Results for “Acceleration Data” — Page 28

DOE OSTI2020

Full-Scale Assembly 30 cm Drop Test

Can Spent Nuclear Fuel withstand the shocks and vibrations experienced during normal conditions of transport? This question was the motivation for the multi-modal transportation test (MMTT) (Summer 2017), 1/3-scale cask 30 cm drop test (December 2018), and full-scale assembly 30 cm drop tests (June 2019). The full-scale ENSA ENUN 32P cask with 3 surrogate 17x17 PWR assemblies was used in the MMTT. The 1/3-scale cask was a mockup of this cask. The 30 cm drop tests provided the accelerations on the 1/3-scale dummy assemblies. These data were used to design full-scale assembly drop tests with the goal to quantify the strain fuel rods experience inside a cask when dropped from a height of 30 cm. The drop tests were first done with the dummy and then with the surrogate assembly. This paper presents the preliminary results of the tests.

Transportation↗

DOE OSTI2020

Pumped Storage Hydropower FAST Commissioning Technical Analysis

This report is designed to address barriers and solutions to modern pumped storage hydropower (PSH) development by establishing baseline project development knowledge, defining key aspects of project development, and identifying opportunities to reduce project timelines, costs, and risks. This report’s scope includes post-licensing activities and excludes factors related to permitting or licensing. The U.S. PSH fleet is composed of 43 projects providing the majority (95%) of utility-scale electricity storage in the US. However, only one new PSH facility has become operational in the past 20 years. Several factors contribute to diminishing PSH growth in the US, including the magnitude of project costs and financing interest during development and construction; the length of time from project investment until project revenue; permitting challenges and construction risks; competition from other storage technologies; and unrecognized energy storage valuation. Although innovative PSH concepts (including underground, small, and modular systems) have been investigated, widespread application has yet to occur. In short, the time, cost, and risk associated with modern PSH development has resulted in limited recent growth in the United States, despite the rising energy storage demand from increased deployment of variable renewable technologies. To address these challenges, the US Department of Energy’s (DOE) Water Power Technologies Office initiated the PSH Furthering Advancements to Shorten Time to (FAST) Commissioning project, aimed at catalyzing new solutions, designs, and strategies to accelerate PSH development. This report uses available data from previous license applications, ongoing project cost data, and other global PSH project information based on a typical closed-loop PSH project.

13 HYDRO ENERGY↗

DOE OSTI2021

Online Monitoring System for Concrete Structures Affected by Alkali-Silica Reaction

This report presents a comprehensive study of developing ultrasonic wave and acoustic emission (AE) techniques for long-term monitoring of alkali-silica reaction (ASR) development in concrete. Small, medium, and full-scale concrete specimens were cast, conditioned, and monitored for various periods, from at least one year to 2.4 years. The reactive concrete specimens contain coarse or fine reactive aggregates to study the effects of different types of reactive aggregates. Confinements were also designed to simulate the 2-dimensional confinement effects of reinforcement in the shield building of nuclear power plants. The concrete specimens were stored in a environmental chamber with high humidity and high temperature to accelerate the ASR development. The ultrasonic monitoring data shows high sensitivity to ASR development and could detect cracking initiation well before visible surface cracks occurred. However, the linear ultrasonic analysis based on wave velocity is strongly affected by temperature variation. Therefore, a nonlinear ultrasonic method was proposed to measure thermally induced nonlinear acoustic responses of concrete (thermal modulation of ultrasonic wave). The measured nonlinear acoustic parameter shows a high correlation with ASR expansion across specimens with different reactive aggregates and confinement conditions. The same conclusion was obtained from nonlinear resonance tests on small concrete prisms. Compared to the linear acoustic methods, the nonlinear tests show high sensitivities to ASR damage from internal microcracking initiation at the early stage to visible cracks at the late stage of ASR. The attributes of the thermal modulation of nonlinear ultrasonic method include high sensitivity, immunity to temperature effects, and strong correlation with ASR expansion, which demonstrate great potentials of the nonlinear ultrasonic method for diagnosis of ASR damage and prediction of concrete deterioration process. Acoustic emission is a passive sensing technique for damage assessment, and access to only one surface is needed even for thick and heavily reinforced elements such as the walls utilized for nuclear shield building. Additionally, relatively few sensors are required to monitor the progression of the damage process. Results indicate that the AE data can be related to the damage rating index, which is a petrography-based means of assessing damage due to ASR in reinforced concrete. Furthermore, AE is capable of detecting ASR damage long before surface cracking is visually noticeable. Boundary conditions play in important role in the progression of ASR damage and differences in boundary conditions are reflected in the AE data. Entropy based data assessment methods provide a means to assess the damage state, and convolutional neural network based data assessment provided a means to assess the damage state in real-time. Results indicate that artificial neural network models may be used as a means to predict volumetric expansion based on AE data.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

