Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “grid 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.

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

At least 613 records · Page 34

Design of the Cruise and Flap Airfoil for the X-57 Maxwell Distributed Electric Propulsion Aircraft

A computational and design study on an airfoil and high-lift flap for the X-57 Maxwell Distributed Electric Propulsion (DEP) testbed aircraft was conducted. The aircraft wing sizing study resulted in a wing area of 66.67 sq ft and aspect ratio of 15 with a design requirement of V(stall) = 58 KEAS, at a gross weight of 3,000 lb. To meet this goal an aircraft C(L,max) of 4.0 was required. The design cruise condition is 150 KTAS at 8,000 ft. This resulted in airfoil requirements of c(l) is approximately 0.90 for the cruise condition at Re = 2.35 x 10 (exp 6). A flapped airfoil with a c(l,max) of approximately 2.5 or greater, at Re = 1.0 x 10 (exp 6), was needed to have enough lift to meet the stall requirement with the DEP system. MSES computational analyses were conducted on the GAW-1, GAW-2, and the NACA 5415 airfoil sections, however they had limitations in either high drag or low c(l,max) on the cruise airfoil, which was the impetus for a new design. A design was conducted to develop a low drag airfoil for the X-57 cruise conditions with high c(l,max). The final design was the GNEW5BP93B airfoil with a minimum drag coefficient of c(d) = 0.0053 at c(l) = 0.90 and achieved laminar flow back to 69% chord on the upper surface and 62% chord on the lower surface. With fully turbulent flow, the drag increases to c(d) = 0.0120. The predicted maximum lift with turbulent flow is a c(l,max) of 1.95 at alpha = 19 deg. The airfoil is characterized by relatively flat pressure gradient regions on both surfaces at alpha = 0 deg, and aft camber to get extra lift out of the lower surface concave region. A 25% chord slotted flap was designed and analyzed with MSES for a 30 deg flap deflection. Additional 30 deg and 40 deg flap deflection analyses for two flap positions were conducted with USM3D using several turbulence models, for two angles of attack, to assess near c(l,max) with varied flap position. The maximum c(l) varied between 2.41 and 3.35. An infinite-span powered high-lift study was conducted on a GAW-1 constant chord 40 deg flapped airfoil section with FUN3D to quantify the airfoil lift increment that can be expected from a DEP system. The 16.7 hp/propeller blown wing increases the maximum C(L) from 3.45 to C(L) = 6.43, which is an effective q ratio of 1.86. This indicates that if the unblown high-lift flapped airfoil of the X-57 airplane achieves a c(l,max) of 2.78, then the high-lift augmentation blowing could yield a sectional lift coefficient of approximately 4.95 at c(l,max). Finally, a computational study was conducted with FUN3D on an infinite-span constant chord GAW-1 cruise airfoil to determine the impact of high-lift propeller diameter to wing chord ratio on the lift increment of the DEP system. A constant diameter propeller and nacelle size were used in the study. Three computational grids were made with airfoil chords of 0.5*chord, 1.0*chord, and 2.0*chord. Results of the propeller diameter to wing chord ratio study indicated that the blown to unblown C(L) ratio increased as the chord was decreased. However, because of the increase in relative size of the high-lift nacelle to the wing, which impacted wing lift performance, the study indicated that a propeller diameter to wing chord ratio of 1.0 gives the overall best maximum lift on the wing with the DEP system.

Viken, Jeffrey K.↗

Recent Progress in OVERFLOW Convergence Improvements

Improvements have been made to the implicit symmetric successive overrelaxation algorithm in the OVERFLOW 2.3 structured, overset grid, computational fluid dynamics flow solver. These improvements, consisting of implicit boundary conditions, improved flux Jacobian linearizations, and CFL number ramping, are a series of evolutionary changes to the linear solver that have resulted in increased nonlinear convergence rates and faster time to solution. A series of test cases are presented that demonstrate the effect of the changes through comparison with the original SSOR path and other linear solver implementations within OVERFLOW.

Joseph M Derlaga↗

Coupling a Particle Trajectory Capability to the US3D CFD Flow Solver

The purpose of the current research effort is to incorporate particle trajectory equations into the US3D CFD flow solver. The integrated code will allow higher-fidelity modeling of particle-fluid interactions and allow for the simulation of physical effects such as surface heating augmentation due to particle impacts. The key elements underpinning the current work are a robust tracking algorithm that can identify the encompassing mesh elements for individual particles and a point to point MPI frame-work. These developments combined with accurate property models and high-order time-stepping deliver realistic simulations on larger computational grids for a wide range of particle mass loadings. The coupled US3D-particle code will be applied to the Schiarparelli capsule and Earth-based experimental test cases.

