in showing that wind speed is the input feature with the highest contribution to ML predictions of ice motion, and we discuss inconsistencies in the spatial variability of the explanations. Additionally, we show that the CNN relies on both linear and nonlinear relationships between the inputs and uses nonlocal information to make predictions. LRP shows that wind speed over land is highly relevant for predicting ice motion offshore. This provides a framework to show how knowledge of environmental variables (i.e., wind) on land could be useful for predicting other properties (i.e., sea ice velocity) elsewhere. Significance Statement Explainable artificial intelligence (XAI) is useful for understanding predictions made by machine learning models. Our research establishes trustability in a novel implementation of an explainable AI method known as layerwise relevance propagation for Earth science applications. To do this, we provide a comparative evaluation of a suite of explainable AI methods applied to machine learning models that make 1-day predictions of Arctic sea ice velocity. We use explainable AI outputs to understand how the input features are used by the machine learning to predict ice motion. Additionally, we show that a convolutional neural network uses nonlinear and nonlocal information in making its predictions. We take advantage of the nonlocality to investigate the extent to which knowledge of wind on land is useful for predicting sea ice velocity elsewhere.

Hoffman, Lauren [Scripps Institution of Oceanograp↗
DOE OSTI2025

McCormick envelopes in mixed-integer PDE-constrained optimization

McCormick envelopes are a standard tool for deriving convex relaxations of optimization problems that involve polynomial terms. Such McCormick relaxations provide lower bounds, for example, in branch-and-bound procedures for mixed-integer nonlinear programs but have not gained much attention in PDE-constrained optimization so far. This lack of attention may be due to the distributed nature of such problems, which on the one hand leads to infinitely many linear constraints (generally state constraints that may be difficult to handle) in addition to the state equation for a pointwise formulation of the McCormick envelopes and renders bound-tightening procedures that successively improve the resulting convex relaxations computationally intractable. We analyze McCormick envelopes for a model problem class that is governed by a semilinear PDE involving a bilinearity and integrality constraints. We approximate the nonlinearity and in turn the McCormick envelopes by averaging the involved terms over the cells of a partition of the computational domain on which the PDE is defined. This yields convex relaxations that underestimate the original problem up to an a priori error estimate that depends on the mesh size of the discretization. These approximate McCormick relaxations can be improved by means of an optimization-based bound-tightening procedure. We show that their minimizers converge to minimizers to a limit problem with a pointwise formulation of the McCormick envelopes when driving the mesh size to zero. We provide a computational example, for which we certify all of our imposed assumptions. The results point to both the potential of the methodology and the gaps in the research that need to be closed. Our methodology provides a framework first for obtaining pointwise underestimators for nonconvexities and second for approximating them with finitely many linear inequalities in an infinite-dimensional setting.

Approximations and Expansions↗
DOE OSTI2025

Tiny Bubbles: Combined HR(S)TEM and 4D-STEM Analysis of Sub-Nanometer He Bubbles in Au

Irradiation produces a distribution of defect sizes in materials, with the smallest defects often below one nanometer in size and approaching the scale of a single unit cell in metals. While high-resolution scanning transmission electron microscopy (STEM)-based imaging can directly image structures at this level, techniques such as four-dimensional STEM (4D-STEM) enable characterization of materials across large fields of view, capturing a more representative volume that can be valuable for quantifying defects, their distributions, and the associated strain fields. Here we present a combined HRSTEM and 4D-STEM approach to study the model system of He bubble implantation in an Au thin film. The present work is of general interest for the study of materials in extreme environments, as it demonstrates an effective way to characterize even the tiniest sub-nanometer sized He bubbles in addition to larger irradiation defects.

atomic-resolution STEM↗
DOE OSTI2025

Integrable higher-spin deformations of sigma models from auxiliary fields

We construct a new infinite family of integrable deformations of the principal chiral model (PCM) parametrized by an interaction function of several variables, which extends the formalism of [C. Ferko and L. Smith, An infinite family of integrable sigma models using auxiliary fields, .] and includes deformations of the PCM by functions of both the stress tensor and higher-spin conserved currents. We show in detail that every model in this class admits a Lax representation for its equations of motion, and that the Poisson bracket of the Lax connection takes the Maillet form, establishing the existence of an infinite set of Poisson-commuting conserved charges. We argue that the non-Abelian T-dual of any model in this family is classically integrable, and that T-duality “commutes” with a general deformation in this class, in a sense which we make precise. Finally, we demonstrate that these higher-spin auxiliary field deformations can be extended to accommodate the addition of a Wess-Zumino term, and we exhibit the Lax connection in this case. Published by the American Physical Society 2025

Bielli, Daniele (ORCID:0009000640034768)↗
DOE OSTI2025

Energy Scheduling-based Operating Envelopes including a Distribution System Branch Screening Algorithm

This paper presents an energy scheduling-based formulation for computing operating envelopes including a distribution branch screening algorithm, termed DBS-ES. The contribution of the paper is two-fold: firstly, it presents an innovative methodology for calculating operating envelopes using energy scheduling (baseline), and secondly, it enhances this methodology by incorporating a custom distribution branch screening algorithm (DBS-ES). The custom algorithm leverages power system knowledge to reduce both model build time and total processing time while maintaining the same scheduling results as the baseline. The effectiveness of the proposed approach is demonstrated through experiments on the IEEE13, IEEE123, and EPRI Secondary test feeders. Results highlight a 24.5% decrease in model build time and an 8.17% decrease in total processing time when using DBS-ES compared to the baseline, specifically for the IEEE123 test feeder. Additionally, the paper briefly discusses the influence of utility-controlled storage on computing operating envelopes, noting a general incre

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2024

Toward digital design at the exascale: An overview of project ICECap

High performance computing has entered the Exascale Age. Capable of performing over 1018 floating point operations per second, exascale computers, such as El Capitan, the National Nuclear Security Administration's first, have the potential to revolutionize the detailed in-depth study of highly complex science and engineering systems. However, in addition to these kind of whole machine “hero” simulations, exascale systems could also enable new paradigms in digital design by making petascale hero runs routine. Currently, untenable problems in complex system design, optimization, model exploration, and scientific discovery could all become possible. Motivated by the challenge of uncovering the next generation of robust high-yield inertial confinement fusion (ICF) designs, project ICECap (Inertial Confinement on El Capitan) attempts to integrate multiple advances in machine learning (ML), scientific workflows, high performance computing, GPU-acceleration, and numerical optimization to prototype such a future. Built on a general framework, ICECap is exploring how these technologies could broadly accelerate scientific discovery on El Capitan. In addition to our requirements, system-level design, and challenges, we describe some of the key technologies in ICECap, including ML replacements for multiphysics packages, tools for human-machine teaming, and algorithms for multifidelity design optimization under uncertainty. As a test of our prototype pre-El Capitan system, we advance the state-of-the art for ICF hohlraum design by demonstrating the optimization of a 17-parameter National Ignition Facility experiment and show that our ML-assisted workflow makes design choices that are consistent with physics intuition, but in an automated, efficient, and mathematically rigorous fashion.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