SEARCH · Engineering Papers

Results for “Complex geometry”

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 1,693 records · Page 94

All sources NASA NTRS DOE OSTI

NASA NTRS2001

Unsteady Spherical Diffusion Flames in Microgravity

The absence of buoyancy-induced flows in microgravity (mu-g) and the resulting increase in the reactant residence time significantly alters the fundamentals of many combustion processes. Substantial differences between normal gravity (ng) and (mu-g) flames have been reported in experiments on candle flames, flame spread over solids, droplet combustion, and others. These differences are more basic than just in the visible flame shape. Longer residence times and higher concentration of combustion products in the flame zone create a thermochemical environment that changes the flame chemistry and the heat and mass transfer processes. Processes such as flame radiation, that are often ignored in ng, become very important and sometimes even controlling. Furthermore, microgravity conditions considerably enhance flame radiation by: (i) the build-up of combustion products in the high-temperature reaction zone which increases the gas radiation, and (ii) longer residence times make conditions appropriate for substantial amounts of soot to form which is also responsible for radiative heat loss. Thus, it is anticipated that radiative heat loss may eventually extinguish the "weak" (low burning rate per unit flame area) mu-g diffusion flame. Yet, space shuttle experiments on candle flames show that in an infinite ambient atmosphere, the hemispherical candle flame in mu-g will burn indefinitely. This may be because of the coupling between the fuel production rate and the flame via the heat-feedback mechanism for candle flames, flames over solids and fuel droplet flames. Thus, to focus only on the gas-phase phenomena leading to radiative extinction, aerodynamically stabilized gaseous diffusion flames are examined. This enables independent control of the fuel flow rate to help identify conditions under which radiative extinction occurs. Also, spherical geometry is chosen for the mu-g experiments and modeling because: (i) It reduces the complexity by making the problem one-dimensional; (ii) The spherical diffusion flame completely encloses the soot which is formed on the fuel rich side of the reaction zone. This increases the importance of flame radiation because now both soot and gaseous combustion products co-exist inside the high temperature spherical diffusion flame. (iii) For small fuel injection velocities, as is usually the case for a pyrolyzing solid, the diffusion flame in mu-g around the solid naturally develops spherical symmetry. Thus, spherical diffusion flames are of interest to fires in mu-g and identifying conditions that lead to radiation-induced extinction is important for spacecraft fire safety.

Atreya, Arvind↗

NASA NTRS2002

Computational Systems for Multidisciplinary Applications

In this paper, we briefly describe our efforts to develop complex simulation systems. We focus first on four key infrastructure items: enterprise computational services, simulation synthesis, geometry modeling and mesh generation, and a fluid flow solver for arbitrary meshes. We conclude by presenting three diverse applications developed using these technologies.

Soni, Bharat↗

NASA NTRS2006

Validation of Multi-Dimensional Stirling Engine Design Codes: Measurements in the 90-Degree Turn Test Section

The work to be presented herein was motivated largely by a desire to improve the understanding of oscillatory fluid mechanics inside a Stirling engine. To this end, a CFD project was undertaken at Cleveland State University with the goal of accurately predicting the fluid dynamics within an engine or engine component. Along with the CFD efforts, a code validation project was undertaken at the University of Minnesota. The material covered herein consists of four main parts. In section 1, an experimental investigation of a small aspect ratio impinging jet is discussed. Included in this discussion is a description of the test facilities and instrumentation. A presentation of the collected data is given and comments are made. Next, in section 2, a parallel experimental investigation is presented in which the same geometry as that of section 1 is used, but the flow conditions are changed from steady unidirectional flow to sinusoidally oscillating flow. In section Two, collected data are presented and comments are made. In section 3, a comparison is made between the results of sections 1 and 2, namely, sinusoidally oscillating flow results are compared to steady, unidirectional flow results from the same geometry. Finally, in section 4, a comparison is made between experimentally collected data (the main subject of this work) and CFD generated results. Furthermore, in appendix A, an introductory description of the primary measurement tool used in the experimental process the hot wire anemometer is given for the unfamiliar. The anemometer calibration procedure is described in appendix B. A portfolio of data reduction and data processing codes is provided in appendix C and lastly, a DVD and a roadmap of its contents is provided in an appendix D. 1.0 Unidirectional Flow Investigations 1.1 Introduction This unidirectional experimental program was undertaken to complement an oscillatory flow investigation conducted at the University of Minnesota. The oscillatory investigation is discussed thoroughly in section 2. We defer the description of the motivation behind these experiments until the introduction of section 2. The work that is discussed in this thesis began (chronologically) with oscillatory flow visualization experiments. It was decided that it would be valuable and important to investigate the flow under unidirectional conditions in the same geometry as that of the oscillatory experiments. The thought was that the unidirectional case would be less complicated to model with a CFD program (a moving boundary would be replaced with a steady state boundary condition). Thus, a series of unidirectional experiments were carried out to capture the important features of the flow within the test section. The purpose of these experiments was to provide a data set for comparison to CFD generated velocity fields. Hot-wire anemometry data were taken and flow visualization was conducted as a standard for code validation. The flow geometry was simple, such that it could be easily gridded in a CFD program. However, the geometry provided separation and transition zones, shear layers and recirculation zones. These characteristics made the flow complex and challenging for CFD computation. We comment that the order of experiments that produced this report is as follows: experimental flow visualization under oscillatory flow conditions was carried out; this was followed by unidirectional flow visualization and hot wire anemometry; finally, oscillatory hot wire anemometry was conducted. We present the results out of chronological order for the following reason: the unidirectional results are easier