A Sahai↗

Computational Evaluation of an OML-based Heat Exchanger Concept for HEATheR

Feasibility of outer mold line (OML) heat exchangers for electrified aircraft is computationally explored for three different aircraft concepts within The High-efficiency Electrified Aircraft Thermal Research (HEATheR) project. OML surface temperature limits were selected based on aluminum-alloy and carbon-fiber composite materials commonly used in transport aircraft. Heat flux distribution over the aircraft OML at the surface temperature limit was evaluated with computational fluid dynamics (CFD) analysis. The candidate OML locations and the size for the implementation of the OML-based heat exchanger is down-selected using this initial analysis and other considerations including the concern for structural integrity and proximity to heat-generating electrical components. The sensitivity of the heat transfer analysis to the computational grid, aircraft angle of attack, surface temperature, and the interaction of individual OML heat exchanger patches were investigated.

CAS↗

Performance and Portability of a Linear Solver Across Emerging Architectures

A linear solver algorithm used by a large-scale unstructured-grid computational fluid dynamics application is examined for a broad range of familiar and emerging architectures. Efficient implementation of a linear solver is challenging on recent CPUs offering vector architectures. Vector loads and stores are essential to effectively utilize available memory bandwidth on CPUs, and maintaining performance across different CPUs can be difficult in the face of varying vector lengths offered by each. A similar challenge occurs on GPU architectures, where it is essential to have coalesced memory accesses to utilize memory bandwidth effectively. In this work, we demonstrate that restructuring a computation, and possibly data layout, with regard to architecture is essential to achieve optimal performance by establishing a performance benchmark for each target architecture in a low level language such as vector intrinsics or CUDA. In doing so, we demonstrate how a linear solver kernel can be mapped to Intel® Xeon™ and Xeon Phi™, Marvell® ThunderX2®, NEC® SX-Aurora™ TSUBASA Vector Engine, and NVIDIA® and AMD® GPUs. We further demonstrate that the required code restructuring can be achieved in higher level programming environments such as OpenACC, OCCA, and Intel® OneAPI™/SYCL, and that each generally results in optimal performance on the target architecture. Relative performance metrics for all implementations are shown, and subjective ratings for ease of implementation and optimization are suggested.

Programming models↗

Implicit Thermochemical Nonequilibrium Flow Simulations on Unstructured Grids using GPUs

Thermochemical nonequilibrium flow simulation capabilities have been previously implemented, verified, and validated for central processing unit (CPU) systems in NASA’s unstructured-grid computational fluid dynamics solver FUN3D. Many exascale-class high-performance computing systems will rely on graphics processing unit (GPU) architectures for high throughput and energy efficiency; thus, CPU-based scientific computing software unable to effectively utilize these systems must be updated. In this work, we present a CUDA C++ implementation of FUN3D’s thermochemical nonequilibrium flow simulation capabilities targeting NVIDIA Tesla GPUs. An overview of the porting and optimization strategy is described and performance comparisons with other recent architectures are presented. Scaling to thousands of GPUs is demonstrated, yielding computational performance equivalent to that of several million CPU cores. The implementation enables efficient, high-fidelity, scale-resolving simulations of thermochemical nonequilibrium flows for many applications including atmospheric entry, hypersonics, and combustion.

GPU↗

A Parallel Symmetric Successive Overrelaxation Method for OVERFLOW

The block Jacobi symmetric successive overrelaxation (SSOR) algorithm has been reformulated as a parallelized algorithm for the OVERFLOWstructured, overset grid, computational fluid dynamics flow solver. Simple changes to the flow solver required to implement the algorithm are discussed. A series of test cases are presented that demonstrate how the addition of implicit overset boundaries has improved the robustness and nonlinear convergence characteristics of the flow solver.

