Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “linear projection”

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 109 records · Page 6
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Batched Sparse Linear Algebra (Final Report for Subcontract B648960)

This report finalizes design specifications for developing batched kernels for small tensor operations for unassembled matrix-free iterative solvers, batched solvers for partially assembled operators, and batched solvers with support for various sparse formats. The outcome of the project milestones is a set of interfaces to Batched Sparse LA solvers running on hardware accelerators for use in ECP Libraries and Applications. It is part of the development of sparse batched kernels, solvers/preconditioners as well as creating interoperability in xSDK libraries with sparse and dense batched functions to benefit ECP applications. The participants included representatives from ECP libraries (not limited to the xSDK project), applications, and vendors (AMD, Intel, and NVIDIA). Batched sparse linear algebra solvers form the new frontier for algorithmic development and performance engineering. Many applications (ECP and non-ECP alike) require simultaneous solutions of small linear systems of equations that are structurally sparse. To move towards high hardware utilization, it is important to provide these applications with appropriate interfaces to efficient batched sparse solvers running on modern hardware accelerators. We present interface designs in use by HPC software libraries supporting batched sparse linear algebra and the development of sparse batched kernel codes for solvers and preconditioners. We also address the potential interoperability opportunities to keep the software portable between the major hardware accelerators from AMD, Intel, and NVIDIA. The presented interface specifications includes batched band, sparse iterative, and sparse direct solvers. This report summarizes progress in Kokkos Kernels and the xSDK libraries MAGMA, Ginkgo, hypre, SUNDIALS, and SuperLU_dist.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Scalable High-H 2 Flux, Robust Thin Film Solid Oxide Electrolyzer

This project was aimed at the development of proton-conducting SOEC (P-SOEC) technology that has the potential to meet key DOE H 2 production targets. A decreased proton resistance of the electrolyte and Faradaic efficiency improvements were sought to increase the fraction of consumed electrolysis power that is used to actually generate H 2 , while simultaneously decreasing cost dramatically. Moreover, the development project was intended to yield maximum durability through the use of a steam protective layer. Furthermore, sputtering was used to overcome processing challenges that have hampered P-SOEC development, while the low-temperature operation goal of 500 °C was expected to aid in mitigating thermally activated long-term degradation. The approach to high-performance, lower-temperature SOECs leveraged our existing SOFC Ni-cermet anode support and extensive thin-film sputtering layer-deposition experience. Rather than an all-in-one, reversible fuel cell approach which has many unacceptable tradeoffs, we focused on the many benefits to hydrogen generating SOECs, including the existence of synergies for reduced manufacturing costs (e.g., SOECs and SOFCs share supporting layers and overall manufacturing processing). The end result of this project was expected to increase current performance at 500 °C from approximately 0.8 A/cm 2 (at 60% Faradaic efficiency) at 1.4 V to > 1 A/cm 2 (at > 95% Faradaic efficiency) with a > 40% reduction in system cost and to enable operation of P-SOECs in steam contents >> 20% for a goal of a > 40,000 hours lifetime. To enable 500 °C operation in a very high steam atmosphere (> 20%), we proposed the use of a sputtered dense thin film (~0.1-1 µm thick) of high-stability Ba(Zr,Y)O 3 (BZY) to protect the Ba(Ce,Zr,Y,Yb)O 3 (BCZYYb) electrolyte. The BZCYYb, in turn, blocks the hole conductivity of the BZY to boost Faradaic efficiency. As FE increases, more of the consumed electricity is used in electrolysis to generate H 2 , rather than being shunted. Additionally, as cell resistance decreases, the voltage required to maintain current decreases, as well as the power required to generate the same amount of H 2 . With the proposed enhancements, these two factors result in the final 46% decrease in power needed to run the system. Likewise, a production rate of 50,000 kg H 2 /day will require 55% less active area, such that a system will need only 650 cells for an 80 cm 2 active area instead of ~1,440. Taking the 2016 DOE projected current cost and modifying the electricity cost and linearly scaling the other costs (except thermal feedstock) based on the cell area improvement, results in a 44% decrease in lifetime system cost, or a decrease from $\$$4.95/kg H 2 to $\$$2.75/kg H 2 . This is well below the 2018 DOE target of $\$$4/kg H 2 . The results from this project showed that we can create a P-SOEC with enhanced steam stability using two different electrolytes (i.e., one on top of the other) and achieve sufficiently low area specific resistance (ASR) to achieve the target performance. Unfortunately, due to extended delays at the beginning of the project and related supply chain and equipment access issues, we were not able to completely show increased Faradaic efficiency for the P-SOEC and therefore were unable to demonstrate the full proof of concept within the first budget period budget. While there are still challenges that remain to be solved, significant progress was made during this project and the concept still has merit that warrants further development.