Simon, Terrence W.↗

NASA NTRS2011

Co-Flow Hollow Cathode Technology

Hall thrusters utilize identical hollow cathode technology as ion thrusters, yet must operate at much higher mass flow rates in order to efficiently couple to the bulk plasma discharge. Higher flow rates are necessary in order to provide enough neutral collisions to transport electrons across magnetic fields so that they can reach the discharge. This higher flow rate, however, has potential life-limiting implications for the operation of the cathode. A solution to the problem involves splitting the mass flow into the hollow cathode into two streams, the internal and external flows. The internal flow is fixed and set such that the neutral pressure in the cathode allows for a high utilization of the emitter surface area. The external flow is variable depending on the flow rate through the anode of the Hall thruster, but also has a minimum in order to suppress high-energy ion generation. In the co-flow hollow cathode, the cathode assembly is mounted on thruster centerline, inside the inner magnetic core of the thruster. An annular gas plenum is placed at the base of the cathode and propellant is fed throughout to produce an azimuthally symmetric flow of gas that evenly expands around the cathode keeper. This configuration maximizes propellant utilization and is not subject to erosion processes. External gas feeds have been considered in the past for ion thruster applications, but usually in the context of eliminating high energy ion production. This approach is adapted specifically for the Hall thruster and exploits the geometry of a Hall thruster to feed and focus the external flow without introducing significant new complexity to the thruster design.

Hofer, Richard R.↗

NASA NTRS2019

Testbed Results for Scalar and Vector Radiative Transfer Computations of Light in Atmosphere-Ocean Systems

We generate and tabulate reflectance values of the Stokes parameters I, Q, and U of upwelling radiance just above a rough ocean surface and at the top of the atmosphere (TOA) for 100 scattering geometries, four atmosphere-ocean systems, and four wavelengths. The atmosphere-ocean systems increase in complexity from (a) a molecular atmosphere above a rough ocean surface (AOS-I model); to (b) a pure water body below a rough ocean surface (AOS-II model); to (c) a fully-coupled simple atmosphere-ocean system (AOS-III model) containing a molecular atmosphere, rough ocean surface, and pure water; to (d) a fully-coupled complex atmosphere-ocean system (AOS-IV model) that includes scattering by molecules, rough ocean surface, pure water, and hydrosols. Our wavelengths (350, 450, 550, and 650 nm) capture the ultraviolet-visible range. Our tables provide radiative transfer (RT) testbed results for atmosphere-ocean systems with an accuracy that surpasses the measurement accuracy of state-of-the-art polarimeters. To validate the accuracy of these tables we performed computations using three independent RT codes that provide deterministic numerical solutions for the RT equation. The agreement is 10(exp –5) for AOS-IV model, and 10(exp –6) for the other models. The degree of linear polarization computed by these RT codes differs by ≤0.2% for 15 isolated cases of tabulated reflectance values, and by ≤0.1% for all remaining cases. We also provide comparisons with results obtained by a stochastic RT code for AOS-I model. The agreement between the deterministic and stochastic results for this model is 10(exp –5) at TOA, and 10(exp –6) above the ocean surface.

