Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “SURROGATE 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 1,441 records · Page 80
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Understanding Power and Energy Utilization in Large Scale Production Physics Simulation Codes

Power is an often-cited reason for moving to advanced architectures on the path to Exascale computing. This is due to the practical concern of delivering enough power to successfully site and operate these machines, as well as concerns over energy usage while running large simulations. Since accurate power measurements can be difficult to obtain, processor thermal design power (TDP) is a possible surrogate due to its simplicity and availability. However, TDP is not indicative of typical power usage while running simulations. Using commodity and advance technology systems at Lawrence Livermore National Laboratory (LLNL) and Sandia National Laboratory, we performed a series of experiments to measure power and energy usage in running simulation codes. These experiments indicate that large scale LLNL simulation codes are significantly more efficient than a simple processor TDP model might suggest.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Denoising diffusion algorithm for inverse design of microstructures with fine-tuned nonlinear material properties

Here we introduce a denoising diffusion algorithm to discover microstructures with nonlinear fine-tuned properties. Denoising diffusion probabilistic models are generative models that use diffusion-based dynamics to gradually denoise images and generate realistic synthetic samples. By learning the reverse of a Markov diffusion process, we design an artificial intelligence to efficiently manipulate the topology of microstructures to generate a massive number of prototypes that exhibit constitutive responses sufficiently close to designated nonlinear constitutive behaviors. To identify the subset of microcstructures with sufficiently precise fine-tuned properties, a convolutional neural network surrogate is trained to replace high-fidelity finite element simulations to filter out prototypes outside the admissible range. Results of this study indicate that the denoising diffusion process is capable of creating microstructures of fine-tuned nonlinear material properties within the latent space of the training data. More importantly, this denoising diffusion algorithm can be easily extended to incorporate additional topological and geometric modifications by introducing high-dimensional structures embedded in the latent space. Numerical experiments are conducted on the open-source mechanical MNIST data set (Lejeune, 2020). Consequently, this algorithm is not only capable of performing inverse design of nonlinear effective media, but also learns the nonlinear structure–property map to quantitatively understand the multiscale interplay among the geometry, topology, and their effective macroscopic properties.

42 ENGINEERING↗
DOE OSTI2023

Development and evaluation of a skeletal mechanism for EHN additized gasoline mixtures in large Eddy simulations of HCCI combustion

Advanced Low Temperature Combustion modes, such as the Sandia proposed Additive-Mixing Fuel Injection (AMFI), can unlock significant potential to boost fuel conversion efficiency and ultimately improve the energy conversion of internal combustion engines. This is a novel improved combustion process that is enabled by supplying small (<5%) variable amounts of autoignition improver to the fuel to enhance the engine operation and control. Common, diesel-fuel ignition-quality enhancing additive, 2-ethylexyl nitrate (EHN), is doped into gasoline to enable Sandia LTGC + AMFI combustion. This manuscript focuses on the development of a reduced sub-mechanism for EHN chemical kinetics at engine relevant conditions that is implemented into a skeletal mechanism for chemical kinetic studies of gasoline surrogate fuels. The mechanism validation utilized zero-dimensional numerical simulations and comparison to shock tube ignition-delay data of pure and EHN-doped n-heptane. Additional validation is presented with Homogeneous Charge Compression-Ignition (HCCI) engine data of pure and EHN-doped research-grade E10 gasoline. Then, the mechanism was deployed in a 3-D computational fluid dynamics (CFD) using Large Eddy Simulations (LES) to model the HCCI engine experiments of 0.4% vol EHN additized E10 gasoline at several equivalence ratios. Simulations showed a very good performance of the mechanism, and the model accurately reproduced (a) the ignition point, (b) combustion phasing, (c) combustion duration, and (d) the peak of the heat release rates of the engine experiments. The results show that EHN promotes Low-Temperature Heat Release, ultimately driving the gasoline to autoignite at thermodynamic conditions where the fuel would not otherwise ignite. Overall, this work demonstrates a viable reduced chemical-kinetic mechanism for EHN and shows that it can be combined with a skeletal gasoline mechanism for CFD-LES analysis of well-mixed LTGC that matches well with experimental results. The CFD-LES analysis also shows the spatial distribution of EHN-fuel interactions that control the autoignition throughout the combustion chamber.