08 HYDROGEN↗
DOE OSTI2020

Multigrid Algorithms with Projection and Prolongation over Elements of the Phase Space for K-Eigenvalue Transport Problems

This paper describes new multilevel acceleration methods for solving the multigroup neu- tron transport eigenvalue problems. These multilevel algorithms use different projection and prolongation operators in the phase space. The Nonlinear Diffusion Acceleration (NDA) method with multiple grids in energy is formulated with the prolongation oper- ator based on multiplication iterative correction and linear-in-energy mapping. Another multilevel NDA method uses the projection operator with coarsening in energy between the high-order transport and low-order NDA equations. The third algorithm is formu- lated with the partial-current based CMFD low-order equations and applies projection operators in space and energy. The numerical results are presented.

Cornejo, Luke↗
DOE OSTI2022

High-Temperature Linear Receiver Enabled by Multicomponent Aerogels

Concentrating solar thermal (CST) technology has significant potential mainly due to its dispatchability and low cost of storage. However, to compete with other sources, including utility-scale solar PV, its final cost (cents/kWh) still needs to be lowered. Cost reduction can be achieved by improving the system level efficiency of the CST plants through the deployment of advanced power cycles, which operate at high temperatures of ~700°C. However, optical, and thermal losses pose a major challenge to the efficiency of such CST systems. The overall aim of this project is to investigate and de-risk a linear solar receiver concept called an Aerogel Insulated Receiver (AIR) that generates high temperatures (up to 700°C) at a low solar concentration ratio (<100) and with a high collection efficiency (optical × receiver). Our prior work has demonstrated the thermal stability1 and optical and heat-insulating properties2,3 of transparent aerogel insulation at a one-inch scale. The focus of this work is on (1) co-optimization of the geometry of the aerogel tile and receiver enclosure to fit a standard parabolic collector (PTC), (2) scale-up of aerogels into 4-inch tiles while preserving key properties, (3) experimental measurement of receiver heat loss (W/m) in a >70-cm test stand and validation of anticipated receiver performance at high temperatures. Regarding (1), appropriate optical and thermal models for a parabolic trough receiver (PTR) are developed and validated. The geometry of the aerogels and the receiver enclosure are co-optimized to maximize the collection efficiency. The model predicts a 54% collection efficiency at 700°C for an AIR design based on flat aerogels. By combining the collection efficiency with the power block efficiency of supercritical CO 2 cycles, we predict >10% improvements in peak plant efficiency relative to existing line-focusing CST systems. The application of curved plasmonic aerogels is predicted to further increase the collection efficiency to 64%. Regarding (2), we demonstrate the successful development of 6-inch-long refractory aerogel tiles with optical, thermal, and stability characteristics consistent with our prior work. This scale-up requires a transition to a larger ALD station and modifying the ALD process variables such as exposure time and the number of precursor doses. Regarding (3), we design and develop an AIR test stand measuring 3 feet in length. Heat loss performance analysis is carried out using the test stand. The results indicate that aerogel insulation can significantly reduce receiver thermal losses at the high operating temperatures required for next-generation PTRs. The experimental results agree with the heat loss performance predicted by our receiver model. Lastly, we conducted preliminary failure mode and effects (FMEA) and techno-economic (TEA) analyses to identify failure mitigation strategies and commercial opportunities, respectively. Overall, this project identifies key opportunities and challenges in deploying aerogel insulating receivers in next-generation line-focusing CST technologies.

