SEARCH · Search NASA
Results for “benchmark problem”
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.
Advanced trust region optimization strategies for glass box / black box models
We present an improved trust region filter (TRF) method for optimization of combined glass box/black box systems. Glass box systems refer to models that are easily expressed in an algebraic modeling language, providing cheap and accurate derivative information. By contrast, black box systems may be computationally expensive and derivatives are unavailable. The TRF method, as first introduced in our previous work (Eason and Biegler, AIChE J . 2016; 62:3124–3136), is able to handle hybrid systems containing both glass and black box components, which can frequently arise in chemical engineering, for example, when a multiphase reactor model is included in a flow sheet optimization problem. We discuss several recent modifications in the algorithm such as the sampling region, which maintains the algorithm's global convergence properties without requiring the trust region to shrink to zero in the limit. To benchmark the development of this optimization method, a test set of problems is generated based on modified problems from the CUTEr and COPS sets. The modified algorithm demonstrates improved performance using the test problem set. Finally, the algorithm is implemented within the Pyomo environment and demonstrated on a rigorous process optimization case study for carbon capture. © 2018 American Institute of Chemical Engineers AIChE J , 64: 3934–3943, 2018
MCS solutions for the TVA Watts Bar Unit 1 multi-physics depletion benchmark
The high-fidelity solution for exercises 1-3 of the recently developed TVA Watts Bar Unit 1 multi-physics and multi-cycle depletion benchmark using MCS was coupled with two different thermal-hydraulics (TH) codes: TH1D and COBRA-TF (CTF), and their solutions were compared. The solutions from MCS/CTF and MCS/TH1D for typical Pressurized Water Reactor (PWR) problems are quite close. And to evaluate MCS solutions for the benchmark, their results were also compared against measured data as well as publicly available high-fidelity solutions. The MCS criticality solutions at Hot Zero Power (HZP) deviate within 50 pcm against the reference solutions. While for the critical boron concentration (CBC) search calculations, the difference on the MCS calculated CBC is about 7 ppm, and the root-mean-squared (RMS) errors for the assembly power distribution and outlet coolant temperature are less than 0.5% and 1 Celsius degree respectively. (authors)
INL Results for Phases I and III of the OECD/NEA MHTGR-350 Benchmark
The Idaho National Laboratory (INL) Very High Temperature Reactor (VHTR) Technology Development Office (TDO) Methods Core Simulation group led the construction of the Organization for Economic Cooperation and Development (OECD) Modular High Temperature Reactor (MHTGR) 350 MW benchmark for comparing and evaluating prismatic VHTR analysis codes. The benchmark is sponsored by the OECD's Nuclear Energy Agency (NEA), and the project will yield a set of reference steady-state, transient, and lattice depletion problems that can be used by the Department of Energy (DOE), the Nuclear Regulatory Commission (NRC), and vendors to assess their code suits. The Methods group is responsible for defining the benchmark specifications, leading the data collection and comparison activities, and chairing the annual technical workshops. This report summarizes the latest INL results for Phase I (steady state) and Phase III (lattice depletion) of the benchmark. The INSTANT, Pronghorn and RattleSnake codes were used for the standalone core neutronics modeling of Exercise 1, and the results obtained from these codes are compared in Section 4. Exercise 2 of Phase I requires the standalone steady-state thermal fluids modeling of the MHTGR-350 design, and the results for the systems code RELAP5-3D are discussed in Section 5. The coupled neutronics and thermal fluids steady-state solution for Exercise 3 are reported in Section 6, utilizing the newly developed Parallel and Highly Innovative Simulation for INL Code System (PHISICS)/RELAP5-3D code suit. Finally, the lattice depletion models and results obtained for Phase III are compared in Section 7. The MHTGR-350 benchmark proved to be a challenging simulation set of problems to model accurately, and even with the simplifications introduced in the benchmark specification this activity is an important step in the code-to-code verification of modern prismatic VHTR codes. A final OECD/NEA comparison report will compare the Phase I and III results of all other international participants in 2014, while the remaining Phase II transient case results will be reported in 2015.
