Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Heterogeneous computing”

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 1,549 records · Page 86

The role of iron-oxide aerosols and sunlight in the atmospheric reduction of Hg(II) species: A DFT+ U study

Experimental and field measurements have shown that, in the presence of both iron-containing aerosols and sunlight, oxidized mercury species such as HgCl 2 and HgBr 2 undergo reduction to elemental mercury (Hg°), which remains in the atmosphere longer than oxidized mercury species due to its higher volatility. We performed density functional theory (DFT, PW91+U) calculations to elucidate the reduction mechanism for atmospheric HgCl 2 and HgBr 2 to Hg° on several iron-oxide aerosol surfaces relevant in the troposphere. On the OH-Fe-R-terminated α-Fe 2 O 3 (0001) surface, predicted to be most prevalent under ambient conditions, we show that: (1) the first Hg-X bond is broken via either thermal or photolytic activation depending on the ambient temperature; (2) photons with an energy of 2.69 eV (461 nm) are required to break the second Hg-X bond; and (3) a photo-induced surface-to-adsorbate charge-transfer process can promote Hg° desorption with an excitation energy of 2.59 eV (479 nm). All the calculated excitation energies are below the threshold value of 3.9 eV (320 nm) for photons in the troposphere, suggesting that sunlight can facilitate mercury reduction on iron-oxide aerosol surfaces. In contrast, the gas-phase reduction of HgCl 2 (HgBr 2 ) involves photoexcitation requiring an energy of 4.98 (4.45) eV (249 (279) nm); therefore, the energy range of sunlight is not suitable for gas-phase reduction. Our computational results provide the first evidence on the detailed mechanism for the combined role of aerosols and photons in the reduction of HgCl 2 and HgBr 2 . In conclusion, our methodology can be adapted to study other photochemical heterogeneous processes in the atmosphere.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

An evaluation of the efficacy of various coal combustion models for predicting char burnout

Coal combustion is comprised of several subprocesses including devolatilization and heterogeneous reactions of the coal char with O 2 , CO 2 , H 2 O and potentially several other species. Much effort has been put forth to develop models for these processes which vary widely in both complexity and computational cost. This work investigates the efficacy of models for devolatilization and char reactions at either end of the complexity and cost spectrums for a range of particle sizes and furnace temperatures and across coal types. The overlap of simulated devolatilization and char consumption is also examined. In the gas phase, a detailed kinetics model based on a reduced version of the GRI 3.0 mechanism is used. The Char Conversion Kinetics and an n th -order Langmuir-Hinshelwood models are considered for char oxidation. The Chemical Percolation and Devolatilization and a two-step model are considered for devolatilization. Results indicate that high-fidelity models perform better at representing particle temperature and mass data across a wide range of O 2 concentrations as well as coal types. A significant overlap in devolatilization and char consumption is observed for both char chemistry and devolatilization models.

01 COAL, LIGNITE, AND PEAT↗

A Perspective on Multiscale Modeling of Explicit Solvation-Enabled Simulations of Catalysis at Liquid–Solid Interfaces

Catalysis at liquid-solid interfaces is profoundly influenced by the interfacial solvent structure, which affects catalytic activity, selectivity, and reaction pathways. This perspective discusses state-of-the-art multiscale modeling methods that integrate quantum mechanics and molecular mechanics approaches to apply explicit solvent molecules to capture these interfacial phenomena. Specifically, the construction of multiscale models, the importance of capturing the interfacial solvent structure, and the computational strategies used to achieve this are explored, and the challenges in balancing chemical accuracy with computational expense are highlighted. Additionally, this perspective addresses the limitations of current methods. Opportunities for integrating machine learning are proposed. Here, by advancing the efficiency and user friendliness of multiscale modeling, it is argued that deeper insights into heterogeneous catalysis in liquid phases can be provided, which will ultimately contribute to the development of more efficient catalytic processes.