14 SOLAR ENERGY↗
DOE OSTI2024

High-Fidelity and High-Performance Computational Simulations for Rapid Design Optimization of Sulfur Thermal Energy Storage (CRADA Report)

NREL and Element 16 collaborated on sulfur thermal energy storage modeling using NREL’s high performance computing (HPC) resources to assist its application in industrial processes. Industrial process heat (IPH) accounts for ~70% of US manufacturing energy use and is primarily produced by fossil fuel combustion. Approximately, 1500 TWht (~60% Terawatt hour thermal) of IPH demand is in the temperature range of 100-300°C. Industrial applications in this temperature range include drying, hydrothermal processing, thermal enhanced oil recovery, food and beverage, bioethanol production, etc. Cost-effective thermal energy storage (TES) that increases the utilization of waste and renewable heat (solar, geothermal, etc.) could provide significant energy savings and reliable heat sources, decrease emissions, and increase US manufacturing competitiveness through reductions in fuel consumption. This HPC4EI project facilitated Element 16’s development of low-cost and high-impact molten sulfur TES for dispatchable IPH. The development of a high-fidelity model validated by experimental data and HPC simulations enabled the successful resolution of the complex interplay between fluid dynamics and heat transfer processes during transient operation of sulfur TES, overcoming the numerical challenges posed by the non-linear temperature-dependent physical properties of sulfur. The project helped accelerate Element 16’s molten sulfur TES product design and support its broad applications.

25 ENERGY STORAGE↗
DOE OSTI2021

Are coupled renewable-battery power plants more valuable than independently sited installations?

Coupled renewable-battery powerplants differ from the traditional concept of independent siting of electricity resources within transmission networks. Prior research on the value proposition and cost savings from coupling did not consider the geographic constraint of co-location. This paper fills the gap by assessing how pricing volatility differences between nodes within electricity markets impact the system value of coupled renewable-battery projects as compared to independent VRE and battery installations. We use wholesale power market prices from 2012–2019 across the seven main U.S. independent system operators (ISOs) with a linear optimization program to compare the electricity market value of coupled projects to the value of the same underlying sub-components, deployed separately. We find that additional value from adding a 4-hour battery sized to 50% of renewable-plant nameplate capacity is $\$$10/MWh across ISOs on average. The highest boost occurs in California ($\$$15/MWh), where the value of adding storage to solar rises over time in tandem with increased solar penetration in the region. If renewables and batteries are deployed independently, we estimate that $\$$12.5/MWh of additional value could be achieved because of more flexibility on battery siting and operation. The $\$$12.5/MWh coupling penalty is reduced to $\$$1.6/MWh when considering alternative approaches to integrating battery storage. This result implies that renewable-battery power plants will play an increasing role in electricity systems if they can be built for $\$$2–$\$$13/MWh less than independent projects of comparable size. However, the wide regional variation in coupling penalties, along with the importance of conditions captured in our sensitivity cases, suggests the tradeoff between coupling penalties and savings will vary by situation. Therefore, roles exist for independent and coupled projects from a system optimization perspective.

14 SOLAR ENERGY↗
DOE OSTI2021

Keep it short: Exploring the impacts of configuration choices on the recent economics of solar-plus-battery and wind-plus-battery hybrid energy plants

Coupled renewable-battery powerplants differ from the traditional concept of independent siting of electricity resources within transmission networks. Prior research on the value proposition and cost savings from coupling did not consider the geographic constraint of co-location. This paper fills the gap by assessing how pricing volatility differences between nodes within electricity markets impact the system value of coupled renewable-battery projects as compared to independent VRE and battery installations. We use wholesale power market prices from 2012–2019 across the seven main U.S. independent system operators (ISOs) with a linear optimization program to compare the electricity market value of coupled projects to the value of the same underlying sub-components, deployed separately. We find that additional value from adding a 4-hour battery sized to 50% of renewable-plant nameplate capacity is $\$10$/MWh across ISOs on average. The highest boost occurs in California ($\$15$/MWh), where the value of adding storage to solar rises over time in tandem with increased solar penetration in the region. If renewables and batteries are deployed independently, we estimate that $\$12.5$/MWh of additional value could be achieved because of more flexibility on battery siting and operation. The $\$12.5$/MWh coupling penalty is reduced to $\$1.6$/MWh when considering alternative approaches to integrating battery storage. This result implies that renewable-battery power plants will play an increasing role in electricity systems if they can be built for $\$2$–$\$13$/MWh less than independent projects of comparable size. However, the wide regional variation in coupling penalties, along with the importance of conditions captured in our sensitivity cases, suggests the tradeoff between coupling penalties and savings will vary by situation. Therefore, roles exist for independent and coupled projects from a system optimization perspective.

