Search NASA⌕ Search

SEARCH · Search NASA

Results for “hardware accelerators”

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 451 records · Page 25

A Relief Zone Architecture for Enhanced Durability of Ultrathin Membranes in Electrochemical Conversion Devices

Premature cell failures in electrochemical conversion systems often result when membrane electrode assemblies (MEAs) are prepared with ultra-thin (= 15 micrometer-thick) polymer electrolyte membranes (PEMs). These high-performance PEMs are susceptible to mechanical degradation from stress concentrations arising from component and device-level integration. Herein, a relief zone was developed to mitigate mechanical degradation by alleviating excess and non-uniform compression across the active area. Relief zones, created through the ablation of carbonaceous diffusion material achieves a seamless adaptation across a range of MEA dimensions and electrochemical applications without need for costly hardware modifications. This concept was demonstrated using fuel cells as a case study. Accelerated stress test (AST) validations yielded a 6-fold improvement for MEAs prepared with relief zones (i.e., lifetime approximately 1500 h) compared to those fabricated using conventional edge-protection techniques, showcasing the utility of this technique in decoupling integration and device-level engineering effects from intrinsic material limitations for durable electrochemical devices.

14 SOLAR ENERGY↗

Accelerating Clean Energy Commercialization. A Strategic Partnership Approach

Technology development in the clean energy and broader clean tech space has proven to be challenging. Long-standing methods for advancing clean energy technologies from science to commercialization are best known for relatively slow, linear progression through research and development, demonstration, and deployment (RDD&D); and characterized by well-known valleys of death for financing. Investment returns expected by traditional venture capital investors have been difficult to achieve, particularly for hardware-centric innovations, and companies that are subject to project finance risks. Commercialization support from incubators and accelerators has helped address these challenges by offering more support services to start-ups; however, more effort is needed to fulfill the desired clean energy future. The emergence of new strategic investors and partners in recent years has opened up innovative opportunities for clean tech entrepreneurs, and novel commercialization models are emerging that involve new alliances among clean energy companies, RDD&D, support systems, and strategic customers. For instance, Wells Fargo and Company (WFC) and the National Renewable Energy Laboratory (NREL) have launched a new technology incubator that supports faster commercialization through a focus on technology development. The incubator combines strategic financing, technology and technical assistance, strategic customer site validation, and ongoing financial support.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Femtosecond MeV Electron Energy-Loss Spectroscopy

Pump-probe electron energy-loss spectroscopy (EELS) with femtosecond temporal resolution will be a transformative research tool for studying nonequilibrium chemistry and electronic dynamics of matter. Here in this article, we propose a concept of femtosecond EELS utilizing mega-electron-volt electron beams from a radio-frequency (rf) photocathode source. The high acceleration gradient and high beam energy of the rf gun are critical to the generation of 10-fs electron beams, which enables an improvement of the temporal resolution by more than 1 order of magnitude beyond the state of the art. In our proposal, the “referencebeam technique” relaxes the energy stability requirement of the rf power source by roughly 2 orders of magnitude. The requirements for the electron-beam quality, photocathode, spectrometer, and detector are also discussed. Supported by particle-tracking simulations, we demonstrate the feasibility of achieving subelectron- volt energy resolution and approximately 10-fs temporal resolution with existing or near-future hardware performance.

43 PARTICLE ACCELERATORS↗

Impact Report: Quantum Systems Accelerator

The Quantum Systems Accelerator (QSA) is a U.S. National Quantum Information Science Research Center established in August 2020 and funded by the Department of Energy (DOE) Office of Science. QSA is composed of 15 partner institutions— universities and national laboratories—bringing together pioneers of many of today’s unique quantum information science (QIS) and engineering capabilities. Led by Lawrence Berkeley National Laboratory (Berkeley Lab), with Sandia National Laboratories (Sandia Labs) as the lead partner, 250+ QSA researchers are catalyzing U.S. leadership in a fast-growing field that seeks solutions to the Nation’s and the world’s most pressing problems by harnessing the laws of quantum mechanics.