Development of CSG-based radiation shielding module for ARCHER: preliminary results for photons - Paper 126
ARCHER is a parallel Monte Carlo radiation transport code being developed for various hardware platforms including multi-core CPUs, the emerging Nvidia graphics processing units (GPUs) and Intel Xeon Phi coprocessors. Previous studies have demonstrated the accuracy and speed for CT imaging and radiotherapy dosimetry applications. Radiation shielding applications, however, require ARCHER to be more flexible in defining and handling geometries, in addition to the voxel phantom geometries that are commonly used in medical physics calculations. In this paper, we present the development of a general geometry module used for radiation protection and shielding problems in ARCHER. The methods of geometry design and particle transport for the GPUs are introduced. Preliminary results comparing against the MCNP code for cases involving Cs-137 0.662 MeV gammas are reported. The radiation shielding design module of ARCHER has been developed, which supports general geometry setups in the CSG format. The code was implemented using CUDA on Nvidia GPUs, and the performance was benchmarked against the MCNP code for a simple radiation shielding problem. Results showed that ARCHER worked properly and achieved impressive speedup due to highly parallel computing nature of the Nvidia GPUs compared to the MCNP code running on multi-core CPUs. Further optimization of the geometry module will be carried out using advanced geometry structures and variance reduction techniques to improve the precision and performance of ARCHER. (authors)
Quantum annealing for combinatorial optimization: a benchmarking study
Quantum annealing (QA) has the potential to significantly improve solution quality and reduce time complexity in solving combinatorial optimization problems compared to classical optimization methods. However, due to the limited number of qubits and their connectivity, the QA hardware did not show such an advantage over classical methods in past benchmarking studies. Recent advancements in QA with more than 5000 qubits, enhanced qubit connectivity, and the hybrid architecture promise to realize the quantum advantage. Here, we use a quantum annealer with state-of-the-art techniques and benchmark its performance against classical solvers. To compare their performance, we solve over 50 optimization problem instances represented by large and dense Hamiltonian matrices using quantum and classical solvers. The results demonstrate that a state-of-the-art quantum solver has higher accuracy (~0.013%) and a significantly faster problem-solving time (~6561×) than the best classical solver. Our results highlight the advantages of leveraging QA over classical counterparts, particularly in hybrid configurations, for achieving high accuracy and substantially reduced problem solving time in large-scale real-world optimization problems.
Benchmarking of collision operators with momentum source corrections
Several linearized collision operator models with parallel momentum conservation enforced by source functions are benchmarked against the correct collision operator based on Rosenbluth potentials. The model operators without energy diffusion allow for an analytic solution of a generalized Spitzer problem. For the other operators under investigation, an integrodifferential equation is solved. The benchmarking is performed both for the parallel conductivity with the ions assumed at rest and for the bootstrap current, where ion and electron flows are collisionally coupled. The accuracy of the results obtained with the operators varies from rather poor in the case of the simplest monoenergetic model to quite satisfactory for the models employing an energy-weighted parallel momentum source function.
Validation of time-dependent shift using the pulsed sphere benchmarks
The detailed behavior of neutrons in a rapidly changing time-dependent physical system is a challenging computational physics problem, particularly when using Monte Carlo methods on heterogeneous high-performance computing architectures. A small number of algorithms and code implementations have been shown to be performant for time-independent (fixed source and k-eigenvalue) Monte Carlo, and there are existing simulation tools that successfully solve the time-dependent Monte Carlo problem on smaller computing platforms. To bridge this gap, a time-dependent version of ORNL’s Shift code has been recently developed. Shift’s history-based algorithm on CPUs, and its event-based algorithm on GPUs, have both been observed to scale well to very large numbers of processors, which motivated the extension of this code to solve time-dependent problems. The validation of this new capability requires a comparison with time-dependent neutron experiments. Lawrence Livermore National Laboratory’s (LLNL) pulsed sphere benchmark experiments were simulated in Shift to validate both the time-independent as well as new time-dependent features recently incorporated into Shift. A suite of pulsed-sphere models was simulated using Shift and compared to the available experimental data and simulations with MCNP. Overall results indicate that Shift accurately simulates the pulsed sphere benchmarks, and that the new time-dependent modifications of Shift are working as intended. Validated exascale neutron transport codes are essential for a wide variety of future multiphysics applications.
Preliminary Validation of the Shift Monte Carlo Code for Fixed-Source Radiation Transport Problems
Simulations of neutron and photon transport using the Shift Monte Carlo radiation transport code are compared with experimental measurements and their corresponding benchmark simulations from several sets of experiments. Overall, Shift results match the calculations made by the benchmark teams quite well and match the measured values, which typically have large uncertainties, fairly well. A variety of attenuation/scattering problems are examined, as well as a streaming problem and a skyshine problem.
