Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “uncertainty modeling”

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 1,297 records · Page 72
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Model Reduction for Steady Hypersonic Aerodynamics via Conservative Manifold Least-Squares Petrov–Galerkin Projection

High-speed aerospace engineering applications rely heavily on computational fluid dynamics (CFD) models for design and analysis. This reliance on CFD models necessitates performing accurate and reliable uncertainty quantification (UQ) of the CFD models, which can be very expensive for hypersonic flows. Additionally, UQ approaches are many-query problems requiring many runs with a wide range of input parameters. One way to enable computationally expensive models to be used in such many-query problems is to employ projection-based reduced-order models (ROMs) in lieu of the (high-fidelity) full-order model (FOM). In particular, the least-squares Petrov–Galerkin (LSPG) ROM (equipped with hyper-reduction) has demonstrated the ability to significantly reduce simulation costs while retaining high levels of accuracy on a range of problems, including subsonic CFD applications. This allows LSPG ROM simulations to replace the FOM simulations in UQ studies, making UQ tractable even for large-scale CFD models. This work presents the first application of LSPG to a hypersonic CFD application, the Hypersonic International Flight Research Experimentation 1 (HIFiRE-1) in a three-dimensional, turbulent Mach 7.1 flow. This paper shows the ability of the ROM to significantly reduce computational costs while maintaining high levels of accuracy in computed quantities of interest.

42 ENGINEERING↗
DOE OSTI2020

Model reduction for hypersonic aerodynamics via conservative LSPG projection and hyper-reduction

High-speed aerospace engineering applications rely heavily on computational fluid dynamics (CFD) models for design and analysis due to the expense and difficulty of flight tests and experiments. This reliance on CFD models necessitates performing accurate and reliable uncertainty quantification (UQ) of the CFD models. However, it is very computationally expensive to run CFD for hypersonic flows due to the fine grid resolution required to capture the strong shocks and large gradients that are typically present. Furthermore, UQ approaches are “many-query” problems requiring many runs with a wide range of input parameters. One way to enable computationally expensive models to be used in such many-query problems is to employ projection-based reduced-order models (ROMs) in lieu of the (high-fidelity) full-order model. In particular, the least-squares Petrov–Galerkin (LSPG) ROM (equipped with hyper-reduction) has demonstrated the ability to significantly reduce simulation costs while retaining high levels of accuracy on a range of problems including subsonic CFD applications. This allows computationally inexpensive LSPG ROM simulations to replace the full-order model simulations in UQ studies, which makes this many-query task tractable, even for large-scale CFD models. This work presents the first application of LSPG to a hypersonic CFD application. In particular, we present results for LSPG ROMs of the HIFiRE-1 in a three-dimensional, turbulent Mach 7.1 flow, showcasing the ability of the ROM to significantly reduce computational costs while maintaining high levels of accuracy in computed quantities of interest.

Fluid Dynamics↗
DOE OSTI2020

Revised and New Proper Motions for Confirmed and Candidate Milky Way Dwarf Galaxies

A new derivation of systemic proper motions of Milky Way satellites is presented and applied to 59 confirmed or candidate dwarf galaxy satellites using Gaia Data Release 2. This constitutes all known Milky Way dwarf galaxies (and likely candidates) as of 2020 May, except for the Magellanic Clouds, the Canis Major and Hydra 1 stellar overdensities, and the tidally disrupting Bootes III and Sagittarius dwarf galaxies. We derive systemic proper motions for the first time for Indus 1, DES J0225+0304, Cetus 2, Pictor 2, and Leo T, but note that the latter three rely on photometry that is of poorer quality than that of the rest of the sample. We cannot resolve a signal for Bootes 4, Cetus 3, Indus 2, Pegasus 3, or Virgo 1. Our method is inspired by the maximum likelihood approach of Pace & Li and examines simultaneously the spatial, color–magnitude, and proper motion distribution of sources. Systemic proper motions are derived without the need to identify confirmed radial velocity members, although the proper motions of these stars, where available, are incorporated into the analysis through a prior on the model. The associated uncertainties on the systemic proper motions are on average a factor of ∼1.4 smaller than existing literature values. Analysis of the implied membership distribution of the satellites suggests that we accurately identify member stars with a contamination rate lower than 1 in 20.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2023

