Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Performance benchmark”

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 1,045 records · Page 58
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Performance Improvements for the Griffin Transport Solvers

Griffin is a Multiphysics Object-Oriented Simulation Environment based reactor multiphysics analysis application jointly developed by Idaho National Laboratory and Argonne National Laboratory. Griffin includes a variety of deterministic radiation transport solvers for fixed source, k-eigenvalue, adjoint, and subcritical multiplication, as well as transient solvers for point-kinetics, improved quasi-static, and spatial dynamics. A code assessment performed in FY-20 identified two significant issues with the transport solvers in Griffin: first, the primary heterogeneous SN (discrete ordinates) transport solver based on continuous finite element methods required significant mesh refinement and higher memory usage compared to solvers based on the method of characteristic for equivalent accuracy. Second, the homogeneous PN (spherical harmonics expansion) transport solver did not adequately support polynomial refinement, which is a feature usually required for problems with spatial homogenization and pronounced streaming, typical in fast or gas-cooled reactor systems. To address the first issue, the development effort focused on the more promising discontinuous finite element method (DFEM)-based SN transport solver in Griffin. The addition of an asynchronous parallel transport sweeper and coarse mesh finite difference (CMFD) acceleration have rendered a superior heterogeneous SN transport capability for multiphysics problems that requires far less computing resources in terms of both CPU time and memory usage. This is demonstrated with typical thermal- and fast-spectrum reactor benchmark problems, including 2D Transient Reactor Test, 3D Advanced Burner Test Reactor (ABTR), and 2D and 3D Empire microreactor. For the second issue, the development effort focused on a new transport solver based on the hybrid finite element PN method (HFEM-PN), equivalent to the variational nodal method, as well as a new diffusion solver based on HFEM-Diffusion. This solver is intended for homogenized domains with multiphysics coupling (i.e., supports mesh displacement, seamless temperature feedback, etc.). Initial calculations with the HFEM-Diffusion implementation show very good parallel efficiency for the residual evaluations with the 2D ABTR benchmark. A future development effort will be centered on further improvements to the CMFD, HFEM-PN, and DFEM diffusion solvers to ensure Griffin meets performance and software quality assurance requirements for advanced reactor design and analysis.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2025

Experiences with SYCL on AMD GPUs with Kokkos

With the recent diversification of the hardware landscape in the high-performance computing (HPC) community, performance-portability solutions are becoming more and more important. One of the most popular choices is Kokkos, which recently became a Linux Foundation project. Most of its development is supported by the US Department of Energy and the French Alternative Energies and Atomic Energy Commission. Kokkos is implemented as a C++ library with multiple backends to support CPUs as well as various GPU architectures. These backends include OpenMP, CUDA, HIP, and also SCYL. This approach enables users to leverage the preferred vendor toolchain for the respective platform (e.g. CUDA, ROCm, OneAPI). The SYCL backend is used to target Intel GPUs, in particular to support the Aurora exascale supercomputer. However, SYCL itself also offers a large degree of portability, and in fact Kokkos’ CI for SYCL has been running on NVIDIA hardware due to a lack of access to Intel GPUs. In this report, we describe our experience with using Kokkos SYCL backend on AMD GPUs targeting the Frontier supercomputer at Oak Ridge National Laboratory. The two major SYCL implementations are DPC++ and AdaptiveCpp. While the Kokkos SYCL backend has been implemented using the former, the latter was the first implementation to target AMD GPUs. We will discuss the experience with both of these SYCL implementations in terms of functionality and performance. Using Kokkos to evaluate SYCL toolchains has a number of benefits. Kokkos’ use of SYCL is fairly complex, exercising features such as graphs, relocatable device functions, atomics – including for non-arithmetic types, as well as pinned and page migratable memory allocations. Kokkos also needs to implement capabilities such as Kokkos’ hierarchical parallelism that are not a straight-forward mapping to SYCL capabilities. Furthermore, a large number of libraries and applications that represent diverse use cases are implemented in Kokkos, providing readily available test cases for a toolchain evaluation. Preliminary results show that support for AMD GPUs in DPC++ is much less mature than for NVIDIA GPUs or Intel GPUs. While the situation has improved significantly over the last year, we still encounter many runtime failures, dispatching problems, and code generation issues. With AdaptiveCpp the challenges arise even earlier in the evaluation process. Since Kokkos’ SYCL implementation is largely focused on supporting Intel GPUs, we opted to leverage SYCL extensions which are available in DPC++ but not in AdaptiveCpp. Furthermore, AdaptiveCpp appears to be less conformant with the SYCL2020 standard which Kokkos relies on. In some cases, we are able to work around the lack of feature support, in other cases we have to disable certain Kokkos capabilities to evaluate the toolchain. Our evaluation will leverage Kokkos’ unit tests to establish basic functionality and feature completeness. We then use simple benchmarks for components of a CG implementation as a measure of usability and performance of the SYCL toolchains.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Packaging HEP Heterogeneous Mini-apps for Portable Benchmarking and Facility Evaluation on Modern HPCs