14 SOLAR ENERGY↗
DOE OSTI2022

Adoption of biofuels for marine shipping decarbonization: A long‐term price and scalability assessment

Abstract This study assessed the long‐term annual biofuel production capacity potential and price in the United States and shed light on the prospect of biofuel adoption for marine propulsion. A linear programming model was developed to assist the projections and provide insightful analyses. The projected long‐term (2040) maximum annual capacity of biofuels in the United States is 245 million metric tons (Mt) or 65 billion gallons of heavy fuel oil gallon equivalent (HFOGE) when based on the median feedstock availability. Between 2022 (near‐term) and 2040, the potential biofuel capacity increases by over 40%, attributed to increased feedstock availability. At a price range up to $500/t, biodiesel is the main product, and the annual capacity (12 Mt) is limited to feedstock availability constraints. Biodiesel and corn ethanol are the main biofuels at a price range up to $750/t. At a higher price point (above $750/t), the biofuel types and annual capacities increase substantially (218 Mt per year). Biofuels above this price include gasoline‐, jet‐, and diesel‐range blendstocks, as well as bio‐methanol, bio‐propane, and biogas. This study concludes that the US domestic feedstock availability coupled with advanced conversion technologies can produce substantial amounts of biofuels to achieve a critical mass and be impactful as alternative marine fuels. There is also a need to improve the biofuel price for marine shipping adoption. Policies and economic incentives that provide temporary financial support would help facilitate maritime biofuel adoption. © 2022 Alliance for Sustainable Energy, LLC. Biofuels, Bioproducts and Biorefining published by Society of Industrial Chemistry and John Wiley & Sons Ltd.

09 BIOMASS FUELS↗
DOE OSTI2023

Thermohydraulic Design Analysis of the Target Assembly in the Material Plasma Exposure Experiment Facility

The Material Plasma Exposure eXperiment (MPEX) project seeks to design a steady-state linear plasma facility at Oak Ridge National Laboratory that will be used to study plasma-material interactions (PMIs) at fusion prototypic levels, supporting the evaluation and development of materials for the next generation of fusion devices. This study is focused on PMI exposure of small-size neutron-irradiated specimens, which are clamped onto an actively cooled component. A thermohydraulic evaluation of a new MPEX target assembly design to assess the appropriate operation during MPEX operation is presented. To further guide the design and assess the structural integrity of the components under expected loads, preliminary thermomechanical stress analyses were also conducted. To ensure good thermal contact between the components, thermal interface materials, such as silver flexible graphite, were used in the assembly. It was found that the maximum target temperatures of 1572, 1463, and 1315 K were obtained for Grafoil thicknesses of 0.61, 0.38, and 0.25 mm, respectively. The distribution of the axial deformation at high heat fluxes showed that there are no gaps between components, indicating good contact at material interfaces. Moreover, the contact pressure between the target and other components indicated that very good contact was established at these interfaces. Furthermore, the stress-strain conditions for the target will be further used to assess the appropriate operation during MPEX experiments and gain insight into materials science phenomena during PMI experiments.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2022

Prospects for Machine Learning and Pulse Shaping on the Scorpius Accelerator [Poster]