97 MATHEMATICS AND COMPUTING↗

NDCX-II project commencing at LBNL

NDCX-II is the second-generation Neutralized Drift Compression eXperiment, capable of accelerating and strongly bunching tens of nanoCoulombs of non-relativistic ions, for applications requiring nanosecond-scale pulses with short stopping ranges. As with the existing NDCX-I at Lawrence Berkeley National Laboratory (LBNL), the new machine is based on the technique of neutralized drift compression, whereby a head-to-tail velocity gradient is imparted to the beam, which then shortens as it drifts in a neutralizing plasma that suppresses space-charge forces. The figure shows the layout of the machine, to be sited at LBNL. It will make extensive use of induction cells and other parts from the decommissioned Advanced Test Accelerator (ATA) at Lawrence Livermore National Laboratory (LLNL). It will be extensible and reconfigurable; in the configuration that has received the most emphasis, each pulse will deliver 30-50 nC of Li+ ions at 3 MeV into a mm-scale spot onto a thin-foil target. Pulse compression to {approx} 1 ns begins in the accelerator and finishes in the drift compression line; the beam is manipulated using suitably tailored voltage waveforms in the accelerating gaps. NDCX-II employs novel beam dynamics. To use the 200 kV Blumlein pulsed power from ATA (blue cylinders in the figure), the pulse duration must first be reduced from an initial 500 ns to less than 70 ns. This shortening is accomplished in an initial stage of non-neutral drift compression, downstream of the injector and the first few induction cells (note the spaces between induction cells at the left end of the figure). The compression is rapid enough that fewer than ten long-pulse waveform generators are needed, with Blumleins powering the rest of the acceleration. Extensive particle-in-cell simulation studies have enabled an attractive physics design that meets the stringent cost goal. Snapshots from a simulation video are shown in the figure. Studies on a dedicated test stand are characterizing the performance of the ATA hardware, and of pulsed solenoids that will provide transverse beam confinement (ions require much stronger fields than the electrons accelerated by ATA). Applications of this facility will include studies of: the basic physics of the poorly understood 'warm dense matter' regime of temperatures around 10,000 K and densities near solid, using uniform, volumetric ion heating of thin foil targets; ion energy coupling into an ablating plasma (such as that which occurs in an inertial fusion target) using beams with time-varying kinetic energy; space-charge-dominated ion beam dynamics; and beam focusing and pulse compression in neutralizing plasma. The machine will complement facilities at GSI in Darmstadt, Germany, but will employ lower ion kinetic energies and commensurately shorter stopping ranges in matter. Much of this research will contribute directly toward the collaboration's ultimate goal of electric power production via heavy-ion beam-driven inertial confinement fusion ('Heavy-Ion Fusion,' or HIF). In inertial fusion, a target containing fusion fuel is heated by energetic 'driver' beams, and undergoes a miniature thermonuclear explosion. Currently the largest U.S. research program in inertial confinement is at Livermore's National Ignition Facility (NIF), a multibillion-dollar, stadium-sized laser facility that was developed for studying physics relevant to nuclear stockpile stewardship. NIF is expected to establish the fundamental feasibility of fusion ignition on the laboratory scale, and to thereby advance this approach to fusion energy. Heavy ion accelerators offer efficiency and longevity, and can use magnetic fields for final focusing onto a target. These attributes make them attractive candidates as inertial fusion energy drivers. The space-charge-dominated beams in such a system will require manipulation and control similar to that being pioneered on NDCX-II.

43 PARTICLE ACCELERATORS↗

EJFAT Scientific Perspective

Presented new computing model to the test by deploying the EJFAT system alongside a data-stream processing framework running the production-level CLAS12 event reconstruction application. In this experiment, a continuous stream of CLAS12 Level-1 identified events was processed in real-time using the EJFAT load balancer, distributing the workload across 90 computing nodes located across the U.S. This marks the first-ever large-scale, real-time distributed data stream processing experiment, demonstrating that scientific data-streaming pipelines can efficiently scale across four dimensions, thanks to EJFAT’s advanced hardware and software capabilities.

Gyurjyan, Vardan [Thomas Jefferson National Accele↗

Measuring Q₀ in LCLS-II Cryomodules Using Helium Liquid Level

The nitrogen-doped cavities used in the Linac Coherent Light Source II (LCLS-II) cryomodules have shown an unprecedented high Q₀ in vertical and cryomodule testing compared with cavities prepared with standard methods. While demonstration of high Q₀ in the test stand has been achieved, maintaining that performance in the linac is critical to the success of LCLS-II and future accelerator projects. The LCLS-II cryomodules required a novel method of measuring Q₀, due to hardware incompatibilities with existing procedures. Initially developed at Jefferson Lab during cryomodule acceptance testing before being used in the tunnel at SLAC, we use helium liquid level data to estimate the heat generated by cavities. We first establish the relationship between the rate of helium evaporation from known heat loads using electric heaters, and then use that relationship to determine heat from an RF load. Here we present the full procedure along with the development process, lessons learned, and reproducibility while demonstrating for the first time that world record Q₀ can be maintained within the real accelerator environment.