Evaluation and Testing of the ADVANTG Code on SNM Detection
Pacific Northwest National Laboratory (PNNL) has been tasked with evaluating the effectiveness of ORNL’s new hybrid transport code, ADVANTG, on scenarios of interest to our NA-22 sponsor, specifically of detection of diversion of special nuclear material (SNM). PNNL staff have determined that acquisition and installation of ADVANTG was relatively straightforward for a code in its phase of development, but probably not yet sufficient for mass distribution to the general user. PNNL staff also determined that with little effort, ADVANTG generated weight windows that typically worked for the problems and generated results consistent with MCNP. With slightly greater effort of choosing a finer mesh around detectors or sample reaction tally regions, the figure of merit (FOM) could be further improved in most cases. This does take some limited knowledge of deterministic transport methods. The FOM could also be increased by limiting the energy range for a tally to the energy region of greatest interest. It was then found that an MCNP run with the full energy range for the tally showed improved statistics in the region used for the ADVANTG run. The specific case of interest chosen by the sponsor is the CIPN project from Las Alamos National Laboratory (LANL), which is an active interrogation, non-destructive assay (NDA) technique to quantify the fissile content in a spent fuel assembly and is also sensitive to cases of material diversion. Unfortunately, weight windows for the CIPN problem cannot currently be properly generated with ADVANTG due to inadequate accommodations for source definition. ADVANTG requires that a fixed neutron source be defined within the problem and cannot account for neutron multiplication. As such, it is rendered useless in active interrogation scenarios. It is also interesting to note that this is a difficult problem to solve and that the automated weight windows generator in MCNP actually slowed down the problem. Therefore, PNNL had determined that there is not an effective tool available for speeding up MCNP for problems such as the CIPN scenario. With regard to the Benchmark scenarios, ADVANTG performed very well for most of the difficult, long-running, standard radiation detection scenarios. Specifically, run time speedups were observed for spatially large scenarios, or those having significant shielding or scattering geometries. ADVANTG performed on par with existing codes for moderate sized scenarios, or those with little to moderate shielding, or multiple paths to the detectors. ADVANTG ran slower than MCNP for very simply, spatially small cases with little to no shielding that run very quickly anyway. Lastly, ADVANTG could not solve problems that did not consist of fixed source to detector geometries. For example, it could not solve scenarios with multiple detectors or secondary particles, such as active interrogation, neutron induced gamma, or fission neutrons.
A Comprehensive Comparative Study of Active Learning Schemes for Nanophotonics Design
We present a benchmarking study of active learning (AL) schemes for designing planar multilayer nanophotonic metamaterials, where the design tasks are formulated as binary optimization problems. Different surrogate models, including factorization machine (FM), Gaussian process regression (GPR), and convolutional neural network (CNN), combined with different optimization methods, including exhaustive enumeration, discrete particle swarm optimization (DPSO), quantum annealing (QA), hybrid QA, and simulated annealing are studied. The benchmark cases investigated range from small problems with short binary lengths (N = 25) to large problems with N up to 100, focusing on the design of two classes of photonic structures, including antireflective coatings for the long-wavelength infrared region and transparent radiative coolers. For small problems, CNN coupled with DPSO in AL achieves the best performance. As N increases, FM with QA outperforms GPR and CNN. For FM-based AL, hybrid QA yields the best optimization results, particularly in high-dimensional cases (N = 100). These results demonstrate that the optimization method can significantly affect in AL performance as N increases, and that QA-based optimization can provide practical routes for mitigating the optimization bottleneck in high-dimensional problems.
OpenGraphGym: A Parallel Reinforcement Learning Framework for Graph Optimization Problems
This paper presents an open-source, parallel AI environment (named OpenGraphGym) to facilitate the application of reinforcement learning (RL) algorithms to address combinatorial graph optimization problems. This environment incorporates a basic deep reinforcement learning method, and several graph embeddings to capture graph features, it also allows users to rapidly plug in and test new RL algorithms and graph embeddings for graph optimization problems. This new open-source RL framework is targeted at achieving both high performance and high quality of the computed graph solutions. This RL framework forms the foundation of several ongoing research directions, including 1) benchmark works on different RL algorithms and embedding methods for classic graph problems; 2) advanced parallel strategies for extreme-scale graph computations, as well as 3) performance evaluation on real-world graph solutions.
The inhibiting effect of dislocation helices on the stress-induced orientation of S' precipitates in Al–Cu–Mg alloy
The phenomenon of restrained stress-induced preferential orientation of S′ precipitates is investigated using a single-crystal of Al–1.23Cu–0.43 Mg alloy. Al–1.23Cu–0.43 Mg single-crystal specimens are subjected to stress aging, and the microstructure is analyzed by transmission electron microscopy (TEM). It is found that the stress-induced preferential orientation of S′ precipitates is restrained owing to the dislocations produced by a higher stress. The effect of dislocations on the oriented precipitates depends on the total length of the intersection lines for precipitate habit planes and dislocation glide planes. This investigation not only provides important insight into solving the anisotropy problem attributed to precipitation strengthening, but also offers a benchmark for choosing the appropriate stress range in manufacturing of Al–Cu–Mg alloys. - Highlights: • Single crystals of an Al–Cu–Mg alloy were prepared for the investigations. • A phenomenon of restrained stress-induced preferential orientation of S′ precipitates was found. • The influence of dislocation helices on precipitation during stress-aging was studied. • Difference of orientation degree of S′ precipitates and θ′ precipitates was explained. • A basis for choosing the appropriate stress range in manufacturing of Al–Cu–Mg alloys is provided.