DOE OSTI2024

Machine Learning for Joint Quality Control

The use of lightweight material combinations has been highly demanded in manufacturing automotive structures. However, making robust dissimilar material joints of such lightweight materials is still challenging. A significant barrier to achieving high-quality and repeatable joint performance is a deficient understanding of the relationship between the welding process, joint attributes, and joint performance. In this context, welding factors refer to material, equipment, environment, and process parameters, while joint features comprise specific microstructural attributes of the weld such as nugget size, heat affected zone (HAZ) topology, intermetallic layer thickness, and sheet thickness reduction. Joint performance is quantified in terms of strength (e.g., tensile shear, coach peel, cross-tension), weld size, and hardness, among other factors. While there have been many attempts to establish this process-structure-property relationship by developing a model derived from the associated physics and first principles, the complexity of the joining processes compounded by the complex interactions with different materials in an automotive assembly line environment, has hindered the usefulness of such attempts. The complexity is further exacerbated using different stacking materials, especially comprising dissimilar material combinations. In practice, the common approach has been the laborious process of creating welds, characterizing them, and then physically testing them through experimentation. With the emergence of artificial intelligence (AI) methods, an alternative pathway to eliciting the desired process-structure-property relationship at an accelerated pace is to use a data-driven approach by employing machine-learning (ML) techniques. This approach is benefitted by the availability of large streams of data, generated through years of research and testing by original equipment manufacturers, in the form of material, process, environmental, equipment, microstructural, and bulk-scale performance information from multimodal, multiscale sensors making measurements from laboratory-scale to production-scale processes. During Phase I efforts, which ended in fiscal year (FY) 2021, the Oak Ridge National Laboratory and Pacific Northwest National Laboratory (ORNL/PNNL) team demonstrated the effectiveness of different ML/AI frameworks in modeling complex relationships between resistance spot welding (RSW) process parameters, weld attributes, and joint properties using a subset of data from General Motors (GM). In FY 2022, the project team further refined and expanded their respective ML models to analyze additional welds with new weld stack-ups and materials to enhance the ML model predictive capability. ORNL extended its unified deep neural networks (DNN) ML training and prediction framework with new data streams of process parameters, and PNNL extended its model describing RSW process parameters’ associations with weld attributes. In FY 2023, the project team completed the development of the AI/ML architecture for analyzing aluminum/steel joints manufactured by GM via RSW and transitioned into the inline welding quality monitoring task for steel/steel RSW joints provided by GM.

36 MATERIALS SCIENCE↗

DOE OSTI2022

Performance Testing of a Moving Bed Gasifier Using Coal, Biomass, and Waste Plastic Blends to Generate White Hydrogen