Polarimeters↗

NASA NTRS1990

Navier-Stokes computations about complex configurations including a complete F-16 aircraft

Transonic Navier-Stokes (TNS) code solutions gathered from the literature for three-dimensional geometries, including two different wings and a complete F-16A aircraft, are presently discussed. The TNS codes use a zonal grid approach whose number of zones vary from four to 54. The Euler equations are solved in zones away from no-slip surfaces, and the thin-layer TNS equations are solved in all zones immediately adjacent to no-slip surfaces. In the case of 'corner' zones possessing no-slip boundary conditions on two different surfaces, a thin-layer formulation along both directions is employed. Employing these features, a zonal construction with the requisite set of boundary conditions can be devised for almost any application.

Holst, Terry L.↗

DOE OSTI2022

S K-edge XAS of Cu Ⅱ , Cu Ⅰ , and Zn Ⅱ oxidized Dithiolene complexes: Covalent contributions to structure and the Jahn-Teller effect

We report reduced dithiolene ligands are bound to high valent Mo centers in the active site of the oxotransferase family of enzymes. Related model complexes have been studied with great insight by Prof. Holm and his colleagues. This study focuses on the other limit of dithiolene chemistry: an investigation of the 2-electron oxidized dithiolene bound to low-valent late transition metal (TM) ions (Zn II , Cu I , and Cu II ). The bonding descriptions of the oxidized dithiolene [N,N-dimethyl piperazine 2,3-dithione (Me 2 Dt 0 )] complexes are probed using S K-edge X-ray absorption spectroscopy (XAS) and the results are correlated to density functional theory (DFT) calculations. These experimentally supported calculations are then extended to explain the different geometric structures of the three complexes. The Zn II (Me 2 Dt 0 ) 2 complex has only ligand-ligand repulsion so it is stabilized at the D 2d symmetry limit. The Cu I (Me 2 Dt 0 ) 2 complex has additional weak backbonding thus distorts somewhat from D 2d toward D 2h symmetry. The Cu II (Me 2 Dt 0 ) 2 complex has a strong σ donor bond that leads to both a large Jahn-Teller stabilization to D 2h and an additional covalent contribution to the geometry. The combined strong stabilization results in the square planar, D 2h structure. This study quantifies the competition between the ligand-ligand repulsion and the change in electronic structures in determining the final geometric structures of the oxidized dithiolene complexes, and provides quantitative insights into the Jahn-Teller stabilization energy and its origin.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

NASA NTRS2015

The Multi-Layer Variable Absorbers in NGC 1365 Revealed by XMM-Newton and NuSTAR

Between 2012 July and 2013 February, NuSTAR and XMM-Newton performed four long-look joint observations of the type 1.8 Seyfert, NGC 1365. We have analyzed the variable absorption seen in these observations in order to characterize the geometry of the absorbing material. Two of the observations caught NGC 1365 in an unusually low absorption state, revealing complexity in the multi-layer absorber that had previously been hidden. We find the need for three distinct zones of neutral absorption in addition to the two zones of ionized absorption and the Compton-thick torus previously seen in this source. The most prominent absorber is likely associated with broad-line region clouds with column densities of around approximately 10 (sup 23) per square centimeter and a highly clumpy nature as evidenced by an occultation event in 2013 February. We also find evidence of a patchy absorber with a variable column around approximately 10 (sup 22) per square centimeter and a line-of-sight covering fraction of 0.3-0.9, which responds directly to the intrinsic source flux, possibly due to a wind geometry. A full-covering, constant absorber with a low column density of approximately 1 by 10 (sup 22) per square centimeter is also present, though the location of this low density haze is unknown.

galaxies: active aEuro" galaxies: individual (NGC ↗

NASA NTRS2010

Evolution of Geometric Sensitivity Derivatives from Computer Aided Design Models

The generation of design parameter sensitivity derivatives is required for gradient-based optimization. Such sensitivity derivatives are elusive at best when working with geometry defined within the solid modeling context of Computer-Aided Design (CAD) systems. Solid modeling CAD systems are often proprietary and always complex, thereby necessitating ad hoc procedures to infer parameter sensitivity. A new perspective is presented that makes direct use of the hierarchical associativity of CAD features to trace their evolution and thereby track design parameter sensitivity. In contrast to ad hoc methods, this method provides a more concise procedure following the model design intent and determining the sensitivity of CAD geometry directly to its respective defining parameters.

