Search NASA⌕ Search

SEARCH · Search NASA

Results for “extrapolate”

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 523 records · Page 29
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Heterogeneous and Framework-Bound Copper Species Contribute to Catalytic Partial Methane Oxidation in Cu–Chabazite Zeolites

The relationship between continuous partial methane oxidation (PMO) rates and Cu site speciation in Cu-CHA zeolite catalysts is explored through differential rate measurements across a series of samples of varying compositions combined with density functional theory, first-principles thermodynamics, and statistical models that characterize Cu speciation. Under continuous PMO conditions (573 K, 0.07 kPa O 2 , 3 kPa H 2 O), Cu ions are shown to anchor to the CHA framework in both monomeric (Z 2 Cu and Z 2 CuH 2 O) and dimeric (O- and OH-bridged Cu) forms that are sensitive to the identity of the local framework anchoring site. Consequently, across the studied compositional range, Cu-CHA catalysts are predicted to contain a mixture of monomeric and dimeric Cu sites. Cu-normalized CH 3 OH formation rates extrapolated to zero conversion reflect contributions from multiple site types. Predicted CH 3 OH formation rates indicate that the Cu site reactivity toward CH 4 is influenced by zeolite composition and is likely limited by the reduction half-cycle.

03 NATURAL GAS↗
DOE OSTI2024

Influence of Water Sorption on Ionic Conductivity in Polyether Electrolytes at Low Hydration

Ion-containing polymers are subject to a wide range of hydration conditions across electrochemical and water treatment applications. Significant work on dry polymer electrolytes for batteries and highly swollen membranes for water purification has informed our understanding of ion transport under extreme conditions. However, knowledge of intermediate conditions (i.e., low hydration) is essential to emerging applications (e.g., electrolyzers, fuel cells, and lithium extraction). Ion transport under low levels of hydration is distinct from the extreme conditions typically investigated, and the relevant physics cannot be extrapolated from existing knowledge, stifling materials design. In this study, we conducted ion transport measurements in LiTFSI-doped polyethers that were systematically hydrated from dry conditions. A semiautomated apparatus that performs parallel measurements of water uptake and ionic conductivity in thin-film polymers under controlled humidity was developed. For the materials and swelling range considered in this study (i.e., <0.07 g water/g dry polymer electrolyte), ionic conductivity depends nonlinearly on water uptake, with the initial sorbed water weakly affecting conductivity. With additional increases in swelling, more significant increases in conductivity were observed. Remarkably, changes in conductivity induced by water sorption were correlated with the number of water molecules per lithium ion, with the normalized molar conductivity of different samples effectively collapsing onto one another until this unit of hydration exceeded the solvation number of lithium ions under aqueous conditions. Furthermore, these results provide important knowledge regarding the effects of trace water contamination on conductivity measurements in polymer electrolytes and demonstrate that the lithium-ion solvation number marks a key transition point regarding the influence of water on ion transport in ion-containing polymers.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Resolving Nonequilibrium Shape Variations among Millions of Gold Nanoparticles

Nanoparticles, exhibiting functionally relevant structural heterogeneity, are at the forefront of cutting-edge research. Now, high-throughput single-particle imaging (SPI) with X-ray free-electron lasers (XFELs) creates opportunities for recovering the shape distributions of millions of particles that exhibit functionally relevant structural heterogeneity. To realize this potential, three challenges have to be overcome: (1) simultaneous parametrization of structural variability in real and reciprocal spaces; (2) efficiently inferring the latent parameters of each SPI measurement; (3) scaling up comparisons between 10 5 structural models and 10 6 XFEL-SPI measurements. Here, we describe how we overcame these three challenges to resolve the nonequilibrium shape distributions within millions of gold nanoparticles imaged at the European XFEL. These shape distributions allowed us to quantify the degree of asymmetry in these particles, discover a relatively stable “shape envelope” among nanoparticles, discern finite-size effects related to shape-controlling surfactants, and extrapolate nanoparticles’ shapes to their idealized thermodynamic limit. Ultimately, these demonstrations show that XFEL SPI can help transform nanoparticle shape characterization from anecdotally interesting to statistically meaningful.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Density Measurements of Molten LiF–BeF 2 and LiF–BeF 2 –LaF 3 Salt Mixtures by Neutron Radiography