The objective of this DOE-funded project by the Electric Power Research Institute, Inc. (EPRI), Hamilton Maurer International (HMI) and Sotacarbo S.p.A. (Sotacarbo), is to qualify coal, biomass, and plastic waste blends based on performance testing of selected pellet recipes in a pilot-scale updraft moving-bed gasifier. The testing provides relevant data to advance the commercial-scale design of the moving-bed gasifier to be able to successfully use these feedstocks to produce hydrogen. In particular, the effects of the waste plastics on feedstock development (i.e., blending and pelletizing) and the resulting products (i.e., syngas compositions, organic condensate production, and ash characteristics) are a focus. The gasifier is Hamilton Mauer International, Inc. (HMI)’s moving-bed gasifier, which has been proven capable of gasifying nearly all coal ranks. It has also shown the ability in prior testing work to gasify wood chips. However, mixtures of these fuels with plastic wastes had not yet been prepared and gasified together. The feedstocks were prepared by California Pellet Mill (CPM) under contract to HMI. The technical tasks for this research project included: • Feed Procurement and Preparation: Nine different feedstocks were prepared from varying compositions of PRB coal, corn stover biomass, and car fluff waste plastics. Fuel pellets were produced by California Pellet Mill (CPM) and shipped to Sotacarbo’s test facility in Italy. • Test Plan Development: A test plan was created to define the test runs to be performed. The test plan detailed the different tests that were run, instrumentation used, extractive samples taken, and relevant figures of merit. • Gasifier Testing: Tests were performed in the pilot-scale gasifier at Sotacarbo using nine different fuel feedstocks generated from varying mixtures of coal, biomass, and plastic wastes. The testing provides information on gasification reaction efficiency/performance, yielding relevant data for models used to scale up the gasifier design. This task will also included work to reassemble the gasifier at Sotacarbo and perform a baseline 100% coal run. • Data Analysis and Reporting: Review of the data, determination of figures of merit, and interpretation of the results was reported. The results can be used to specify the range of feedstock blends that can be successfully gasified as well as quantify gasifier outputs based on specific blends. The tri-fuel pelletizing conducted at CPM and gasification testing results from Sotacarbo’s 30mm up draft moving bed gasifier are significant. Providing data for an established gasifier to help accelerate its updated design to be able to accommodate feedstocks composed of coal, biomass, and plastic waste. This should ultimately lead to development and commercialization of a lower cost, white hydrogen generation system.

01 COAL, LIGNITE, AND PEAT↗

DOE OSTI2023

Performance Testing of a Moving-Bed Gasifier Using Coal, Biomass, and Waste Plastic Blends to Generate White Hydrogen

The objective of this DOE-funded project by the Electric Power Research Institute, Inc. (EPRI), Hamilton Maurer International (HMI) and Sotacarbo S.p.A. (Sotacarbo), is to qualify coal, biomass, and plastic waste blends based on performance testing of selected pellet recipes in a pilot-scale updraft moving-bed gasifier. The testing provides relevant data to advance the commercial-scale design of the moving-bed gasifier to be able to successfully use these feedstocks to produce hydrogen. In particular, the effects of the waste plastics on feedstock development (i.e., blending and pelletizing) and the resulting products (i.e., syngas compositions, organic condensate production, and ash characteristics) are a focus. The gasifier is Hamilton Mauer International, Inc. (HMI)’s moving-bed gasifier, which has been proven capable of gasifying nearly all coal ranks. It has also shown the ability in prior testing work to gasify wood chips. However, mixtures of these fuels with plastic wastes had not yet been prepared and gasified together. The feedstocks were prepared by California Pellet Mill (CPM) under contract to HMI. The technical tasks for this research project included: • Feed Procurement and Preparation: Nine different feedstocks were prepared from varying compositions of PRB coal, corn stover biomass, and car fluff waste plastics. Fuel pellets were produced by California Pellet Mill (CPM) and shipped to Sotacarbo’s test facility in Italy. • Test Plan Development: A test plan was created to define the test runs to be performed. The test plan detailed the different tests that were run, instrumentation used, extractive samples taken, and relevant figures of merit. • Gasifier Testing: Tests were performed in the pilot-scale gasifier at Sotacarbo using nine different fuel feedstocks generated from varying mixtures of coal, biomass, and plastic wastes. The testing provides information on gasification reaction efficiency/performance, yielding relevant data for models used to scale up the gasifier design. This task will also included work to reassemble the gasifier at Sotacarbo and perform a baseline 100% coal run. • Data Analysis and Reporting: Review of the data, determination of figures of merit, and interpretation of the results was reported. The results can be used to specify the range of feedstock blends that can be successfully gasified as well as quantify gasifier outputs based on specific blends. The tri-fuel pelletizing conducted at CPM and gasification testing results from Sotacarbo’s 30mm up draft moving bed gasifier are significant. Providing data for an established gasifier to help accelerate its updated design to be able to accommodate feedstocks composed of coal, biomass, and plastic waste. This should ultimately lead to development and commercialization of a lower cost, white hydrogen generation system.

