Search NASA⌕ Search

SEARCH · Search NASA

Results for “Spatial load balancing”

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 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2017

Scalable and fast heterogeneous molecular simulation with predictive parallelization schemes

Multiscale and inhomogeneous molecular systems are challenging topics in the field of molecular simulation. In particular, modeling biological systems in the context of multiscale simulations and exploring material properties are driving a permanent development of new simulation methods and optimization algorithms. In computational terms, those methods require parallelization schemes that make a productive use of computational resources for each simulation and from its genesis. Here, we introduce the heterogeneous domain decomposition approach, which is a combination of an heterogeneity-sensitive spatial domain decomposition with an a priori rearrangement of subdomain walls. Within this approach and paper, the theoretical modeling and scaling laws for the force computation time are proposed and studied as a function of the number of particles and the spatial resolution ratio. We also show the new approach capabilities, by comparing it to both static domain decomposition algorithms and dynamic load-balancing schemes. Specifically, two representative molecular systems have been simulated and compared to the heterogeneous domain decomposition proposed in this work. Finally, these two systems comprise an adaptive resolution simulation of a biomolecule solvated in water and a phase-separated binary Lennard-Jones fluid.

36 MATERIALS SCIENCE↗
DOE OSTI2017

AMRZone: A Runtime AMR Data Sharing Framework For Scientific Applications

Frameworks that facilitate runtime data sharing across multiple applications are of great importance for scientific data analytics. Although existing frameworks work well over uniform mesh data, they can not effectively handle adaptive mesh refinement (AMR) data. Among the challenges to construct an AMR-capable framework include: (1) designing an architecture that facilitates online AMR data management; (2) achieving a load-balanced AMR data distribution for the data staging space at runtime; and (3) building an effective online index to support the unique spatial data retrieval requirements for AMR data. Towards addressing these challenges to support runtime AMR data sharing across scientific applications, we present the AMRZone framework. Experiments over real-world AMR datasets demonstrate AMRZone's effectiveness at achieving a balanced workload distribution, reading/writing large-scale datasets with thousands of parallel processes, and satisfying queries with spatial constraints. Moreover, AMRZone's performance and scalability are even comparable with existing state-of-the-art work when tested over uniform mesh data with up to 16384 cores; in the best case, our framework achieves a 46% performance improvement.

Zhang, Wenzhao↗
DOE OSTI2018

The Building ENergy Demand (BEND) Model: A Comprehensive Composite Building Energy Modeling Framework

The Building ENergy Demand (BEND) model was developed at Pacific Northwest National Laboratory (PNNL) to simulate climate-dependent hourly building energy demand with the ability to disaggregate down to or aggregate up to any geographic area of interest, including counties, states, electric utility service territories, and census regions. The BEND modeling framework is designed to study the effects of climate change, economics, technology advancement, and various mitigation policies on aggregate building energy demand. To make the process computationally feasible, BEND reduces granularity in building characteristics and weather forcing while maintaining enough detail to minimize systematic differences (biases) between BEND results and data from the Energy Information Administration (EIA). Based on the Commercial Building Energy Consumption Survey (CBECS) and Residential Energy Consumption Survey (RECS) databases, BEND establishes a set of representative buildings with predefined sets of building types, sizes, and vintages. These representative buildings are defined for each of the four U.S. census regions and the five climate zones used by the EIA for building surveys (hereafter “EIA climate zones”). BEND’s representative buildings are based on a statistical analysis of the CBECS and RECS databases that infers detailed parameters such as construction type, thermal properties, internal loads, equipment and mechanical systems, and operation schedules from the high-level descriptors in those databases (building type, vintage, size, and location). The representative buildings are simplified and consolidated into a set of prototype buildings for the purpose of simulating building energy demand. Combining the prototype buildings with weather files from selected locations and years, BEND creates a set of simulation-ready building models to be run using the EnergyPlus simulation engine (EnergyPlus). The base spatial unit of each BEND simulation is a subregion that is the intersection of an International Energy Conservation Code climate zone, a census region, and an EIA climate zone; analysis for each spatial unit is influenced by the weather file BEND selects to represent the climate-similar area. The major simulation outputs of interest to BEND are hourly energy consumption values for each simulated year, possibly for multiple fuel types. Based on building-stock weights derived from the CBECS/RECS data, the hourly output streams are aggregated to a single, 8760-hour load profile for each of the major building categories—residential and commercial—with the possibility of segregating by fuel types. To correct for biases inherent in the use of a limited suite of representative buildings, BEND is calibrated against observational targets of total annual building energy consumption and hourly load profiles (including peak demand) obtained for the decision-relevant geographic scales of interest (e.g., total electric load data for individual balancing authorities available from EIA). Based on available data such as population distribution, BEND is able to aggregate both up and down scales, allowing the study of building energy demand over a range of spatial scales (e.g., states or counties within the Western Electricity Coordinating Council, Eastern Interconnection, or the entire U.S.) and/or subsets of building types (e.g., office buildings with particular size ranges).

