SEARCH · Search NASA

Results for “high performance and distributed computing”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,045 records · Page 58

All sources NASA NTRS DOE OSTI

DOE OSTI2018

OpenMP in VASP: Threading and SIMD

Abstract The Vienna Ab initio Simulation Package (VASP) is a widely used electronic structure code that originally exploits process‐level parallelism through the Message Passing Interface (MPI) for work distribution within and across nodes. Architectural changes of modern parallel processors urge programmers to address thread‐ and data‐level parallelism as well to benefit most from the available compute resources within a node. We describe for VASP how to approach for an MPI + OpenMP parallelization including data‐level parallelism through OpenMP SIMD constructs together with a generic high‐level vector coding scheme. We can demonstrate an improved scalability of VASP and more than 20% gain over the MPI‐only version as well as a 2× increased performance of collective operations using the multiple‐endpoint MPI feature. The high‐level vector coding scheme applied to VASP's general gradient approximation routine gives up 9× performance gain on AVX512 platforms with the Intel compiler.

Wende, Florian↗

DOE OSTI2022

Identification of efficient and stable solid-solid electrochemical interfaces for Li-S batteries. Final Report

The objective of this work is to develop a better understanding of the interactions between battery components that drive the observed battery performance. Using computational tools at the atomistic level, we elucidate molecular, crystal, and amorphous structures, interfacial structures and reactions, and electrochemical behavior. We focus on Li metal anodes, liquid electrolytes, solid electrolytes, S cathodes, and metal-oxide cathodes, and their corresponding interfaces. Within liquid electrolytes, we study their bulk structure and dynamics, as a function of salt concentration. The analysis allows to understand the formation of localized high concentrated regions as the solvent is substituted by a weakly interacting diluent species. Analysis of the electron distribution yields the reactive behavior at interfaces. Moreover, we employ tools that allow the characterization of ion and electron transport at interfaces, during the processes of solvation and desolvation occurring while cycling. For Sulfur cathodes we elucidate synthesis processes and the role of carbon-sulfur and nitrogen-sulfur interactions not only to provide electronic conductivity but also to define alternative reaction pathways during discharge of Li-S batteries. Full-cell simulations offer new insights regarding the integration of all the battery components. Our studies also include solid electrolytes and their interfaces with Li metal, S-C, and metal-oxide electrodes.

25 ENERGY STORAGE↗

DOE OSTI2024

Experimental and Theoretical Evaluation of Feed Flow Collar Design for Shell Fed Hollow Fiber Membrane Modules

An experimental and theoretical study of module collar design is presented here. Hollow fiber membranes are prepared by dip coating a poly(vinylidene) (PVDF) support with a polydimethylsiloxane (PDMS) gutter layer and a Pebax 2533 selective layer. Fiber bundles with a well-defined fiber packing are prepared using a 3D printed module. A parallel fiber bundle consisting of 4-9 uniformly spaced fibers is created with printed tabs that align the fibers and create a tubesheet. The tabs are sealed within a printed case that possesses a series of external ports for gas introduction and removal. Uniquely, both port location and the use of a collar to assist fluid distribution in the shell can be varied for the same fiber bundle. Experimental measurements are compared to computational fluid dynamics (CFD) simulations. The experimental module design allows high-fidelity representation of the fiber bundle and module case in the simulations. Comparisons between experiment and simulation are in good agreement over a broad range of experimental conditions. The detrimental effect of having ports located too close, leading to stagnation regions, is captured as well as the beneficial effects of using a collar for shell-side fluid distribution around the fiber bundle. Such results help validate the use of CFD to develop high-performance module designs.