Guleria, Gaurav↗
DOE OSTI2023

Methods to Account for CAP-88 PC-Omitted Nuclides in Radioactive Air Emissions From DOE Facilities

Routine research and development activities at US Department of Energy facilities can result in the release of radioactive emissions, potentially exposing the public and the environment. Such emissions are subject to certain Clean Air Act regulations, specifically those outlined in 40 CFR Part 61, Subpart H, which are enforced by the US Environmental Protection Agency. Compliance is determined in part with a dose standard. Doses to members of the public from most US Department of Energy facility airborne emissions are modeled using CAP-88 PC software, the latest version of which contains more than 1,200 radionuclides. Radioactive effluents from US Department of Energy facilities may contain radioisotopes that are not available in CAP-88 PC, and those radioisotopes must still be evaluated to determine whether doses to members of the public are below established limits. CAP-88 PC-omitted radionuclides can be accounted for using alternate methods, including the use of surrogate radionuclides. Here, this paper elucidates the process of accounting for CAP-88 PC-omitted radionuclides by providing criteria and rationale for surrogate selection, a compilation of surrogate radionuclides used in the past by US Department of Energy facilities, a discussion of alternate methods used to account for CAP-88 PC-omitted radionuclides and a comparison of methods and impacts on receptor doses. Ultimately, this paper aims to aid in the process of surrogate selection and consequently to simplify and expedite compliance with Clean Air Act regulations.

61 RADIATION PROTECTION AND DOSIMETRY↗
DOE OSTI2020

Comparison of machine learning techniques to optimize the analysis of plutonium surrogate material via a portable LIBS device

The utilization of machine learning techniques has become commonplace in the analysis of optical emission spectra. These methods are often limited to variants of principal components analysis (PCA), partial-least squares (PLS), and artificial neural networks (ANNs). A plethora of other techniques exist and are well established in the world of data science, yet are seldom investigated for their use in spectroscopic problems. In this study, machine learning techniques were used to analyze optical emission spectra of laser-induced plasma from ceria pellets doped with silicon in order to predict silicon content. Additionally, a boosted regression ensemble model was created, and its predictive accuracy was compared to that of traditional PCA, PLS, and ANN regression models. Boosted regression tree ensembles yielded fits with R-squared (R2) values as high as 0.964 and mean-squared errors of prediction (MSEPs) as low as 0.074, providing the most accurate predictive model. Neural networks performed with slightly lower R2 values and higher MSEPs compared to the ensemble methods, thus indicating susceptibility to overfitting.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

Response of a Pressurized Water Reactor Dashpot Region to Commercial Drying Cycles

The purpose of this report is to document updates to the simulation of commercial vacuum drying procedures at the Nuclear Energy Work Complex at Sandia National Laboratories. Validation of the extent of water removal in a dry spent nuclear fuel storage system based on drying procedures used at nuclear power plants is needed to close existing technical gaps. Operational conditions leading to incomplete drying may have potential impacts on the fuel, cladding, and other components in the system. A general lack of data suitable for model validation of commercial nuclear canister drying processes necessitates additional, well-designed investigations of drying process efficacy and water retention. Scaled tests that incorporate relevant physics and well-controlled boundary conditions are essential to provide insight and guidance to the simulation of prototypic systems undergoing drying processes. This report documents testing updates for the Dashpot Drying Apparatus (DDA), an apparatus constructed at a reduced scale with multiple Pressurized Water Reactor (PWR) fuel rod surrogates and a single guide tube dashpot. This apparatus is fashioned from a truncated 5×5 section of a prototypic 17×17 PWR fuel skeleton and includes the lowest segment of a single guide tube, often referred to as the dashpot region. The guide tube in this assembly is open and allows for insertion of a poison rod (neutron absorber) surrogate.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2026

Evaluation of methods and improvement of predictions for specification properties of petroleum-based and alternative aviation fuels

To support our research and process modeling for liquid fuels, including blends, from petroleum and synthetic sources such as from biomass intermediates, we evaluated composition-based prediction methods and improved predictions for five key specification properties of petroleum-based and alternative aviation fuels, namely distillation temperatures (10 % distilled, t 10 , and final boiling point, t FBP ), density, flash point, net heat of combustion, and freezing point. The types of fuels included were petroleum-based jet fuels, jet-fuel surrogate mixtures, synthetic blending components obtained from different sources, and blends of Jet A with many synthetic blending components. Expanded datasets to update associated parameters allowed significant improvements for one of the prediction methods used in earlier work, namely the Modified Weighted Average method published initially by Shi et al. By considering the importance of lighter compounds for flash points and heavier compounds for freezing points, the revised Modified Weighted Average method was further improved. For liquid density, the revised Modified Weighted Average method gave the best overall results. The revised Modified Weighted Average method, the American Society for Testing and Materials D7215 method, and the D7215 method modified by another group gave comparable results for flash point, while the revised Modified Weighted Average and D3338 methods gave the best results for net heat of combustion. Freezing point was well predicted using the revised Modified Weighted Average method and showed the most significant improvements over current predictions. Distillation temperature t 10 was not well predicted, while t FBP was predicted with a mean absolute error comparable to experimental reproducibility.