The Advanced Sources and Detectors (ASD) project aims to build Scorpius, a multi pulse linear induction accelerator capable of delivering a 1.4 kA electron beam at energies up to 24 MeV. One of the primary advancements of Scorpius is the use of solid state pulsed power (SSPP) to provide flexibility in pulse shaping by independently triggering 45 individual stages stacked in each of 984 line replaceable units (LRU), with 168 LRUs dedicated to the injector. By leveraging circuit modeling of each LRU stage, a machine learning model of the SSPP will be developed to allow for optimization of the pulse shape, including pulse flattening and reflection mitigation. Particle-in-cell simulations of Scorpius have, for example, demonstrated that reducing reflections during multi-pulse operation mitigates beam spill by preventing the production of off-energy electrons between pulses, thereby abating stimulated ion desorption from the wall and beam charge neutralization. This machine learning model will be validated and tuned with experimental data collected from the Scorpius injector and Integrated Test Stand

43 PARTICLE ACCELERATORS↗
DOE OSTI2023

Leverage Score Sampling for Parametric PDEs (Final Technical Report)

This final technical report summarizes the accomplishments of work performed under DOE Office of Science Award DE-SC0022266, which is titled “Leverage Score Sampling for Parametric PDEs”. The goal of the project was to extend methods from Randomized Numerical Linear Algebra (RandNLA) to tackle central computational challenges in model order reduction and uncertainty quantification (UQ) for parametric partial differential equations (PDEs). In particular, we sought to use importance sampling methods originally developed for RandNLA to develop sample efficient active learning algorithms for approximating high-dimensional scalar functions, e.g. by polynomials, Gaussian process models, and simple neural networks. Such methods can be immediately applied to developing surrogate models or to approximating quantity of interest (QoI) surfaces. In the context of PDEs, each sample used for learning equates to the solution of the differential equation for a particular set of parameters, so sample efficiency translates to improved computational efficiency for a variety of downstream tasks.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

PID-Regulated Heating System for PIP-II Reference Line

The Proton Improvement Project-2 centers on building a new superconducting linear particle accelerator (Linac) at Fermilab. At the heart of the accelerator is the reference line, a critical system that defines the ideal path for the particle beam as it passes through magnets, RF cavities, and other beamline elements. Temperature stability is crucial for the reliable operation of RF components, such as mixers and filters. Fluctuations affect key performance parameters like conversion loss, isolation, and linearity. To mitigate any drift caused by ambient temperature changes, a heating plate assembly is utilized to maintain key components at a controlled temperature of 40°C. The system utilizes an aluminum 36”x36”x0.5” heat plate powered by a MOSFET-based control circuit, delivering approximately 460 W of thermal energy through a resistor array. Real-time temperature feedback is provided by a PT100 Resistance Temperature Detector (RTD), which interfaces with a Proportional–Integral–Derivative (PID) control algorithm to maintain closed-loop temperature regulation. The control signal actively modulates the gate voltage of an N channel MOSFET, dynamically adjusting power delivery in response to deviations from the temperature setpoint. Simulations and LTspice models validate the functionality and responsiveness of the circuit under varying conditions. The prototype has successfully demonstrated stable thermal control, paving the way for integration into the PIP-II infrastructure. The final design will feature an expanded resistor array, as well as communication with a PLC for continuous data acquisition and diagnostics. This work directly supports Fermilab’s broader mission by contributing to the stability and reliability of core accelerator systems, enhancing the precision of particle beam delivery for future physics experiments.

Mosher, Alexander [Fermilab]↗
DOE OSTI2021

Upgrade of the ILC Cryomodule

Results of testing modified components for the cryomodule of the International Linear Collider (ILC) are summarized. To reduce the ILC project cost, it is proposed to replace titanium cryomodule components with stainless steel (SS) ones. New bimetallic transitions (Ti/SS, Nb/SS) have been produced by a unique method based on explosion welding. Successive upgrading of these components to the latest version of the Nb/Ti/SS transition element has led to improvement of the ILC cryomodule. This new component resolves problems of residual stress, and its specific design prevents the possibility of a shift due to the difference in the linear expansion coefficients of the constituent metals. Furthermore, leak tests with the He gas revealed no leaks at the background rate of ≈0.2 × 10$^{–10}$ atm cc/s. The test results are very encouraging. The up-to-date design of trimetallic Nb/Ti/SS element promises technologically simpler and less expensive manufacture. Investigations have shown that explosion welding allows unique trimetallic components to be made not only for cryogenic units of accelerators but also for laboratory equipment and for general engineering applications.