The densities of eutectic (LiF) 2 –BeF 2 and mixtures of this salt (FLiBe) with LaF 3 were measured by dilatometry and by neutron attenuation from 673 K to 1,073 K. Because LaF 3 has a limited solubility in FLiBe, it was necessary to determine the amount of LaF 3 in solution before the density could be determined. The FLiBe density determination was favorably benchmarked against the literature data. A simple comparison was not available for the LaF 3 –FLiBe mixtures, so extrapolation of published data was necessary based on analysis using the Molten Salt Thermal Properties Database-Thermochemistry, or MSTDB-TC, developed by the US Department of Energy. Solubilities for LaF 3 in FLiBe ranged from 1 to 4 mol % over 673 to 1,073 K. The salt system was heated and cooled over 24 h to evaluate potential changes in composition and hysteresis during the measurement. Changes in the meniscus were observed, and these were included in the correction for density determinations. Salt surface tension may have led to supersaturation of LaF 3 in the salt because the solubility curve was nonlinear with respect to the inverse temperature, as would be expected for an ideal system. Surface tension measurements are currently underway to test this hypothesis.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2026

Henry’s Solubility and Diffusion Coefficients for 29 Volatile Organic Compounds in Polydimethylsiloxane Sylgard 184 at 293 K

Two-dimensional (2D) inverse gas chromatography (IGC) enables simultaneous determination of Henry’s solubility and Fickian diffusion coefficients for volatile organic compounds (VOCs) in polymer films. This technique offers a significant advantage over traditional cylindrical column IGC by providing precise control and measurement of the film thickness (here, 0.064 ± 0.002 mm), which is the critical length scale for accurate diffusivity determination. We apply this methodology to characterize VOC transport in Sylgard 184, a widely used polydimethylsiloxane (PDMS)-based polymer containing substantial silica filler content. At room temperature (20 °C), we measured solubility and diffusion coefficients for 29 common VOCs spanning diverse chemical functionalities, including alkanes, aromatics, chlorinated solvents, ketones, esters, and alcohols. Comparison with literature data for pure PDMS reveals that VOC solubility in Sylgard 184 is generally higher; for most non-hydrogen-bonding compounds it remains within a factor of 2 of pure PDMS, whereas alcohols are enhanced by roughly 1.8 to 3.7 times, which we attribute to favorable interactions with residual silanol groups on the silanized silica filler. Diffusion coefficients range from 1.0 × 10 –6 cm 2 /s (n-undecane) to 8.9 × 10–5 cm 2 /s (acetonitrile) and align well with extrapolated literature values for PDMS. This study provides essential thermodynamic and transport data for predicting VOC permeation in Sylgard 184 while demonstrating the utility of 2D IGC as a robust technique for characterizing rubbery polymer membranes across diverse industrial applications.

organic↗
DOE OSTI2025

Interactions of Polar and Nonpolar Groups of Alcohols in Zeolite Pores

Understanding the quantitative interactions among zeolite pore walls, Bro̷nsted acid sites, and molecules with both polar and nonpolar regions is essential for scoping out the potential of zeolites as sorbents and catalysts. Purely siliceous zeolites (MFI and Beta in the present study) are hydrophobic, whereas those containing aluminum are considered hydrophilic, preferentially adsorbing organic molecules even in aqueous environments. To characterize these interactions, we use primary alcohols of increasing molecular weight, quantifying their specific interactions in the confined pore space of the alkyl (CH x ) and OH groups. Three types of interactions were identified: (i) alkyl CH x groups interacting with the zeolite pore walls (approximately 10 kJ mol −1 per carbon), (ii) alcohol OH groups interacting with the pore walls (30−35 kJ mol −1 ), and (iii) alcohol OH groups interacting with Bro̷nsted acid sites (37 kJ mol −1 ). All three interactions were well mirrored by computational simulations. The contribution of the alkyl CH x groups was inferred from the incremental increase in sorption enthalpy with increasing molecular weight; the interaction strength of the OH groups was determined by extrapolating the global adsorption enthalpy of the alcohols to a hypothetical OH group without an alkyl group. This value was identical to the adsorption enthalpy of water. The experiments demonstrated that only water has an adsorption enthalpy on zeolite pore walls lower than its condensation enthalpy (30−35 kJ mol −1 vs 45 kJ mol −1 ), limiting the concentration of water that can be adsorbed.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Estimates of Lake Nitrogen, Phosphorus, and Chlorophyll‐ a Concentrations to Characterize Harmful Algal Bloom Risk Across the United States

