Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “architecture optimization”

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 685 records · Page 38
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

An agent-based blackboard system for multi-objective optimization

In the field of multi-objective optimization, there are a multitude of algorithms from which to choose. Each algorithm has strengths and weaknesses associated with the mechanics for finding the Pareto front. Recently, researchers have begun to examine how multi-agent environments can be used to help solve multi-objective optimization problems. In this work, we propose a multi-objective optimization algorithm based on a multi-agent blackboard system (MABS). The MABS framework allows for multiple agents to read and write pertinent optimization problem data to a central blackboard agent. Agents can stochastically search the design space, use previously discovered solutions to explore local optima, or update and prune the Pareto front. A centralized blackboard framework allows the optimization problem to be solved in a cohesive manner and permits stopping, restarting, or updating the optimization problem. The MABS framework is tested against three alternative optimization algorithms across a suite of engineering design problems and typically outperforms the other algorithms in discovering the Pareto front. A parallelizability study is performed where we find that the MABS is able to evaluate a set number of designs, which require an evaluation time ranging from 0 to 300 seconds, quicker than a traditional optimization algorithm: this fact becomes more apparent the longer it takes to evaluate a design. To provide context for the benefits provided by MABS, a real-world nuclear engineering design problem is examined. MABS is used to examine the placement of experiments in a nuclear reactor, where we are able to evaluate hundreds of configurations for experimental placement while maintaining a strict set of safety constraints.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
NASA NTRS2011

Hybrid Power Management-Based Vehicle Architecture

Hybrid Power Management (HPM) is the integration of diverse, state-of-the-art power devices in an optimal configuration for space and terrestrial applications (s ee figure). The appropriate application and control of the various power devices significantly improves overall system performance and efficiency. The basic vehicle architecture consists of a primary power source, and possibly other power sources, that provides all power to a common energy storage system that is used to power the drive motors and vehicle accessory systems. This architecture also provides power as an emergency power system. Each component is independent, permitting it to be optimized for its intended purpose. The key element of HPM is the energy storage system. All generated power is sent to the energy storage system, and all loads derive their power from that system. This can significantly reduce the power requirement of the primary power source, while increasing the vehicle reliability. Ultracapacitors are ideal for an HPM-based energy storage system due to their exceptionally long cycle life, high reliability, high efficiency, high power density, and excellent low-temperature performance. Multiple power sources and multiple loads are easily incorporated into an HPM-based vehicle. A gas turbine is a good primary power source because of its high efficiency, high power density, long life, high reliability, and ability to operate on a wide range of fuels. An HPM controller maintains optimal control over each vehicle component. This flexible operating system can be applied to all vehicles to considerably improve vehicle efficiency, reliability, safety, security, and performance. The HPM-based vehicle architecture has many advantages over conventional vehicle architectures. Ultracapacitors have a much longer cycle life than batteries, which greatly improves system reliability, reduces life-of-system costs, and reduces environmental impact as ultracapacitors will probably never need to be replaced and disposed of. The environmentally safe ultracapacitor components reduce disposal concerns, and their recyclable nature reduces the environmental impact. High ultracapacitor power density provides high power during surges, and the ability to absorb high power during recharging. Ultracapacitors are extremely efficient in capturing recharging energy, are rugged, reliable, maintenance-free, have excellent lowtemperature characteristic, provide consistent performance over time, and promote safety as they can be left indefinitely in a safe, discharged state whereas batteries cannot.

Eichenberg, Dennis J.↗
DOE OSTI2022

Open-source FPGA-ML codesign for the MLPerf Tiny Benchmark

We present our development experience and recent results for the MLPerf Tiny Inference Benchmark on field-programmable gate array (FPGA) platforms. We use the open-source hls4ml and FINN workflows, which aim to democratize AI-hardware codesign of optimized neural networks on FPGAs. We present the design and implementation process for the keyword spotting, anomaly detection, and image classification benchmark tasks. The resulting hardware implementations are quantized, configurable, spatial dataflow architectures tailored for speed and efficiency and introduce new generic optimizations and common workflows developed as a part of this work. The full workflow is presented from quantization-aware training to FPGA implementation. The solutions are deployed on system-on-chip (Pynq-Z2) and pure FPGA (Arty A7-100T) platforms. The resulting submissions achieve latencies as low as 20 $\mu$s and energy consumption as low as 30 $\mu$J per inference. We demonstrate how emerging ML benchmarks on heterogeneous hardware platforms can catalyze collaboration and the development of new techniques and more accessible tools.