High Energy Physics (HEP) experiments are making increasing use of GPUs and GPU dominated High Performance Computer facilities. Both the software and hardware of these systems are rapidly evolving, creating challenges for experiments to make informed decisions as to where they wish to devote resources. In its first phase, the High Energy Physics Center for Computational Excellence (HEP-CCE) produced portable versions of a number of heterogeneous HEP mini-apps, such as p2r, FastCaloSim, Patatrack and the WireCell Toolkit, that exercise a broad range of GPU characteristics, enabling cross platform and facility benchmarking and evaluation. However, these miniapps still require a significant amount of manual intervention to deploy on a new facility. We present our work in developing turn-key deployments of these mini-apps, where by means of containerization and automated configuration and build techniques such as Spack, we are able to quickly test new hardware, software, environments and entire facilities with minimal user intervention, and then track performance metrics over time.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2025

Antenna arrays for neutrino mass measurements with cyclotron radiation emission spectroscopy

Cyclotron Radiation Emission Spectroscopy (CRES) is a technique for precision measurements of kinetic energies of charged particles, pioneered by the Project 8 experiment to measure the neutrino mass using the tritium end-point method. It was recently employed for the first time to measure the molecular tritium spectrum and place a limit on the neutrino mass using a cubic-centimeter-scale detector. Future direct neutrino mass experiments are developing the technique to overcome the systematic and statistical limitations of current detectors. Here, this paper describes one such approach, namely the use of antenna arrays for CRES in free space. Phenomenology, detector design, simulation, and performance estimates are discussed, culminating with an example design with a projected sensitivity of 𝑚 𝛽 < ⁢0.04 eV/𝑐 2 . Prototype antenna array measurements are also shown for a demonstrator-scale setup as a benchmark for the simulation. By consolidating these results, this paper serves as a comprehensive reference for the development and performance of antenna arrays for CRES.

Physics - Nuclear physics and radiation physics↗
DOE OSTI2022

CSEWG Meeting Minutes

Work has been done at LANL to validate various updated evaluations from ENDF/BVIII.1b0 including: Fe, Cu, Si, U, Pu. Continuous energy tables were processed with NJOY2016 and output to ACE files. For non-Pu legacy critical benchmarks, all calculations were within the experimental benchmark uncertainty, with some benchmarks having shown lower bias using ENDF/B-VIII.1b0. In HST systems, ENDF/B-VIII.0 shows reasonable results above the thermal leakage fraction. ENDF/B-VIII.0.1 is doing slightly worse than ENDF/B- VIII.0. b0.2 and it also includes a new 19 F evaluation, so more work is needed to examine its impact. 233 U thermal and intermediate bare and water-reflected benchmarks agree well with ENDF/B-VIII.0. From ENDF/B-VIII.0 to ENDF/B-VIII.0b0.1, HMF-007 C/E are increasing and the new 19 F evaluation is helping the HMF-007 cases containing Teflon which have historically had high C/E. The Pu legacy benchmark suite is within experimental uncertainty with the exception of THOR. The INDEN evaluation followed by LANL gave the lowest mean absolute bias, but it remains difficult to make a determination on which Pu evaluation is the best. INDEN performs slightly better for the JEZEBEL series, while the LANL Pu evaluation is better for THOR and flattop benchmarks. For the PMI benchmark series INDEN outperforms the LLNL evaluation. For the PST benchmark series ENDF/B-VIII.0 and LLNL Pu evaluations are better than the INDEN evaluation. C/E ratios for Jezebel and Godiva reaction rate ratios show all the Pu evaluations are very close to one another, making it again difficult to say which is best. When examining neutron reactor kinetics parameters, each Pu evaluation is once again very close. Each individual Pu evaluation seems to have advantages in their own applications.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2025