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2017

MPACT Development Activities at University of Michigan

This report documents the efforts of general development tasks in MPACT performed primarily at the University of Michigan during FY17. These tasks included: 1. Unstructured CMFD - The unstructured CMFD supports the RIA and BWR challenge problems. It allows MPACT to simulate mixed lattice cores (e.g. 8x8 next to 9x9). This design feature is common to operating BWRs and also the SPERT experiments being used to validate VERA-CS for RIA. 2. Higher order spatial sources for MOC - The higher order sources are for improving the overall performance of MPACT. Higher order spatial sources in MOC provide accurate solutions with potentially significantly fewer mesh regions, thus speeding up every calculation. This method is also particularly important for BWRs. 3. Improving fine source region visualization - The improvement of the visualization of the flat source region mesh in MPACT is to support users and communication of models and results from MPACT. VERAView provides visualization of pin-wise information, however MPACT resolves the spatial solution to intra pin level. The capability provided here will let users visualize the full model geometry and components of the solution at the highest spatial resolution available from MPACT. 4. Automated spatial decomposition - The automated spatial decomposition removes the burden from the user to know how to spatially decompose the model for MPACT; a task that has been rather tricky in the past. Instead the user may give some number of processors that is a multiple of the axial mesh and MPACT will determine a nearly optimal decomposition that gives better load balance than most decompositions used currently. 5. Optimal relaxation factors for feedback - The development of optimal relaxation factor is to improve the stability and robustness of problems with feedback. Presently the default feedback values are 0.5, which may or may not be optimal. Without relaxation the fixed point iteration that is currently used is known to be unstable for large problems. The optimal relaxation factors are derived by utilizing a previous Fourier analysis. From the original milestone there was a scope change where the work on the Intel MIC was deferred for work on the automated spatial decomposition.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2025

Deployment Potential of Concentrating Solar Power Technologies in California

As states within the United States respond to future grid development goals, there is a growing demand for reliable and resilient nighttime generation that can be addressed by low-cost, long-duration energy storage solutions. This report studies the potential of including concentrating solar power (CSP) in the technology mix to support California’s goals as defined in Senate Bill 100. A joint agency report study that determined potential pathways to achieve the renewable portfolio standard set by the bill did not include CSP, and our work provides information that could be used as a follow-up. This study uses a capacity expansion model configured to have nodal spatial fidelity in California and balancing-area fidelity in the Western Interconnection outside of California. The authors discovered that by applying current technology cost projections CSP fulfills nearly 15% of the annual load while representing just 6% of total installed capacity in 2045, replacing approximately 30 GWe of wind, solar PV, and standalone batteries compared to a scenario without CSP included. The deployment of CSP in the results is sensitive to the technology’s cost, which highlights the importance of meeting cost targets in 2030 and beyond to enable the technology’s potential contribution to California’s carbon reduction goals.

