Search NASA⌕ Search

SEARCH · Search NASA

Results for “two-stage problems”

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 145 records · Page 8

Investigation of Microbunching Instabilities in Modern Recirculating Accelerators

Particle accelerators are machines to accelerate and store charged particles, such as electrons or protons, to the energy levels for various scientific applications. A collection of charged particles usually forms a particle beam. There are three basic types of particle accelerators: linear accelerators (linac), storage-ring (or circular) accelerators, and recirculating accelerators. In a linac, particles are accelerated and pass through once along a linear or straight beamline. Storage-ring accelerators propel particles around a circular track and repetitively append the energy to the stored beam. The third type, also the most recent one in chronology, the recirculating accelerator, is designed to accelerate the particle beam in a short section of linac, circulate the beam, and then either continue to accelerate for energy boost or decelerate it for energy recovery. The beam properties of a linac machine are set at best by the initial particle sources. For storage rings, the beam equilibria are instead determined by the overall machine design. The modern recirculating machines share with linacs the advantages to both accelerate and preserve the beam with high beam quality, as well as efficiently reuse the accelerating components. The beamline design in such a machine configuration can however be much more complicated than that of linacs. As modern accelerators push toward the high-brightness or high-intensity frontier by demanding particles in a highly charged bunch (about nano-Coulomb per bunch) to concentrate in an ever-decreasing beam phase space (transverse normalized emittance about 1 μm and relative energy spread of the order of 10^-5 in GeV beam energy), the interaction amongst particles via their self-generated electromagnetic fields can potentially lead to coherent instabilities of the beam and thus pose significant challenges to the machine design and operation. In the past decade and a half, microbunching instability (MBI) has been one of the most challenging issues for such high-brightness or high-intensity beam transport, as it would degrade lasing performance in the fourth-generation light sources, reduce cooling efficiency in electron cooling facilities, and compromise the luminosity of colliding beams in lepton or lepton-hadron colliders. The dissertation work will focus on the MBI in modern recirculating electron accelerators. It has been known that the collective interactions, the coherent synchrotron radiation (CSR) and the longitudinal space charge (LSC) forces, can drive MBI. The CSR effect is a collective phenomenon in which the electrons in a curved motion, e.g. a bending dipole, emit radiation at a scale comparable to the micro-bunched structure of the bunch distribution. The LSC effect stems from non-uniformity of the charge distribution, acts as plasma oscillation, and can eventually accumulate an amount of energy modulation when the beam traverses a long section of a beamline. MBI can be seeded by non-uniformity or shot noise of the beam, which originates from granularity of the elementary charge. Through the aforementioned collective effects, the modulation of the bunch sub-structure can be amplified and, once the beam-wave interaction formed a positive feedback, can result in MBI. The problem of MBI has been intensively studied for linac-based facilities and for storage-ring accelerators. However, systematic studies for recirculation machines are still very limited and form a knowledge gap. Because of the much more complicated machine configuration of the recirculating accelerators than that of linacs, the existing MBI analysis needs to be extended to accommodate the high-brightness particle beam transport in modern recirculating accelerators. This dissertation is focused on theoretical investigation of MBI in such machine configuration in the following seven themes: (1) Development and generalization of MBI theory The theoretical formulation has been extended so as to be applicable to a general linear beamline lattice including horizontal and vertical transport bending elements, and beam acceleration or deceleration. These featured generalizations are required for MBI analysis in recirculation accelerators. (2) Construction of CSR impedance models In addition to the steady-state CSR interaction, it has been found that the exit transient effect (or CSR drift) can even result in more serious MBI in high-brightness recirculation arcs. The onedimensional free-space CSR impedances, especially the exit transients, are derived. The steady-state CSR impedance is also extended to non-ultrarelativistic beam energy for MBI analysis of low-energy merger sections in recirculating accelerators. (3) Numerical implementation of the derived semi-analytical formulation This includes the development of a semi-analytical Vlasov solver for MBI analysis, and also benchmarking of the solver against massive particle tracking simulations. (4) Exploration of multistage amplification behavior of CSR microbunching development The CSR-induced MBI acts as an amplifier, which amplifies the sub-bunch modulation of a beam. The amplification is commonly quantified by the amplification gain. A beam transport system can be considered as a cascaded amplifier. Unlike the two-stage amplification of four-dipole bunch compressor chicanes employed in linacs, the recirculation arcs, which are usually constituted by several tens of bending magnets, show a distinguishing feature of up to six-stage microbunching amplification for our example arc lattices. That is, the maximal CSR amplification gain can be proportional to the peak bunch current up to sixth power. A method to compare lattice performance has been developed in terms of gain coefficients, which nearly depend on the lattice properties only. This method has also proven to be an effective way to quantify the current dependence of the maximal (5) Control of CSR MBI in multibend transport or recirculation arcs The existing mitigation schemes of MBI mostly aim to linac-based accelerators and may not be practical to the recirculating accelerator facilities. Thus a set of conditions for suppression of CSR MBI was proposed and examined for example lattices from low (~100 MeV) to high (~1 GeV) energies. (6) Study of more aspects of microbunched structures in beam phase spaces For a cascaded amplifier in circuit electronics, the total amplification gain can be estimated as the product of individual gains. In a beam transport line of an accelerator, the (scalar) gain multiplication was examined and found to under-estimate the overall microbunching amplification. The concept of gain matrix was developed, which includes the density, energy and transverse-longitudinal modulations in a beam phase space, and used to analyze MBI for a proposed recirculating machine. Throughout the gain matrix approach, it reasonably gives the upper limit of spectral MBI gain curves. This extended analysis can be employed to study multi-pass recirculation. (7) Study of MBI for magnetized beams Driven by a recent energy-recovery-linac based cooler design for electron cooling at Jefferson Lab Electron-Ion Collider Project, the generalized theoretical formulation for MBI to a transversely coupled beam has been developed and applied to this study. A magnetized beam in general features non-zero canonical angular momentum, thus considered to be a transversely coupled beam. A novel idea of utilizing magnetized beam transport was proposed for improvement of cooling efficiency and possible mitigation of collective effects. A concern of MBI regarding this design was studied and excluded. The large transverse beam size associated with the beam magnetization is found to help suppress MBI via the transverse-longitudinal correlation.