Packaging HEP Heterogeneous Mini-apps for Portable Benchmarking and Facility Evaluation on Modern HPCs

High Energy Physics (HEP) experiments are making increasing use of GPUs and GPU dominated High Performance Computer facilities. Both the software and hardware of these systems are rapidly evolving, creating challenges for experiments to make informed decisions as to where they wish to devote resources. In its first phase, the High Energy Physics Center for Computational Excellence (HEP-CCE) produced portable versions of a number of heterogeneous HEP mini-apps, such as \ptor, FastCaloSim, Patatrack and the WireCell Toolkit, that exercise a broad range of GPU characteristics, enabling cross platform and facility benchmarking and evaluation. However, these mini-apps still require a significant amount of manual intervention to deploy on a new facility. We present our work in developing turn-key deployments of these mini-apps, where by means of containerization and automated configuration and build techniques such as Spack, we are able to quickly test new hardware, software, environments and entire facilities with minimal user intervention, and then track performance metrics over time.

Atif, Mohammad [Brookhaven] (ORCID:000000026889770↗
DOE OSTI2025

Comprehensive assessment of deep reinforcement learning approaches for economic dispatch in nuclear-driven microgrids

As the electrical grid integrates more variable renewable energy sources such as wind and solar, the demand for distributed and flexible systems to address this increased variability becomes critical. Nuclear-driven microgrids provide a promising solution by offering stable generation to complement intermittent renewables, ensuring grid reliability and operating efficiency. This paper proposes a recurrent deep reinforcement learning framework for optimal economic dispatch in a nuclear-powered microgrid integrating renewable energy sources, small modular reactors, battery storage systems, and balance-of-plant dynamics. A three-agent control architecture is developed, where demand and renewable energy agents act as forecasters, and a reinforcement learning-based dispatch agent performs real-time energy allocation. A nonlinear programming formulation is first used to generate an optimal baseline for benchmarking. The proposed dispatch controller, based on Proximal Policy Optimization enhanced with Long Short-Term Memory networks, exploits temporal correlations in system dynamics by taking advantage of the time series used as inputs to improve policy robustness under uncertainty. Comparative analysis against established deep reinforcement learning methods, including Proximal Policy Optimization with a feedforward architecture, Soft Actor-Critic, and Twin Delayed Deep Deterministic Policy Gradient, demonstrates superior performance. Numerical results indicate that the proposed controller achieves a 0.39% cost reduction relative to the nonlinear programming benchmark and outperforms other learning-based methods by generating additional revenue of up to 0.35%. All reinforcement learning controllers compute dispatch actions in less than 0.3 s, resulting in a computational speedup of more than three orders of magnitude over the nonlinear programming baseline. The findings of this paper highlight their applicability for real-time operation and control in nuclear-integrated microgrids under volatile operating conditions.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2023

Multiphysics analysis of fuel fragmentation, relocation, and dispersal susceptibility–Part 3: Thermal hydraulic evaluation of large break LOCA under high-burnup conditions

Increasing the peak rod average burnup of pressurized water reactor (PWR) fuel beyond 62 GWd/tU may increase fuel fragmentation, relocation, and dispersal (FFRD) susceptibility during a large break loss of coolant accident (LBLOCA). TRACE thermal hydraulic (TH) LBLOCA analyses were performed for a realistic 24-month high-burnup PWR equilibrium cycle, to inform subsequent transient BISON high-burnup FFRD susceptibility evaluations. Realistic LBLOCA systems behavior was first established by configuring to and comparing with the BEMUSE OECD LBLOCA benchmark. Fuel and operating conditions were then applied from high-burnup VERA depletion calculations. LBLOCA simulations were performed for 281 selected high-burnup rods, for which transient TH boundary conditions were collected for later use in BISON. The TRACE results indicated that rod linear heat rate (rather than burnup) is the main predictor of peak cladding temperature (PCT) during the event. PCT typically occurred at a local burnup lower than the rod-average burnup, especially for twice-burned fuel.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2021

Benchmarking near-term quantum devices with the variational quantum eigensolver and the Lipkin-Meshkov-Glick model

