Topology-Aware Performance Optimization and Modeling of Adaptive Mesh Refinement Codes for Exascale.
Abstract not provided.
Chan, Cy P↗
SEARCH · Engineering Papers
Results for “code optimization”
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.
TRANSP-based optimization towards tokamak scenario development
An optimization approach that incorporates the predictive transport code TRANSP is proposed for tokamak scenario development. Optimization methods are often employed to develop open-loop control strategies to aid access to high performance tokamak scenarios. Generally, the optimization approaches use control-oriented models, i.e. models that are reduced in complexity and prediction accuracy as compared to physics-oriented transport codes such as TRANSP. In the presented approach, an optimization procedure using the TRANSP code to simulate the tokamak plasma is considered for improved predictive capabilities. As a test case, the neutral beam injection (NBI) power is optimized to develop a control strategy that maximizes the noninductive current fraction during the ramp-up phase for NSTX-U. Simulation studies towards the achievement of noninductive ramp-up in NSTX-U have already been carried out with the TRANSP code. The optimization-based approach presented in this work is used to maximize the noninductive current fraction during ramp-up in NSTX-U, demonstrating that the scenario development task can be automated. An additional test case considers optimization of the current ramp rate in DIII-D for obtaining a stationary plasma characterized by a flat loop voltage profile in the flattop phase.
70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
TRANSP-based Optimization Towards Tokamak Scenario Development
An optimization approach that incorporates the predictive transport code TRANSP is proposed for tokamak scenario development. Optimization methods are often employed to develop open-loop control strategies to aid access to high performance tokamak scenarios. In general, the optimization approaches use control-oriented models, i.e. models that are reduced in complexity and prediction accuracy as compared to physics-oriented transport codes such as TRANSP. In the presented approach, an optimization procedure using the TRANSP code to simulate the tokamak plasma is considered for improved predictive capabilities. As a test case, the neutral beam injection (NBI) power is optimized to develop a control strategy that maximizes the non-inductive current fraction during the ramp-up phase for NSTX-U. Simulation studies towards the achievement of non-inductive ramp up in NSTX-U have already been carried out with the TRANSP code. The optimization-based approach proposed in this work is used to maximize the non-inductive current fraction during ramp-up in NSTX-U, demonstrating that the scenario development task can be automated. An additional test case considers optimization of the current ramp rate in DIII-D for obtaining a stationary plasma characterized bya flat loop voltage profile in the flattop phase.
DIII-D↗
Light Water Reactor LEU+ Lattice Optimization
Commercial light water reactor (LWR) operators and fuel vendors in the United States are exploring potential changes to nuclear fuel that include low-enriched uranium plus (LEU+) designs to further improve operational economics (e.g., extend cycle length). LEU+ fuel is fuel with a maximum enrichment between 5 wt% and 10 wt% 235 U; it allows for higher assembly burnup but likely requires additional reactivity control, e.g., increased burnable absorbers. This report examines possible LEU+ fuel lattice design changes using the lattice physics code, SCALE/Polaris. An optimization driver called the Metaheuristic Optimization Tool (MOT) is used to automate domain space exploration and optimization of LEU+ lattice designs. Heuristics from previous LWR lattice optimization studies were used to construct the objective function and define the domain space for optimization. This work successfully demonstrated that the optimization algorithms of MOT can generate feasible, nonproprietary LEU+ lattice designs (GE14 10 × 10 and Westinghouse 17 × 17) that meet the constraints of traditional LWR lattices while extending cycle length.
21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
Implementation of a Particle Accelerator Beam Dynamics Code on Multi-Node GPUs
Particle accelerators play an important role in a wide range of scientific discoveries and industrial applications. The self-consistent multi-particle simulation based on the particle-in-cell (PIC) method has been used to study charged particle beam dynamics inside those accelerators. However, the PIC simulation is time-consuming and needs to use modern parallel computers for high-resolution applications. In this paper, we implemented a parallel beam dynamics PIC code on multi-node hybrid architecture computers with multiple Graphics Processing Units (GPUs). We used two methods to parallelize the PIC code on multiple GPUs and observed that the replication method is a better choice for moderate problem size and current computer hardware while the domain decomposition method might be a better choice for large problem size and more advanced computer hardware that allows direct communications among multiple GPUs. Using the multi-node hybrid architectures at Oak Ridge Leadership Computing Facility (OLCF), the optimized GPU PIC code achieves a reasonable parallel performance and scales up to 64 GPUs with 16 million particles.