Tran, Thien↗

DOE OSTI2023

Bayes goes fast: Uncertainty quantification for a covariant energy density functional emulated by the reduced basis method

A covariant energy density functional is calibrated using a principled Bayesian statistical framework informed by experimental binding energies and charge radii of several magic and semi-magic nuclei. The Bayesian sampling required for the calibration is enabled by the emulation of the high-fidelity model through the implementation of a reduced basis method (RBM)—a set of dimensionality reduction techniques that can speed up demanding calculations involving partial differential equations by several orders of magnitude. The RBM emulator we build—using only 100 evaluations of the high-fidelity model—is able to accurately reproduce the model calculations in tens of milliseconds on a personal computer, an increase in speed of nearly a factor of 3,300 when compared to the original solver. Besides the analysis of the posterior distribution of parameters, we present model calculations for masses and radii with properly estimated uncertainties. We also analyze the model correlation between the slope of the symmetry energy L and the neutron skin of 48 Ca and 208 Pb. The straightforward implementation and outstanding performance of the RBM makes it an ideal tool for assisting the nuclear theory community in providing reliable estimates with properly quantified uncertainties of physical observables. Such uncertainty quantification tools will become essential given the expected abundance of data from the recently inaugurated and future experimental and observational facilities.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

NASA NTRS1985

Relative advantages of thin-layer Navier-Stokes and interactive boundary-layer procedures

Numerical procedures for solving the thin-shear-layer Navier-Stokes equations and for the interaction of solutions to inviscid and boundary-layer equations are described and evaluated. To allow appraisal of the numerical and fluid dynamic abilities of the two schemes, they have been applied to one airfoil as a function of angle of attack at two slightly different Reynolds numbers. The NACA 0012 airfoil has been chosen because it allows comparison with measured lift, drag, and moment and with surface-pressure distributions. Calculations have been performed with algebraic eddy-viscosity formulations, and they include consideration of transition. The results are presented in a form that allows easy appraisal of the accuracy of both procedures and of the relative costs. The interactive procedure is computationally efficient but restrictive relative to the thin-layer Navier-Stokes procedure. The latter procedure does a better job of predicting drag than does the former. In both procedures, the location of transition is crucial for accurate or detailed computations, particularly at high angles of attack. When the upstream influence of pressure field through the shear layer is important, the thin-layer Navier-Stokes procedure has an edge over the interactive procedure.

Mehta, U.↗

NASA NTRS1986

A comparison of interactive boundary-layer and thin-layer Navier-Stokes procedures

Mehta, U.↗

NASA NTRS1989

An advanced probabilistic structural analysis method for implicit performance functions

In probabilistic structural analysis, the performance or response functions usually are implicitly defined and must be solved by numerical analysis methods such as finite element methods. In such cases, the most commonly used probabilistic analysis tool is the mean-based, second-moment method which provides only the first two statistical moments. This paper presents a generalized advanced mean value (AMV) method which is capable of establishing the distributions to provide additional information for reliability design. The method requires slightly more computations than the second-moment method but is highly efficient relative to the other alternative methods. In particular, the examples show that the AMV method can be used to solve problems involving non-monotonic functions that result in truncated distributions.

Wu, Y.-T.↗

NASA NTRS1990

Scattering of sound by atmospheric turbulence predictions in a refractive shadow zone

According to ray theory, regions exist in an upward refracting atmosphere where no sound should be present. Experiments show, however, that appreciable sound levels penetrate these so-called shadow zones. Two mechanisms contribute to sound in the shadow zone: diffraction and turbulent scattering of sound. Diffractive effects can be pronounced at lower frequencies but are small at high frequencies. In the short wavelength limit, then, scattering due to turbulence should be the predominant mechanism involved in producing the sound levels measured in shadow zones. No existing analytical method includes turbulence effects in the prediction of sound pressure levels in upward refractive shadow zones. In order to obtain quantitative average sound pressure level predictions, a numerical simulation of the effect of atmospheric turbulence on sound propagation is performed. The simulation is based on scattering from randomly distributed scattering centers ('turbules'). Sound pressure levels are computed for many realizations of a turbulent atmosphere. Predictions from the numerical simulation are compared with existing theories and experimental data.