The variational quantum eigensolver is a promising algorithm for noisy intermediate scale quantum (NISQ) computation. Verification and validation of NISQ algorithms' performance on NISQ devices is an important task. Here, we consider the exactly diagonalizable Lipkin-Meshkov-Glick (LMG) model as a candidate for benchmarking NISQ computers. We use the Bethe Ansatz to construct eigenstates of the trigonometric LMG model using quantum circuits inspired by the LMG's underlying algebraic structure. We construct circuits with depth $\mathcal{O}$(N) and $\mathcal{O}$(log 2 N) that can prepare any trigonometric LMG eigenstate of N particles. The number of gates required for both circuits is $\mathcal{O}$(N). The energies of the eigenstates can then be measured and compared to the exactly known answers.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2020

Benchmark Modeling and Simulation of the FFTF LOFWOS Test #13 Using SAM

The Fast Flux Test Facility (FFTF) was a 400 MW thermal powered, oxide-fueled, liquid sodium cooled test reactor, built to assist development and testing of advanced fuels and materials for fast breeder reactors. In July 1986, a series of unprotected Loss of Flow Without Scram (LOFWOS) transients were performed in FFTF as part of the Passive Safety Testing (PST) program. The LOFWOS Test #13, which was initiated at 50% power and 100% flow with the pump pony motors left off, has been chosen as a benchmark case by IAEA to support collaborative efforts within international partnerships on the validation of simulation tools and models in the area of sodium fast reactor passive safety in an IAEA Coordinated Research Project (CRP), launched in October 2018. The System Analysis Module (SAM) is an advanced and modern system analysis tool under development at Argonne National Laboratory for advanced non-LWR safety analysis. It utilizes the object-oriented application framework MOOSE to leverage the modern software environment and advanced numerical methods. The capabilities of SAM are being extended to enable the transient modeling, analysis, and design of various advanced nuclear reactor systems. To participate the IAEA CRP and enhance the SAM validation base for advanced reactor transient safety analysis, benchmark simulations of the FFTF LOFWOS Test #13 are performed using the SAM code. In this first phase of the validation effort, the thermal-hydraulic behavior of the reactor system is the focus and the reactor kinetics is not considered in the SAM FFTF model. Instead, the results of Argonne’s neutronics calculations are directly used, including the power shape of the active core region and the power history during the transient. The simulation results of FFTF at steady state agreed well with the measured data from the test. During the transient, reasonably good agreement were also obtained. Future work to improve the model will focus on introducing the reactivity predictions into the model, as well as better understanding or resolving the current discrepancies with the measured data.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2023

Improving E3SM Land Model Photosynthesis Parameterization via Satellite SIF, Machine Learning, and Surrogate Modeling

The parameterization of key photosynthesis parameters is one of the key uncertain sources in modeling ecosystem gross primary productivity (GPP). Solar-induced chlorophyll fluorescence (SIF) offers a good proxy for GPP since it marks the actual process of photosynthesis; while machine learning (ML) provides a robust approach to model the GPP-SIF relationship. Here, we trained the boosted regressing tree (BRT) and the Random Forest ML models with Greenhouse Gases Observing Satellite SIF data and in situ GPP observations from 49 eddy covariance towers. These trained ML GPP-SIF models were fed into the Energy Exascale Earth System Model (E3SM) Land Model (ELM) to generate ELM-simulated global SIF estimates, which were then benchmarked against satellite SIF observations with a surrogate modeling approach. Our results indicated good modeling performance of the ML-based GPP-SIF relationship. The ELM model when fed with the ML GPP-SIF models also can well predict the spatial-temporal variations in SIF. We also found high model accuracy for the surrogate modeling. Model parameter sensitivity analysis suggested that the fraction of leaf nitrogen in RuBisCO (flnr) is the most sensitive parameter to the SIF; other sensitive parameters include the Ball-Berry stomatal conductance slope (mbbopt) and the vcmax entropy (vcmaxse). The posterior uncertainty in simulated GPP was greatly reduced after benchmarking, and the model produced improved spatial patterns of mean GPP relative to FLUXCOM GPP. Our integrated approach provides a new avenue for improving land models and using remote-sensing SIF, which can be further improved in the future with more ground- and satellite-based observations.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

HEU Metal Delayed Critical Experiments with 10 to 19 Inch Thick Graphite Reflectors