01 COAL, LIGNITE, AND PEAT↗

DOE OSTI2021

Gasification of Mixed Blends of Coal, Biomass, and Plastic Waste

The objective of this DOE-funded project by the Electric Power Research Institute, Inc. (EPRI), Hamilton Maurer International (HMI) and Sotacarbo S.p.A. (Sotacarbo), is to qualify coal, biomass, and plastic waste blends based on performance testing of selected pellet recipes in a pilot-scale updraft moving-bed gasifier. The testing provides relevant data to advance the commercial-scale design of the moving-bed gasifier to be able to successfully use these feedstocks to produce hydrogen. In particular, the effects of the waste plastics on feedstock development (i.e., blending and pelletizing) and the resulting products (i.e., syngas compositions, organic condensate production, and ash characteristics) are a focus. The gasifier is Hamilton Mauer International, Inc. (HMI)’s moving-bed gasifier, which has been proven capable of gasifying nearly all coal ranks. It has also shown the ability in prior testing work to gasify wood chips. However, mixtures of these fuels with plastic wastes had not yet been prepared and gasified together. The feedstocks were prepared by California Pellet Mill (CPM) under contract to HMI. The technical tasks for this research project included: • Feed Procurement and Preparation: Nine different feedstocks were prepared from varying compositions of PRB coal, corn stover biomass, and car fluff waste plastics. Fuel pellets were produced by California Pellet Mill (CPM) and shipped to Sotacarbo’s test facility in Italy. • Test Plan Development: A test plan was created to define the test runs to be performed. The test plan detailed the different tests that were run, instrumentation used, extractive samples taken, and relevant figures of merit. • Gasifier Testing: Tests were performed in the pilot-scale gasifier at Sotacarbo using nine different fuel feedstocks generated from varying mixtures of coal, biomass, and plastic wastes. The testing provides information on gasification reaction efficiency/performance, yielding relevant data for models used to scale up the gasifier design. This task will also included work to reassemble the gasifier at Sotacarbo and perform a baseline 100% coal run. • Data Analysis and Reporting: Review of the data, determination of figures of merit, and interpretation of the results was reported. The results can be used to specify the range of feedstock blends that can be successfully gasified as well as quantify gasifier outputs based on specific blends. The tri-fuel pelletizing conducted at CPM and gasification testing results from Sotacarbo’s 30mm up draft moving bed gasifier are significant. Providing data for an established gasifier to help accelerate its updated design to be able to accommodate feedstocks composed of coal, biomass, and plastic waste. This should ultimately lead to development and commercialization of a lower cost, white hydrogen generation system.

01 COAL, LIGNITE, AND PEAT↗

DOE OSTI2026

Integrated Direct Air Capture and H₂-Free CO₂ Valorization

This project advances fundamental understanding of a novel integrated direct air capture (DAC) and CO₂ conversion process that valorizes atmospheric CO₂ without external H₂. The research encompasses four critical components: (1) design of task-specific ionic liquids for efficient CO₂ capture under ambient conditions, (2) development of H₂-free tandem catalytic systems using ethane as a reductant, (3) advanced operando characterization to elucidate capture and conversion mechanisms, and (4) data science-driven predictive computation to accelerate material discovery. Over the project period, we developed five high-performance DAC sorbent systems—including CaO/superbase ionic liquid composites, Ni-MOF/Ionic Liquid (IL) hybrids, fluorinated covalent organic frameworks with ion-pair functional groups, defect-engineered UiO-66, and a validated kinetic model for humid-condition operation, achieving CO₂ capacities up to 1.86 mmol/g at 400 ppm with excellent cycling stability. For H₂-free conversion, we constructed atomically synergistic Zn–O–Cr binuclear catalytic sites that achieve 100% ethylene selectivity, ~9.6% ethane conversion, and 99% CO₂ utilization in equimolar co-conversion of ethane and CO₂. We further demonstrated downstream valorization pathways converting CO and C₂H₄ into polyketones and C₃ chemicals. These advances strengthen the scientific foundation for producing value-added materials from ambient CO₂.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2022