14 SOLAR ENERGY↗
DOE OSTI2017

Flexible Spatial Partitions in MPACT Through Module-Based Data Passing

As part of the Consortium for Advanced Simulation of Light Water Reactors (CASL), the Virtual Environment for Reactor Applications (VERA) is being developed to provide high-fidelity multiphysics simulations of nuclear reactor cores. The MPACT code being developed collaboratively by Oak Ridge National Laboratory and the University of Michigan is the primary deterministic neutron transport solver available within VERA. MPACT employs the 2D/1D method to solve 3D problems using 2D method of characteristics (MOC) radial transport solvers with 1D nodal expansion method-based P3 (NEM-P3) axial transport solvers, which are coupled using axial and radial transverse leakages. A 3D coarse mesh finite difference (CMFD) solver is used to provide accelerated convergence and stability. Quarter-core problems in MPACT are typically run on moderately sized clusters ({approx}500-10,000 cores). As such, it relies heavily on spatial decomposition, both radially and axially, using MPI data communication along spatial domain boundaries to pass angular flux boundary condition and neighboring data. The initial spatial decomposition strategy implemented into MPACT overlaid a Cartesian grid over the core to form the domain boundaries. With this approach, a single buffer spanning across the entire boundary between two spatial domains is passed to a domain's neighbor. Additionally, only one neighboring domain index was allowed in each direction. For example, the entire eastern boundary of a domain must be the entire western boundary of the corresponding neighbour's domain. This implementation helped to execute cases on several thousand cores, where substantial radial decomposition was used. The default for several years was to use assembly decomposition, which yielded 73 radial partitions in the Watts Bar Unit 1 (WBN1) models being analyzed. However, as CASL formed stronger connections with industry partners, there was a strong push to allow MPACT to run on fewer cores, preferably in the 500-1,000 core range. Such cases require much less radial decomposition, typically with 8-16 radial domains on each plane. Unfortunately, the previous decomposition strategy is very limited in its ability to yield well-balanced domains, and it does not permit partitions with an arbitrary number of domains, particularly with small numbers of radial domains (3, 5, 7, 11, 13, etc.). With these limitations, a more flexible approach that provided more even-load balancing was needed. The details of this new strategy are outlined here, along with insights to its advantages. The examples provided are for the WBN1 core, which has been the focus of several initial CASL analyses. It is worth noting that spatial decomposition strategies have been employed in many codes with focus on a number of different transport methodologies. The references cited are not exhaustive as this has been very important research topic across many areas.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2017

Galactos: computing the anisotropic 3-point correlation function for 2 billion galaxies

The nature of dark energy and the complete theory of gravity are two central questions currently facing cosmology. A vital tool for addressing them is the 3-point correlation function (3PCF), which probes deviations from a spatially random distribution of galaxies. However, the 3PCF's formidable computational expense has prevented its application to astronomical surveys comprising millions to billions of galaxies. We present Galactos, a high-performance implementation of a novel, O( N 2 ) algorithm that uses a load-balanced k-d tree and spherical harmonic expansions to compute the anisotropic 3PCF. Our implementation is optimized for the Intel Xeon Phi architecture, exploiting SIMD parallelism, instruction and thread concurrency, and significant L1 and L2 cache reuse, reaching 39% of peak performance on a single node. Galactos scales to the full Cori system, achieving 9.8 PF (peak) and 5.06 PF (sustained) across 9636 nodes, making the 3PCF easily computable for all galaxies in the observable universe.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2026

Knowledge-guided graph machine learning for spatially distributed prediction of daily discharge and nitrogen export dynamics

Spatially distributed prediction of streamflow and nitrogen export dynamics is essential for precision management of agricultural watersheds. While temporal deep learning models such as Long Short-Term Memory (LSTM) have shown strong performance at basin scales, their ability to generalize spatially is limited by insufficient representation of spatial dependencies and flow paths, particularly under data-scarce conditions. To address this gap, we propose HydroGraphNet, a knowledge-guided graph machine learning framework that integrates process-based knowledge and explicit spatial learning into temporal modeling. This framework incorporates directed graph topology to encode watershed connectivity and upstream inflows, with mass balance constraints to improve physical consistency. To enhance generalization in sparsely monitored regions, HydroGraphNet is pretrained on synthetic data generated by the SWAT+ (Soil and Water Assessment Tool Plus) model. We evaluated HydroGraphNet in the Upper Sangamon River Basin (44 HUC-12 subwatersheds, 2001–2020) against two LSTM baselines: a lumped basin-level model and a distributed variant. When benchmarked on SWAT+ simulations in pretraining, HydroGraphNet improved test NSEs by 8.9% (discharge) and 13.7% (NO₃–N load) in temporal extrapolation, and by 27.1% and 34.7% in spatial extrapolation, relative to the Lumped LSTM baseline. After fine-tuning with USGS monitoring data, the model achieved mean test NSE (KGE) scores of 0.768 (0.861) for discharge and 0.626 (0.664) for NO₃–N load, substantially outperforming baselines. Attribution analysis further highlighted the importance of upstream inflow representation and graph-based spatial learning in capturing cross-subwatershed dependencies. The model also reproduced seasonal hydrological and biogeochemical patterns consistent with known processes, demonstrating its robustness and process fidelity for spatially distributed prediction. Altogether, HydroGraphNet advances the integration of physical knowledge and spatially explicit learning in hydrological modeling, offering a generalizable framework for distributed modeling to support spatially targeted water quality management in data-scarce watersheds.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2019