Cosmic Supernova Neutrino and Gamma-Ray Backgrounds in the MeV Regime

Cosmic multimessenger backgrounds include relic diffuse components created in the early universe and contributions from individual sources. Here, we study both type Ia supernovae (SNe Ia) and core-collapse supernovae (CCSNe) contributions to the diffuse neutrino and gamma-ray backgrounds in the MeV regime referred to as DSNB and DSGB respectively. We show that the diffuse SN Ia background is 10 6 times lower (for $\overline{v}_e$) than the CCSN background making it negligible. Our predicted DSNB $\overline{v}_e$ flux at earth in the 19.3–32 MeV regime is 0.36 ν cm -2 s -1 . We also find that the DSNB flux in the energy range from 11.3 to 32 MeV varies by +29% with a change in the SFRD model from Madau & Fragos which yielded a minimum predicted flux, to the extragalactic background light reconstruction model (maximum predicted flux). The diffuse SN Ia gamma-ray background and its dependence on the progenitor supernova delay time distribution are also evaluated. Furthermore, we address the origin of the CGB (Cosmic Gamma-ray Background) in the 0.1–7 MeV regime by adding contributions from sources such as SNe Ia, CCSNe, radio-quiet Active Galactic Nuclei, Flat spectrum radio quasars (FSRQs) and Neutron star—neutron star mergers. We find that our modeled background (including uncertainties) matches the observed CGB above 1.0 MeV, but is a factor ≈2 lower than the observed flux in the 0.1–1.0 MeV range, highlighting the need for future MeV missions to establish the CGB spectrum more reliably, and to possibly identify additional sources or even source classes in the underexplored MeV band.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

Particle phase-state variability in the North Atlantic free troposphere during summertime is determined by atmospheric transport patterns and sources

Abstract. Free tropospheric aerosol particles have important but poorly constrained climate effects due to transformations of their physicochemical properties during long-range transport. In this study, we investigate the chemical composition and provide an overview of the phase states of individual particles that have undergone long-range transport over the North Atlantic Ocean in June and July 2014, 2015, and 2017 to the Observatory of Mount Pico (OMP) in the Azores. The OMP is an ideal site for studying long-range-transported free tropospheric particles because local emissions have a negligible influence and contributions from the boundary layer are rare. We used the FLEXible PARTicle Lagrangian particle dispersion model (FLEXPART) to determine the origins and transport trajectories of sampled air masses and found that most of them originated from North America and recirculated over the North Atlantic Ocean. The FLEXPART analysis showed that the sampled air masses were highly aged (average plume age >10 d). Size-resolved chemical compositions of individual particles were probed using computer-controlled scanning electron microscopy with an energy-dispersive X-ray spectrometer (CCSEM-EDX) and scanning transmission X-ray microscopy with near-edge X-ray absorption fine structure spectroscopy (STXM-NEXAFS). CCSEM-EDX results showed that the most abundant particle types were carbonaceous (∼ 29.9 % to 82.0 %), sea salt (∼ 0.3 % to 31.6 %), and sea salt with sulfate (∼ 2.4 % to 31.5 %). We used a tilted stage interfaced within an environmental scanning electron microscope (ESEM) to determine the phase states of individual submicron particles. We found that most particles (∼ 47 % to 99 %) were in the liquid state at the time of collection due to inorganic inclusions. Moreover, we also observed substantial fractions of solid and semisolid particles (∼ 0 % to 30 % and ∼ 1 % to 42 %, respectively) during different transport patterns and events, reflecting the particles' phase-state variability for different atmospheric transport events and sources. Combining phase state measurements with FLEXPART CO tracer analysis, we found that wildfire-influenced plumes can result in particles with a wide range of viscosities after long-range transport in the free troposphere. We also used temperature and RH values extracted from the Global Forecast System (GFS) along the FLEXPART-simulated path to predict the phase state of the particles during transport and found that neglecting internal mixing with inorganics would lead to an overestimation of the viscosity of free tropospheric particles. Our findings warrant future investigation aiming at the quantitative assessment of the influence of internal mixing on the phase states of the individual particles. This study also provides insights into the chemical composition and phase state of free tropospheric particles, which can help models to reduce uncertainties about the effects of ambient aerosol particles on climate.