Seismic Waveform Inversion Capability on Resource-Constrained Edge Devices

Seismic full wave inversion (FWI) is a widely used non-linear seismic imaging method used to reconstruct subsurface velocity images, however it is time consuming, has high computational cost and depend heavily on human interaction. Recently, deep learning has accelerated it’s use in several data-driven techniques, however most deep learning techniques suffer from overfitting and stability issues. In this work, we propose an edge computing-based data-driven inversion technique based on supervised deep convolutional neural network to accurately reconstruct the subsurface velocities. Deep learning based data-driven technique depends mostly on bulk data training. In this work, we train our deep convolutional neural network (DCN) (UNet and InversionNet) on the raw seismic data and their corresponding velocity models during the training phase to learn the non-linear mapping between the seismic data and velocity models. The trained network is then used to estimate the velocity models from new input seismic data during the prediction phase. The prediction phase is performed on a resource-constrained edge device such as Raspberry Pi. Raspberry Pi provides real-time and on-device computational power to execute the inference process. In addition, we demonstrate robustness of our models to perform inversion in the presence on noise by performing both noise-aware and no-noise training and feeding the resulting trained models with noise at different signal-to-noise (SNR) ratio values. We make great efforts to achieve very feasible inference times on the Raspberry Pi for both models. Specifically, the inference times per prediction for UNet and InversionNet models on Raspberry Pi were 22 and 4 s respectively whilst inference times for both models on the GPU were 2 and 18 s which are very comparable. Finally, we have designed a user-friendly interactive graphical user interface (GUI) to automate the model execution and inversion process on the Raspberry Pi.

Manu, Daniel (ORCID:0000000154982677)↗

DOE OSTI2020

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

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

Bi, Sirui↗

DOE OSTI2021

The Spallation Neutron Source Normal Conducting Linac RF System Design for the Proton Power Upgrade Project

The Proton Power Upgrade (PPU) project at the Spallation Neutron Source will double the available proton beam power from 1.4 to 2.8 MW by increasing the beam energy from 1.0 to 1.3 GeV and the beam current from 26 to 38 mA. The increase in beam current resulted in the need to redesign the existing normal conducting linac (NCL) RF Systems. High-power testing of the existing NCL RF Systems configured to accelerate PPU-level beam provided the data used to make the final design decisions. This paper describes the development and execution of those in-situ tests and the subsequent results.

Moss, John↗

DOE OSTI2024

Application of performance portability solutions for GPUs and many-core CPUs to track reconstruction kernels

Next generation High-Energy Physics (HEP) experiments are presented with significant computational challenges, both in terms of data volume and processing power. Using compute accelerators, such as GPUs, is one of the promising ways to provide the necessary computational power to meet the challenge. The current programming models for compute accelerators often involve using architecture-specific programming languages promoted by the hardware vendors and hence limit the set of platforms that the code can run on. Developing software with platform restrictions is especially unfeasible for HEP communities as it takes significant effort to convert typical HEP algorithms into ones that are efficient for compute accelerators. Multiple performance portability solutions have recently emerged and provide an alternative path for using compute accelerators, which allow the code to be executed on hardware from different vendors. We apply several portability solutions, such as Kokkos, SYCL, C++17 std::execution::par and Alpaka, on two mini-apps extracted from the mkFit project: p2z and p2r. These apps include basic kernels for a Kalman filter track fit, such as propagation and update of track parameters, for detectors at a fixed z or fixed r position, respectively. The two mini-apps explore different memory layout formats. We report on the development experience with different portability solutions, as well as their performance on GPUs and many-core CPUs, measured as the throughput of the kernels from different GPU and CPU vendors such as NVIDIA, AMD and Intel.