Performance analysis and optimization for scalable deployment of deep learning models for country-scale settlement mapping on Titan supercomputer

Here, we present a scalable object detection workflow for detecting objects, such as settlements, from remotely sensed (RS) imagery. We have successfully deployed this workflow on Titan supercomputer and utilized it for the task of mapping human settlement at a country scale. The performance of various stages in the workflow was analyzed before making it operational. The workflow implemented various strategies to address issues such as suboptimal resource utilization and long-tail effects due to unbalanced image workload, data loss due to runtime failures, and maximum wall-time constraints imposed by Titan's job scheduling policy. A mean shift clustering–based static load balancing strategy was implemented, which partitions the image load such that each partition contained similar-sized images. Furthermore, a checkpoint-restart strategy was added in the workflow as a fault-tolerance mechanism to prevent the data losses due to unforeseen runtime failures. The performance of the above-mentioned strategies was observed in various scenarios, such as node failure, exceeding wall time, and successful completion. Using this workflow, we have examined an RS data set that has a spatial resolution of 0.31 m and is comprised of 685 675 km 2 of area of the Republic of Zambia in under six hours using 5426 nodes of the Titan supercomputer.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2017

Numerical investigation of plasma edge transport and limiter heat fluxes in Wendelstein 7-X startup plasmas with EMC3-EIRENE

Here the results of a first systematic assessment of plasma edge transport processes for the limiter startup configuration at Wendelstein 7-X are presented. This includes an investigation of transport from intrinsic and externally injected impurities and their impact on the power balance and limiter heat fluxes. The fully 3D coupled plasma fluid and kinetic neutral transport Monte Carlo code EMC3-EIRENE is used. The analysis of the magnetic topology shows that the poloidally and toroidally localized limiters cause a 3D helical scrape-off layer (SOL) consisting of magnetic flux tubes of three different connection lengths L C . The transport in the helical SOL is governed by L C as topological scale length for the parallel plasma loss channel to the limiters. A clear modulation of the plasma pressure with L C is seen. The helical flux tube topology results in counter streaming sonic plasma flows. The heterogeneous SOL plasma structure yields an uneven limiter heat load distribution with localized peaking. Assuming spatially constant anomalous transport coefficients, increasing plasma density yields a reduction of the maximum peak heat loads from 12 MWm -2 to 7.5 MWm -2 and a broadening of the deposited heat fluxes. The impact of impurities on the limiter heat loads is studied by assuming intrinsic carbon impurities eroded from the limiter surfaces with a gross chemical sputtering yield of 2 . The resulting radiative losses account for less than 10% of the input power in the power balance with marginal impact on the limiter heat loads. It is shown that a significant mitigation of peak heat loads, 40–50%, can be achieved with controlled impurity seeding with nitrogen and neon, which is a method of particular interest for the later island divertor phase.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2022

A Micromorphic Length-Scale Coupling Framework for the Determination of Higher-Order Constitutive Models and the Multi-Scale Simulation of Heterogeneous Materials [Thesis]