Jones, William T.↗

DOE OSTI2024

Tunable Metamaterials for Impact Mitigation

Traditional methods of shielding fragile goods and human tissues from impact energy rely on isotropic foam materials. The mechanical properties of these foams are inferior to an emerging class of metamaterials called plate lattices, which have predominantly been fabricated in simple 2.5-dimensional geometries using conventional methods that constrain the feasible design space. Here in this work, additive manufacturing is used to relax these constraints and realize plate lattice metamaterials with nontrivial, locally varying geometry. The limitations of traditional computer-aided design tools are circumvented and allow the simulation of complex buckling and collapse behaviors without a manual meshing step. By validating these simulations against experimental data from tests on fabricated samples, sweeping exploration of the plate lattice design space is enabled. Numerical and experimental tests demonstrate plate lattices absorb up to six times more impact energy at equivalent densities relative to foams and shield objects from impacts ten times more energetic while transmitting equivalent peak stresses. In contrast to previous investigations of plate lattice metamaterials, designs with nonuniform geometric prebuckling in the out-of-plane direction is explored and showed that these designs exhibit 10% higher energy absorption efficiency on average and 25% higher in the highest-performing design.

36 MATERIALS SCIENCE↗

DOE OSTI2022

Small angle neutron scattering and lipidomic analysis of a native, trimeric PSI-SMALP from a thermophilic cyanobacteria

The use of styrene-maleic acid copolymers (SMAs) to produce membrane protein-containing nanodiscs without the initial detergent isolation has gained significant interest over the last decade. We have previously shown that a Photosystem I SMALP from the thermophilic cyanobacterium, Thermosynechococcus elongatus (PSI-SMALP), has much more rapid energy transfer and charge separation in vitro than detergent isolated PSI complexes. In this study, we have utilized small-angle neutron scattering (SANS) to better understand the geometry of these SMALPs. These techniques allow us to investigate the size and shape of these particles in their fully solvated state. Further, the particle's proteolipid core and detergent shell or copolymer belt can be interrogated separately using contrast variation, a capability unique to SANS. Here we report the dimensions of the Thermosynechococcus elongatus PSI-SMALP containing a PSI trimer. At ~1.5 MDa, PSI-SMALP is the largest SMALP to be isolated; our lipidomic analysis indicates it contains ~1300 lipids/per trimeric particle, >40-fold more than the PSI-DDM particle and > 100 fold more than identified in the 1JB0 crystal structure. Interestingly, the lipid composition to the PSI trimer in the PSI-SMALP differs significantly from bulk thylakoid composition, being enriched ~50 % in the anionic sulfolipid, SQDG. Finally, utilizing the contrast match point for the SMA 1440 copolymer, we also can observe the ~1 nm SMA copolymer belt surrounding this SMALP for the first time, consistent with most models of SMA organization.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2025

Capturing Ring Opening in Photoexcited Enolic Acetylacetone upon Hydrogen Bond Dissociation by Ultrafast Electron Diffraction

Photoinduced biological and chemical reactions are often based on key structural transformations of a molecule driven across multiple electronic states. Acetylacetone (AcAc) is a prototypical system for complex chemical pathways involving several conical intersections (CI) and singlet–triplet intersystem crossings (ISC) characterized by distinct geometries. In the gas phase, AcAc is predominantly in a planar ring-like enolic form stabilized by a strong intramolecular O–H···O hydrogen bond. Following excitation into the S 2 (ππ*) state at 266 nm, acetylacetone undergoes rapid internal conversion followed by intersystem crossing. Such relaxation pathways are associated with structural changes including ring opening, deplanarization, and bond elongation. In this work, ultrafast electron diffraction (UED) at the SLAC MeV-UED setup is employed as a direct structural probe with a time resolution of 160 fs. Together with trajectory surface hopping simulations, analysis of the UED data provides a new perspective on the early time nuclear dynamics in acetylacetone. Specifically, AcAc is observed to undergo ring opening, deplanarization, and bond elongation all within the first 700 fs after photoexcitation. The monitored dynamics is associated mainly with the nuclear motion on the S 1 potential energy surface, formed after very rapid transfer from S 2 to S 1 , allowing AcAc to reach the conical intersection to intersystem crossing. Such time scales of nuclear motion are contrasted with the time scales of electronic transitions in AcAc that were previously characterized with spectroscopic methods, specifically internal conversion (<100 fs) and intersystem crossing (∼1.5 ps).