43 PARTICLE ACCELERATORS↗

Disentangling Beam Losses in The Fermilab Main Injector Enclosure Using Real-Time Edge AI

The Fermilab Main Injector enclosure houses two accelerators, the Main Injector and Recycler Ring. During normal operation, high intensity proton beams exist simultaneously in both. The two accelerators share the same beam loss monitors (BLM) and monitoring system. Deciphering the origin of any of the 260 BLM readings is often difficult. The (Accelerator) Real-time Edge AI for Distributed Systems project, or READS, has developed an AI/ML model, and implemented it on fast FPGA hardware, that disentangles mixed beam losses and attributes probabilities to each BLM as to which machine(s) the loss originated from in real-time. The model inferences are then streamed to the Fermilab accelerator controls network (ACNET) where they are available for operators and experts alike to aid in tuning the machines.

43 PARTICLE ACCELERATORS↗

First Results of the SRF Wafer Test Cavity for the Characterization of Superconductors

The wafer test cavity was designed as a short sample test system that could create a reproducible environment for the testing of superconducting materials above the Bardeen-Cooper- Schrieffer limit of niobium. The results of the sapphire test cavity showed that the dielectric was too lossy, and thus, the original design had to be altered to make operation feasible with current hardware and achieve ~200 mT. The new design was fabricated at Thomas Jefferson National Accelerator Facility and was cryogenically tested. After four tests, the cavity was able to produce a 6.6-mT field with a Q of 3.96 * 108. Although lower than anticipated, in comparison to other TE01 cavities, this result is quite encouraging. Multipacting and coupling were limitations, but current work is pursuing the elimination of these complications. This document will expound upon the new design, mathematical simulations, testing of the cavity, complications, results, and future work.

Pogue, Nathaniel J.↗

Captan+X Data Converter Integration

Fermi National Accelerator Laboratory's CAPTAN (Compact And Programmable daTa Acquisition Node) series provides a flexible hardware platform for data acquisition across a range of experiments and facilities. The latest iteration, CAPTAN+X, is built around a Kintex-7 FPGA supporting four FPGA Mezzanine Card (FMC) connections. As part of a broader laboratory effort to bring facility systems under a Model-Based Systems Engineering (MBSE) framework, CAPTAN+X is one of several systems slated to be incorporated into this modeling environment in the near term. A necessary step toward that goal is incorporating the platform's core functionality, which centers on integration with the LXD31K4 FMC, a data converter module combining dual AD9652 analog-to-digital converters and dual AD9142A digital-to-analog converters. Achieving compatibility required resolving pin-mapping conflicts between the LXD31K4's High Pin Count connector and the CAPTAN+X's available pin types, adapting a Board Support Project originally written for an UltraScale-class evaluation board to the Kintex-7 architecture, replacing incompatible primitives, restructuring clock distribution, and manually configuring chip initialization in place of an unsupported soft-processor-based approach. Functional verification of the ADC and DAC channels, followed by closed-loop testing combining both converters with real-time filtering, confirmed correct operation of the integrated system. These results establish a working hardware and firmware baseline for the CAPTAN+X platform, positioning it for future inclusion in the laboratory's growing MBSE modeling effort.

Espinoza, David [Illinois U., Urbana (main)]↗

CAPTAN+X Data Converter Integration

Fermi National Accelerator Laboratory's CAPTAN (Compact And Programmable daTa Acquisition Node) series provides a flexible hardware platform for data acquisition across a range of experiments and facilities. The latest iteration, CAPTAN+X, is built around a Kintex-7 FPGA supporting four FPGA Mezzanine Card (FMC) connections. As part of a broader laboratory effort to bring facility systems under a Model-Based Systems Engineering (MBSE) framework, CAPTAN+X is one of several systems slated to be incorporated into this modeling environment in the near term. A necessary step toward that goal is incorporating the platform's core functionality, which centers on integration with the LXD31K4 FMC, a data converter module combining dual AD9652 analog-to-digital converters and dual AD9142A digital-to-analog converters. Achieving compatibility required resolving pin-mapping conflicts between the LXD31K4's High Pin Count connector and the CAPTAN+X's available pin types, adapting a Board Support Project originally written for an UltraScale-class evaluation board to the Kintex-7 architecture, replacing incompatible primitives, restructuring clock distribution, and manually configuring chip initialization in place of an unsupported soft-processor-based approach. Functional verification of the ADC and DAC channels, followed by closed-loop testing combining both converters with real-time filtering, confirmed correct operation of the integrated system. These results establish a working hardware and firmware baseline for the CAPTAN+X platform, positioning it for future inclusion in the laboratory's growing MBSE modeling effort.