Borras, Hendrik↗
DOE OSTI2020

CCAMP: An Integrated Translation and Optimization Framework for OpenACC and OpenMP

Heterogeneous computing and exploration into specialized accelerators are inevitable in current and future supercomputers. Although this diversity of devices is promising for performance, the array of architectures presents programming challenges. High-level programming strategies have emerged to face these challenges, such as the OpenMP offloading model and OpenACC. The varying levels of support for these standards, however, within vendor-specific and open-source tools, as well as the lack of performance portability across devices, have prevented the standards from achieving their goals. To address these shortcomings, we present CCAMP, an OpenMP and OpenACC interoperable framework. CCAMP provides two primary facilities: language translation between the two standards and device-specific directive optimization within each standard. We show that by using the CCAMP framework, programmers can easily transplant non-portable code into new ecosystems for new architectures. Additionally, by using CCAMP device-specific directive optimizations, users can achieve optimized performance across architectures using a single source code.

Lambert, Jacob↗
DOE OSTI2020

Characterization of an architecture for front-end pixel binning in an integrating pixel array detector

Optimization of an area detector involves compromises between various parameters like frame rate, read noise, dynamic range and pixel size. We have implemented and tested a novel front-end binning design in a photon-integrating hybrid pixel array detector using the MM-PAD- 2.0 pixel architecture. In this architecture, the pixels can be optionally binned in a 2 x 2 pixel configuration using a network of switches to selectively direct the output of 4 sensor pixels to a single amplifier input. Doing this allows a trade-off between frame rate and spatial resolution. Tests show that the binned pixels perform well, but with some degradation on performance as compared to an un-binned pixel. The increased parasitic input capacitance does reduce the signal collected per x-ray as well as increases the noise of the pixel. Here, the increase in noise is, however, less than the factor of 2 increase one would observe for binning in post-processing. Spatial scans across the binned pixels show that no measured signal intensity is lost at the inner binning unit boundaries. In the high flux regime, at a 2 x 2 pixel wide beam spot (FWHM) size, binned mode responds linearly up to a photon flux of 10 7 x-rays/s, and performs comparably with un-binned mode up to a photon flux of 10 8 x-rays/s. While this study demonstrates a proof of concept for front-end binning in integrating detectors, we also identify changes to this early-stage prototype which can further improve the performance of binning pixel structures.

47 OTHER INSTRUMENTATION↗
NASA NTRS1998

Aerospace Power Systems Design and Analysis (APSDA) Tool

The conceptual design of space and/or planetary electrical power systems has required considerable effort. Traditionally, in the early stages of the design cycle (conceptual design), the researchers have had to thoroughly study and analyze tradeoffs between system components, hardware architectures, and operating parameters (such as frequencies) to optimize system mass, efficiency, reliability, and cost. This process could take anywhere from several months to several years (as for the former Space Station Freedom), depending on the scale of the system. Although there are many sophisticated commercial software design tools for personal computers (PC's), none of them can support or provide total system design. To meet this need, researchers at the NASA Lewis Research Center cooperated with Professor George Kusic from the University of Pittsburgh to develop a new tool to help project managers and design engineers choose the best system parameters as quickly as possible in the early design stages (in days instead of months). It is called the Aerospace Power Systems Design and Analysis (APSDA) Tool. By using this tool, users can obtain desirable system design and operating parameters such as system weight, electrical distribution efficiency, bus power, and electrical load schedule. With APSDA, a large-scale specific power system was designed in a matter of days. It is an excellent tool to help designers make tradeoffs between system components, hardware architectures, and operation parameters in the early stages of the design cycle. user interface. It operates on any PC running the MS-DOS (Microsoft Corp.) operating system, version 5.0 or later. A color monitor (EGA or VGA) and two-button mouse are required. The APSDA tool was presented at the 30th Intersociety Energy Conversion Engineering Conference (IECEC) and is being beta tested at several NASA centers. Beta test packages are available for evaluation by contacting the author.