Ab initio molecular dynamics↗

Heterogenized Iridium Water-Oxidation Catalyst from a Silatrane Precursor

Here, a pentamethylcyclopentadienyl (Cp*) iridium water-oxidation precatalyst was modified to include a silatrane functional group for covalent attachment to metal oxide semiconductor surfaces. The heterogenized catalyst was discovered to perform electrochemically driven water oxidation at an overpotential of 462 mV with a turnover number of 304 and turnover frequency of 0.035 s –1 in a 0.1 M KNO3 electrolyte at pH 5.8. Computational modeling of the experimental IR spectra suggests that the catalyst retains its Cp* group during the first hour of catalysis and likely remains monomeric.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Feasibility study of endoscopic x-ray luminescence computed tomography: Simulation demonstration and phantom application

Even though endoscopic fluorescence diffuse optical tomography (eFDOT) exhibits significant potential, currently its application is limited due to the existence of strong autofluorescence and the imaging inaccuracy caused by a very short source-detector distance. Motivated by the emerging X-ray luminescence computed tomography (XLCT) technology, we presented an endoscopic XLCT (eXLCT) methodology. In the methodology, the aperture angle of the objective was incorporated into the forward model, providing a more accurate description of light propagation. Numerical simulation with a heterogeneous geometry and an imaging experiment with a physical phantom were illustrated to demonstrate the feasibility of the presented eXLCT methodology.

36 MATERIALS SCIENCE↗

A Block-Based Triangle Counting Algorithm on Heterogeneous Environments

Triangle counting is a fundamental building block in graph algorithms. In this article, we propose a block-based triangle counting algorithm to reduce data movement during both sequential and parallel execution. Our block-based formulation makes the algorithm naturally suitable for heterogeneous architectures. The problem of partitioning the adjacency matrix of a graph is well-studied. Our task decomposition goes one step further: it partitions the set of triangles in the graph. By streaming these small tasks to compute resources, we can solve problems that do not fit on a device. We demonstrate the effectiveness of our approach by providing an implementation on a compute node with multiple sockets, cores and GPUs. The current state-of-the-art in triangle enumeration processes the Friendster graph in 2.1 seconds, not including data copy time between CPU and GPU. Using that metric, our approach is 20 percent faster. When copy times are included, our algorithm takes 3.2 seconds. This is 5.6 times faster than the fastest published CPU-only time.

97 MATHEMATICS AND COMPUTING↗

SU-E-P-05: Is Routine Treatment Planning System Quality Assurance Necessary?

Purpose: To evaluate the variation of dose calculations using a treatment planning system (TPS) over a two year period and assessment of the need for TPS QA on regular intervals. Methods: Two phantoms containing solid water and lung- and bone-equivalent heterogeneities were constructed in two different institutions for the same brand treatment planning system. Multiple plans, consisting of photons and electron beams, including IMRT and VMAT ones, were created and calculated on the phantoms. The accuracy of dose computation in the phantoms was evaluated at the onset by dose measurements within the phantoms. The dose values at up to 24 points of interest (POI) within the solid water, lung, and bone slabs, as well as mean doses to several regions of interest (ROI), were re-calculated over a two-year period which included two software upgrades. The variations in POI and ROI dose values were analyzed and evaluated. Results: The computed doses vary slightly month-over-month. There are noticeable variations at the times of software upgrade, if the upgrade involves remodeling and/or re-commissioning of the beams. The variations are larger in certain points within the phantom, usually in the buildup region or near interfaces, and are almost non-existent for electron beams. Conclusion: Routine TPS QA is recommended by AAPM and other professional societies, and is often required by accreditation organizations. The frequency and type of QA, though, is subject to debate. The results presented here demonstrate that the frequency of these tests could be at longer intervals than monthly. However, it is essential to perform TPS QA at the time of commissioning and after each software upgrade.

60 APPLIED LIFE SCIENCES↗

A Block-Based Triangle Counting Algorithm on Heterogeneous Environments

Triangle counting is a fundamental building block in graph algorithms. In this paper, we propose a block-based triangle counting algorithm to reduce data movement during both sequential and parallel execution. Our block-based formulation makes the algorithm naturally suitable for heterogeneous architectures. The problem of partitioning the adjacency matrix of a graph is well-studied. Our task decomposition goes one step further: it partitions the set of triangles in the graph. By streaming these small tasks to compute resources, we can solve problems that do not fit on a device. We demonstrate the effectiveness of our approach by providing an implementation on a compute node with multiple sockets, cores and GPUs. The current state-of-the-art in triangle enumeration processes the Friendster graph in 2.1 seconds, not including data copy time between CPU and GPU. Using that metric, our approach is 20 percent faster. When copy times are included, our algorithm takes 3.2 seconds. This is 5.6 times faster than the fastest published CPU-only time.