Abstract Excess nutrient pollution contributes to the formation of harmful algal blooms (HABs) that compromise fisheries and recreation and that can directly endanger human and animal health via cyanotoxins. Efforts to quantify the occurrence, drivers, and severity of HABs across large areas is difficult due to the resource intensive nature of field monitoring of lake nutrient and chlorophyll‐aconcentrations. To better characterize how nutrients interact with other environmental factors to produce algal blooms in freshwater systems, we used spatially explicit and temporally matched climate, landscape, in‐lake characteristic, and nutrient inventory data sets to predict nutrients and chlorophyll‐aacross the conterminous US (CONUS). Using a nested modeling approach, three random forest (RF) models were trained to explain the spatiotemporal variation in total nitrogen (TN), total phosphorus (TP), and chlorophyll‐aconcentrations across US EPA's National Lakes Assessment (n = 2,062). Concentrations of TN and TP were the most important predictors and, with other variables, the RF model accounted for 68% of variation in chlorophyll‐a. We then used these RF models to extrapolate lake TN and TP predictions to lakes without nutrient observations and predict chlorophyll‐afor ∼112,000 lakes across the CONUS. Risk for high chlorophyll‐aconcentrations is highest in the agriculturally dominated Midwest, but other areas of risk emerge in nutrient pollution hot spots across the country. These catchment and lake‐specific results can help managers identify potential nutrient pollution and chlorophyll‐ahot spots that may fuel blooms, prioritize at‐risk lakes for additional monitoring, and optimize management to protect human health and other environmental end goals.

Environmental Sciences & Ecology↗
DOE OSTI2024

Velocity Dependence of Rate‐And‐State Friction in Granular Fault Gouge and Implications for Slow‐Slip Events

Abstract The standard rate‐and‐state friction (RSF) has extensively captured frictional behaviors, but it fails to explain the velocity dependence of frictional stability transition and widespread slow‐slip events (SSEs) in experiments and nature adequately. An alternative microphysical Chen‐Niemeijer‐Spiers (CNS) model can well describe the velocity dependence of frictional behaviors of granular gouges. Using the original CNS model, standard RSF parameters can be quantified microphysically. However, some micro‐parameters are not easy to estimate quantitatively, making it difficult to extrapolate to natural and experimental conditions. Here, we simplify the microphysically‐derived RSF parameters including direct effect a , evolution effect b , and critical slip distance D c , as well as equivalent values ( a eq , b eq , and D eq ). The simplified friction parameters directly illustrate their velocity dependence, namely the essentially constant a , a eq , and D c , negatively velocity‐dependent b and b eq , as well as varying D eq for different laws. They are roughly consistent with experimental results in various fault gouges. A modified CNS model is further derived from the original CNS model, establishing a direct link between the standard RSF and CNS models. The modified CNS model exhibits virtually identical frictional behaviors to the original CNS, but differs from the standard RSF at large velocity perturbations. Moreover, the linearized stability analysis indicates that the critical stiffness for the modified CNS model is velocity‐dependent. Compared with the standard RSF, the modified CNS model not only explains the velocity dependence of frictional stability transition, but also exhibits a more gradual transition for SSEs with a broader range of stiffness ratios.

Geochemistry & Geophysics↗
DOE OSTI2024

High Pressure Melting Curve of Fe‐Si: Implication for the Thermal Properties in Mercury's Core

The motion of liquid iron (Fe) alloy materials in the outer core drives the dynamo, which generates Mercury's magnetic field. The assessment of core models requires laboratory measurements of the melting temperature of Fe alloys at high pressure. Here, we experimentally determined the melting curve of Fe9wt%Si and Fe17wt%Si up to 17 GPa using in situ and ex situ measurements of intermetallic fast diffusion that serves as the melting criterion in a large-volume press. Our determined melting slopes are comparable with previous studies up to about 17 GPa. However, when extrapolated, our melting slopes significantly deviate from previous studies at higher pressures. For Mercury's core with a model composition of Fe9wt%Si, the melting temperature-depth profile determined in our study is lower by ∼150–250 K when compared with theoretical calculations. Using the new melting curve of Fe9wt%Si and the electrical resistivity values from a previous study of Fe8.5wt%Si, we estimate that the electronic thermal conductivity of liquid Fe9wt%Si is 30 Wm −1 K −1 at the Mercury's CMB pressure of 5 GPa and 37 Wm −1 K −1 at an assumed ICB of 21 GPa, corresponding to heat flux values of 23 mWm −2 and 32 mWm −2 , respectively. These values provide new constraints on the core models.