Truong, Long V.↗
NASA NTRS1995

A open loop guidance architecture for navigationally robust on-orbit docking

The development of an open-hop guidance architecture is outlined for autonomous rendezvous and docking (AR&D) missions to determine whether the Global Positioning System (GPS) can be used in place of optical sensors for relative initial position determination of the chase vehicle. Feasible command trajectories for one, two, and three impulse AR&D maneuvers are determined using constrained trajectory optimization. Early AR&D command trajectory results suggest that docking accuracies are most sensitive to vertical position errors at the initial conduction of the chase vehicle. Thus, a feasible command trajectory is based on maximizing the size of the locus of initial vertical positions for which a fixed sequence of impulses will translate the chase vehicle into the target while satisfying docking accuracy requirements. Documented accuracies are used to determine whether relative GPS can achieve the vertical position error requirements of the impulsive command trajectories. Preliminary development of a thruster management system for the Cargo Transfer Vehicle (CTV) based on optimal throttle settings is presented to complete the guidance architecture. Results show that a guidance architecture based on a two impulse maneuvers generated the best performance in terms of initial position error and total velocity change for the chase vehicle.

Chern, Hung-Sheng↗
DOE OSTI2018

Research three-dimensional hierarchical graphite architectures for anodes for fast charging

With current lithium ion batteries optimized for performance under relatively low charge rate conditions, implementation of XFC has been hindered by drawbacks including Li plating, kinetic polarization, and heat dissipation. This project will utilize model-informed design of 3-D hierarchical electrodes to tune key XFC related variables like 1) bulk porosity/tortuosity 2) vertical pore diameter, spacing, and lattice 3) crystallographic orientation of graphite particles relative to exposed surfaces 4) interfacial chemistry of the graphite surfaces through "artificial sEr formation using ALD 5) current collector surface roughness (aspect ratio, roughness factor, etc.). A key aspect of implementing novel electrodes is characterizing them in relevant settings. For this project, ultimately led out of University of Michigan by Neil Dasgupta, that includes both coin cell and 2+ Ah pouch cell testing, as well as comparison testing against baselines. Sandia National Labs will be conducting detailed cell characterization on iterative versions/improvements of the model-based hierarchical electrodes, as well as COTS cells for baseline comparisons. Key metrics include performance under fast charge conditions, as well as the absence or degree of lithium plating. Sandia will use their unique high precision cycling and rapid EIS capabilities to accurately characterize performance and any lithium plating during 6C charging and beyond, coupling electrochemical observations with cell teardown. Sandia will also design custom fixturing to cool cells during rapid charge, to decouple any kinetic effects brought about by cell heating and allow comparisons between different cells and charge rates. Using these techniques, Sandia will assess HOH electrodes from the University of Michigan, as well as aiding in iterative model and electrode design.

25 ENERGY STORAGE↗
DOE OSTI2020

Research three-dimensional hierarchical graphite architectures for anodes for fast charging

With current lithium ion batteries optimized for performance under relatively low charge rate conditions, implementation of XFC has been hindered by drawbacks including Li plating, kinetic polarization, and heat dissipation. This project will utilize model-informed design of 3-D hierarchical electrodes to tune key XFCrelated variables like 1) bulk porosity/tortuosity 2) vertical pore diameter, spacing, and lattice 3) crystallographic orientation of graphite particles relative to exposed surfaces 4) interfacial chemistry of the graphite surfaces through "artificial sEr' formation using ALD 5) current collector surface roughness (aspect ratio, roughness factor, etc.). A key aspect of implementing novel electrodes is characterizing them in relevant settings. For this project, ultimately led out of University of Michigan by Neil Dasgupta, that includes both coin cell and 2+ Ah pouch cell testing, as well as comparison testing against baselines. Sandia National Labs will be conducting detailed cell characterization on iterative versions/improvements of the model-based hierarchical electrodes, as well as COTS cells for baseline comparisons. Key metrics include performance under fast charge conditions, as well as the absence or degree of lithium plating. Sandia will use their unique high precision cycling and rapid EIS capabilities to accurately characterize performance and any lithium plating during 6C charging and beyond, coupling electrochemical observations with cell teardown. Sandia will also design custom fixturing to cool cells during rapid charge, to decouple any kinetic effects brought about by cell heating and allow comparisons between different cells and charge rates. Using these techniques, Sandia will assess HOH electrodes from the University of Michigan, as well as aiding in iterative model and electrode design.