43 PARTICLE ACCELERATORS↗

Galvanic Transformation Dynamics in Heterostructured Nanoparticles

Corrosion is a significant problem for the stability of structural metals and potentially for functional nanomaterials in operating environments. When two metals with different electrochemical potentials form a junction, galvanic corrosion occurs, resulting in the sacrificial dissolution of the metal with a higher oxidation potential (lower electrode potential). In this report it is shown that bimetallic hetero-nanostructures composed of phase-segregated metals undergo galvanic corrosion in aqueous environments. Such selective etching of the sacrificial metal in heterojunction particles leads to the formation of unusual and kinetically stabilized half-spheroid particles. By using a fluid cell and in situ scanning transmission electron microscopy, a two-stage corrosion process can be observed where the Cu experiences a fractal breakdown before the Ag corrodes due to the lack of a protective oxide layer. However, when treated with a mild Ar plasma, the stability of these structures against corrosion is enhanced due to the conversion of the amorphous native oxide to a denser, thin layer of CuO on the Cu surface. Taken together, this work highlights the importance of considering the effects of galvanic corrosion on the stability of multicomponent nanoparticles, and it shows how mass transport in a nanoscale system is influenced by redox processes.

36 MATERIALS SCIENCE↗

Multicollinearity in simultaneous equations system: evaluation of estimation performance of two-parameter estimator

In simultaneous equations model, two-stage least squares estimator is easy to apply and commonly preferred. When multicollinearity exists, two-stage least squares estimator has some drawbacks and it is no longer favorable. In this context, biased estimation methods are recommended. Two-parameter estimator of Özkale and Kaçıranlar (Commun Stat Theory Methods 36(15):2707–2725, 2007) had been established to be superior to the ordinary least squares estimator under some conditions in linear regression model suffering from multicollinearity. In this paper, the idea of two-parameter estimation in linear regression model is carried out to the simultaneous equations model. For this model, two-stage two-parameter estimator is proposed to remedy the problem of multicollinearity. Estimation performance of this new estimator is evaluated by means of two real-life data analyses. In addition to the numerical example, an extensive Monte Carlo experiment is conducted.