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2026

Electron–Ion Covariance Reveals Geometry-Resolved Inner-Shell Spectra in CS 2 Photodissociation

The chemical shifts of inner-shell atomic orbitals are highly sensitive to chemical bonding and molecular structure. In evolving systems, however, the spectra of distinct molecular species and geometries overlap, obscuring the underlying chemical dynamics. Here we demonstrate the use of electron–ion covariance analysis to combine the structural sensitivity of Coulomb explosion imaging with inner-shell spectroscopy, yielding geometry-specific spectra of transient and product species. We apply this approach to the excited state dynamics of CS 2 probed by ionization above the S 2p edge. Electron–ion covariance with time- and momentum-selected S + and S 2+ ions isolates distinct S 2p photoelectron spectra for ground-state CS 2 , bent photoexcited CS 2 , the CS photoproduct, and bare atomic sulfur ─ species whose spectra overlap strongly in the channel-averaged measurement. Clear chemical shifts are observed in the covariance photoelectron spectrum for each of these species, all of which are consistent with high-level calculations. Here, by extracting the atomic S contribution to the photoelectron spectrum in a finely time-resolved manner we can disentangle this contribution to the overall time-resolved photoelectron spectrum as the photodissociation proceeds. These results demonstrate the promise of electron–ion covariance as a general approach to geometry-resolved inner-shell spectroscopy, opening a route to tracking structural evolution through chemical shifts in complex photoexcited molecules.

Ionization↗

DOE OSTI2020

Insight into the Catalytic Mechanism of GH11 Xylanase: Computational Analysis of Substrate Distortion Based on a Neutron Structure

The reaction mechanism of biomass decomposition by xylanases remains the subject of debate. Here, to clarify the mechanism, we investigated the glycosylation step of GH11 xylanase, an enzyme that catalyzes the hydrolysis of lignocellulosic hemicellulose (xylan). Making use of a recent neutron crystal structure, which revealed the protonation states of relevant residues, we used ab initio quantum mechanics/molecular mechanics (QM/MM) calculations to determine the detailed reaction mechanism of the glycosylation step. In particular, our focus is on the controversial question of whether or not an oxocarbenium ion intermediate is formed on the reaction pathway. The calculations support the validity of a basic retaining mechanism within a double-displacement scheme. The estimated free energy barrier of this reaction is ~18 kcal/mol with QM/MM-CCSD(T)/6-31(+)G**//MP2/6-31+G**/AMBER calculations, and the rate-determining step of the glycosylation is scission of the glycosidic bond after proton transfer from the acidic Glu177. The estimated lifetime of the oxocarbenium ion intermediate (on the order of tens of ps) and the secondary kinetic isotope effect suggest that there is no accumulation of this intermediate on the reaction path, although the intermediate can be transiently formed. In the enzyme–substrate (ES) complex, the carbohydrate structure of the xylose residue at the -1 subsite has a rather distorted (skewed) geometry, and this xylose unit at the active site has an apparent half-chair conformation when the oxocarbenium ion intermediate is formed. The major catalytic role of the protein environment is to orient residues that take part in the initial proton transfer. Because of a fine alignment of catalytic residues, the enzyme can accelerate the glycosylation reaction without paying a reorganization energy penalty.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2021

Fully variational incremental CASSCF

We report the complete-active-space self-consistent field (CASSCF) method is a canonical electronic structure theory that holds a central place in conceptualizing and practicing first principles simulations. For application to realistic molecules, however, the CASSCF must be approximated to circumvent its exponentially scaling computational costs. Applying the many-body expansion - also known as the method of increments - to CASSCF (iCASSCF) has been shown to produce a polynomially scaling method that retains much of the accuracy of the parent theory and is capable of treating full valence active spaces. Due to an approximation made in the orbital gradient, the orbital parameters of the original iCASSCF formulation could not be variationally optimized, which limited the accuracy of its nuclear gradient. Herein, a variational iCASSCF is introduced and implemented, where all parameters are fully optimized during energy minimization. This method is able to recover electronic correlations from the full valence space in large systems, produce accurate gradients, and optimize stable geometries as well as transition states. Demonstrations on challenging test cases, such as the oxoMn(salen)Cl complex with 84 electrons in 84 orbitals and the automerization of cyclobutadiene, show that the fully variational iCASSCF is a powerful tool for describing challenging molecular chemistries.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2025

From disorganized data to emergent dynamic models: Questionnaires to partial differential equations