25 ENERGY STORAGE↗
NASA NTRS1994

An incremental database access method for autonomous interoperable databases

We investigated a number of design and performance issues of interoperable database management systems (DBMS's). The major results of our investigation were obtained in the areas of client-server database architectures for heterogeneous DBMS's, incremental computation models, buffer management techniques, and query optimization. We finished a prototype of an advanced client-server workstation-based DBMS which allows access to multiple heterogeneous commercial DBMS's. Experiments and simulations were then run to compare its performance with the standard client-server architectures. The focus of this research was on adaptive optimization methods of heterogeneous database systems. Adaptive buffer management accounts for the random and object-oriented access methods for which no known characterization of the access patterns exists. Adaptive query optimization means that value distributions and selectives, which play the most significant role in query plan evaluation, are continuously refined to reflect the actual values as opposed to static ones that are computed off-line. Query feedback is a concept that was first introduced to the literature by our group. We employed query feedback for both adaptive buffer management and for computing value distributions and selectivities. For adaptive buffer management, we use the page faults of prior executions to achieve more 'informed' management decisions. For the estimation of the distributions of the selectivities, we use curve-fitting techniques, such as least squares and splines, for regressing on these values.

Roussopoulos, Nicholas↗
NASA NTRS2017

Four Bed Molecular Sieve - Exploration (4BMS-X) Virtual Heater Design and Optimization

A 4BMS-X (Four Bed Molecular Sieve - Exploration) design and heater optimization study for CO2 sorbent beds in proposed exploration system architectures is presented. The primary objectives of the study are to reduce heater power and thermal gradients within the CO2 sorbent beds while minimizing channeling effects. Some of the notable changes from the ISS (International Space Station) CDRA (Carbon Dioxide Removal Assembly) to the proposed exploration system architecture include cylindrical beds, alternate sorbents and an improved heater core. Results from both 2D and 3D sorbent bed thermal models with integrated heaters are presented. The 2D sorbent bed models are used to optimize heater power and fin geometry while the 3D models address end effects in the beds for more realistic thermal gradient and heater power predictions.

Schunk, R. Gregory↗
DOE OSTI2023

ZeoNet: 3D convolutional neural networks for predicting adsorption in nanoporous zeolites

Zeolites are one of the most widely used materials in the chemical industry due to their nanometer-sized pores that can adsorb and react upon molecules selectively. With hundreds of known framework topologies and hundreds of thousands of computationally predicted structures, the ability to rapidly predict zeolite performance allows researchers to prioritize their efforts on the most promising structures for a given application. Although the accuracy of forcefield-based atomistic simulations has advanced significantly in the past two decades, these simulations can be computationally expensive, especially for long-chain, complex molecules. Here, we present ZeoNet, a representation learning framework using convolutional neural networks (ConvNets) and 3D volumetric representations for predicting adsorption in zeolites. ZeoNet was trained on the task of predicting Henry's constants for adsorption, k H , of n-octadecane in more than 330 000 known and predicted zeolite materials. Employing a 3D grid based on the distances to solvent-accessible surfaces, a volumetric representation that can be generated efficiently, the best-performing ZeoNet achieved a correlation coefficient r 2 = 0.977 and a mean-squared error MSE = 3.8 in ln k H , which corresponds to an error of 9.3 kJ mol -1 in adsorption free energy. In comparison, a model based on hand-designed geometric features has values of r 2 = 0.783 and MSE = 35.7. ZeoNet is also relatively efficient and can process ≈8 structures per second on an Nvidia RTX 2080TI GPU, orders of magnitude faster than forcefield-based simulations. A systematic analysis was conducted to investigate how the choice of ConvNet architectures, the linear dimension (L) and spatial resolution (Δd) of the distance grids, batch size, optimizer, and learning rate impact the model performance. We found that ConvNets based on the ResNet architecture offer the best tradeoff between expressiveness and efficiency. The performance for all models reaches a plateau at L = 30–45 Å and depends less sensitively on grid resolution, with a small benefit around Δd = 0.30–0.45 Å. Finally, saliency maps were visualized to identify which regions of the materials contributed the most to model predictions. It was found, interestingly, that the predictions are driven primarily by the accessible pore volume rather than the region occupied by the framework atoms.