97 MATHEMATICS AND COMPUTING↗

An adaptive and stability-promoting layerwise training approach for sparse deep neural network architecture

This work presents a two-stage adaptive framework for progressively developing deep neural network (DNN) architectures that generalize well for a given training data set. In the first stage, a layerwise training approach is adopted where a new layer is added each time and trained independently by freezing parameters in the previous layers. We impose desirable structures on the DNN by employing manifold regularization, sparsity regularization, and physics-informed terms. We introduce a ε – δ – stability-promoting concept as a desirable property for a learning algorithm and show that employing manifold regularization yields a ε – δ stability-promoting algorithm. Further, we also derive the necessary conditions for the trainability of a newly added layer and investigate the training saturation problem. In the second stage of the algorithm (post-processing), a sequence of shallow networks is employed to extract information from the residual produced in the first stage, thereby improving the prediction accuracy. Numerical investigations on prototype regression and classification problems demonstrate that the proposed approach can outperform fully connected DNNs of the same size. Moreover, by equipping the physics-informed neural network (PINN) with the proposed adaptive architecture strategy to solve partial differential equations, we numerically show that adaptive PINNs not only are superior to standard PINNs but also produce interpretable hidden layers with provable stability. As a result, we also apply our architecture design strategy to solve inverse problems governed by elliptic partial differential equations.

42 ENGINEERING↗

The structure and propagation of laminar flames under autoignitive conditions

The laminar flame speed s l is an important reference quantity for characterising and modelling combustion. Experimental measurements of laminar flame speed require the residence time of the fuel/air mixture (τ f ) to be shorter than the autoignition delay time (τ). This presents a considerable challenge for conditions where autoignition occurs rapidly, such as in compression ignition engines. As a result, experimental measurements in typical compression ignition engine conditions do not exist. Simulations of freely propagating premixed flames, where the burning velocity is found as an eigenvalue of the solution, are also not well posed in such conditions, since the mixture ahead of the flame can autoignite, leading to the so called “cold boundary problem”. In this paper, a numerical method for estimating a reference flame speed, s R , is proposed that is valid for laminar flame propagation at autoignitive conditions. Two isomer fuels are considered to test this method: ethanol, which in the considered conditions is a single-stage ignition fuel; and dimethyl ether, which has a temperature-dependent single- or two-stage ignition and a negative temperature coefficient regime for τ. Calculations are performed for the flame position in a one-dimensional computational domain with inflow-outflow boundary conditions, as a function of the inlet velocity U I and for stoichiometric fuel–air premixtures. The response of the flame position, L F , to U I shows distinct stabilisation regimes. For single-stage ignition fuels, at low U I the flame speed exceeds U I and the flame becomes attached to the inlet. Above a critical U I value, the flame detaches from the inlet and L f becomes extremely sensitive to U I until, for sufficiently high U I , the sensitivity decreases and L f corresponds to the location expected from a purely autoignition stabilised flame. The transition from the attached to the autoignition regimes has a corresponding peak dL f /dU I value which is proposed to be a unique reference flame speed s R for single-stage ignition fuels. For two-stage ignition fuels, there is an additional stable regime where a high-temperature flame propagates into a pool of combustion intermediates generated by the first stage of autoignition. This results in two peaks in dL f /dU I and therefore two reference flame speed values. The lower value corresponds to the definition of s R for single-stage ignition fuels, while the higher value exists only for two-stage ignition fuels and corresponds to a high temperature flame propagating into the first stage of autoignition and is denoted s R ' . Finally, a transport budget analysis for low- and high-temperature radical species is also performed, which confirms that the flame structures at U I = s R and U I = s R ' do indeed correspond to premixed flames (deflagrations), as opposed to spontaneous ignition fronts which do not have a unique propagation speed.