Kwok, Ka Martin↗

DOE OSTI2023

FPGA Acceleration of GCN in Light of the Symmetry of Graph Adjacency Matrix

Graph Convolutional Neural Networks (GCNs) are widely used to process large-scale graph data. Different from deep neural networks (DNNs), GCNs are sparse, irregular, and unstructured, posing unique challenges to hardware acceleration with regular processing elements (PEs). In particular, the adjacency matrix of a GCN is extremely sparse, leading to frequent but irregular memory access, low spatial/temporal data locality and poor data reuse. Furthermore, a realistic graph usually consists of unstructured data (e.g., unbalanced distributions), creating significantly different processing times and imbalanced workload for each node in GCN acceleration. To overcome these challenges, we propose an end-to-end hardware-software co-design to accelerate GCNs on resource-constrained FPGAs with the features including: (1) A custom dataflow that leverages symmetry along the diagonal of the adjacency matrix to accelerate feature aggregation for undirected graphs. We utilize either the upper or the lower triangular matrix of the adjacency matrix to perform aggregation in GCN to improve data reuse. (2) Unified compute cores for both aggregation and transform phases, with full support to the symmetry-based dataflow. These cores can be dynamically reconfigured to the systolic mode for transformation or as individual accumulators for aggregation in GCN processing. (3) Preprocessing of the graph in software to rearrange the edges and features to match the custom dataflow. This step improves the regularity in memory access and data reuse in the aggregation phase. Moreover, we quantize the GCN precision from FP32 to INT8 to reduce the memory footprint without losing the inference accuracy. We implement our accelerator design in Intel Stratix10 MX FPGA board with HBM2, and demonstrate 1.3x-110.5x improvement in end-to-end GCN latency as compared to the state-of the-art FPGA implementations, on the graph datasets of Cora, Pubmed, Citeseer and Reddit.

Nair, Gopikrishnan R.↗

DOE OSTI2025

A Roadmap for the Future of Systems Biology in Cancer Research

Cancer systems biology seeks to understand how cancer arises as a system of interconnected molecules, cells, and tissues, with the goal of understanding, predicting, and controlling the disease. In the last decade, the field has rapidly grown as advances in experimental, computational, and analytic technologies have improved our ability to capture and recapitulate the complexities of cancer at multiple scales. However, the field’s promise to understand how specific molecular changes give rise to altered cancer outcomes remains incompletely fulfilled. Fortunately, an opportunity exists to accelerate progress by better coordinating modeling and data-gathering efforts across the cancer systems biology community. This will create the foundation for building accurate, multiscale cancer models that can better predict and identify improved therapeutic interventions. Here, in this study, we outline some of the current challenges in cancer systems biology research, how they can be addressed, and actions that the community can take to accelerate progress in the field.

Modeling & Simulation↗

DOE OSTI2022

Designing a Magnetic Measurement Data Acquisition and Control System With Reuse in Mind: A Rotating Coil System Example

Accelerator magnet test facilities frequently need to measure different magnets on differently equipped test stands and with different instrumentation. Designing a modular and highly reusable system that combines flexibility built-in at the architectural level as well as on the component level addresses this need. Specification of the backbone of the system, with the interfaces and dataflow for software components and core hardware modules, serves as a basis for building such a system. The design process and implementation of an extensible magnetic measurement data acquisition and control system are described, including techniques for maximizing the reuse of software. The discussion is supported by showing the application of this methodology to constructing two dissimilar systems for rotating coil measurements, both based on the same architecture and sharing core hardware modules and many software components. The first system is for production testing 10 m long cryo-assemblies containing two MQXFA quadrupole magnets for the high-luminosity upgrade of the Large Hadron Collider and the second for testing IQC conventional quadrupole magnets in support of the accelerator system at Fermilab.

43 PARTICLE ACCELERATORS↗

DOE OSTI2026

Development and Validation of the Near-Miss Safety Score (NMSS) Framework for Heavy-Duty Vehicle Safety Assessment

Heavy-duty commercial vehicles present unique safety challenges due to their size, articulation dynamics, and operational complexity. As Advanced Driver Assistance Systems (ADAS) and Automated Driving Systems (ADS) become more common in Class 8 tractor-trailers, traditional crash-based metrics are no longer sufficient to evaluate safety performance. This study introduces the Near-Miss Safety Score (NMSS)—a quantitative, physics-informed framework developed as a leading indicator of safety for advanced commercial vehicle technologies. NMSS quantifies how close a vehicle or operator comes to a collision or safety-critical event by integrating vehicle kinematics (relative distance, velocity, and acceleration) with driver or system response latency and time-to-collision. A modifier function adjusts the base score for vehicle-specific and environmental factors such as trailer articulation, load distribution, braking condition, and roadway environment. The framework enables systematic evaluation of ADAS/ADS performance under a range of operational and degraded conditions. By capturing near-miss dynamics rather than relying on crash data, NMSS provides a proactive approach to risk assessment, accelerates technology validation, and enhances interpretability for regulators and fleet operators. The proposed NMSS was validated using data collected from a motorcoach platform, demonstrating the framework’s applicability to heavy-duty safety evaluation and performance benchmarking. Results and key insights are presented in this paper.