58 GEOSCIENCES↗
DOE OSTI2024

Phase Transitions and Thermal Equation of State of Fe‐9wt.%Si Applied to the Moon and Mercury

Abstract Accurate knowledge of the phase transitions and thermoelastic properties of candidate iron alloys, such as Fe‐Si alloys, is essential for understanding the nature and dynamics of planetary cores. The phase diagrams of some Fe‐Si alloys between 1 atm and 16 GPa have been back‐extrapolated from higher pressures, but the resulting phase diagram of Fe 83.6 Si 16.4 (9 wt.% Si) is inconsistent with temperature‐induced changes in its electrical resistivity between 6 and 8 GPa. This study reports in situ synchrotron X‐ray diffraction (XRD) measurements on pre‐melted and powder Fe 83.6 Si 16.4 samples from ambient conditions to 60 GPa and 900 K using an externally heated diamond‐anvil cell. Upon compression at 300 K, the bcc phase persisted up to ∼38 GPa. The hcp phase appeared near 8 GPa in the pre‐melted sample, and near 17 GPa in the powder sample. The appearance of the hcp phase in the pre‐melted sample reconciles the reported changes in electrical resistivity of a similar sample, thus resolving the low‐pressure region of the phase diagram. The resulting high‐temperature Birch‐Murnaghan equation of state (EoS) and thermal EoS based on the Mie‐Gruneisen‐Debye model of the bcc and hcp structures are consistent with, and complement the literature data at higher pressures. The calculated densities based on the thermal EoS of Fe‐9wt.%Si indicate that both bcc and hcp phases agree with the reported core density estimates for the Moon and Mercury.

Berrada, Meryem↗
DOE OSTI2025

Control Mechanisms for Self‐Sealing in Activated Clay‐Rich Faults Through Controlled Hydraulic Injection Experiment

Abstract In a high‐pressure injection fault activation experiment conducted at the Mont Terri underground research laboratory in Switzerland, the transmissivity of the Opalinus Clay fault significantly increased due to opening and shearing. The fluid injection, spanning a few hours, generated a 10 m radius fault activation patch. Subsequent pressure pulse tests conducted bi‐weekly for a year revealed the gradual return of fault transmissivity to its initial state. The study utilized fluid pressure decay analysis, optical fiber monitoring, continuous active source seismic measurements and borehole displacement sensors for measuring fault displacements. The fault zone exhibited a dilation of approximately 1.4 mm, associated with both normal and tangential movements during activation, resulting in a sudden transmissivity increase from 1 × 10 −12 to 3.2 × 10 −7 m 2 /s. Early post‐activation, transient compaction and the subsequent slow compaction were observed, transitioning to an extension regime. The pressure pulse tests demonstrated a rapid transmissivity drop by more than two orders of magnitude within the first 10 days, followed by a gradual and less pronounced decrease. Plastic shear and compaction dominated the transmissivity evolution until 70 days after injection ended, followed by a period where additional factors, such as clay mineral swelling, influenced the behavior. Extrapolation suggested a sealing process taking at least 50 years after the initial activation. Plain Language Summary A field‐scale fault activation experiment offers valuable insights into the elasto‐plastic processes governing the sealing of shale faults. The experiment reveals a rapid increase in the fault's transmissivity by approximately five orders of magnitude during activation. Subsequent observations show a gradual transmissivity decrease by about three orders of magnitude post‐activation, with slow long‐term plastic shear and compaction of the fault competing against secondary processes, notably clay mineral swelling. All conceptual models employed to interpret these field data converge on the estimation that the fault's return to its initial low transmissivity state would require a minimum of 50 years. Key Points High‐pressure injection fault activation experiment at the Mont Terri underground research laboratory Continuous transmissivity measurements record self‐sealing inside a clay‐rich fault zone Transmissivity undergoes a phase of domination by slow plastic compaction and shearing during the initial post‐activation period, with mineral swelling exerting its influence over the long term