Espinoza, David [Illinois U., Urbana (main)]↗

Reliability of Operation at SLAC in the LCLS Era

LCLS hardware availability has been above 90% for the first two commissioning runs of the accelerator. In this paper we compare the reliability data for LCLS (availability, MTBF and MTTR) to those of PEP-II, the e{sup +}e{sup -} collider operating previously at SLAC. It may be seen that the linac availability is not significantly different now than it was before, while the availability of the whole LCLS facility is significantly higher than that of the PEP-II facility as a whole (which was about 87%). Most of the improvement is in the MTTR. Ways to improve availability towards the goal of 95% are discussed.

43 PARTICLE ACCELERATORS↗

XploreNAS : Explore Adversarially Robust and Hardware-efficient Neural Architectures for Non-ideal Xbars

Compute In-Memory platforms such as memristive crossbars are gaining focus as they facilitate acceleration of Deep Neural Networks (DNNs) with high area and compute efficiencies. However, the intrinsic non-idealities associated with the analog nature of computing in crossbars limits the performance of the deployed DNNs. Furthermore, DNNs are shown to be vulnerable to adversarial attacks leading to severe security threats in their large-scale deployment. Thus, finding adversarially robust DNN architectures for non-ideal crossbars is critical to the safe and secure deployment of DNNs on the edge. This work proposes a two-phase algorithm-hardware co-optimization approach called XploreNAS that searches for hardware efficient and adversarially robust neural architectures for non-ideal crossbar platforms. We use the one-shot Neural Architecture Search approach to train a large Supernet with crossbar-awareness and sample adversarially robust Subnets therefrom, maintaining competitive hardware efficiency. Our experiments on crossbars with benchmark datasets (SVHN, CIFAR10, CIFAR100) show up to ~8–16% improvement in the adversarial robustness of the searched Subnets against a baseline ResNet-18 model subjected to crossbar-aware adversarial training. We benchmark our robust Subnets for Energy-Delay-Area-Products (EDAPs) using the Neurosim tool and find that with additional hardware efficiency–driven optimizations, the Subnets attain ~1.5–1.6× lower EDAPs than ResNet-18 baseline.

97 MATHEMATICS AND COMPUTING↗

FPGA-Accelerated Machine Learning Inference as a Service for Particle Physics Computing

Large-scale particle physics experiments face challenging demands for high-throughput computing resources both now and in the future. New heterogeneous computing paradigms on dedicated hardware with increased parallelization, such as Field Programmable Gate Arrays (FPGAs), offer exciting solutions with large potential gains. The growing applications of machine learning algorithms in particle physics for simulation, reconstruction, and analysis are naturally deployed on such platforms. We demonstrate that the acceleration of machine learning inference as a web service represents a heterogeneous computing solution for particle physics experiments that potentially requires minimal modification to the current computing model. As examples, we retrain the ResNet-50 convolutional neural network to demonstrate state-of-the-art performance for top quark jet tagging at the LHC and apply a ResNet-50 model with transfer learning for neutrino event classification. Using Project Brainwave by Microsoft to accelerate the ResNet-50 image classification model, we achieve average inference times of 60 (10) ms with our experimental physics software framework using Brainwave as a cloud (edge or on-premises) service, representing an improvement by a factor of approximately 30 (175) in model inference latency over traditional CPU inference in current experimental hardware. A single FPGA service accessed by many CPUs achieves a throughput of 600–700 inferences per second using an image batch of one, comparable to large batch-size GPU throughput and significantly better than small batch-size GPU throughput. Deployed as an edge or cloud service for the particle physics computing model, coprocessor accelerators can have a higher duty cycle and are potentially much more cost-effective.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

hls4ml: An Open-Source Codesign Workflow to Empower Scientific Low-Power Machine Learning Devices

Accessible machine learning algorithms, software, and diagnostic tools for energy-efficient devices and systems are extremely valuable across a broad range of application domains. In scientific domains, real-time near-sensor processing can drastically improve experimental design and accelerate scientific discoveries. To support domain scientists, we have developed hls4ml, an open-source software-hardware codesign workflow to interpret and translate machine learning algorithms for implementation with both FPGA and ASIC technologies. We expand on previous hls4ml work by extending capabilities and techniques towards low-power implementations and increased usability: new Python APIs, quantization-aware pruning, end-to-end FPGA workflows, long pipeline kernels for low power, and new device backends include an ASIC workflow. Taken together, these and continued efforts in hls4ml will arm a new generation of domain scientists with accessible, efficient, and powerful tools for machine-learning-accelerated discovery.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