Coannihilation scenarios and particle spectroscopy in SU(4){sub c}xSU(2){sub L}xSU(2){sub R}
We identify a variety of coannihilation scenarios in a supersymmetric SU(4){sub c}xSU(2){sub L}xSU(2){sub R} model with discrete left-right symmetry. Nonuniversal gaugino masses, compatible with the gauge symmetry, play an essential role in realizing gluino and bino-wino coannihilation regions that are consistent with the WMAP dark matter constraints. We also explore regions of the parameter space in which the little hierarchy problem is partially resolved. We present several phenomenologically interesting benchmark points and the associated sparticle and Higgs mass spectra.
Conservative High-Order Time Integration for Lagrangian Hydrodynamics
In tis study, we develop novel time integration methods for the compressible Euler equations in the Lagrangian frame that are of arbitrary high order and exactly preserve the mass, momentum, and total energy of the system. The equations are considered in nonconservative form, that is, common for staggered grid hydrodynamics (SGH) methods; namely, the evolved quantities are mass, momentum, and internal energy. A general family of time integration schemes is formulated, and practical pairs for orders three and four are derived. Numerical results on standard hydrodynamics benchmarks confirm the high-order convergence on smooth problems and the exact numerical preservation of all physically conserved quantities.
Enhancing Gaussian Process Surrogates for Optimization and Posterior Approximation via Random Exploration
This paper proposes novel noise-free Bayesian optimization strategies that rely on a random exploration step to enhance the accuracy of Gaussian process surrogate models. The new algorithms retain the ease of implementation of the classical GP-UCB algorithm, but the additional random exploration step accelerates their convergence, nearly achieving the optimal convergence rate. Furthermore, to facilitate Bayesian inference with intractable likelihoods, we propose to utilize optimization iterates for maximum a posteriori estimation to build a Gaussian process surrogate model for the unnormalized log-posterior density. We provide bounds for the Hellinger distance between the true and the approximate posterior distributions in terms of the number of design points. We demonstrate the effectiveness of our Bayesian optimization algorithms in nonconvex benchmark objective functions, in a machine learning hyperparameter tuning problem, and in a black-box engineering design problem. The effectiveness of our posterior approximation approach is demonstrated in two Bayesian inference problems for parameters of dynamical systems.
CLOMP_TM
The CLOMP_TM benchmark was developed to measure the overheads of various mechanisms used to guarantee the generation of correct results in threaded code even when multiple threads might be accessing and/or updating the same memory location at the same time. This includes widely supported mechanisms like OpenMP Atomic and Open MP Critical as well as new mechanisms like Transactional Memory (TM) and Speculative Execution (SE) of threads. The CLOMP_TM benchmark is highly configurable to allow a variety of problem sizes, atomic update region sizes, and potential race condition opportunities to be studied. A wide variety of checks are done on the results to verify the expected answer was actually generated. Thus CLOMP_TM can be used both to measure the overhead for various mechanisms and to check the correctness of their implementation.
Fracture Capabilities in Grizzly with the extended Finite Element Method (X-FEM)
Efforts are underway to develop fracture mechanics capabilities in the Grizzly code to enable it to be used to perform deterministic fracture assessments of degraded reactor pressure vessels (RPVs). A capability was previously developed to calculate three-dimensional interaction- integrals to extract mixed-mode stress-intensity factors. This capability requires the use of a finite element mesh that conforms to the crack geometry. The eXtended Finite Element Method (X-FEM) provides a means to represent a crack geometry without explicitly fitting the finite element mesh to it. This is effected by enhancing the element kinematics to represent jump discontinuities at arbitrary locations inside of the element, as well as the incorporation of asymptotic near-tip fields to better capture crack singularities. In this work, use of only the discontinuous enrichment functions was examined to see how accurate stress intensity factors could still be calculated. This report documents the following work to enhance Grizzly’s engineering fracture capabilities by introducing arbitrary jump discontinuities for prescribed crack geometries; X-FEM Mesh Cutting in 3D: to enhance the kinematics of elements that are intersected by arbitrary crack geometries, a mesh cutting algorithm was implemented in Grizzly. The algorithm introduces new virtual nodes and creates partial elements, and then creates a new mesh connectivity; Interaction Integral Modifications: the existing code for evaluating the interaction integral in Grizzly was based on the assumption of a mesh that was fitted to the crack geometry. Modifications were made to allow for the possibility of a crack front that passes arbitrarily through the mesh; and Benchmarking for 3D Fracture: the new capabilities were benchmarked against mixed-mode three-dimensional fracture problems with known analytical solutions.