09 BIOMASS FUELS↗
DOE OSTI2026

Robust Design Under Uncertainty in Quantum Error Mitigation

Error mitigation techniques are crucial to achieving near-term quantum advantage. Classical postprocessing of quantum computation outcomes is a popular approach for error mitigation, which includes methods, such as zero noise extrapolation, virtual distillation, and learning-based error mitigation. However, these techniques have limitations due to the propagation of uncertainty resulting from the finite shot number of a quantum measurement. In this work, we introduce general and unbiased methods for quantifying the uncertainty and error of error-mitigated observables based on the strategic sampling of error mitigation outcomes. We then extend our approach to demonstrate the optimization of performance and robustness of error mitigation under uncertainty. To illustrate our methods, we apply them to zero noise extrapolation and Clifford date regression in the ground state of the XY model simulated using depolarizing and International Business Machines Corporation (IBM) Toronto noise models, respectively. In particular, we optimize the choice of noise levels and the allocation of shots for zero noise extrapolation and the distribution of the training circuits for Clifford data regression. While our methods are readily applicable to any postprocessing-based error mitigation approach, in practice they must not be prohibitively expensive—even though they perform optimizations of the error mitigation hyperparameters requiring sampling of a statistical distribution of error mitigation outcomes. By leveraging surrogate-based optimization, we show that our methods can efficiently perform optimal design for a zero noise extrapolation implementation. We then further demonstrate the transferability of learned zero noise extrapolation hyperparameters to other similar circuits.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

MOOSE ProbML: Parallelized probabilistic machine learning and uncertainty quantification for computational energy applications

Here, this paper presents the development and demonstration of massively parallel probabilistic machine learning (ML) and uncertainty quantification (UQ) capabilities within the Multiphysics Object-Oriented Simulation Environment (MOOSE), an open-source computational platform for parallel finite element and finite volume analyses. In addressing the computational expense and uncertainties inherent in complex multiphysics simulations, this paper integrates Gaussian process (GP) variants, active learning, Bayesian inverse UQ, adaptive forward UQ, Bayesian optimization, evolutionary optimization, and Markov chain Monte Carlo (MCMC) within MOOSE. It also elaborates on the interaction among key MOOSE systems — Sampler, MultiApp, Reporter, and Surrogate — in enabling these capabilities. The modularity offered by these systems enables development of a multitude of probabilistic ML and UQ algorithms in MOOSE. Example code demonstrations include parallel active learning and parallel Bayesian inference via active learning. The impact of these developments is illustrated through five applications relevant to computational energy applications: UQ of nuclear fuel fission product release, using parallel active learning Bayesian inference; very rare events analysis in nuclear microreactors using active learning; advanced manufacturing process modeling using multi-output GPs (MOGPs) and dimensionality reduction; fluid flow using deep GPs (DGPs); and tritium transport model parameter optimization for fusion energy, using batch Bayesian optimization. These capabilities are part of the MOOSE framework.

97 - MATHEMATICS AND COMPUTING↗
DOE OSTI2026

GP Cosmology Surrogate v1.0

GP Cosmology Surrogate is a Python library for building and training a generalized multi-output Gaussian process (GP) framework of @takhtaganov2021cosmic. In this approach, the surrogate is constructed sequentially, guided by a Bayesian optimization acquisition function that targets reduction of emulation error in the regions most consistent with the observational data. This adaptive design concentrates computational resources where they have the greatest impact on inference accuracy. The library supports efficient training for separable GP kernels, which allows the use of Kronecker algebra to handle high-dimensional input spaces and large numbers of correlated outputs. This makes it well suited for applications such as modeling cosmological power spectra, large-scale physical simulations, and multi-output hyperparameter tuning. By combining scalable multi-output GP modeling with data-driven adaptive sampling, GPsurrogate enables parameter inference and optimization with substantially fewer simulations than conventional space-filling designs.