97 MATHEMATICS AND COMPUTING↗

SmartApps: Middle-ware for Adaptive Applications on Reconfigurable Platforms

There are several reasons why the performance of current distributed and heterogeneous systems is often disappointing. For example, the characteristics of the application may be input sensitive and evolve during execution causing dramatic changes in memory reference patterns, resource requirements, or degree of concurrency between different phases of the computation. Or, the system may change dynamically with nodes failing or appearing, some network links severed and other links established with different latencies and bandwidths. Another important reason for poor performance is the fairly compartmentalized approach to optimization: applications, compilers, operating systems and hardware configurations are designed and optimized in isolation and without the knowledge of instance specific information and needs of a running application. There is too little information flow across these boundaries and no global optimization is even attempted. For example, most operating systems services like paging, virtual-to-physical page mapping, I/O, or data layout in disks, provide little or no application customization. Similarly, the off-the-shelf hardware used by most commercial systems is optimized to give best average-case performance. To address this problem, we have proposed application-centric computing, or Smart Applications (SAS). In the SAS executable, the compiler embeds most run-time system services, and a performance-optimizing feedback loop that monitors the application's performance and adaptively reconfigures the application and the OS/system platform. At run-time, after incorporating the code's input and determining the system's resources and state, the SAS performs an *instance* specific optimization, which is more tractable than a global generic optimization between application, OS and system. The overriding philosophy of SAS is ``measure, compare, and adapt if beneficial.'' That is, the application will continually monitor its performance and the available resources to determine if, and by how much, performance could be improved if the application was restructured. Then, if the potential performance benefit outweighs the projected overhead costs, the application will restructure itself and the underlying system accordingly. The SAS framework includes performance monitoring and modeling components and mechanisms for performing the actual restructuring at various levels including: (i) algorithmic adaptation, (ii) run-time software optimization (e.g., input sensitivity analysis, etc.), (iii) tuning reconfigurable OS services (scheduling policy, page size, etc), and (iv) system configuration (e.g., selecting which, and how many, computational resources to use). SmartApps is being developed in the STAPL infrastructure. STAPL (the Standard Template Adaptive Parallel Library) is a framework for developing highly-optimizable, adaptable, and portable parallel and distributed applications. It consists of a relatively new and still evolving collection of generic parallel algorithms and distributed containers and a run-time system (RTS) through which the application and compiler interact with the OS and hardware.

97 MATHEMATICS AND COMPUTING↗

Towards Lightweight Data Integration Using Multi-Workflow Provenance and Data Observability

Modern large-scale scientific discovery requires multidisciplinary collaboration across diverse computing facilities, including High Performance Computing (HPC) machines and the Edge-to-Cloud continuum. Integrated data analysis plays a crucial role in scientific discovery, especially in the current AI era, by enabling Responsible AI development, FAIR, Reproducibility, and User Steering. However, the heterogeneous nature of science poses challenges such as dealing with multiple supporting tools, cross-facility environments, and efficient HPC execution. Building on data observability, adapter system design, and provenance, we propose MIDA: an approach for lightweight runtime Multi-workflow Integrated Data Analysis. MIDA defines data observability strategies and adaptability methods for various parallel systems and machine learning tools. With observability, it intercepts the dataflows in the background without requiring instrumentation while integrating domain, provenance, and telemetry data at runtime into a unified database ready for user steering queries. We conduct experiments showing end-to-end multi-workflow analysis integrating data from Dask and MLFlow in a real distributed deep learning use case for materials science that runs on multiple environments with up to 276 GPUs in parallel. We show near-zero overhead running up to 100,000 tasks on 1,680 CPU cores on the Summit supercomputer.

Santos Souza, Renan↗

Verification suite for the application of the limiting surface density method to arrays of 9975 packages

Criticality Safety Engineering at the Savannah River Site's K-Area Complex seeks to reduce the computational burden of analysis for the K-Area Complex. By applying an adapted Limiting Surface Density hand calculation method, the need for lengthy detailed Monte Carlo models of the storage arrays may be reduced or eliminated. After promising initial results using models of idealized but heterogeneous packages, application of the method to large arrays of highly heterogeneous 9975 shipping packages is attempted. A SCALE KENO-VI model of the 9975 package, with true dimensions and radial and axial heterogeneity is developed. A suite of array arrangements representing both arrays with equal number of packages per side (package-cubic) and arrays with unequal number of packages per side (non-cubic) are evaluated in SCALE. The cubic arrangements are used to derive the 9975-specific material and geometry constants for the LSD method. These constants are then used to implement the LSD method and compare the LSD- predicted critical mass per package to SCALE-predicted critical mass per package. Sensitivities to certain material and geometric nuances are evaluated for inclusion in the verification suite. Finally, the ability of the method to predict neutron multiplication changes induced by fissile mass changes is examined as a segue for evaluating credible abnormal conditions with the LSD method. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Spall damage of a Ta particle-reinforced metallic glass matrix composite under high strain rate loading