Mcbride, Walton E.↗

NASA NTRS2005

Implementation of a High-Speed FPGA and DSP Based FFT Processor for Improving Strain Demodulation Performance in a Fiber-Optic-Based Sensing System

NASA's Aviation Safety and Security Program is pursuing research in on-board Structural Health Management (SHM) technologies for purposes of reducing or eliminating aircraft accidents due to system and component failures. Under this program, NASA Langley Research Center (LaRC) is developing a strain-based structural health-monitoring concept that incorporates a fiber optic-based measuring system for acquiring strain values. This fiber optic-based measuring system provides for the distribution of thousands of strain sensors embedded in a network of fiber optic cables. The resolution of strain value at each discrete sensor point requires a computationally demanding data reduction software process that, when hosted on a conventional processor, is not suitable for near real-time measurement. This report describes the development and integration of an alternative computing environment using dedicated computing hardware for performing the data reduction. Performance comparison between the existing and the hardware-based system is presented.

Farley, Douglas L.↗

DOE OSTI2021

Can artificial intelligence and data-driven machine learning models match or even replace process-driven hydrologic models for streamflow simulation?: A case study of four watersheds with different hydro-climatic regions across the CONUS

With recent developments in computational techniques, Data-driven Machine Learning Models (DMLs) have shown great potential in simulating streamflow and capturing the rainfall-runoff relationship in given watersheds, which are traditionally fulfilled by Process-based Hydrologic Models (PHMs). There are debates on whether the DMLs can outperform and possibly replace the classical PHMs for streamflow simulation and river forecasting, but no clear conclusions have been made. This study aims to investigate whether the newer DMLs have any potential in further improving the simulation accuracy of classical PHMs, and vice versa. To do this, we compared a few popular PHMs and DMLs over four watersheds across the Continental US (CONUS) that are associated with different input, climate, and regional conditions. A total of five hydrologic models were chosen, including (1) two classical lumped models, i.e., the Sacramento Soil Moisture Accounting (SAC-SMA) and Xinanjiang (XAJ); (2) one modern distributed model, termed Coupled Routing and Excess Storage (CREST); (3) and two DMLs including an Artificial Neural Networks (ANN) and a deep learning model, termed Long Short Term Memory (LSTM). Our results demonstrated that the DMLs still significantly biased when using the baseline input scenario with the PHMs. However, the DMLs fed with delayed input scenarios had great potential and can reach high simulation accuracy. The DMLs, especially the ANN, outperformed other employed models under the rainfall-runoff relationship in which rainfall dominantly drives. Furthermore, the DMLs also showed better performance in the high-flow regime, while the PHMs had a better performance for the low-flow regime, implying both PHMs and DMLs have their own merits and are worthy of joint development. In general, our study indicated a great potential of using DMLs to simulate streamflow, but further studies are still needed to verify the transferability and scalability of DMLs in large-scale experiments, such as the Distributed Model Intercomparison Projects 1&2 conducted by National Weather Services but to compare modern DMLs and PHMs.

58 GEOSCIENCES↗

DOE OSTI2017

High temperature high velocity direct power extraction using an open-cycle oxy-combustion system