36 MATERIALS SCIENCE↗
NASA NTRS1978

MBASIC batch processor architectural overview

The MBASIC (TM) batch processor, a language translator designed to operate in the MBASIC (TM) environment is described. Features include: (1) a CONVERT TO BATCH command, usable from the ready mode; and (2) translation of the users program in stages through several levels of intermediate language and optimization. The processor is to be designed and implemented in both machine-independent and machine-dependent sections. The architecture is planned so that optimization processes are transparent to the rest of the system and need not be included in the first design implementation cycle.

Reynolds, S. M.↗
NASA NTRS2004

Thermal design trades for SAFIR architecture concepts

SAFIR is a IO-meter, 4 K space telescope optimized for wavelengths between 20 microns and 1 mm. The combination of aperture diameter and telescope temperature will provide a raw sensitivity improvement of more than a factor of 1000 over presently-planned missions. The sensitivity will be comparable to that of the JWST and ALMA, but at the critical far-IR wavelengths where much of the universe's radiative energy has emerged since the origin of stars and galaxies. We examine several of the critical technologies for SAFIR which enable the large cold aperture, and present results of studies examining the telescope optics and the spacecraft thermal architecture. Both the method by which the aperture is filled, and the overall optical design for the telescope can impact the potential scientific return of SAFIR. Thermal architecture that goes far beyond the sunshades developed for the James Webb Space Telescope will be necessary to achieve the desired sensitivity of SAFIR. By combining active and passive cooling at critical points within the observatory, a significant reduction of the required level of active cooling can be obtained.

infrared telescopes↗
DOE OSTI2022

Distributed Training and Optimization of Neural Networks

Deep learning models are yielding increasingly better performances thanks to multiple factors. To be successful, model may have large number of parameters or complex architectures and be trained on large dataset. This leads to large requirements on computing resource and turn around time, even more so when hyperparameter optimization is done (e.g. search over model architectures). While this is a challenge that goes beyond particle physics, we review the various ways to do the necessary computations in parallel, and put it in the context of high-energy physics.

Vlimant, Jean-Roch↗
NASA NTRS2010

Integration of Advanced Concepts and Vehicles Into the Next Generation Air Transportation System: Introduction, Key Messages, and Vehicle Attributes - Volume 1

Raytheon, in partnership with NASA, is leading the way in ensuring that the future air transportation continues to be a key driver of economic growth and stability and that this system provides an environmentally friendly, safe, and effective means of moving people and goods. A Raytheon-led team of industry and academic experts, under NASA contract NNA08BA47C, looked at the potential issues and impact of introducing four new classes of advanced aircraft into the next generation air transportation system -- known as NextGen. The study will help determine where NASA should further invest in research to support the safe introduction of these new air vehicles. Small uncrewed or unmanned aerial systems (SUAS), super heavy transports (SHT) including hybrid wing body versions (HWB), very light jets (VLJ), and supersonic business jets (SSBJ) are the four classes of aircraft that we studied. Understanding each vehicle's business purpose and strategy is critical to assessing the feasibility of new aircraft operations and their impact on NextGen's architecture. The Raytheon team used scenarios created by aviation experts that depict vehicles in year 2025 operations along with scripts or use cases to understand the issues presented by these new types of vehicles. The information was then mapped into the Joint Planning and Development Office's (JPDO s) Enterprise Architecture to show how the vehicles will fit into NextGen's Concept of Operations. The team also identified significant changes to the JPDO's Integrated Work Plan (IWP) to optimize the NextGen vision for these vehicles. Using a proven enterprise architecture approach and the JPDO s Joint Planning Environment (JPE) web site helped make the leap from architecture to planning efficient, manageable and achievable. Very Light Jets flying into busy hub airports -- Supersonic Business Jets needing to climb and descend rapidly to achieve the necessary altitude Super-heavy cargo planes requiring the shortest common flight path -- are just a few of the potential new operations in the future National Airspace System. To assess the impact of these new scenarios on overall national airspace operations, the Raytheon team used the capabilities of a suite of tools such as NASA's Airspace Concepts Evaluation System (ACES), the Flight Optimization System (FLOPS), FAA's Aviation Environmental Design Tool (AEDT), Intelligent Automations Kinematic Trajectory Generator (KTG) and the Aviation Safety Risk Model (ASRM). Detailed metroplex modeling, surface delay models for super heavy transports, prioritized routing and corridors for supersonics business jets, and VLJ demand models are some of the models developed by the Raytheon team to study the effect of operating these new vehicles in the future NAS. Using this suite of models, several trade studies were conducted to evaluate these effects in terms of delays, equity in access, safety, and the environment. Looking at the impact of each vehicle, a number of critical issues were identified. The Raytheon team concluded that strict compliance to NextGen's 4-dimensional trajectory (4DT) management will be required to accommodate these vehicles unique operations and increased number of flights in the future air space system. The next section provides a discussion of this and the other key findings from our study.