43 PARTICLE ACCELERATORS↗
Yearly report for the period Jan. 2018 – Feb. 2019 IC Project: w17_faultprediction “Critical Stress in Earth”
The consequence of earthquake is usually catastrophic. Currently, we cannot determine when a fault is approaching failure due to either anthropogenic activities or tectonic movement. In an anthropogenically perturbed setting, the rate of physical changes on faults far exceeds most natural processes with currently unpredictable and unmeasurable results. In a tectonic setting we generally do not know where a fault is in the earthquake cycle beyond noting how long has passed since the last rupture, which is itself sometimes unknown. The Laboratory has therefore funded a multi million multi-year LDRD-DR program that comprises experiments, numerical modeling, and machine learning in order to develop and test a highly novel approach to probe the earthquake cycle and in particular, to detect and locate stressed faults that are approaching failure. For this program, modeling of a series of sheared granular layers undergoing stick-slip behavior using the EES-17’s Combined Finite-Discrete Element Method (FDEM) based code HOSS (Hybrid Optimization Software Suite) plays a pivotal role since numerical simulations of granular fault gouge allow for analysis of the mechanical behavior of the system at a level of spatial and temporal resolution not accessible experimentally, and also offer the possibility of identifying the most important parameters governing earthquake dynamics. The simulated data also provide very detailed stress, strain and motion information for machine learning purpose to predict earthquake. The code – HOSS – has been developed in Geophysics Group in LANL during the past ten years and is used to simulate the stick-slip behavior of granular fault gouge. One of the key feature of HOSS is that it is superior to pure finite element method and discrete element method and can provide very detailed evolvement information of granular fault gouge. Additionally, its parallel computing capability makes large scale earthquake gouge simulation possible. Taking these facts into account, and also because few other fully parallel FDEM code exists, this IC project allows LANL to stay at the forefront of earthquake gouge and granular material simulation in the world.
42 ENGINEERING↗
New method to design stellarator coils without the winding surface
Finding an easy-to-build coils set has been a critical issue for stellarator design for decades. Conventional approaches assume a toroidal 'winding' surface, but a poorly chosen winding surface can unnecessarily constrain the coil optimization algorithm, This article presents a new method to design coils for stellarators. Each discrete coil is represented as an arbitrary, closed, one-dimensional curve embedded in three-dimensional space. A target function to be minimized that includes both physical requirements and engineering constraints is constructed. The derivatives of the target function with respect to the parameters describing the coil geometries and currents are calculated analytically. A numerical code, named flexible optimized coils using space curves (FOCUS), has been developed. Furthermore, applications to a simple stellarator configuration, W7-X and LHD vacuum fields are presented.
70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
SPEARS: A Database-Invariant Spectral modeling API
The Spectral Physics Environment for Advanced Remote Sensing (SPEARS) application programming interface (API) is a Python-based, line-by-line, local thermal equilibrium (LTE) spectral modeling code which is optimized for simultaneously synthesizing optical spectra from any combination of fundamental spectroscopic databases. In this article, we contribute two novel spectral modeling techniques to the scientific literature. First we describe how SPEARS integrates a physics-based collisional model for calculating pressure broadening in the absence of available broadening coefficients. With this collisional model implementation, a generalized approach to fundamental spectroscopic databases can be achieved across multiple databases. We also detail our adaptive grid mesh algorithm developed to make the code scalable for simulating large spectral bandwidths at high spectral fidelity using intuitive grid parameters. Here, we present comparisons to other modeling tools, experiments, and provide a discussion on the SPEARS user interface.