Guglielmi, Yves↗
DOE OSTI2025

A Decadal Hybrid GCM Simulation Using Deep‐Learning‐Based Cloud and Convection Parameterization Generalized to a Warm Climate

A critical challenge for machine‐learning (ML) parameterization in global climate models (GCMs) is to achieve stable, accurate simulations under climates not seen during training. Previous studies have demonstrated promising offline performance and year‐long online stability in aquaplanet simulations but have encountered difficulties in real geography and under climate warming. Here we report that a GCM with real geography configuration using neural‐network‐based cloud and convection parameterization, trained exclusively with present‐day climate data, successfully performs a stable, decade‐long simulation of a warm climate with +4 K sea surface temperature (SST). The neural network (NN) is based on Han et al. (2023, https://doi.org/10.1029/2022ms003508 ) with additional inputs. The simulation captures the global precipitation distribution, surface temperatures, vertical atmospheric structures, and extreme precipitation very well, closely matching simulations from both the superparameterized CAM (SPCAM) and the conventional CAM5 in the warm climate without accuracy degradation compared to those in the baseline climate. Moreover, it produces a climate response to +4 K SST in atmospheric thermodynamic states and circulations similar to those from SPCAM and CAM5. Prognostic ablation tests on NN input variables show that the NN without convective memory as input suffers from numerical instability, and the NN without considering radiative variables and land fraction as input, or with reduced training samples produce less accurate results. To our knowledge, this is the first time an ML parameterization successfully achieves online extrapolation to a warm climate without using additional warm‐climate data for training. It demonstrates the potential of ML‐driven parameterizations for credible long‐term climate projections.

Atmosphere model↗
DOE OSTI2026

Integrating Maximum Entropy Production Theory and Machine Learning to Improve Global Evapotranspiration Modeling

Accurate estimation of terrestrial evapotranspiration (ET) is vital for understanding global water and energy cycles. However, current global ET estimations are not well constrained. This study introduces an integrated framework combining the Maximum Entropy Production (MEP) theory with Random Forest (RF) model to improve global ET estimation. Specifically, in contrast to direct ET estimation by the RF model, the integrated framework (MEP‐RF) trains to predict error of MEP‐simulated ET. MEP‐RF outperforms RF in spatiotemporal extrapolation. Attribution analysis with in situ observations reveals that the inputs of MEP are the most critical variables for the ET process, including net radiation, vegetated area, soil moisture, and surface temperature. We further drive MEP‐RF with global reanalysis and satellite data sets of these four inputs, yielding a global mean terrestrial ET of 548 mm/year, with 77% attributed to transpiration. The global ET increased at a rate of 0.85 mm/year per year during 2003–2021, primarily due to vegetation greening rather than rising temperature, while decreasing soil moisture led to decreasing regional ET. The integrated framework provides a novel approach for the estimation of global ET without the need for hard‐to‐obtain and thus uncertain inputs, such as wind speed, surface roughness, aerodynamic and canopy stomatal resistance. Therefore, MEP‐RF offers an independent method on existing global ET products. It represents a promising physically based approach that can be incorporated into Earth System Models to enhance water and energy cycle simulations.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Machine learning the electric field response of condensed phase systems using perturbed neural network potentials

Abstract The interaction of condensed phase systems with external electric fields is of major importance in a myriad of processes in nature and technology, ranging from the field-directed motion of cells (galvanotaxis), to geochemistry and the formation of ice phases on planets, to field-directed chemical catalysis and energy storage and conversion systems including supercapacitors, batteries and solar cells. Molecular simulation in the presence of electric fields would give important atomistic insight into these processes but applications of the most accurate methods such as ab-initio molecular dynamics (AIMD) are limited in scope by their computational expense. Here we introduce Perturbed Neural Network Potential Molecular Dynamics (PNNP MD) to push back the accessible time and length scales of such simulations. We demonstrate that important dielectric properties of liquid water including the field-induced relaxation dynamics, the dielectric constant and the field-dependent IR spectrum can be machine learned up to surprisingly high field strengths of about 0.2 V Å −1 without loss in accuracy when compared to ab-initio molecular dynamics. This is remarkable because, in contrast to most previous approaches, the two neural networks on which PNNP MD is based are exclusively trained on molecular configurations sampled from zero-field MD simulations, demonstrating that the networks not only interpolate but also reliably extrapolate the field response. PNNP MD is based on rigorous theory yet it is simple, general, modular, and systematically improvable allowing us to obtain atomistic insight into the interaction of a wide range of condensed phase systems with external electric fields.