The implementation of oxy-fuel technology in fossil-fuel power plants may contribute to increased system efficiencies and a reduction of pollutant emissions. One technology that has potential to utilize the temperature of undiluted oxy-combustion flames is open-cycle magnetohydrodynamic (MHD) power generators. These systems can be configured as a topping cycle and provide high enthalpy, electrically conductive flows for direct conversion of electricity. This report presents the design and modeling strategies of a MHD combustor operating at temperatures exceeding 3000 K. Throughout the study, computational fluid dynamics (CFD) models were extensively used as a design and optimization tool. A lab-scale 60 kWth model was designed, manufactured and tested as part of this project. A fully-coupled numerical method was developed in ANSYS FLUENT to characterize the heat transfer in the system. This study revealed that nozzle heat transfer may be predicted through a 40% reduction of the semi-empirical Bartz correlation. Experimental results showed good agreement with the numerical evaluation, with the combustor exhibiting a favorable performance when tested during extended time periods. A transient numerical method was employed to analyze fuel injector geometries for the 60-kW combustor. The ANSYS FLUENT study revealed that counter-swirl inlets achieve a uniform pressure and velocity ratio when the ports of the injector length to diameter ratio (L/D) is 4. An angle of 115 degrees was found to increase distribution efficiency. The findings show that this oxy-combustion concept is capable of providing a high-enthalpy environment for seeding, in order to render the flow to be conductive. Based on previous findings, temperatures in the range of 2800-3000 K may enable magnetohydrodynamic power extraction. The heat loss fraction in this oxy-combustion system, based on CFD and analytical calculations, at optimal operating conditions, was estimated to be less than 10 percent. Furthermore, the heat transfer design removed approximately 7 MW/m2. The results observed in the lab-scale system were employed to develop a 1-MW scaled prototype. Scaling methods were based on critical design criteria found in similar systems, aimed at replicating combustion flow fields and reducing possible instabilities. A numerical simulation of the combustor wall was developed for a combined thermal steady model and static structural model. This combined model was developed predict combined stress parameters within the wall during testing conditions. Both models were developed within ANSYS FEA software package. The relative accuracy presented as well major performance parameters are discussed to assess the design's validity and ensure safety. The scaled prototype was manufactured through selective laser melting (SLM)-based additive manufacturing to reduce lead times and increase geometrical complexity. Additional CFD models were developed to optimize coolant manifold system parameters and perform a parametric study on channel geometry. An investigation on coolant manifold geometry demonstrated improvements in channel flow distribution when enlarging manifold lengths and increasing the number of tubes feeding into the flow. A three-dimensional model based on a single channel was developed to capture the effect of variable properties and thermal stratification. All cases in the simulation exhibited higher wall temperatures and lower convective coefficients than those determined through 1-D analytical equations. This implies that pressure and velocity safety factors must be implemented during system operation. Overall, the findings made in this investigation are thought to be of value to researchers and industrial practitioners when designing oxy-fuel direct power extraction systems operating at temperatures exceeding 3000 K. In addition to this, the implementation of the developed technology at pilot and commercial scales could result in a significant improvement in the efficiencies of heritage and next-generation power cycles.

03 NATURAL GAS↗

DOE OSTI2019

Parameter inference with deep jointly informed neural networks

Abstract A common challenge in modeling inertial confinement fusion (ICF) experiments with computer simulations is that many of the simulation inputs are unknown and cannot be directly measured. Often, parameters that are measured in the experiment are used to infer the unknown inputs by solving the inverse problem: finding the set of simulation inputs that result in outputs consistent with the experimental observations. In ICF, this process is often referred to as a “post‐shot analysis.” Post‐shot analyses are challenging as the inverse problem is often highly degenerate, the input parameter space is vast, and simulations are computationally expensive. In this work, deep neural network models equipped with model uncertainty estimates are used to train inverse models, which map directly from output to input space, to find the distribution of post‐shot simulations that are consistent with experimental observations. The inverse model approach is compared to Markov chain Monte Carlo (MCMC) sampling of the forward model, which maps from input to output space, for parameter inference tasks of varying complexity. The inverse models perform best when searching vast parameter spaces for post‐shot simulations that are consistent with a large number of observables, where MCMC sampling can be prohibitively expensive. We demonstrate how augmenting inverse models with autoencoders enable the inclusion of several dozen observables in the inverse mapping, reducing the degeneracy of the model and improving the accuracy of the post‐shot analysis.