Starting with sets of disorganized observations of spatially varying and temporally evolving systems, obtained at different (also disorganized) sets of parameters, we demonstrate the data-driven derivation of parameter dependent, evolutionary partial differential equation (PDE) models capable of generating the data. This tensor type of data is reminiscent of shuffled (multidimensional) puzzle tiles. The independent variables for the evolution equations (their “space” and “time”) as well as their effective parameters are all emergent , i.e. determined in a data-driven way from our disorganized observations of behavior in them. We use a diffusion map based questionnaire approach to build a smooth parametrization of our emergent space/time/parameter space for the data. This approach iteratively processes the data by successively observing them on the “space,” the “time” and the “parameter” axes of a tensor. Once the data become organized, we use machine learning (here, neural networks) to approximate the operators governing the evolution equations in this emergent space. Our illustrative examples are based (i) on a simple advection–diffusion model; (ii) on a previously developed vertex-plus-signaling model of Drosophila embryonic development; and (iii) on two complex dynamic network models (one neuronal and one coupled oscillator model) for which no obvious smooth embedding geometry is known a priori. This allows us to discuss features of the process like symmetry breaking, translational invariance, and autonomousness of the emergent PDE model, as well as its interpretability.

generative models↗

DOE OSTI2020

Compare linear-system solver and preconditioner stacks with emphasis on GPU performance and propose phase-2 NGP solver development pathway

The goal of the ExaWind project is to enable predictive simulations of wind farms comprised of many megawatt-scale turbines situated in complex terrain. Predictive simulations will require computational fluid dynamics (CFD) simulations for which the mesh resolves the geometry of the turbines and captures the rotation and large deflections of blades. Whereas such simulations for a single turbine are arguably petascale class, multi-turbine wind farm simulations will require exascale-class resources. The primary physics codes in the ExaWind project are Nalu-Wind, which is an unstructured-grid solver for the acoustically incompressible Navier-Stokes equations, and OpenFAST, which is a whole-turbine simulation code. The Nalu-Wind model consists of the mass-continuity Poisson-type equation for pressure and a momentum equation for the velocity. For such modeling approaches, simulation times are dominated by linear-system setup and solution for the continuity and momentum systems. For the ExaWind challenge problem, the moving meshes greatly affect overall solver costs as reinitialization of matrices and recomputation of preconditioners is required at every time step. In this report we evaluated GPU-performance baselines for the linear solvers in the Trilinos and hypre solver stacks using two representative Nalu-Wind simulations: an atmospheric boundary layer precursor simulation on a structured mesh, and a fixed-wing simulation using unstructured overset meshes. Both strong-scaling and weak-scaling experiments were conducted on the OLCF supercomputer Summit and similar proxy clusters. We focused on the performance of multi-threaded Gauss-Seidel and two-stage Gauss-Seidel that are extensions of classical Gauss-Seidel; of one-reduce GMRES, a communication-reducing variant of the Krylov GMRES; and algebraic multigrid methods that incorporate the afore-mentioned methods. The team has established that AMG methods are capable of solving linear systems arising from the fixed-wing overset meshes on CPU, a critical intermediate result for ExaWind FY20 Q3 and Q4 milestones. For the fixed-wing strong-scaling study (model with 3M grid-points), the team identified that Nalu-Wind simulations with the new Trilinos and hypre solvers scale to modest GPU counts, maintaining above 70% efficiency up to 6 GPUs. However, there still remain significant bottlenecks to performance: matrix assembly (hypre), AMG setup (hypre and Trilinos) In the weak-scaling experiments (going from 0.4M to 211M gridpoints), it's shown that the solver apply phases are faster on GPUs, but that Nalu-Wind simulation times grow, primarily due to the multigrid-setup process. Finally, based on the report outcomes, we propose a linear solver path-forward for the remainder of the ExaWind project. Near term, the NREL team will continue their work on GPU-based linear-system assembly. They will also investigate how the use of alternatives to the NVIDIA UVM (unified virtual memory) paradigm affects performance. Longer term, the NREL team will evaluate algorithmic performance on other types of accelerators and merge their improvements back to the main hypre repository branch. Near term, the Trilinos team will address performance bottlenecks identified in this milestone, such as implementing a GPU-based segregated momentum solve and reusing matrix graphs across linear-system assembly phases. Longer term, the Trilinos team will do detailed analysis and optimization of multigrid setup.

17 WIND ENERGY↗

DOE OSTI2020

Demonstration and performance testing of extreme-resolution simulations with static meshes on Summit (CPU & GPU) for a parked-turbine configuration and an actuator-line (mid-fidelity model) wind farm configuration (ECP-Q4 FY2020 Milestone Report)

17 WIND ENERGY↗