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

Aircraft-derived particle fluxes distinguish entrainment zone and decoupled layer nucleation in marine boundary layers

The vertical distribution of freshly nucleated aerosol particles in the marine boundary layer remains poorly constrained, limiting our ability to represent new particle formation in climate models. Here we characterize 3–10 nm particle events, termed small particle events (SPEs), by deriving their vertical turbulent fluxes from aircraft measurements during the Aerosol and Cloud Experiments in the Eastern North Atlantic (ACE-ENA) campaign. To overcome stationarity limitations of traditional eddy covariance methods, we applied continuous wavelet transform analysis to data collected during June–July 2017 and January–February 2018 flights over the Azores. Our flux-based analysis revealed two distinct SPE scenarios with different vertical structures and spatial extents. The first featured nucleation in the entrainment zone, where free tropospheric air entrains into the boundary layer. The second showed nucleation in the decoupled layer, a stratified region between the well-mixed surface layer and cloud-topped upper boundary layer. In both cases, convergence of air masses from different layers diluted preexisting aerosol surface area to very low levels, creating conditions favorable for nucleation and generating strong downward particle fluxes. SPEs occurred in 15 % of flights, challenging prevailing theoretical expectations that new particle formation should rarely occur in marine boundary layers due to high condensation and coagulation sink capacity of sea spray aerosols. Aircraft-derived particle fluxes provide first observational constraints on the vertical location and source strength of likely nucleation regions in the remote marine boundary layer, improving aerosol source representations in climate models and reducing uncertainties in aerosol-cloud interactions.

Rasheeda Satheesh, Ajmal [North Carolina State Uni↗
DOE OSTI2022

The sensitivity of the Fitch wind farm parameterization to a three-dimensional planetary boundary layer scheme

Abstract. Wind plant wake impacts can be estimated with a number of simulation methodologies, each with its own fidelity and sensitivity to model inputs. In turbine-free mesoscale simulations, hub-height wind speeds often significantly vary with the choice of a planetary boundary layer (PBL) scheme. However, the sensitivity of wind plant wakes to a PBL scheme has not been explored because, as of the Weather Research and Forecasting model v4.3.3, wake parameterizations were only compatible with one PBL scheme. We couple the Fitch wind farm parameterization with the new NCAR 3DPBL scheme and compare the resulting wakes to those simulated with a widely used PBL scheme. We simulate a wind plant in pseudo-steady states under idealized stable, neutral, and unstable conditions with matching hub-height wind speeds using two PBL schemes: MYNN and the NCAR 3DPBL. For these idealized scenarios, average hub-height wind speed losses within the plant differ between PBL schemes by between −0.20 and 0.22 m s−1, and correspondingly, capacity factors range between 39.5 %–53.8 %. These simulations suggest that PBL schemes represent a meaningful source of modeled wind resource uncertainty; therefore, we recommend incorporating PBL variability into future wind plant planning sensitivity studies as well as wind forecasting studies.

17 WIND ENERGY↗
DOE OSTI2019

Mercury reallocation in thawing subarctic peatlands

Warming Arctic temperatures have led to permafrost thaw that threatens to release previously sequestered mercury (Hg) back into the environment. Mobilisation of Hg in permafrost waters is of concern, as Hg methylation produced under water-saturated conditions results in the neurotoxin, methyl Hg (MeHg). Thawing permafrost may enhance Hg export, but the magnitude and mechanisms of this mobilisation within Arctic ecosystems remain poorly understood. Such uncertainty limits prognostic modelling of Hg mobilisation and impedes a comprehensive assessment of its threat to Arctic ecosystems and peoples. Here, we address this knowledge gap through an assessment of Hg dynamics across a well-studied permafrost thaw sequence at the peak of the growing season in biologically active peat overlying permafrost, quantifying total gaseous mercury (TGM) fluxes, total mercury (Hg Tot ) in the active layer peat, porewater MeHg concentrations, and identifying microbes with the potential to methylate Hg. During the initial thaw, TGM is liberated, likely by photoreduction from permafrost where it was previously stored for decades to centuries. As thawing proceeds, TGM is largely driven by hydrologic changes as evidenced by Hg accumulation in water-logged, organic-rich peat sediments in fen sites. MeHg in porewaters increase across the thaw gradient, a pattern coincident with increases in the relative abundance of microbes possibly containing genes allowing for methylation of ionic Hg. Findings suggest that under changing climate, frozen, well-drained habitats will thaw and collapse into saturated landscapes, increasing the production of MeHg and providing a significant source of the toxic, bioaccumulative contaminant.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Measurement of the Higgs boson production via vector boson fusion and its decay into bottom quarks in proton-proton collisions at $\sqrt{s}$ = 13 TeV