33 ADVANCED PROPULSION SYSTEMS↗

Hydropower flexibility valuation tool for flow requirement evaluation

Timing of generation is becoming more and more valuable. This creates greater potential tension between environmental and power system objectives since both systems require their own flow patterns. Identifying win–win outcomes in this context requires being able to discuss the value of flexibility across stakeholder groups. This research proposes a two-stage optimization method to understand hydropower flexibility to meet both environmental and power system requirements. The tool simulates the two-settlement market process in the U.S. by maximizing revenues from both the day-ahead and real-time markets, subject to plant operational limits, regulatory flow and ramping requirements, and uncertainties associated with water availability and market prices. The model is formulated as linear programming problems and solved using IBM ILOG CPLEX optimizer. By examining a range of flow requirements, ramping constraints, and storage capacities, the proposed tool shows how to make more informed decisions to weigh the cost of specific flow requirements in the context of the overall license requirements. Results from the case study show that revenue is more sensitive to the ramping constraints than the minimum flow constraints. We also demonstrate that removing flow constraints in a dry month increases monthly revenue by up to 118%, as opposed to only 1% in a wet month. In addition, our results suggest that using a learning-based water flow forecast results in an increase of monthly revenue up to 6.4% compared with persistence forecast.

13 HYDRO ENERGY↗

Resilient co-expansion planning between gas and electric distribution networks against natural disasters

Resiliently designed and constructed integrated gas-electric distribution networks (GEDNs) against natural disasters are crucial to social welfare. In this study, a two-stage robust optimisation-based co-expansion planning model is proposed to attain an integrated GEDN with a given resilience level, by optimising the investment strategies of hardening and selective expansion of power distribution feeders and natural gas pipelines, as well as the location and capacity of natural-gas-fired distributed generation. In the first stage, the overall annual investment and operation cost is minimised under normal operation conditions while in the second stage, the feasibility of the investment decisions under the identified worst-case natural disaster scenario is checked with an adjustable load shedding cost criterion. The proposed model is formulated as a mixed integer second-order cone programming problem with the column and constraint generation algorithm employed to seek the optimal solution. Case studies on two integrated GEDNs demonstrate the performance of the proposed methodology.

Zou, Bo↗

Two-stage Raman compression of laser pulses with controllable phase fronts

The phase front of an ultrashort laser pulse undergoing Raman amplification and compression in inhomogeneous plasma can be controlled such that it is unaffected by density gradients and thus remains focusable. This is achieved by performing the Raman backscattering in two stages. At the first stage, the standard nonlinear Raman compression of a seed wave takes place and produces a short ultraintense pulse, which yet may be poorly focusable. At the second stage, this amplified pulse is scattered again, now serving as a pump, off a second copy of the initial seed. This stage, which utilizes a denser and shorter plasma, is intended not for compression but rather for passing a significant fraction of the energy to the second seed quickly. Then, the output pulse that is produced is not just short and ultraintense, but also has the smooth phase front of the original seed.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

A flexible and differentiable coil proxy for stellarator equilibrium optimization

Balancing plasma performance and coil cost is a significant challenge when designing a stellarator power plant. Most current stellarator designs are produced through two-stage optimization: stage-1 for the equilibrium and stage-2 for a coil design that reproduces its magnetic configuration. Because few proxies connect both stages, two-stage optimization can produce plasmas that have high-quality physical properties but overly complex coils. In recent years, single-stage optimization has increasingly been used to optimize the plasma and coils simultaneously in order to improve the plasma–coil balance. However, all existing single-stage tools are specialized for filament coils, cannot model coil systems containing permanent magnets or dipole arrays, and continue to be challenged by numerical problems. The quasi-single-stage optimization finds a middle ground by integrating a coil optimization subproblem into stage-1 optimization. We present a flexible, differentiable coil complexity proxy based on the newly developed QUADCOIL coil optimization code. QUADCOIL is fast and can target realistic coil metrics and constraints that are unavailable to codes with comparable speed. We demonstrate the effectiveness and flexibility of the QUADCOIL proxy by presenting two QSS optimization studies. The first study produces a permanent magnet solution for the MUSE stellarator with 34% fewer magnets than previous solutions. The second study produces a coil solution for the ARIES-CS stellarator with 30% reductions in both peak and root-mean-square force.