We investigate deformation and damage of a Zr-based bulk metallic glass (BMG) and its Ta particle-reinforced composite (MGMC) under impact loading, as well as quasi-static tension for comparison. Yield strength, spall strength, and damage accumulation rate are obtained from free-surface velocity histories, and MGMC appears to be more damage-resistant. Scanning electron microscopy, electron back scattering diffraction and x-ray computed tomography, are utilized for characterizing microstructures, which show features consistent with macroscopic measurements. Different damage and fracture modes are observed for BMG and MGMC. Multiple well-defined spall planes are observed in BMG, while isolated and scattered cracking around reinforced particles dominates fracture of MGMC. Particle–matrix interface serves as the source and barrier to crack nucleation and propagation under both quasi-static and impact loading. Finally, deformation twinning and grain refinement play a key role in plastic deformation during shock loading but not in quasi-static loading. In addition, 3D cup-cone structures are resolved in BMG, but not in MGMC due to its heterogeneous stress field.

36 MATERIALS SCIENCE↗

Clinical Significance of Accounting for Tissue Heterogeneity in Permanent Breast Seed Implant Brachytherapy Planning

Purpose: The inhomogeneity correction factor (ICF) method provides heterogeneity correction for the fast calculation TG43 formalism in seed brachytherapy. This study compared ICF-corrected plans to their standard TG43 counterparts, looking at their capacity to assess inadequate coverage and/or risk of any skin toxicities for patients who received permanent breast seed implant (PBSI). Methods and Materials: Two-month postimplant computed tomography scans and plans of 140 PBSI patients were used to calculate dose distributions by using the TG43 and the ICF methods. Multiple dose-volume histogram (DVH) parameters of clinical target volume (CTV) and skin were extracted and compared for both ICF and TG43 dose distributions. Short-term (desquamation and erythema) and long-term (telangiectasia) skin toxicity data were available on 125 and 110 of the patients, respectively, at the time of the study. The predictive value of each DVH parameter of skin was evaluated using the area under the receiver operating characteristic (ROC) curve for each toxicity endpoint. Results: Dose-volume histogram parameters of CTV, calculated using the ICF method, showed an overall decrease compared to TG43, whereas those of skin showed an increase, confirming previously reported findings of the impact of heterogeneity with low-energy sources. The ICF methodology enabled us to distinguish patients for whom the CTV V{sub 100} and V{sub 90} are up to 19% lower compared to TG43, which could present a risk of recurrence not detected when heterogeneity are not accounted for. The ICF method also led to an increase in the prediction of desquamation, erythema, and telangiectasia for 91% of skin DVH parameters studied. Conclusions: The ICF methodology has the advantage of distinguishing any inadequate dose coverage of CTV due to breast heterogeneity, which can be missed by TG43. Use of ICF correction also led to an increase in prediction accuracy of skin toxicities in most cases.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Effect of Hydrophilicity-Imparting Substituents on Exciton Delocalization in Squaraine Dye Aggregates Covalently Templated to DNA Holliday Junctions

Molecular aggregates exhibit emergent properties, including the collective sharing of electronic excitation energy known as exciton delocalization, that can be leveraged in applications such as quantum computing, optical information processing, and light harvesting. In a previous study, we found unexpectedly large excitonic interactions (quantified by the excitonic hopping parameter J m,n ) in DNA-templated aggregates of squaraine (SQ) dyes with hydrophilic-imparting sulfo and butylsulfo substituents. Here, we characterize DNA Holliday junction (DNA-HJ) templated aggregates of an expanded set of SQs and evaluate their optical properties in the context of structural heterogeneity. Specifically, we characterized the orientation of and J m,n between dyes in dimer aggregates of non-chlorinated and chlorinated SQs. Here, three new chlorinated SQs that feature a varying number of butylsulfo substituents were synthesized and attached to a DNA-HJ via a covalent linker to form adjacent and transverse dimers. Various characteristics of the dye, including its hydrophilicity (in terms of log P o/w ) and surface area, and of the substituents, including their local bulkiness and electron withdrawing capacity, were quantified computationally.