Zellweger, Andres↗
DOE OSTI2023

Co-Design of Multijunction Photoelectrochemical Devices for Unassisted CO 2 Reduction to Multicarbon Products

Photoelectrochemical (PEC) CO 2 reduction (PEC CO 2 R) is a prospective approach for utilizing solar energy to synthesize a variety of carbon-containing chemicals and fuels, the most valuable of which are multicarbon (C 2+ ) products, such as ethylene and ethanol. While these products can be produced with high faradaic efficiency using Cu, this occurs over a relatively narrow potential range, which, in turn, imposes constraints on the design of a device for PEC CO 2 R. Herein, we used continuum-scale modeling to simulate the solar-to-C 2+ (STC 2+ ) efficiency of PEC CO 2 R devices fed with CO 2 -saturated, 0.1 M CsHCO 3 . We then explored how cell architecture and the use of single or dual photoelectrode(s) alters the optimal combination of photoelectrode bandgaps for high STC 2+ efficiency. Ultimately, this work provides guidance for the co-design of the device architecture and photoelectrode bandgaps required to achieve high STC 2+ efficiency. The insights gained are then used to identify systems that yield the highest amount of C 2+ products throughout the day and year.

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

Design-Technology Co-Optimization for NVM-based Neuromorphic Processing Elements

An emerging use-case of machine learning (ML) is to train a model on a high-performance system and deploy the trained model on energy-constrained embedded systems. Neuromorphic hardware platforms, which operate on principles of the biological brain, can significantly lower the energy overhead of a machine learning inference task, making these platforms an attractive solution for embedded ML systems. In this paper, we present a design-technology tradeoff analysis to implement such inference tasks on the processing elements (PEs) of a Non-Volatile Memory (NVM)-based neuromorphic hardware. Through detailed circuit-level simulations at scaled process technology nodes, we show the negative impact of technology scaling on the information-processing latency, which impacts the quality-of-service (QoS) of an embedded ML system. At a finer granularity, the latency inside a PE depends on 1) the delay introduced by parasitic components on its current paths, and 2) the varying delay to sense different resistance states of its NVM cells. Based on these two observations, we make the following three contributions. First, on the technology front, we propose an optimization scheme where the NVM resistance state that takes the longest time to sense is set on current paths having the least delay, and vice versa, reducing the average PE latency, which improves the QoS. Second, on the architecture front, we introduce isolation transistors within each PE to partition it into regions that can be individually power-gated, reducing both latency and energy. Finally, on the system-software front, we propose a mechanism to leverage the proposed technological and architectural enhancements when implementing a machine-learning inference task on neuromorphic PEs of the hardware. Evaluations with a recent neuromorphic hardware architecture show that our proposed design-technology co-optimization approach improves both performance and energy efficiency of machine-learning inference tasks without incurring high cost-per-bit.

42 ENGINEERING↗