Approximately 100 graphite-reflected highly enriched uranium (HEU, 93.14 wt % 235 U) metal annular and cylindrical critical experiments were performed in the early 1960s at the Oak Ridge Critical Experiments Facility (ORCEF). This report presents details from experiment logbooks, experimental data sheets and the author's memory for 44 HEU metal (93.14 wt % 235 U) critical assemblies with graphite reflectors varying from 10 to 19 in. thick, outside diameters varying from 7 to 15 in., inside diameters varying from 7 to 13 in. and critical HEU metal masses varying from 20.4 to 69.0 kg. The data from the 44 experiments described in this report are acceptable for use as criticality safety benchmark experiments for the International Criticality Safety Evaluation Program (ICSBEP) once the uncertainty analysis on the measured k eff is completed. Based on previous ICSBEP benchmarks with this HEU metal at ORCEF, the uncertainties in the measured k eff are expected to be as low as ±0.0004. Preparation of this report is part of an effort at Oak Ridge National Laboratory (ORNL) to document more than 15 undocumented series of critical and subcritical experiments enumerated in Critical and Subcritical NEA Benchmark Possibilities for Measurements at ORCEF and Other US DOE Facilities (Mihalzo, ORNL/TM-2019/1188, 2019) and performed by ORNL at ORCEF and other US Department of Energy critical experiments facilities. More than 500 operational days of critical facility time were used, not including setup and dismantlement time. This documentation for a part of one series of graphite reflected highly enriched uranium metal critical experiments, that used 50 operational days of ORCEF time, was performed using funding received from the DOE Office of Nuclear Energy’s Nuclear Energy University Programs at the University of Tennessee Nuclear Engineering Department. This documentation was also supported by the Nuclear Criticality, Radiation Transport, and Safety programs at ORNL.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2022

Setup and verification of a SCALE/KENO platform for generic FHR benchmark calculations

The work presented in this article is preliminary to downstream analysis of a generic fluoride salt-cooled high-temperature reactor (gFHR) core performed by the University of Tennessee in collaboration with Kairos Power (KP). A Monte Carlo transport model of the publicly available gFHR equilibrium core is developed in SCALE/KENO with multigroup energy treatment. Several output quantities of interest are used to verify the simulations against a benchmark model developed by KP using the continuous energy Monte Carlo code Serpent 2. Good agreement is seen in flux and fission rate profiles with a maximum relative difference of 1.4% and 2.8% respectively. Furthermore, an effective multiplication factor bias of 44 pcm was observed between the two simulations. The fuel temperature reactivity coefficient calculated with SCALE is within uncertainty to the reference model. This verification acts as a publicly reproducible benchmark for the gFHR in SCALE/KENO. A simplified depletion model is also presented where a single fuel pebble is depleted to discharge burnup through the equilibrium core while the equilibrium core is assumed to be invariant. This method produces results that intercept the equilibrium core concentrations in every case, however, an interesting artifact of this particular depletion model is uncovered. The phenomenon is shown to be a fundamental feature of the differential rate equations and inspires questions about how this system behaves when the time evolution of the equilibrium core is considered. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2021

Determining the Overpotential of Electrochemical Fuel Synthesis Mediated by Molecular Catalysts: Recommended Practices, Standard Reduction Potentials, and Challenges

Molecular catalysts capable of performing electrochemical transformations are frequently utilized in the synthesis of fuels. A common benchmark for evaluating catalysts for electrochemical reactions is overpotential (η) and minimizing η remains an active goal in catalysis. This Review details approaches for the determination of thermodynamic potentials for common fuel-forming reactions and for the determination of electrochemical overpotential, and underscores the need to employ methods that enable meaningful comparisons between molecular catalysts. Strategies for minimizing overpotential while maintaining high activity are highlighted.

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

Stable parallel training of Wasserstein conditional generative adversarial neural networks

In this work, we propose a stable, parallel approach to train Wasserstein conditional generative adversarial neural networks (W-CGANs) under the constraint of a fixed computational budget. Differently from previous distributed GANs training techniques, our approach avoids inter-process communications, reduces the risk of mode collapse and enhances scalability by using multiple generators, each one of them concurrently trained on a single data label. The use of the Wasserstein metric also reduces the risk of cycling by stabilizing the training of each generator. We illustrate the approach on the CIFAR10, CIFAR100, and ImageNet1k datasets, three standard benchmark image datasets, maintaining the original resolution of the images for each dataset. Performance is assessed in terms of scalability and final accuracy within a limited fixed computational time and computational resources. To measure accuracy, we use the inception score, the Fréchet inception distance, and image quality. An improvement in inception score and Fréchet inception distance is shown in comparison to previous results obtained by performing the parallel approach on deep convolutional conditional generative adversarial neural networks as well as an improvement of image quality of the new images created by the GANs approach. Weak scaling is attained on both datasets using up to 2000 NVIDIA V100 GPUs on the OLCF supercomputer Summit.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