Lukic, Zarija [Lawrence Berkeley National Laborato↗
DOE OSTI2022

Monitoring Noble Gases (Xe and Kr) and Aerosols (Cs and Rb) in a Molten Salt Reactor Surrogate Off-Gas Stream Using Laser-Induced Breakdown Spectroscopy (LIBS)

In this study with surrogate materials we show that laser-induced breakdown spectroscopy (LIBS) is a robust tool with promising capability toward monitoring gaseous (Xe and Kr) and aerosol (Cs and Rb) species in an off-gas stream from a molten salt reactor (MSR). MSRs will continually evolve fission products into the cover gas flowing across the reactor headspace. The cover gas entrains Xe and Kr gases, along with aerosol particles, before passing into an off-gas treatment system. Univariate models of Xe and Kr peaks showed a strong correlation to concentration indicated by their coefficients of determination of 0.983 and 0.997, respectively. Multivariate models were built for all four analytes using partial least squares regression coupled with preprocessing steps including normalization, trimming, and/or genetic algorithm derived filters. The models were evaluated by predicting the concentrations of the analytes in four validation samples, in which all calibration models were successfully validated at a confidence interval of 99.9%. Finally, pressure controllers were used to regulate the mass flow rate of Kr flowing into the measurement cell in sinusoidal and stepwise waveforms to test the real-time monitoring capabilities of the regression models. Both univariate and partial least squares Kr models were able to successfully quantify the gas concentration in the real-time evaluation. The root mean squared error of prediction (RMSEP) values for these real-time tests were calculated to be 0.051, 0.060, and 0.121 mol% demonstrating the measurement systems’ capability to perform online monitoring with acceptable accuracy.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2022

Numerical analysis of soot emissions from gasoline-ethanol and gasoline-butanol blends under gasoline compression ignition conditions

In the present work, computational fluid dynamics (CFD) simulations of a single-cylinder gasoline compression ignition (GCI) engine were performed to investigate the impact of blending two biofuels, ethanol and n-butanol, with gasoline on the trade-off between combustion phasing and soot emissions under low load conditions. Here, in order to represent market gasoline (RD5-87), a four-component toluene primary reference fuel (TPRF) + ethanol (ETPRF) surrogate (with 20% ethanol by mole; E20) was formulated using a neural network based octane predictor such that the surrogate had the same ethanol content, Research Octane Number (RON) and Octane Sensitivity (S). In addition, a novel skeletal kinetic mechanism for ETPRF and TPRF + n-butanol (BTPRF) blends, incorporating polycyclic aromatic hydrocarbon (PAH) chemistry, was developed. A three-dimensional (3D) engine CFD formulation employing the skeletal mechanism, adaptive mesh refinement (AMR), finite-rate chemistry approach, and hybrid method of moments (HMOM) was adopted to capture the in-cylinder combustion phenomena and soot emissions. The engine CFD model was validated against RD5-87 experimental data for a broad range of start-of-injection (SOI) timings (-21/-27/-36/-45 crank angle degrees (CAD) after top-dead center (aTDC)), with respect to in-cylinder pressure, heat release rate, combustion phasing, and soot emissions. The closed-cycle simulation results were analyzed to elucidate the non-monotonic trend of soot emissions versus SOI timing: SOI-36 > SOI-45 > SOI-21 > SOI-27. Thereafter, the validated CFD model was employed to simulate the combustion of a gasoline-ethanol blend with 45% (by mole) ethanol (E45) and a gasoline-butanol blend with 45% (by mole) n-butanol (B45) under the same operating conditions to study the effects of fuel composition and SOI timing on combustion phasing and soot emissions. The sooting propensity followed the trend: B45 > E20 > E45 at all SOI timings. Overall, it was observed that the autoignition propensity was primarily related to fuel chemistry. On the other hand, sooting propensity showed strong coupling with both fuel chemistry and physical properties, with greater impact of fuel physical properties at advanced SOI timings.