43 PARTICLE ACCELERATORS↗
DOE OSTI2025

Towards Seamless Interoperability of MPI-OpenMP Applications

A chasm exists between mathematical software libraries written for MPI-based applications and those written for OpenMP applications. Recently, however, PETSc enables the simple use of its MPI-based linear solvers from OpenMP applications. Separately, the MPICH MPI development team has started a new project to allow almost seamless MPI use in OpenMP applications. Both proposed approaches would result in a similar user experience. Here, we discuss the reasons for these projects and their potential for providing more numerical library choices for OpenMP applications, including the unlimited assortment of linear solvers available in PETSc. In addition, we present the performance of an application using the first approach, demonstrating its efficacy.

MPI↗
DOE OSTI2020

Residue Specific Characterization of Electrostatics, Conformational Heterogeneity, and Dynamics of Electron Transfer Proteins with Linear and Nonlinear Infrared Spectroscopy. Final Report

The goal of this research project was to further our understanding of biological electron transfer through the development and application of new methods for studying proteins and their complexes with high spatial and temporal detail. The approach was based on incorporation of reporter bonds with distinct vibrational frequencies at specific locations in proteins and their characterization through the interaction of the bonds with pulses of infrared light. The project focused on the proteins plastocyanin (Pc) and cytochrome (cyt) f, two electron transfer partners in photosynthesis. The first thrust of the project was the incorporation of carbon deuterium bonds as non-perturbative probes at a ligand to the copper ion of Pc and their use to detect changes in the bonding due to the number of electrons at the metal and how the bonding changes upon association with cyt f. Toward developing more advanced approaches to characterize proteins, we incorporated a stronger reporter, the cyano group of cyanophenylalanine, at several locations in Pc and measured the frequency dynamics in the absence and presence of cyt f. The acquired data improved the current model for the protein complex. Notably, the data indicated that the majority of the population is involved in a loosely bound complex, providing insight into the nature of the ensembles formed by electron transfer protein complexes that is not available via other approaches. In addition to learning about the protein complex, we made strides toward developing and improving the new experimental approach to study highly dynamic behaviors of proteins.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

Wide-Bandgap Semiconductor Amplifiers for Fusion Plasma Heating and Control

This paper discusses power electronics developed under the ARPA-E GAMOW program to support nuclear fusion power production. The goal of this project was to develop and assess the potential for wide-bandgap (WBG) semiconductor devices in power electronics to enable high-efficiency and high-voltage solid-state systems for fusion plasma generation, heating, and control. The power electronics use an architecture in which multiple high-power boards can be combined to produce megawatt-level power, where using multiple boards provides high reliability. Two main areas of power electronics boards are developed in this project for fusion plasma heating and control applications: (1) pulse generation and control and (2) radiofrequency generation. The first area is for boards capable of driving high-voltage millisecond pulses at high duty cycles. The envisioned application of these pulses is in plasma control of magnetohydrodynamic instabilities, plasma position, and edge-localized modes. Pulse-width modulation allows for the implementation of a wide variety of linear and nonlinear control systems. The boards developed for this project could actuate control coils based on digital input signals and can be parallelized to provide megawatts of output power. The design of the pulse generator is a low-side load switch. A load switch was designed and constructed that utilized 2-kV-rated field-effect transistor (FET)-based cascodes developed by Qorvo under this project to perform initial testing of these cascodes. The second area is being implemented using class E amplifiers with WBG devices and a reactance steering network to handle inductive or capacitive plasma loads. Applications include ion cyclotron resonance heating (ICRH) and high-harmonic fast-wave (HHFW) heating. A class E reactance steering network is demonstrated in modeling and experiment with a resistive-inductive load that models an inductively-coupled plasma. Power combining of boards with class E reactance steering networks is also simulated and demonstrated experimentally, to enable scaling up to high power. Modeling of high-power-density cooling and remaining useful life is conducted to enable reliable, effectively cooled high-power electronics for fusion applications.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
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↗