Heterogeneous materials and materials with complex microstructures pose a unique challenge in the development of accurate models of their response to external stimuli. These difficulties principally arise due to the difficulty in characterizing and modeling the constituents and their interactions. It is usually possible, though non-trivial, to construct an explicit representation of the microstructure (a direct numeric simulation or DNS) but the method by which the complex modes of deformation and other processes can be homogenized to a reduced-order approximation is frequently unclear. Many homogenization approaches are ad-hoc and lack a strong justification beyond ease of computation. Furthermore, the homogenization approach can, in some cases, not utilize the full breadth of information available in the computation of the macro-scale stresses and deformation measures. It is also noteworthy that homogenization, by its very nature, will tend to obfuscate details of processes occurring at the lower length-scale. It is therefore of interest to include as much information as possible in the construction of the reduced-order model so as to be predictive in a variety of loading environments. We here present a length-scale bridging technique based upon the micromorphic continuum mechanics of Eringen which incorporates volume and surface area averages as a part of its construction. This approach enforces the balance equations at the micro-scale and then studies the effect of the spatially varying nature on the macro-scale. This leads, naturally, to further balance equations which are solved at the macro-scale. This work details a homogenization framework which arises naturally from the micromorphic construction of Eringen attempting to introduce no definitions beyond which are justifiable from micro-structural considerations. One of the results of this effort is the so-called “micromorphic filter” which has been developed and applied to several DNS to demonstrate its effectively. In order to determine the macroscopic degrees of freedom we utilize the special case of an overlap coupling technique where the macro-scale is fully constrained to the micro-scale. This enables us to study the resulting material properties as expected but also allows us to further study the boundary conditions on the additional degrees of freedom at the macro-scale.

42 ENGINEERING↗
DOE OSTI2017

Modeling plug-in electric vehicle charging demand with BEAM: the framework for behavior energy autonomy mobility

This report summarizes the BEAM modeling framework (Behavior, Energy, Mobility, and Autonomy) and its application to simulating plug-in electric vehicle (PEV) mobility, energy consumption, and spatiotemporal charging demand. BEAM is an agent-based model of PEV mobility and charging behavior designed as an extension to MATSim (the Multi-Agent Transportation Simulation model). We apply BEAM to the San Francisco Bay Area and conduct a preliminary calibration and validation of its prediction of charging load based on observed charging infrastructure utilization for the region in 2016. We then explore the impact of a variety of common modeling assumptions in the literature regarding charging infrastructure availability and driver behavior. We find that accurately reproducing observed charging patterns requires an explicit representation of spatially disaggregated charging infrastructure as well as a more nuanced model of the decision to charge that balances tradeoffs people make with regards to time, cost, convenience, and range anxiety.

33 ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2024

Super Resolving Unrolled Neural Networks for Remote Sensing

In remote sensing systems, the capabilities of the system are constrained by the complex interactions between size, weight, and power (SWAP) of potential designs. In electro-optical (EO) systems, examples of these critical parameters include the system’s sensitivity and resolution. Those parameters can be increased by ever larger optical apertures and focal planes but at the cost of more SWAP. Multi-image super resolution (MISR) techniques allow resolution to be enhanced via computation rather than more sophisticated optical hardware. These algorithms combine multiple images together into a single, higher resolution image, trading temporal resolution and computation for spatial resolution. Fielded MISR techniques, such as Drizzle, can require several hundred images to create a single super resolved image, implying reduced temporal resolution, increased data acquisition load, and limiting mission applications. Iterative techniques, such as model-based image reconstruction and compressive sensing, have been shown to create super resolved images using fewer images than Drizzle. They do this by posing an optimization problem that balances accuracy between a highly accurate physical model and an image model. In the case of super resolution, the physical model is defined by the relation between low resolution input images and the desired high resolution output image. The image model encodes some assumptions about the super resolved image. These assumptions are meant to suppress reconstruction artifacts that arise due to deterministic physical model error, stochastic measurement noise, and potential undersampling. In practice, the performance of iterative methods are limited by imaging models compatible with optimization. Deep learning-based methods can effectively learn image models of arbitrary complexity, but lack the theoretical explainability and robustness of iterative techniques. Consensus equilibrium (CE) generalizes the iterative techniques beyond optimization, enabling blackbox algorithms such as traditional and neural image denoisers to be used as the image model. CE-based approaches retain much of the explainability and robustness of iterative techniques while allowing the expressiveness of machine learning image models to be used. Additionally, by unrolling iterations of CE with an embedded image denoiser, the image denoiser can be further trained and specialized to the specific application with potentially higher quality reconstructions. Under this project, we demonstrated the feasibility of training an unrolled neural network based upon CE. While we didn’t train one, we showed that the CE process is differentiable and its gradient can be tractably computed. We also explored the usage of a variants of CE akin to generative neural works. Most importantly, we applied the CE framework to a number of problems including non-blind deconvolution, upsampling, single-image super resolution, MISR, event-based sensing, and saturated deconvolution. Our MISR prototype creates high quality reconstructions with an order of magnitude fewer images than previous approaches and, critically, produces these reconstructions fast enough for practical usage.