Humbird, Kelli D.↗

DOE OSTI2019

SimCCS: An open-source tool for optimizing CO 2 capture, transport, and storage infrastructure

Commercial-scale carbon capture and storage (CCS) technology will involve deploying infrastructure on a massive and costly scale. This effort will require careful and comprehensive planning to ensure that capture locations, storage sites, and the dedicated CO 2 distribution pipelines are selected in a robust and cost-effective manner. Presented in 2009, SimCCS is an optimization model for integrated system design that enables researchers, stakeholders, and policy makers to design CCS infrastructure networks. SimCCS 2.0 is a complete, ground-up redesign that is now a portable software package, useable and shareable by the CCS research, industrial, policy, and public communities. SimCCS 2.0 integrates multiple new capabilities including a refined optimization model, novel candidate network generation techniques, and optional integration with high-performance computing platforms. Accessing user-provided CO 2 source, sink, and transportation data, SimCCS 2.0 creates candidate transportation routes and formalizes an optimization problem that determines the most cost-effective CCS system design. This optimization problem is further solved either through a high-performance computing interface, or through third-party software on a local desktop computing platform. Finally, SimCCS 2.0 employs an open-access geographic information system framework to enable analysis and visualization capabilities. SimCCS 2.0 is written in Java and is publicly available via GitHub to encourage collaboration, modification, and community development.

42 ENGINEERING↗

DOE OSTI2025

Simultaneous inference of equation of state parameters and unknown data errors with uncertainty quantification via hierarchical Bayesian posterior maximization

Equations of state (EOSs) are a key component in running hydrodynamic simulations as they relate the thermodynamic states for the material. The Davis reactants EOS is commonly used for modeling high explosives (HEs), and the EOS model parameters are calibrated using material specific data. The calibrations are often performed with uncertainty quantification via Bayesian inference to account for uncertainty in the data and generate ensembles of likely parameters. However, there are relatively few HE data sets to use for calibration and many are historical and lack error information. In this work, we simultaneously calibrate the Davis reactants EOS model parameters and unknown data error terms for the high explosive PBX 9501. To quantify the uncertainty in the models and the data, we use a Bayesian framework for the calibration and compute the hierarchical Bayesian posterior distribution with both a posteriori maximization approach and Markov Chain Monte Carlo. In general, we find that, given our assumptions, the two approaches result in similar calibrated parameters, posterior covariance matrices, and insights about the parameters but that the posterior maximization requires far less computational resources.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2019

PaKman: Scalable Assembly of Large Genomes on Distributed Memory Machines

De novo genome assembly is a fundamental problem in the field of bioinformatics, that aims to assemble the DNA sequence of an unknown genome from numerous short DNA fragments (aka reads) obtained from it. With the advent of high-throughput sequencing technologies, billions of reads can be generated in a matter of hours, necessitating efficient parallelization of the assembly process. While multiple parallel solutions have been proposed in the past, conducting a large- scale assembly at scale remains a challenging problem because of the inherent complexities associated with data movement, and irregular access footprints of memory and I/O operations. In this paper, we present a novel algorithm, called PaKman, to address the problem of performing large-scale genome assemblies on a distributed memory parallel computer. Our approach focuses on improving performance through a combination of novel data structures and algorithmic strategies for reducing the communication and I/O footprint during the assembly process. A key aspect of our algorithm is its graph data structure, which comprises fat nodes (or what we call "macro-nodes") that reduce the communication burden during contig generation. We present an extensive performance and qualitative evaluation of our algorithm, including comparisons to other state-of-the-art parallel assemblers. Our results demonstrate the ability to achieve near-linear speedups on up to 8K cores (tested); outperform state-of-the-art distributed memory and shared memory tools in performance while delivering comparable (if not better) quality; and reduce time to solution significantly. For instance, PaKman is able to complete an assembly of the full human genome in just over a minute on 8K cores.