A measurement of the Higgs boson (H) production via vector boson fusion (VBF) and its decay into a bottom quark-antiquark pair ($\mathrm{b\bar{b}}$) is presented using proton-proton collision data recorded by the CMS experiment at $\sqrt{s}$ = 13 TeV and corresponding to an integrated luminosity of 90.8 fb$^{-1}$. Treating the gluon-gluon fusion process as a background and constraining its rate to the value expected in the standard model (SM) within uncertainties, the signal strength of the VBF process, defined as the ratio of the observed signal rate to that predicted by the SM, is measured to be $\mu^\text{qqH}_\mathrm{Hb\bar{b}}$ = 1.01$^{+0.55}_{-0.46}$. The VBF signal is observed with a significance of 2.4 standard deviations relative to the background prediction, while the expected significance is 2.7 standard deviations. Considering inclusive Higgs boson production and decay into bottom quarks, the signal strength is measured to be $\mu^\text{incl.}_\mathrm{Hb\bar{b}}$ = 0.99$^{+0.48}_{-0.41}$, corresponding to an observed (expected) significance of 2.6 (2.9) standard deviations.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2024

DeepUQ: Assessing the Aleatoric Uncertainties from two Deep Learning Methods

Assessing the quality of aleatoric uncertainty estimates from uncertainty quantification (UQ) deep learning methods is important in scientific contexts, where uncertainty is physically meaningful and important to characterize and interpret exactly. We systematically compare aleatoric uncertainty measured by two UQ techniques, Deep Ensembles (DE) and Deep Evidential Regression (DER). Our method focuses on both zero-dimensional (0D) and two-dimensional (2D) data, to explore how the UQ methods function for different data dimensionalities. We investigate uncertainty injected on the input and output variables and include a method to propagate uncertainty in the case of input uncertainty so that we can compare the predicted aleatoric uncertainty to the known values. We experiment with three levels of noise. The aleatoric uncertainty predicted across all models and experiments scales with the injected noise level. However, the predicted uncertainty is miscalibrated to $\rm{std}(\sigma_{\rm al})$ with the true uncertainty for half of the DE experiments and almost all of the DER experiments. The predicted uncertainty is the least accurate for both UQ methods for the 2D input uncertainty experiment and the high-noise level. While these results do not apply to more complex data, they highlight that further research on post-facto calibration for these methods would be beneficial, particularly for high-noise and high-dimensional settings.

Nevin, Rebecca↗
DOE OSTI2025

End-to-end differentiable digital twin for the IOTA/FAST facility

As the design complexity of modern accelerators grows, there is more interest in using controllable-fidelity simulations that have fast execution time or can yield additional insights about accelerator state. One notable example of additional information are gradients of physical observables with respect to design parameters produced by differentiable simulations. The IOTA/FAST facility has recently begun a program to implement and experimentally validate an end-to-end digital twin to serve as a virtual accelerator test stand, allowing for rapid prototyping of new software and experiments with minimal beam time costs. In this contribution we will discuss our plans and progress. Specifically, we will cover the selection and benchmarking of both physics and ML codes for linac and ring simulation, the development of generic interfaces between surrogate and physics-based sections, and presenting the control interface as either a deterministic event loop or a fully asynchronous EPICS soft input/output controller. We will also discuss challenges in model calibration and uncertainty quantification, as well as future plans to implement larger proton accelerators like PIPII and Booster.

Kuklev, N. [Fermilab]↗
DOE OSTI2023

Code and Data for "Quantifying Uncertainties in Greenhouse Gas Savings and Mitigation Costs with Cellulosic Biofuels"

Use of cellulosic biofuels from non-feedstocks are modeled using the BEPAM (Biofuel and Environmental Policy Analysis Model) model to quantifying the uncertainties about induced land use change effects, net greenhouse gas saving potential, and economic costs. The code is in GAMS, general algebraic modeling language. NOTE: Column 3 is titled "BAU" in "merged_BAU.gdx", "merged_RFS.gdx", and "merged_CEM.gdx", but contains "RFS" data in "merged_RFS.gdx" and "CEM" data in "merged_CEM.gdx".