47 OTHER INSTRUMENTATION↗
Resilient Operation of Networked Community Microgrids with High Solar Penetration
This project, funded by the US Department of Energy’s Solar Energy Technologies Office (SETO), focused on the operation of microgrids as a coordinated network. The primary objective, which was successfully achieved, was to develop both control strategies and hardware solutions to support the resilient and efficient operation of networked microgrids with high solar penetration. The work was structured around the following four main tasks: • Development of distributed and scalable optimization algorithms for AC-coupled networked microgrids. • Design and implementation of a novel DC interconnection hardware to enable precise power exchange between microgrids. • Laboratory operational validation of the developed technologies using 480 V testbeds and commercially available hardware. • Field operational validation of the complete solution in Adjuntas, Puerto Rico, interconnecting two kW-scale, split-phase microgrids of Casa Pueblo’s microgrids. This project addressed multiple technical challenges across the domains of optimization, control, hardware interconnection, and protection. One of its key contributions was delivering tangible, real-world solutions for networking microgrids. In contrast to purely theoretical or simulation-based work, this project included full-scale hardware operational validation both in the lab and in the field. The work conducted as part of this project—in collaboration with the University of Puerto Rico; the University of Tennessee, Knoxville; the University of Central Florida; and Casa Pueblo—has advanced the state of the art in networked microgrids. Key contributions include the development of distributed control strategies, practical solutions for real-world implementation challenges, and the introduction of a novel DC interlink approach for microgrid interconnection. The project featured both laboratory and field validation using commercial off-the-shelf components. The field deployment successfully validated that a group of microgrids can operate in a coordinated manner, enabling precise power flow between systems and mutual support during extreme events. This project resulted in 15 journal publications and 15 conference papers; 5 graduate students and 15 undergraduate students were supported. The codes of distributed optimization and forecasting were made open-source through OSTI.gov for distributed optimization and forecasting. All the publications are available in the ORNL-hosted project landing page. The DC interlink with state-of-charge balancing control was operationally validated in Adjuntas by interconnecting two real-world, 240 V split-phase microgrids. To the best knowledge of the team, this represents the first operational validation of AC microgrids interconnected via DC-interlinks. As a culmination of this project, a follow-on grant was awarded to support the technology transfer of the distributed optimization framework to a commercial microgrid controller, Stellar Edge, developed by the California-based company New Sun Road.
14 SOLAR ENERGY↗
Comprehensive Margin Assessment of the ASME Section III, Division 5, Class A Primary Load Design Rules
This report provides a comprehensive margin assessment of the ASME Section III, Division 5, Subsection HB, Subpart B rules for the design of high temperature reactor components against load controlled stress limits. These rules, often called the Code primary load design rules, provide protection against creep rupture and plastic collapse under steady conditions. The method adopted here is to compare the deterministic ASME design life, according to the primary load design rules, against the expected, statistical service life of the component, considering creep rupture as the relevant failure mode. The report provides the design margin of the ASME rules in terms of the probability of premature failure -- the probability that the component fails in service before reaching the ASME design life. To complete this assessment, the report describes the development of a statistical creep life assessment procedure accounting for variations in the component loading, material creep rate, and the material rupture time. This includes a novel method for correlating creep rupture data using a Gaussian process accounting for heat-to-heat variation in the Larson-Miller time-temperature parameter. The report applies the complete margin assessment process to quantify the design margin inherent in the ASME primary load design rules as a function of temperature for relevant component geometries and materials. The overall conclusion is that the ASME rules are conservative, providing designs with very small probability of premature failure, but some optimization of the Code design rules is possible to provide a more uniform design margin.
22 GENERAL STUDIES OF NUCLEAR REACTORS↗
Ocean power technology design optimization
For this study, the National Renewable Energy Laboratory and Ocean Power Technologies (OPT) conducted a collaborative code validation and design optimization study for OPT's PowerBuoy wave energy converter (WEC). NREL utilized WEC-Sim, an open-source WEC simulator, to compare four design variations of OPT's PowerBuoy. As an input to the WEC-Sim models, viscous drag coefficients for the PowerBuoy floats were first evaluated using computational fluid dynamics. The resulting WEC-Sim PowerBuoy models were then validated with experimental power output and fatigue load data provided by OPT. The validated WEC-Sim models were then used to simulate the power performance and loads for operational conditions, extreme conditions, and directional waves, for each of the four PowerBuoy design variations, assuming the wave environment of Humboldt Bay, California. And finally, ratios of power-to-weight, power-to-fatigue-load, power-to-maximum-extreme-load, power-to-water-plane-area, and power-to-wetted-surface-area were used to make a final comparison of the potential PowerBuoy WEC designs. Lastly, the design comparison methodologies developed and presented in this study are applicable to other WEC devices and may be useful as a framework for future WEC design development projects.