Ghosh, Priyanka↗

DOE OSTI2023

Integrated Large-Scale Data Management Platform for Photovoltaic Power Conversion Equipment (PCE) Reliability Data

To meet the demand for accuracy and real-time capability of PV system degradation evaluation, massive volume data is needed to run high-fidelity and high-efficiency simulations and perform advanced data analysis. However, PV farm operators have a series of difficulties with PV inverter data, such as data collection from multiple channels, massive data storage, data management and massive data analysis. To address these challenges, we developed an integrated data management platform capable of data acquisition, processing, storage, query, and performing big data analysis utilizing AI algorithms. The platform can also achieve data correctness verification and provide an effective distributed data management solution to retrieve massive data and establish a connection to distributed computational frameworks.

data management platform↗

DOE OSTI2018

On The Theory of Speculative Checkpointing: Time and Energy Considerations

Collective checkpoint-rollback is the most popular approach to dealing with fail-stop errors on high-performance computing (HPC) platforms. Prior work has focused on choosing checkpoint intervals that minimize the total cost of checkpoint-rollback. In this paper, we introduce the notion of speculative checkpointing, where we probabilistically skip some checkpoints. Careful choice of check- points to be taken or skipped has the potential to further reduce the total checkpoint-rollback overhead. We mathematically formulate the overall checkpoint-rollback cost in the presence of speculation. We consider the choice of speculation being a fixed probability or a probability distribution. We formulate two criteria to be minimized: total execution time and approximate total energy. We derive the criteria for beneficial speculative checkpointing for exponential and arbitrary failure distributions. Further, we analyze the joint optimization of energy and time to mathematically express the trade-offs.

Subasi, Omer↗

DOE OSTI2025

Design and Production for Maximum Structural Efficiency With Respect to Fiber Orientation With Increased Understanding of Hybrid Fiber Flow Behavior

Discontinuous fiber-reinforced thermoplastic composites have gained considerable attention in automotive, aerospace, and other industries, due to their high-rate of production combined with their ability to attain complex and intricate shapes. Among other high-rate thermoplastic manufacturing processes, injection-molding is one of the most common manufacturing methods due to fast production and high surface finishing of complex geometries. Fiber orientation in discontinuous fiber composites plays a pivotal role in determining the mechanical, electrical, and thermomechanical properties, underscoring the necessity to comprehend fiber orientation in injection molded parts. Among different fiber types, glass and carbon fibers are most common in the composite industries. The recent trend of hybrid composites comprising both glass fiber (GF) and carbon fiber (CF) is also gaining importance in the automotive industry. Hybrid fiber options allow designers to optimize the balance between glass and carbon fibers by leveraging the high durability and low cost of GF while the strength and lightweight properties of CF. Consequently, comparing the fiber oriented distribution (FOD) of injection molded composites containing GF, CF, and a hybrid of GF/CF is critical to investigating the local mechanical properties of intricate structures for high-end applications. In Phase I of this project, FOD in injection-molded panels with respect to distance from the gate was analyzed using X-ray computed tomography (X-CT) for GF, CF, and hybrid CF/GF (CGF) reinforced nylon 66. To understand the reason behind the FOD with different fiber types, computational fluid dynamics (CFD) and rheology were performed. Samples were extracted at three locations: near the gate, center, and opposite end. Thickness of the layers of typical skin-shell-core type FOD varies with fiber type and location. GF achieved flow direction alignment (in shell) earlier than viscous CF and CGF near the gate, whereas CF showed the highest flow-direction alignment at the center due to shear induced orientation. At the opposite end, GF experienced more backflow than others indicating faster mold filling owing to its lower viscosity. Hybrid CGF exhibited GF-dominated center and CF-dominated end region. The numerical model used to obtain FOD and rheological predictions for the CF and GF composites served to corroborate the trends observed in the experimental trials. The FOD responses across fiber types and location were reflected in their longitudinal and transverse properties. Only GF showed higher longitudinal modulus over transverse modulus near the gate attributed to rapid alignment, whereas CF and CGF exhibited opposite trend. However, fountain flow enhanced the longitudinal modulus over transverse modulus with the distance for all, particularly for CF. This study offers insights into mold filling behavior of different fibers which are critical in optimizing injection molding conditions for tailored final properties.

36 MATERIALS SCIENCE↗