BEPAM↗
DOE OSTI2024

Unveiling the Essential Parameters Driving Mineral Reactions during CO2 Storage in Carbonate Aquifers through Proxy Models

Numerical simulation is a commonly employed technique for studying carbon dioxide (CO2) storage processes in porous media, particularly saline aquifers. It enables the representation of diverse trapping mechanisms and the assessment of CO2 retention capacity within the subsurface. The intricate physicochemical phenomena involved necessitate the incorporation of multiphase flow, accurate depiction of fluid and rock properties, and their interactions. Among these factors, geochemical reaction rates and mechanisms are pivotal for successful CO2 trapping in carbonate reactive rocks. However, research on kinetic parameters and the influence of lithology on CO2 storage remains limited. This limitation is partly due to the challenges faced in laboratory experiments, where the time scale of the reactions and the lack of in situ conditions hinder accurate measurement of mineral reaction rates. This study employs proxy models constructed using response surfaces calibrated with simulation results to address uncertainties associated with geochemical reactions. Monte Carlo simulation is utilized to explore a broader range of parameters and identify influential factors affecting CO2 mineralization. The findings indicate that an open database containing kinetic parameters can support uncertainty assessment. Additionally, the proxy models effectively represent objective functions related to CO2 injectivity and mineralization, with calcite dissolution playing a predominant role. pH, calcite concentration, and CO2 injection rate significantly impact dolomite precipitation, while quartz content remains unaffected.

58 GEOSCIENCES↗
DOE OSTI2022

Theoretical and calculable dependent variables and their covariance in nuclear data libraries [Slides]

This presentation begins by defining observables, including theoretical observable, calculable observable, and measured observable. It also provides definitions and examples of experimental effects. Additionally, uncertainty in evaluated libraries and a metric to probe evaluated uncertainty is presented. In conclusion, being related to theoretical quantities such as resonance parameters, the uncertainty in ENDF libraries (e.g., ENDF/B-VIII.0) can be overestimated because it is usually evaluated under the guidance of experimental uncertainty: lim Δσexp.corr.→0 Δσ = Δσ theoretical model . The presentation states that a clear distinction of the uncertainty between nuclear theoretical models and experimental corrections should be revisited and this is important since (e.g., transport) simulations need theoretical quantities convoluted with specific operational parameters and material configurations. It also states that strong coupling between uncertainty quantification methodologies and optimization procedures exists, and it is necessary to develop methodologies to obtain uncertainty on theoretical models from physical and mathematical constraints.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2022

Inverse Modeling of Hydrologic Parameters in CLM4 via Generalized Polynomial Chaos in the Bayesian Framework

In this work, generalized polynomial chaos (gPC) expansion for land surface model parameter estimation is evaluated. We perform inverse modeling and compute the posterior distribution of the critical hydrological parameters that are subject to great uncertainty in the Community Land Model (CLM) for a given value of the output LH. The unknown parameters include those that have been identified as the most influential factors on the simulations of surface and subsurface runoff, latent and sensible heat fluxes, and soil moisture in CLM4.0. We set up the inversion problem in the Bayesian framework in two steps: (i) building a surrogate model expressing the input–output mapping, and (ii) performing inverse modeling and computing the posterior distributions of the input parameters using observation data for a given value of the output LH. The development of the surrogate model is carried out with a Bayesian procedure based on the variable selection methods that use gPC expansions. Our approach accounts for bases selection uncertainty and quantifies the importance of the gPC terms, and, hence, all of the input parameters, via the associated posterior probabilities.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

EPIsembleVis: A geo-visual analysis and comparison of the prediction ensembles of multiple COVID-19 models