30 DIRECT ENERGY CONVERSION↗
DOE OSTI2023

Fuel-rich oxidation of gasoline surrogate components in an atmospheric flow reactor

Fuel-rich oxidation of three typical gasoline surrogate components, toluene, isooctane, and n-heptane, was investigated in an atmospheric-pressure flow reactor at mean gas temperatures from 1050 to 1350 K, equivalence ratio of 9.0, and residence times of 0.45 and 1.2 s. Not only polycyclic aromatic hydrocarbons (PAHs) up to 3 ring structure but also small intermediate products from C 1 to C 5 were quantified by a gas chromatograph mass spectrometry coupled with photon ionization and gas chromatograph with flame ionization detector, respectively. The kinetic model recently developed by Lawrence Livermore National Laboratory was revised to reflect the results of many recent investigations. Basically, the updated model could satisfactorily reproduce the experimental mole fractions of many species. The experimental and simulated results showed that PAH mole fractions produced were in the order of toluene, isooctane, and n-heptane. The kinetic analysis using the model was carried out to explore PAH formation pathways, especially focusing on naphthalene, acenaphthylene, and phenanthrene. Through rate of production analysis, it was found that the main formation pathways of many PAHs were affected by the fuels. Although resonantly stabilized radicals, such as benzyl and fulvenallenyl radicals, played a crucial role in the formation pathways of many PAHs in every fuel, they were produced through hydrogen elimination of toluene in toluene fuel, while they were formed from small intermediate products in isooctane and n-heptane fuels. Sensitivity analysis revealed the difference and similarity of the reactions with large positive coefficients according to the fuels studied here. In conclusion, the molecular growth reactions of aromatic species were influential in PAH production in every fuel, whereas the ring formation reactions from small species and the reactions involving toluene had large positive sensitivity coefficients in isooctane/n-heptane and toluene fuels, respectively.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2025

Design of an Out-Of-Pile Experimental Facility to Demonstrate the Feasibility of In Situ Thermal Conductivity Measurements of Nuclear Fuels Under Irradiation

There is substantial merit in quantifying nuclear fuel performance under irradiation. At Oak Ridge National Laboratory (ORNL), the MiniFuel irradiation platform has become the primary test vehicle for conducting separate-effects fuel performance irradiation experiments. The MiniFuel experiment is a passively controlled capsule design deployed in the High Flux Isotope Reactor (HFIR) through which fuel performance data is collected post-irradiation. Separate effects fuels irradiation capabilities are being expanded at ORNL by developing instrumented capsule designs that aim to capture fuel performance phenomena in-situ. One such capsule will specifically target fuel specimen thermal conductivity changes as a function of fuel burnup. Due to the complexity of making this measurement on nuclear fuel in-pile, this paper describes the necessary out-of-pile testing conducted on the thermal conductivity capsule (TCC) design. The measurement is ascertained via a thermopile system with heat transferred unidirectionally through a surrogate fuel specimen sandwiched between two conductive materials. The capsules investigated in this study are representative of the in-pile design, with the primary departure from irradiation conditions being the distribution of heat generation within the capsule. In the out-of-pile experiment, an external heater was used to drive heat through the conductive slug materials and into the specimen. This paper expounds the design of the out-of-pile experimental system and the thermal conductivity measurement technique. Predictive models used to determine the sensitivity of the measurement to variables governing thermal contact conductance between the specimen and slug materials and to predict experimental results are also described. Data from the out-of-pile experiment will be used to validate the readiness of the design for insertion into HFIR for irradiation.

Parker, Trevor [ORNL]↗
DOE OSTI2020

Scalable Deep-Learning-Accelerated Topology Optimization for Additively Manufactured Materials

Topology optimization (TO) is a popular and powerful computational approach for designing novel structures, materials, and devices. Two computational challenges have limited the applicability of TO to a variety of industrial applications. First, a TO problem often involves a large number of design variables to guarantee sufficient expressive power. Second, many TO problems require a large number of expensive physical model simulations, and those simulations cannot be parallelized. To address these issues, we propose a general scalable deep-learning (DL) based TO framework, referred to as SDL-TO, which utilizes parallel schemes in high performance computing (HPC) to accelerate the TO process for designing additively manufactured (AM) materials. Unlike the existing studies of DL for TO, our framework accelerates TO by learning the iterative history data and simultaneously training on the mapping between the given design and its gradient. The surrogate gradient is learned by utilizing parallel computing on multiple CPUs incorporated with a distributed DL training on multiple GPUs. The learned TO gradient enables a fast online update scheme instead of an expensive update based on the physical simulator or solver. Using a local sampling strategy, we achieve to reduce the intrinsic high dimensionality of the design space and improve the training accuracy and the scalability of the SDL-TO framework. The method is demonstrated by benchmark examples and AM materials design for heat conduction. The proposed SDL-TO framework shows competitive performance compared to the baseline methods but significantly reduces the computational cost by a speed up of around 8.6x over the standard TO implementation.