Computational Fluid Dynamics↗

High-Order Finite-Difference Large-Eddy Simulations of a Supersonic Nozzle

Large-eddy simulations of the flow emanating from supersonic nozzles were carried out for the 6th Propulsion Aerodynamics Workshop to be held at the AIAA SciTech Conference. The purpose of the work is to detail the ability of a high-order of accuracy finite-difference based numerical scheme to accurately compute this flow. Results for both a perfectly expanded and over-expanded flow case are presented. Several complicating factors are discussed: 1) the limitations on the computational grid imposed by the numerical method, 2) the effect of additional dissipation necessary to resolve shock waves and 3) the effect of the nozzle boundary layer on the flow. Results indicate that the simulations replicate the basic flow physics, but there is room for improvement in accuracy. Areas for future study are outlined.

turbulence↗

Implicit Large-Eddy Simulations of Compressible Mixing Layers

Implicit large-eddy simulations of the self-similar regions of two compressible mixing layers at high Reynolds number and convective Mach numbers of 0.381 and 0.690 were carried out. Experimental data was used as the input into a synthetic eddy method turbulent inflow boundary condition to initiate simulations. Computational grids with nearly isotropic cells were used and the grid counts ranged from 24 to 270 million points. Agreement with experimental velocity and Reynolds stress data was good given the simplifications of the simulation. The decrease in mixing layer growth rate with increasing compressibility compares well with the literature. Budgets for the Reynolds stress transport equation were extracted from the simulations. It was demonstrated that the grids used did not resolve the dissipation term at the experimental Reynolds number. However, low Reynolds number simulations indicate that the numerical dissipation replaces the physical dissipation and the dissipation term in the transport equation can be represented by the summation term that quantifies the imbalance in the budget. The budgets for the high Reynolds number mixing layers agree well with previous low Reynolds number studies; indicating that the primary mechanism for the reduced growth rate and increased anistropy of the Reynolds stress tensor due to compressibility is a reduction in the pressure-strain and production terms. A new scaling based on the magnitude of the Reynolds stress tensor is proposed. This scaling provides better relative comparisons between the data and shows that the decrease in pressure-strain in the shear component is the primary driver in reducing the transverse normal stress.

turbulence↗

Assessment of UMUSCL Scheme for DNS of Turbulent Flows

Direct Numerical Simulations (DNS) are performed using the FUN3D code ( https://fun3d.larc.nasa.gov) for three validation cases: (1) flow through a plane channel, (2) flow through a channel with a constriction, and (3) flow over a flat plate. FUN3D is a node-centered finite-volume code developed at the NASA Langley Research Center that solves the three-dimensional compressible Navier-Stokes equations on unstructured computational grids. The simulations are performed employing the 2nd-order unstructured monotonic upstream scheme for conservation laws (UMUSCL). The results are compared with available experimental and numerical data. The effect of the UMUSCL reconstruction parameter (κ) is assessed, and the results indicate that κ = 0.9 yields satisfactory results in terms of accuracy and robustness compared to available data. Further analyses of the results, along with additional test cases and grids will be presented in the final manuscript.

Direct Numerical Simulation↗

A GridFTP transport driver for Globus XIO.

GridFTP is a high-performance, reliable data transfer protocol optimized for high-bandwidth wide-area networks. Based on the Internet FTP protocol, it defines extensions for high-performance operation and security. The Globus implementation of GridFTP provides a modular and extensible data transfer system architecture suitable for wide area and high-performance environments. GridFTP is the de facto standard in projects requiring secure, robust, high-speed bulk data transport. For example, the high energy physics community is basing its entire tiered data movement infrastructure for the Large Hadron Collider computing Grid on GridFTP; the Laser Interferometer Gravitational Wave Observatory routinely uses GridFTP to move 1 TB a day during production runs; and GridFTP is the recommended data transfer mechanism to maximize data transfer rates on the TeraGrid. Commonly used GridFTP clients include globus-url-copy, uberftp, and the Globus Reliable File Transfer service. In this paper, we present a Globus XIO based client to GridFTP that provides a simple Open/Close/Read/Write (OCRW) interface to the users. Such a client greatly eases the addition of GridFTP support to third-party programs, such as SRB and MPICH-G2. Further, this client provides an easier and familiar interface for applications to efficiently access remote files. We compare the performance of this client with that of globus-url-copy on multiple endpoints in the TeraGrid infrastructure. We perform both memory-to-memory and disk-to-disk transfers and show that the performance of this OCRW client is comparable to that of globus-url-copy. We also show that our GridFTP client significantly outperforms the GPFS WAN on the TeraGrid.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Welcome

A Welcome presentation for CERN, The Globe, following a successful first test of the LHC computing grid on September 10, 2008. A mention of issues with the grid is also included.