In this work, we present EPIsembleVis, a web-based comparative visual analysis tool for evaluating the consistency of multiple COVID-19 prediction models. Our approach analyzes a collection of COVID-19 predictions from different epidemiological models as an ensemble and utilizes two metrics to quantify model performance. These metrics include (a) prediction uncertainty (represented as the dispersion of predictions in each ensemble) and (b) prediction error (calculated by comparing individual model predictions with the recorded data). Through an interactive visual interface, our approach provides a data-driven workflow for (a) selecting and constructing the COVID-19 model prediction ensemble based on the spatiotemporal overlap of available predictions of multiple epidemiological models, (b) quantifying the model performance using both the uncertainty of each model prediction ensemble, and the error of each ensemble member that represents individual model predictions, and (c) visualizing the spatiotemporal variability in the projection performance of individual models using a suite of novel ensemble visualization techniques, such as the data availability map, a spatiotemporal textured-tile calendar, multivariate rose chart, and time-series leaflet glyph. We demonstrate the capability of our ensemble visual interface through a case study that investigates the performance of weekly COVID-19 predictions, which are provided through the COVID-19 Forecast Hub UMass-Amherst Influenza Forecasting Center of Excellence [47] for the United States and United States Territories. The EPIsembleVis tool is implemented using open-source web technologies and adaptive system design, rendering it interoperable with Elasticsearch and Kibana for automatically ingesting COVID-19 predictions from online repositories, and it is generalizable for analyzing worldwide projections from more epidemiological models.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2021

National Opacity Program: Analysis of Opacity-relevant X-ray Emission Spectra (Milestone Report ID 7121)

We have produced high energy density iron plasmas at temperatures above 1 keV and electron densities exceeding 1023 cm -3 (~ 1 g/cm 3 ) using the Orion laser at the Atomic Weapons Establishment. These plasmas were created by irradiating 50 µm diameter layered targets with frequency doubled (λ = 527 nm), 1 ps laser pulses focused to a 100 µm diameter producing an irradiance of ~ 2 x 10 18 W/cm 2 . The buried layer targets consist of 160 nm iron sulfide (FeS), 60 nm potassium chloride (KCl), and 15 nm carbon. The combined layers are tamped on both sides with 3 µm of parylene-N. The x-ray emission from the plasma was measured using two time-resolved, and four time-integrated Bragg crystal spectrometers, as well as one time-integrated imaging system. One time-resolved x-ray spectrometer measured emission from L-shell transitions in highly charged iron, the other from K-shell transitions in helium-like S 14+ and hydrogen-like S 15+ . The time-integrated spectrometers are intensity-calibrated and measured emission from K-shell transitions in sulfur, potassium, chlorine, and both K-shell and L-shell transitions in iron. The density and temperature of the plasma were determined by modeling the x-ray spectra using different spectral and hydrodynamic modeling packages. A brief overview of the uncertainties associated with the measurements and models are presented. We also give an overview of our 1-D HYDRA-DCA radiation-hydrodynamics model and improvements for future work. Our results aid in assessing experimental uncertainties associated with plasma uniformity and with x-ray emission employed as diagnostics in opacity experiments at temperatures and densities not achievable elsewhere and represent a significant step in creating and diagnosing plasmas near LTE. These results are summarized as part of the completion requirements for milestone 7121

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2025

Surrogate-driven Variance-based Sensitivity Analysis of Thermal Storage Tanks in Integrated Energy Systems

Sensitivity analysis and uncertainty quantification are essential steps for enhancing the accuracy of computational models by identifying and mitigating uncertainties. This study focuses on these steps for the Thermal Energy Delivery System at Idaho National Laboratory, specifically targeting the thermocline tank. Using a Modelica/Dymola simulation model, the study perturbed various design parameters and boundary conditions, including shape factor, porosity, outlet temperature, inlet mass flow rate, and system pressure, to predict and quantify uncertainty in the tank’s ax- ial temperature. A dataset of over 1,000 simulations was generated, and surrogate models were developed using the pyMAISE (Michigan Artificial Intelligence Standard Environment) library, which is an Automatic Machine Learning library for nuclear engineering applications. The optimal model, a feedforward neural network with two hidden layers, achieved an R2 score above 0.99 and a mean absolute error below 1 Kelvin. Sensitivity analyses using Sobol indices and Fourier amplitude sensitivity testing methods on this surrogate model revealed that the inlet mass flow rate at initial timestamps and porosity significantly impacts predicted temperatures across all sensors and time steps.

22 - GENERAL STUDIES OF NUCLEAR REACTORS↗