Bi, Sirui↗
DOE OSTI2021

A magnetic particle time-of-flight (MagPTOF) diagnostic for measurements of shock and compression-bang time at the NIF [National Ignition Facility]

A magnetic particle time-of-flight (MagPTOF) diagnostic has been designed to measure shock- and compression-bang time using DHe-fusion protons and DD-fusion neutrons, respectively, at the National Ignition Facility (NIF). This capability, in combination with shock-burn weighted areal density measurements, will significantly constrain the modeling of the implosion dynamics. This design is an upgrade to the existing particle time-of-flight (pTOF) diagnostic, which records bang times using DD or DT neutrons with an accuracy better than ±70 ps [H. G. Rinderknecht et al., Rev. Sci. Instrum. 83, 10D902 (2012)]. The inclusion of a deflecting magnet will increase DHe-proton signal-to-background by a factor of 1000, allowing for the first time simultaneous measurements of shock and compression-bang times in DHe-filled surrogate implosions at the NIF.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2025

The molten salt tritium transport experiment: A pumped fluoride salt loop for hydrogen isotope experimentation

Molten salt reactors (MSRs) and fusion reactors propose to use molten salts as coolants and breeder blanket materials, respectively. Tritium, however, poses safety concerns in both reactor types due to its ability to permeate through reactor materials and potential for environmental release. This manuscript addresses the tritium transport phenomena in molten salts and presents the design and analysis of the Molten Salt Tritium Transport Experiment (MSTTE). MSTTE is a forced-convection fluoride salt loop intended to measure hydrogen isotope permeation through structural materials in a flowing salt system. In the first phase, MSTTE will use FLiNaK salt and deuterium as surrogates for FLiBe and tritium, with future plans to utilize tritium and FLiBe. MSTTE couples a Copenhagen Atomics pumped salt loop with an external test section that introduces hydrogen isotopes into the loop and measures transport phenomena. The Hydrogen Injection System (HIS) controls hydrogen isotope introduction into the molten salt loop. Here, the permeation test section measures the permeation rate through stainless steel tubing in contact with flowing salt. Computational fluid dynamics (CFD) analysis ensures fully developed salt flow in the permeation test section. MSTTE is modeled with MELCOR-TMAP to predict the permeation rate as a function of experimental variables such as source term, salt flow rate, and salt temperature. Results indicate that the source term is the only parameter with a significant effect on the permeation rate. Pressure drop analysis suggests that the loop should operate below 200 LPM to maintain a pressure drop below 200 kPa. Additionally, finite-element analysis assesses thermal stress during loop operation to ensure the experiment's safe design. MSTTE will provide semi-integral data on tritium transport phenomena in molten salts and serve as a testbed for advancing molten salt technology.

11 - NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2025

How does ion temperature gradient turbulence depend on magnetic geometry? Insights from data and machine learning

Magnetic geometry has a significant effect on the level of turbulent transport in fusion plasmas. Here, we model and analyse this dependence using multiple machine learning methods and a dataset of >200 000 nonlinear gyrokinetic simulations of ion-temperature-gradient turbulence in diverse non-axisymmetric geometries. The dataset is generated using a large collection of both optimised and randomly generated stellarator equilibria. At fixed gradients and other input parameters, the turbulent heat flux varies between geometries by several orders of magnitude. Trends are apparent among the configurations with particularly high or particularly low heat flux. Regression and classification techniques from machine learning are then applied to extract patterns in the dataset. Due to a symmetry of the gyrokinetic equation, the heat flux and regressions thereof should be invariant to translations of the raw features in the parallel coordinate, similar to translation invariance in computer vision applications. Multiple regression models including convolutional neural networks (CNNs) and decision trees can achieve reasonable predictive power for the heat flux in held-out test configurations, with highest accuracy for the CNNs. Using Spearman correlation, sequential feature selection and Shapley values to measure feature importance, it is consistently found that the most important geometric lever on the heat flux is the flux surface compression in regions of bad curvature. The second most important geometric feature relates to the magnitude of geodesic curvature. These two features align remarkably with surrogates that have been proposed based on theory, while the methods here allow a natural extension to more features for increased accuracy. The dataset, released with this publication, may also be used to test other proposed surrogates, and we find that many previously published proxies do correlate well with both the heat flux and stability boundary.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