47 OTHER INSTRUMENTATION↗
DOE OSTI2025

ASEAN Technical Exchange Workshop for System Operators, Regulators, and Policymakers

This presentation provides an in-depth exploration of power system planning, cross-border electricity trading, and battery energy storage systems (BESS), offering actionable insights for system operators, regulators, and policymakers. The first section delves into power system planning and analysis, focusing on capacity expansion models and resource adequacy studies, including their role in optimizing system efficiency, managing emissions, and addressing system reliability risks. Key considerations, such as integration of transmission into generation planning and the forecasting versus optimization of customer distributed energy resources (DER) technologies, are explored. The session highlights critical trade-offs in spatial granularity and model runtimes, as well as the feasibility of aligning distribution investments with capacity expansion efforts. The second section examines cross-border electricity trading, with an emphasis on resource adequacy concepts such as reliability targets, loss of load expectation (LOLE), and planning reserve margins (PRM). Case studies on reserve market design and coordination across US regions provide insights into improving reserve deliverability and managing interregional power balance and congestion. This section also addresses market-to-market congestion management, including advanced strategies for high-voltage direct current (HVDC) optimization and ancillary service delivery. Finally, the presentation covers the rapid evolution of Battery Energy Storage Systems (BESS), highlighting their operational growth, regulatory frameworks, and use cases in grid flexibility, energy storage, and reliability. The discussion focuses on the benefits of BESS for system stability, resilience, and integration of renewable energy, offering insights into its role as a vital component in the transition toward a more sustainable and flexible grid. Key performance parameters, such as throughput, round-trip efficiency, and state of charge, are also examined.

25 ENERGY STORAGE↗
DOE OSTI2022

Effect of dust on rainfall over the Red Sea coast based on WRF-Chem model simulations

Water is the single most important element of life. Rainfall plays an important role in the spatial and temporal distribution of this precious natural resource, and it has a direct impact on agricultural production, daily life activities, and human health. One of the important elements that govern rainfall formation and distribution is atmospheric aerosol, which also affects the Earth's radiation balance and climate. Therefore, understanding how dust compositions and distributions affect the regional rainfall pattern is crucial, particularly in regions with high atmospheric dust loads such as the Middle East. Although aerosol and rainfall research has garnered increasing attention as both an independent and interdisciplinary topic in the last few decades, the details of various direct and indirect pathways by which dust affects rainfall are not yet fully understood. Here, we explored the effects of dust on rainfall formation and distribution as well as the physical mechanisms that govern these phenomena, using high-resolution WRF-Chem simulations (~1.5 km × 1.5 km) configured with an advanced double-moment cloud microphysics scheme coupled with a sectional eight-bin aerosol scheme. Our model-simulated results were realistic, as evaluated from multiple perspectives including vertical profiles of aerosol concentrations, aerosol size distributions, vertical profiles of air temperature, diurnal wind cycles, and spatio-temporal rainfall patterns. Rainfall over the Red Sea coast is mainly caused by warm rain processes, which are typically confined within a height of ~6 km over the Sarawat mountains and exhibit a strong diurnal cycle that peaks in the evening at approximately 18:00 local time under the influence of sea breezes. Numerical experiments indicated that dust could both suppress or enhance rainfall. The effect of dust on rainfall was calculated as total, indirect, and direct effects, based on 10-year August-average daily-accumulated rainfall over the study domain covering the eastern Red Sea coast. For extreme rainfall events (domain-average daily-accumulated rainfall of ≥ 1.33 mm), the net effect of dust on rainfall was positive or enhancement (6.05 %), with the indirect effect (4.54 %) and direct effect (1.51 %) both causing rainfall increase. At a 5 % significance level, the total and indirect effects were statistically significant whereas the direct effect was not. For normal rainfall events (domain-average daily-accumulated rainfall < 1.33 mm), the indirect effect enhanced rainfall (4.76 %) whereas the direct effect suppressed rainfall (-5.78 %), resulting in a negative net suppressing effect (-1.02 %), all of which were statistically significant. We investigated the possible physical mechanisms of the effects and found that the rainfall suppression by dust direct effects was mainly caused by the scattering of solar radiation by dust. The surface cooling induced by dust weakens the sea breeze circulation, which decreases the associated landward moisture transport, ultimately suppressing rainfall. For extreme rainfall events, dust causes net rainfall enhancement through indirect effects as the high dust concentration facilitates raindrops to grow when the water vapor is sufficiently available. Our results have broader scientific and environmental implications. Specifically, although dust is considered a problem from an air quality perspective, our results highlight the important role of dust on sea breeze circulation and associated rainfall over the Red Sea coastal regions. Our results also have implications for cloud seeding and water resource management.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Scalable multiscale modeling of platelets with 100 million particles