Fu, Lanke [Princeton Plasma Physics Laboratory (PP↗

Two-stage nonlinear compression of high-power femtosecond laser pulses

Two-stage compression of laser pulses with a power of 250 TW is experimentally realised by broadening their spectrum during self-phase modulation in fused silica and subsequent dispersion compensation upon reflection from chirping mirrors. A five-fold decrease in the duration is demonstrated, from 75 to 15 fs, with a B-integral value of about 5 at each stage. It is possible to avoid small-scale self-focusing due to self-filtering of the laser beam during free propagation in vacuum. With optimal parameters of the dispersive mirror, the pulse can be compressed to a duration of less than 5 fs. (paper)

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Optimizing Service Restoration in Distribution Systems With Uncertain Repair Time and Demand

This paper proposes a novel method to co-optimize the distribution system operation and repair crew routing for outage restoration after extreme weather events. A two-stage stochastic mixed integer linear program is developed. The first stage is to dispatch the repair crews to the damaged components. The second stage is distribution system restoration using distributed generators, and reconfiguration. We consider demand uncertainty in terms of a truncated normal forecast error distribution, and model the uncertainty of the repair time using a lognormal distribution. A new decomposition approach, combined with the progressive hedging algorithm, is developed for solving large-scale outage management problems in an effective and timely manner. Furthermore, the proposed method is validated on modified IEEE 34- and 8500-bus distribution test systems.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Chance-Constrained Energy Management System for Power Grids With High Proliferation of Renewables and Electric Vehicles

This paper proposes a two-stage energy management system (EMS) for power grids with massive integration of electric vehicles (EVs) and renewable energy resources. The first stage economic dispatch determines the optimal operating points of charging stations and battery swapping stations (BSS) for EVs under plug-in and battery swapping modes, respectively. The proposed stochastic model predictive control (SMPC) problem in this stage is characterized through a chance-constrained optimization formulation that can effectively capture the system and the forecast uncertainties. A distributed algorithm, the alternating direction method of multipliers (ADMM), is applied to accelerate the optimization computation through parallel computing. The second stage is aimed in coordinating the EV charging mechanisms to continuously follow the first-stage solutions, i.e., the target operating points, and meeting the EV customers' charging demands captured via the Advanced Metering Infrastructure (AMI). In this work, the proposed solution offers a holistic control strategy for large-scale centralized power grids in which the aggregated individual parameters are predictable and the system dynamics do not vary sharply within a short time-interval.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Fuzzy Energy and Reserve Co-optimization With High Penetration of Renewable Energy

In this study, we propose a fuzzy-based energy and reserve co-optimization model with consideration of high penetration of renewable energy. Under the assumption of a fixed uncertainty set of renewables, a two-stage robust model is proposed for clearing energy and reserves in the first stage and checking the feasibility and robustness of re-dispatches in the second stage. Fuzzy sets and their membership functions are introduced into the optimization model to represent the satisfaction degree of the variable uncertainty sets. The lower bound of the uncertainty set is expressed as fuzzy membership functions. The solutions are obtained by transforming the fuzzy mathematical programming formulation into traditional mixed integer linear programming problems.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Technique for the Formation of Antireflection Coatings Based on ITO Films

A technique for forming transparent conducting indium-tin-oxide films for applications, which require high current densities, are considered. These films have to combine a good surface conductivity with an increased transmittance of radiation within a wide wavelength range. A two-stage process is designed to fabricate such coatings: at the first stage, a dense indium-tin-oxide film with high electrical conductivity is formed; then, an additional antireflection layer of the same material is deposited. Possible methods for optimizing these additional indium-tin-oxide films obtained by forming nanostructured self-assembled coatings on the surface of dense films are investigated. It is shown that the best light extraction over a wide wavelength range is provided by the composite film-deposition technique, which consists in the deposition of self-assembled nanostructured coatings by electron-beam evaporation on a hot substrate and the further filling of voids by magnetron sputtering at room temperature. Since the optical properties of the obtained antireflection coatings depend on the mass ratio of materials deposited at each stage, the technique makes it possible to optimize the film properties on the basis of problems specified by the area of their application. Computer simulation of the behavior of the material distribution in the structure of the studied coatings confirms the ability of the method of combined film deposition to form the smoothest character of the change in the effective refractive index of the coating structure, which leads to the most efficient extraction of optical-range light.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Advanced High Torque Density Magnetically Geared Generator