16 TIDAL AND WAVE POWER↗
Machine learning application to single channel design of molten salt reactor
This study proposes a robust approach to quickly design a nuclear reactor core and explores the best performing machine learning (ML) technique for predicting feature parameters of the core. Here we implemented the approach into a hypothetical channel of molten salt reactors to demonstrate the applicability of the method. We prepared a Python tool, named Plankton, which couples to a reactor physics code and an optimization tool, and imports ML methods. The tool performs three consecutive phases: reactor database generation, machine learning application, and design optimization. We identified the extra trees method as the best performing estimator. With the estimator, we found nine optimum designs in total, one for each fuel-salt pair, and estimated all the performance metrics of the designs with a <5% prediction error compared to their actual values. U-Pu-NaCl fuel-salt gave promising results with the highest conversion ratio, the most negative feedback coefficient, and the lowest fast flux.
21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
A Green's function quantum average atom model
A quantum average atom model is reformulated using Green's functions. This allows integrals along the real energy axis to be deformed into the complex plane. The advantage being that sharp features such as resonances and bound states are broadened by a Lorentzian with a half-width chosen for numerical convenience. An implementation of this method therefore avoids numerically challenging resonance tracking and the search for weakly bound states, without changing the physical content or results of the model. A straightforward implementation results in up to a factor of 5 speed-up relative to an optimized orbital based code.
70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
Constructing stellarators with quasisymmetry to high order
Here, a method is given to rapidly compute quasisymmetric stellarator magnetic fields for plasma confinement, without the need to call a three-dimensional magnetohydrodynamic equilibrium code inside an optimization iteration. The method is based on direct solution of the equations of magnetohydrodynamic equilibrium and quasisymmetry using Garren & Boozer’s expansion about the magnetic axis ( Phys Fluids B, vol. 3, 1991, p. 2805), and it is several orders of magnitude faster than the conventional optimization approach. The work here extends the method of Landreman et al. ( J. Plasma Phys. , vol. 85, 2019, 905850103), which was limited to flux surfaces with elliptical cross-section, to higher order in the aspect-ratio expansion. As a result, configurations can be generated with strong shaping that achieve quasisymmetry to high accuracy. Using this construction, we give the first numerical demonstrations of Garren and Boozer’s ideal scaling of quasisymmetry breaking with the cube of the inverse aspect ratio. We also demonstrate a strongly non-axisymmetric configuration (vacuum rotational transform $\unicode[STIX]{x1D704} >0.4$ ) in which symmetry-breaking mode amplitudes throughout a finite volume are ${<}2\times 10^{-7}$ , the smallest ever reported. To generate boundary shapes of finite-minor-radius configurations, a careful analysis is given of the effect of substituting a finite minor radius into the near-axis expansion. In conclusion, the approach here can provide analytic insight into the space of possible quasisymmetric stellarator configurations, and it can be used to generate good initial conditions for conventional stellarator optimization.
70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
Isochoric heating with laser-accelerated proton beams
Numerical simulations are used to study acceleration of protons by intense laser pulses and the induced heating of a secondary target using this proton beam. Laser-plasma interaction and ion acceleration are modeled with a multidimensional, relativistic particle-in-cell code. Proton slowing-down and secondary target heating are computed with a two-dimensional Monte Carlo and a one-dimensional Lagrangian hydrodynamics codes. Strategies for optimization of heating uniformity include tailoring the accelerated proton spectrum with specific target geometries or laser pulse parameters. A trade-off must then be found between heating uniformity and efficiency. The parameters (temperature, depth and radius of heated zone, characteristic heating and hydrodynamics times) that could be obtained on current short pulse laser facilities, using a typical 50 fs Gaussian pulse with an intensity of several 10{sup 19} W/cm{sup 2}, are explored.