An Open Combinatorial Diffraction Dataset Including Consensus Human and Machine Learning Labels with Quantified Uncertainty for Training New Machine Learning Models

Modern machine learning and autonomous experimentation schemes in materials science rely on accurate analysis of the data ingested by these models. Unfortunately, accurate analysis of the underlying data can be difficult, even for domain experts, complicating the training of the models intended to drive experiments. This is especially true when the goal is to identify the presence of weak signatures in diffraction or spectroscopic datasets. In this work, we examine a set of as-obtained diffraction data that track the phase transition from monoclinic to tetragonal in a Nb-doped VO2 film as a function of temperature and dopant concentration. We then task a set of domain experts and a set of machine learning experts with identifying which phase is present in each diffraction pattern manually and algorithmically, respectively; in both cases, the labels can vary dramatically, especially at the phase boundaries. We use the mode of the labels and the Shannon entropy as a method to capture, preserve and propagate consensus labels and their variance. Further we use the expert labels as a benchmark and demonstrate the use of Shannon entropy weighted scoring to test the performance of machine learning generated labels. Finally, we propose a material data challenge centered around generating improved labeling algorithms. This real-world dataset curated with expert labels can act as test bed for new algorithms. The raw data, annotations and code used in this study are all available online at data.gov and the interested reader is encouraged to replicate and improve the existing models

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Convolutional neural networks for intra-hour solar forecasting based on sky image sequences

Accurate and timely solar forecasts play an increasingly critical role in power systems. Compared to longer forecasting timescales, very short-term solar forecasting has lagged behind in both research and practice. In this paper, we propose deep convolutional neural networks (CNNs) to provide operational intra-hour (10-minute-ahead to 60-minute-ahead) solar forecasts. We develop two CNN structures inspired by a widely-used CNN architecture. The CNNs are tailored to our solar forecasting regression tasks and rely solely on sky image sequences. Case studies based on six years of data (over 150,000 data points) demonstrate that the best CNN model has forecast skill scores of 20%-39% over the naive persistence of cloudiness benchmark, even at these very short timescales. The CNNs also have consistently superior performance when compared to shallow machine learning models with meteorological predictors, where the improvement averages around 7%. The sensitivity analyses show that the sky image length, resolution, and weather conditions have impacts on the deep learning model accuracy. In our intra-hour problem with specific setups, two sky images with a 10-minute 128 x 128 resolution yield the most accurate forecasts. Current limitations, future work, and deployment challenges and solutions are also discussed.

14 SOLAR ENERGY↗
DOE OSTI2026

Systematic Evaluation of Atmospheric Forcing, Surface Datasets, and Mesh Effects on Kilometer-Scale Land Surface and River Modeling

Earth system models are advancing toward kilometer-scale resolution to capture local climate impacts and extremes. High-resolution land and river modeling depends on multiple factors, including mesh, surface datasets, and atmospheric forcing, but their relative effects at kilometer scales remain unquantified. We evaluated five Energy Exascale Earth System Model land and river configurations over the Mid-Atlantic region using two mesh (1/8° structured versus variable-resolution unstructured mesh), two surface datasets (default versus newly developed), and three atmospheric forcings (NLDAS2, MSWX, GSWP). Evaluation against satellite, reanalysis, and in situ benchmarks across water, energy, and carbon cycles quantifies how these factors affect model performance. Forcing selection produces the largest bias reductions (12-99% across variables), followed by surface datasets (7-75%) and mesh (up to 21%). Forcing effects vary by variable, with MSWX reducing biases for snow water equivalent, evapotranspiration, albedo, temperature, and gross primary productivity, GSWP for snow cover and runoff, and NLDAS for soil moisture and streamflow. The use of newly developed surface datasets improves gross primary productivity (58% bias reduction) and evapotranspiration but increase soil moisture and albedo biases due to current modeling limitations. Variable-resolution unstructured mesh improves the simulation of small-basin streamflow through better capturing drainage networks, though mesh minimally affects other land variables. These findings provide important guidance for high-resolution modeling development and actionable science.

Land and River modeling↗