Science & Technology - Other Topics↗
DOE OSTI2025

New proton emitter 188 At implies an interaction unprecedented in heavy nuclei

We report the discovery of a new atomic nucleus 188 At, which is the heaviest proton-emitting isotope known to date. The new activity was observed through the 107 Ag( 84 Sr, 3n) 188 At fusion-evaporation reaction using the focal-plane spectrometer of the gas-filled recoil separator in the Accelerator Laboratory of the University of Jyväskylä, Finland. To fully interpret the experimental data, we have expanded the non-adiabatic quasiparticle model to treat nuclei in the beyond-lead region. The description reproduced the measured decay rate and pointed towards emission from an extremely prolate-deformed state with a dominant s 1/2 proton component in the wave function. The Thomas-Ehrman shift can be enhanced in low angular momentum states, but such effects have not been observed in heavy nuclei. The single-proton separation energy of 188 At deviates from that extrapolated from the systematics, which can be interpreted as the first evidence of this effect in heavy nuclei.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2025

Parametric matrix models

We present a general class of machine learning algorithms called parametric matrix models. In contrast with most existing machine learning models that imitate the biology of neurons, parametric matrix models use matrix equations that emulate physical systems. Similar to how physics problems are usually solved, parametric matrix models learn the governing equations that lead to the desired outputs. Parametric matrix models can be efficiently trained from empirical data, and the equations may use algebraic, differential, or integral relations. While originally designed for scientific computing, we prove that parametric matrix models are universal function approximators that can be applied to general machine learning problems. After introducing the underlying theory, we apply parametric matrix models to a series of different challenges that show their performance for a wide range of problems. For all the challenges tested here, parametric matrix models produce accurate results within an efficient and interpretable computational framework that allows for input feature extrapolation.

Computational science↗
DOE OSTI2025

Data-driven organic solubility prediction at the limit of aleatoric uncertainty

Abstract Small molecule solubility is a critically important property which affects the efficiency, environmental impact, and phase behavior of synthetic processes. Experimental determination of solubility is a time- and resource-intensive process and existing methods for in silico estimation of solubility are limited by their generality, speed, and accuracy. This work presents two models derived from the FASTPROP and CHEMPROP architectures and trained on BigSolDB which are capable of predicting solubility at arbitrary temperatures for a wide range of small molecules in organic solvent. Both extrapolate to unseen solutes 2–3 times more accurately than the current state-of-the-art model and we demonstrate that they are approaching the aleatoric limit (0.5–1$$\log S$$ log S ) of available test data, suggesting that further improvements in prediction accuracy require more accurate datasets. The FASTPROP-derived model (called FASTSOLV) and the CHEMPROP-based model are open source, freely accessible via a Python package and web interface, highly reproducible, and up to 2 orders of magnitude faster than current alternatives.

Science & Technology - Other Topics↗
DOE OSTI2025

Learning plasma dynamics and robust rampdown trajectories with predict-first experiments at TCV

The rampdown phase of a tokamak pulse is difficult to simulate and often exacerbates multiple plasma instabilities. To reduce the risk of disrupting operations, we leverage advances in Scientific Machine Learning (SciML) to combine physics with data-driven models, developing a neural state-space model (NSSM) that predicts plasma dynamics during Tokamak à Configuration Variable (TCV) rampdowns. The NSSM efficiently learns dynamics from a modest dataset of 311 pulses with only five pulses in a reactor-relevant high-performance regime. The NSSM is parallelized across uncertainties, and reinforcement learning (RL) is applied to design trajectories that avoid instability limits. High-performance experiments at TCV show statistically significant improvements in relevant metrics. A predict-first experiment, increasing plasma current by 20% from baseline, demonstrates the NSSM’s ability to make small extrapolations. The developed approach paves the way for designing tokamak controls with robustness to considerable uncertainty and demonstrates the relevance of SciML for fusion experiments.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