Source record↗

Simulating the Effects of Stellarator Geometry on Gyrokinetic Drift-Wave Turbulence

Nuclear fusion is a clean, safe form of energy with abundant fuel. In magnetic fusion energy (MFE) experiments, the plasma fuel is confined by magnetic fields at very high temperatures and densities. One fusion reactor design is the non-axisymmetric, torus-shaped stellarator. Its fully-3D fields have advantages over the simpler, better-understood axisymmetric tokamak, including the ability to optimize magnetic configurations for desired properties, such as lower transport (longer confinement time). Turbulence in the plasma can break MFE confinement. While turbulent transport is known to cause a significant amount of heat loss in tokamaks, it is a new area of research in stellarators. Gyrokinetics is a good mathematical model of the drift-wave instabilities that cause turbulence. Multiple gyrokinetic turbulence codes that had great success comparing to tokamak experiments are being converted for use with stellarator geometry. This thesis describes such adaptations of the gyrokinetic turbulence code, GS2. Herein a new computational grid generator and upgrades to GS2 itself are described, tested, and benchmarked against three other gyrokinetic codes. Using GS2, detailed linear studies using the National Compact Stellarator Experiment (NCSX) geometry were conducted. The first compares stability in two equilibria with different beta=(plasma pressure)/(magnetic pressure). Overall, the higher beta case was more stable than the lower beta case. As high beta is important for MFE experiments, this is encouraging. The second compares NCSX linear stability to a tokamak case. NCSX was more stable with a 20% higher critical temperature gradient normalized by the minor radius, suggesting that the fusion power might be enhanced by about 50%. In addition, the first nonlinear, non-axisymmetric GS2 simulations are presented. Finally, linear stability of two locations in a W7-AS plasma were compared. The experimentally-measured parameters used were from a W7-AS shot in which measured heat fluxes match neoclassical theory predictions at inner radii, but are too large for neoclassical predictions at outer radii. Results from GS2 linear simulations show that the outer location has higher gyrokinetic instability growth rates than at the inner one. Mixing-length estimates of the heat flux are within a factor of 3 of the experimental measurements, indicating that gyrokinetic turbulence may be responsible for the higher transport measured by the experiment in the outer regions. Future nonlinear simulations can explore this question in more detail.

Gyrokinetics↗

A New Fast Fluid Dynamics Model for Data-Center Floor Plenums

A raised-floor plenum is often used to distribute cooling airflow in data centers. Traditional Computational Fluid Dynamics (CFD) can accurately model airflow distribution in plenums, but it is expensive (computationally and financially) and is, therefore, often not accessible to the data center designer or operator. This paper presents an alternative CFD methodology based on Fast Fluid Dynamics (FFD), which is faster, more parallelizable, and simpler to program. The improvement in computational speed afforded by FFD is particularly important for multiple-iteration optimization, transient simulations, or applications for which near-real-time performance is required. Here, the potential speed improvements of FFD have been widely documented; however, it has been generally accepted that this benefit is accompanied by a modest loss of accuracy relative to traditional CFD. We show here that this is not necessarily true. Plenum airflow predictions by our FFD implementation, which utilizes a first-order upwind rather than Semi-Lagrangian advection scheme of previous research, are extremely close to those predicted by traditional CFD - provided the same turbulence model and computational grid are used consistently.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

FTS3: Data Movement Service in containers deployed in OKD

The File Transfer Service (FTS3) is a data movement service developed at CERN which is used to distribute the majority of the Large Hadron Collider's data across the Worldwide LHC Computing Grid (WLCG) infrastructure. At Fermilab, we have deployed FTS3 instances for Intensity Frontier experiments (e.g. DUNE) to transfer data in America and Europe, using a container-based strategy. In this article we summarize our experience building docker images based on work from the SLATE project (slateci.io) and deployed in OKD, the community distribution of Red Hat OpenShift. Additionally, we discuss our method of certificate management and maintenance utilizing Kubernetes CronJobs. Finally, we also report on the two different configurations currently running at Fermilab, comparing and contrasting a Docker-based OKD deployment against a traditional RPM-based deployment.

Lobato Pardavila, Lorena↗