Siekmann, Adam [ORNL] (ORCID:0000000284653935)↗

DOE OSTI2022

MICCO: An Enhanced Multi-GPU Scheduling Framework for Many-Body Correlation Functions

Calculation of many-body correlation functions is one of the critical kernels utilized in many scientific computing areas, especially in Lattice Quantum Chromodynamics (Lattice QCD). It is formalized as a sum of a large number of contraction terms each of which can be represented by a graph consisting of vertices describing quarks inside a hadron node and edges designating quark propagations at specific time intervals. Due to its computation- and memory-intensive nature, real-world physics systems (e.g., multi-meson or multi-baryon systems) explored by Lattice QCD prefer to leverage multi-GPUs. Different from general graph processing, many-body correlation function calculations show two specific features: a large number of computation-/data-intensive kernels and frequently repeated appearances of original and intermediate data. The former results in expensive memory operations such as tensor movements and evictions. The latter offers data reuse opportunities to mitigate the data-intensive nature of many-body correlation function calculations. However, existing graph-based multi-GPU schedulers cannot capture these data-centric features, thus resulting in a sub-optimal performance for many-body correlation function calculations. To address this issue, this paper presents a multi-GPU scheduling framework, MICCO, to accelerate contractions for correlation functions particularly by taking the data dimension (e.g., data reuse and data eviction) into account. This work first performs a comprehensive study on the interplay of data reuse and load balance, and designs two new concepts: local reuse pattern and reuse bound to study the opportunity of achieving the optimal trade-off between them. Based on this study, MICCO proposes a heuristic scheduling algorithm and a machine-learning-based regression model to generate the optimal setting of reuse bounds. Specifically, MICCO is integrated into a real-world Lattice QCD system, Redstar, for the first time running on multiple GPUs. The evaluation demonstrates MICCO outperforms other state-of-art works, achieving up to 2.25× speedup in synthesized datasets, and 1.49× speedup in real-world correlation functions.

Wang, Qihan↗

DOE OSTI2024

Pretraining Billion-Scale Geospatial Foundational Models on Frontier

As AI workloads increase in scope, generalization capability becomes challenging for small task-specific models and their demand for large amounts of labeled training samples increases. On the contrary, Foundation Models (FMs) are trained with internet-scale unlabeled data via self-supervised learning and have been shown to adapt to various tasks with minimal fine-tuning. Although large FMs have demonstrated significant impact in natural language processing and computer vision, efforts toward FMs for geospatial applications have been restricted to smaller size models, as pretraining larger models requires very large computing resources equipped with state-of-the-art hardware accelerators. Current satellite constellations collect 100+TBs of data a day, resulting in images that are billions of pixels and multimodal in nature. Such geospatial data poses unique challenges opening up new opportunities to develop FMs. We investigate billion scale FMs and HPC training profiles for geospatial applications by pretraining on publicly available data. We studied from end-to-end the performance and impact in the solution by scaling the model size. Our larger 3B parameter size model achieves up to 30% improvement in top1 scene classification accuracy when comparing a 100M parameter model. Moreover, we detail performance experiments on the Frontier supercomputer, America's first exascale system, where we study different model and data parallel approaches using PyTorch's Fully Sharded Data Parallel library. Specifically, we study variants of the Vision Transformer architecture (ViT), conducting performance analysis for ViT models with size up to 15B parameters. By discussing throughput and performance bottlenecks under different parallelism configurations, we offer insights on how to leverage such leadership-class HPC resources when developing large models for geospatial imagery applications.

Tsaris, Aristeidis (aris)↗