70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
A study of the role of various reactions on the density distribution of hydrogen, silylene, and silyl in SiH{sub 4}/H{sub 2} plasma discharges
Fluid model has been used to study the effect of pressure on the distribution of hydrogen (H), silylene (SiH{sub 2}), and silyl (SiH{sub 3}) in hydrogen silane plasma discharges used for deposition of hydrogenated microcrystalline silicon ({mu}c-Si:H) or hydrogenated amorphous silicon (a-Si:H) thin films for solar cells. Time averaged reaction rates have been calculated to study the influence of various reactions on the density distributions of hydrogen, silylene, and silyl. Change in the distributions of hydrogen and silylene from bell shaped distribution at low pressure (1 Torr) to double humped distribution at high pressure (5 Torr) is explained with the help of time averaged reaction rates. Important reactions have been identified that contribute to the production and consumption of hydrogen (H), silylene (SiH{sub 2}), and silyl (SiH{sub 3}). The hydrogen consumption reactions SiH{sub 4} + H {yields} SiH{sub 3} + H{sub 2} and SiH{sub 3} + H {yields} SiH{sub 2} + H{sub 2} are found to play a central role in deciding the distribution of hydrogen. On the other hand, silylene consumption reaction SiH{sub 2} + H{sub 2}{yields} SiH{sub 4} is found to play a central role in determining the distribution of silylene. The distribution of these species at high pressure has been explained by using time averaged continuity equation. The code has been optimized by identifying 33 reactions (out of 53 reactions which contribute in the production and consumption of H, SiH{sub 2}, and SiH{sub 3}) that have no net effect on the density and distribution of these species. It is observed that dropping of 33 reactions has insignificant effect on the density of all the thin film deposition precursors such as Si, SiH, SiH{sub 2}, SiH{sub 3}, and Si{sub 2}H{sub 5}. This reduced set of 20 reactions can be used instead of 53 reactions to calculate the density and distribution of H, SiH{sub 2}, and SiH{sub 3} in the fluid simulation of SiH{sub 4}/H{sub 2} plasma discharges.
70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
First principles transport coefficients and reaction rates of Ar{sub 2}{sup +} ions in argon for cold plasma jet modeling
Momentum-transfer collision cross-sections and integral collision cross-sections for the collision-induced dissociation are calculated for collisions of ionized argon dimers with argon atoms using a nonadiabatic semiclassical method with the electronic Hamiltonian calculated on the fly via a diatomics-in-molecules semiempirical model as well as inverse-method modeling based on simple isotropic rigid-core potential. The collision cross-sections are then used in an optimized Monte Carlo code for evaluations of the Ar{sub 2}{sup +} mobility in argon gas, longitudinal diffusion coefficient, and collision-induced dissociation rates. A thorough comparison of various theoretical calculations as well as with available experimental data on the Ar{sub 2}{sup +} mobility and collision cross-sections is performed. Good agreement is found between both theoretical approaches and the experiment. Analysis of the role of inelastic processes in Ar{sub 2}{sup +}/Ar collisions is also provided.
37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
Determination of anharmonic free energy contributions: Low temperature phases of the Lennard-Jones system
We investigate a general method to calculate the free energy of crystalline solids by considering the harmonic approximation and quasistatically switching the anharmonic contribution. The advantage of this method is that the harmonic approximation provides an already very accurate estimate of the free energy, and therefore the anharmonic term is numerically very small and can be determined to high accuracy. We further show that the anharmonic contribution to the free energy satisfies a number of exact inequalities that place constraints on its magnitude and allows approximate but fast and accurate estimates. The method is implemented into a readily available general software by combining the code HOODLT (Highly Optimized Object Oriented Dynamic Lattice Theory) for the harmonic part and the molecular dynamics (MD) simulation package HOOMD-blue for the anharmonic part. We use the method to calculate the low temperature phase diagram for Lennard-Jones particles. We demonstrate that hcp is the equilibrium phase at low temperature and pressure and obtain the coexistence curve with the fcc phase, which exhibits reentrant behavior. Furthermore, several implications of the method are discussed.
36 MATERIALS SCIENCE↗