The objective of this research project was to investigate the fundamental torque density limitations of multistage flux focusing magnetic gears and magnetically geared direct-drive generators for wind turbines. Magnetic gears create non-contact speed amplification by modulating the air-gap magnetic field harmonics created by magnetic rotors. By integrating a stator into the magnetic gear structure a very high torque density direct-drive generator can be created. Magnetic geared systems offer unique capabilities that are currently not available using existing mechanically geared systems, such as overload protection. Magnetic gears also offer the potential for higher reliability, higher efficiency and quieter operation than is currently achievable using a mechanically geared generator. By increasing reliability and design life magnetic gears could decrease the overall life-cycle costs when compared with equivalent mechanical or direct-drive drivetrains. This research project investigated and experimentally testing two sub-scale magnetic gearboxes (0.12 m diameter) and two large diameter magnetic gearboxes (0.56m and 0.63m diameter). The goal of the project was to demonstrate a torque density >300Nm/L for a multistage magnetic gearbox with integrated stator. A nested multistage magnetic gear structure was first proposed and studied. However, it was shown that the nested design greatly increased mechanical complexity and therefore a two-stage series typology was utilized. A multi-stage magnetic gear with a 59:1 gear ratio was successfully experimentally tested. A peak torque of 4.25 kN·m was measured and this resulted in an overall active region torque density of 141 N·m/L (when including support rings). The torque density was significantly reduced from what was computed using 2-D electromagnetic analysis. This reduction was due to the 3-D edge effects, increased power losses and the inclusion of additional mechanical support rings as well as other mechanical assembly design changes. At large diameters contiguous lamination parts could not be fabricated and so a segmented laminated rotor design had to be used. This reduced the mechanical strength of the design when compared to small diameter magnetic gears and resulted in the need for the addition of mechanical support rings to prevent deflection. The final experimentally tested design also had a high torque ripple and the power loss within the magnetic gear did not change significantly with load. This therefore resulted in the efficiency being high only at peak loads. Despite not experimentally achieving the 2-D calculated high torque density the analysis has shown that the problems encountered with the scaling can be overcome with careful design changes. For example, understanding the unique mechanical deflection requirements during the electromagnetic design of the large diameter magnetic gears is key to achieving both a high torque density based on the electromagnetic analysis and a high torque density based on the mechanical support needs. Furthermore, ensuring that end plates are made of non-conductive material and minimizing other conductive support elements within the magnetic structure is a key requirement for reducing eddy current losses. In summary, if the mechanical design can be made more robust and eddy current loss mitigation strategies are used in the mechanical design phase then the magnetic gear is capable of being scaled up whilst operating at a very high torque density. This will then enable it to be potentially performance and cost competitive with alternative direct-drive and mechanically geared wind turbine drivetrains.

17 WIND ENERGY↗

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↗

New physical processes for extracting GPDs with a better sensitivity to partonic structure

We introduce a new type of exclusive processes for a better study of generalized parton distributions (GPDs), which we refer to as single-diffractive hard exclusive processes (SDHEPs). We advocate a two-stage framework for picturing SDHEPs based on the separation of scales, which gives a clear description both kinematically and dynamically. We examine the sensitivity of the SDHEP to the parton momentum fraction x -dependence of GPDs, and demonstrate it quantitatively with two specific processes that can be readily measured at J-PARC or AMBER using a pion beam and at JLab using a photon beam, respectively. Both processes are capable of providing enhanced sensitivity to the x -dependence, overcoming the problem of shadow GPDs, and disentangling different types of GPDs with various spin asymmetries.

Qiu, Jianwei↗