Here, we developed the core components of the AI-aided multiple time stepping algorithm for multiscale modeling of cell dynamics. This algorithm was implemented and analyzed on two supercomputer architectures with an application of simulating the aggregation of 250 platelets, or 102 million particles. To scale on these computers with complex memory and network architectures with GPUs, we devised a biomechanics-informed task mapping scheme to optimize load imbalance, communications, and memory utilization. Our simulations, scaling well up to 192 nodes on a Summit-like supercomputer with a peak speed of 11 petaflops, achieved a rate of 423 μs/day which is 500 times faster than the conventional algorithm using static time step and this has enabled studies of record size blood clots at record spatial–temporal resolutions. Additionally, we discovered the sensitive dependence of the scalability and execution time on the methods of decomposition, CPU–GPU coupling, and task mapping.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Upper troposphere dust belt formation processes vary seasonally and spatially in the Northern Hemisphere

Dust aerosols impact global energy balance substantially by acting as efficient ice nuclei to alter cold cloud properties. However, the estimate of dust indirect effect remains uncertain due to simulating dust distributions poorly and lacking reliable dust observations, especially in the upper-troposphere. Here, we characterize and understand upper-troposphere dust sources and transport with an improved dust dataset derived from A-train satellite lidar and radar measurements and an air parcel trajectory model. The distinct upper-troposphere dust belt over the northern hemisphere has seasonally varying base and top heights of 3.65 ± 2.84 and 8.35 ± 1.50 km above mean sea level and its column loading is strongest during spring (March-April-May). The out-of-phase annual cycles of mid-level dust concentration and westerly wind over source regions control the seasonal upper-tropospheric dust loading variations. African deserts contribute the most (46.3%) to the upper-troposphere dust belt in spring and the synoptic trough is the leading (49%) dust lifting mechanism.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

A cross-dimensional analysis of data-driven short-term load forecasting methods with large-scale smart meter data

Electricity load forecasting is essential to utility operation and power grid stability. A wide spectrum of data-driven methods, ranging from linear regression models to more recent deep learning models have been adopted to forecast electric load over the years. However, there still lacks a holistic evaluation of the applicability of conventional statistical and machine learning based algorithms with respect to different temporal and spatial scopes, computational requirements, and sensitivity of model-tuning. Enabled by a large-scale electricity load profile dataset of over 40,000 residential customers in a utility region, we conducted a cross-dimensional analysis of data-driven load forecasting methods. Three regression-based and seven deep learning algorithms with different model configurations were evaluated in terms of their overall and peak load prediction accuracy, and training burdens, across spatial aggregation levels ranging from the transformer, feeder, substation, to neighborhood. We found, first, the load forecasting accuracy is constrained by a predictability boundary, influenced by the forecasting horizon and spatial aggregation level. Specifically, RandomForest, XGBoost, TFT, TSMixer, and TiDE models achieved less than 10 % prediction error for up to 96-h ahead forecasting for district, substation, and feeder levels, while other models struggle at long-horizon predictions; Second, for winter and summer peak load dates, most models were able to predict the peak demand timing within ± 1 h, but the prediction percentage error varied by models, with TFT and TiDE models being the top performers; Third, models with similar prediction accuracy can differ in training burden by an order of magnitude. Therefore, choosing model configurations that balance prediction performance and computational resource is an important practical consideration for large-scale deployment of the machine learning based load forecasting. The outcome of this study can guide researchers and practitioners to choose the proper load forecasting algorithms based on their problem scope, required accuracy, and available resources. The predictability boundary can serve as a benchmark for electricity load forecasting problems with new algorithms and datasets.

Li, Han↗