SU-E-T-226: Correction of a Standard Model-Based Dose Calculator Using Measurement Data

Purpose: To propose a hybrid method that combines advantages of the model-based and measurement-based method for independent dose calculation. Modeled-based dose calculation, such as collapsed-cone-convolution/superposition (CCCS) or the Monte-Carlo method, models dose deposition in the patient body accurately; however, due to lack of detail knowledge about the linear accelerator (LINAC) head, commissioning for an arbitrary machine is tedious and challenging in case of hardware changes. On the contrary, the measurement-based method characterizes the beam property accurately but lacks the capability of dose disposition modeling in heterogeneous media. Methods: We used a standard CCCS calculator, which is commissioned by published data, as the standard model calculator. For a given machine, water phantom measurements were acquired. A set of dose distributions were also calculated using the CCCS for the same setup. The difference between the measurements and the CCCS results were tabulated and used as the commissioning data for a measurement based calculator. Here we used a direct-ray-tracing calculator (ΔDRT). The proposed independent dose calculation consists of the following steps: 1. calculate D-model using CCCS. 2. calculate D-ΔDRT using ΔDRT. 3. combine Results: D=D-model+D-ΔDRT. Results: The hybrid dose calculation was tested on digital phantoms and patient CT data for standard fields and IMRT plan. The results were compared to dose calculated by the treatment planning system (TPS). The agreement of the hybrid and the TPS was within 3%, 3 mm for over 98% of the volume for phantom studies and lung patients. Conclusion: The proposed hybrid method uses the same commissioning data as those for the measurement-based method and can be easily extended to any non-standard LINAC. The results met the accuracy, independence, and simple commissioning criteria for an independent dose calculator.

60 APPLIED LIFE SCIENCES↗

IOTA Run 2 Beam Dynamics Studies in Nonlinear Integrable Systems

Nonlinear integrable optics is a promising design approach for suppressing fast collective instabilities. To study it experimentally, a new storage ring, the Integrable Optics Test Accelerator (IOTA), was built at Fermilab. IOTA has recently completed its second scientific run, incorporating many hardware and instrumentation improvements. This report presents the results of the two integrable optics experiments - the quasi-integrable Henon-Heiles octupole system and the fully integrable Danilov-Nagaitsev system. We demonstrate tune spread and dynamic aperture in agreement with tracking simulations, and a stable crossing of the integer resonance. Based on recovered single-particle phase space dynamics, we show improved invariant jitter consistent with intended effective Hamiltonian. We conclude by outlining future plans and efforts towards proton studies and larger designs.

43 PARTICLE ACCELERATORS↗

Assessment of Vibration and Shock Loading Environments on Steels Used in Railroad Applications Considering Hydrogen Embrittlement

Studies that examine the combined effects of hydrogen and rail loading environments in the context of hydrogen embrittlement of steels have not been performed. Large-scale hydrogen storage requires either gaseous hydrogen stored in pressure vessels and tubing or liquid hydrogen in cryogenic, insulated tanks. Since the focus of this study is on steels, we examine the influence of hydrogen on stainless steel used for high pressure tubing and cryogenic tanks and Cr-Mo steel liners used in gaseous pressure vessels. Literature studies show that despite hydrogen degradation, these steels can be used effectively and safely if the stresses are managed. The shock and vibration environment on railcars present unique environments where stresses and loading rates can be high. A literature review was conducted to understand the accelerations experienced in rail by examining instrumented railcars of nuclear fuel casks. Hydrogen studies were examined that focused on steel pressure hardware with an emphasis on fatigue, fast loading rates, and low temperatures. A simplified fatigue assessment was performed to examine design cycles for three different hydrogen components: stainless steel tubing, stainless steel cryogenic tanks, and gaseous pressure vessels. The assessments were idealized and used bounding cases of accelerations in both normal rail conditions and coupling events. In normal shock and vibration environments when accelerations were less than 2 g, infinite design life was calculated. In extreme shock or coupling environments, finite design life was calculated which was shown to be dependent on unsupported lengths of pressure components. It was shown that adjusting the unsupported length can reduce the bending stress thereby increasing design life.

36 MATERIALS SCIENCE↗