59 BASIC BIOLOGICAL SCIENCES↗

Statistic inversion of multi-zone transition probability models for aquifer characterization in alluvial fans

Understanding the heterogeneity arising from the complex architecture of sedimentary sequences in alluvial fans is challenging. This study develops a statistical inverse framework in a multi-zone transition probability approach for characterizing the heterogeneity in alluvial fans. An analytical solution of the transition probability matrix is used to define the statistical relationships among different hydrofacies and their mean lengths, integral scales, and volumetric proportions. A statistical inversion is conducted to identify the multi-zone transition probability models and estimate the optimal statistical parameters using the modified Gauss–Newton–Levenberg–Marquardt method. The Jacobian matrix is computed by the sensitivity equation method, which results in an accurate inverse solution with quantification of parameter uncertainty. We use the Chaobai River alluvial fan in the Beijing Plain, China, as an example for elucidating the methodology of alluvial fan characterization. The alluvial fan is divided into three sediment zones. In each zone, the explicit mathematical formulations of the transition probability models are constructed with optimized different integral scales and volumetric proportions. The hydrofacies distributions in the three zones are simulated sequentially by the multi-zone transition probability-based indicator simulations. Finally, the result of this study provides the heterogeneous structure of the alluvial fan for further study of flow and transport simulations.

54 ENVIRONMENTAL SCIENCES↗

Modeling flow and transport in fracture networks using graphs

Fractures form the main pathways for flow in the subsurface within low-permeability rock. For this reason, accurately predicting flow and transport in fractured systems is vital for improving the performance of subsurface applications. Fracture sizes in these systems can range from millimeters to kilometers. Although modeling flow and transport using the discrete fracture network (DFN) approach is known to be more accurate due to incorporation of the detailed fracture network structure over continuum-based methods, capturing the flow and transport in such a wide range of scales is still computationally intractable. Furthermore, if one has to quantify uncertainty, hundreds of realizations of these DFN models have to be run. To reduce the computational burden, we solve flow and transport on a graph representation of a DFN. We study the accuracy of the graph approach by comparing breakthrough times and tracer particle statistical data between the graph-based and the high-fidelity DFN approaches, for fracture networks with varying number of fractures and degree of heterogeneity. Due to our recent developments in capabilities to perform DFN high-fidelity simulations on fracture networks with large number of fractures, we are in a unique position to perform such a comparison. We show that the graph approach shows a consistent bias with up to an order of magnitude slower breakthrough when compared to the DFN approach. We show that this is due to graph algorithm's underprediction of the pressure gradients across intersections on a given fracture, leading to slower tracer particle speeds between intersections and longer travel times. We present a bias correction methodology to the graph algorithm that reduces the discrepancy between the DFN and graph predictions. We show that with this bias correction, the graph algorithm predictions significantly improve and the results are very accurate. In conclusion, the good accuracy and the low computational cost, with O(10 4 ) times lower times than the DFN, makes the graph algorithm an ideal technique to incorporate in uncertainty quantification methods.

58 GEOSCIENCES↗

Concurrent multiscale modeling of microstructural effects on localization behavior in finite deformation solid mechanics

The heterogeneity in mechanical fields introduced by microstructure plays a critical role in the localization of deformation. In order to resolve this incipient stage of failure, it is therefore necessary to incorporate microstructure with sufficient resolution. On the other hand, computational limitations make it infeasible to represent the microstructure in the entire domain at the component scale. Here, the authors demonstrate the use of concurrent multiscale modeling to incorporate explicit, finely resolved microstructure in a critical region while resolving the smoother mechanical fields outside this region with a coarser discretization to limit computational cost. The microstructural physics is modeled with a high-fidelity model that incorporates anisotropic crystal elasticity and rate-dependent crystal plasticity to simulate the behavior of a stainless steel alloy. The component-scale material behavior is treated with a lower fidelity model incorporating isotropic linear elasticity and rate-independent J 2 plasticity. The microstructural and component scale subdomains are modeled concurrently, with coupling via the Schwarz alternating method, which solves boundary-value problems in each subdomain separately and transfers solution information between subdomains via Dirichlet boundary conditions. In this study, the framework is applied to model incipient localization in tensile specimens during necking.

36 MATERIALS SCIENCE↗

Theory and Applications of Macroscale Models in Porous Media

Systems dominated by heterogeneity over a multiplicity of scales, like porous media, still challenge our modeling efforts. The presence of disparate length- and time-scales that control dynamical processes in porous media hinders not only models predictive capabilities, but also their computational efficiency. Macrosopic models, i.e., averaged representations of pore-scale processes, are computationally efficient alternatives to microscale models in the study of transport phenomena in porous media at the system, field or device scale (i.e., at a scale much larger than a characteristic pore size). We present an overview of common upscaling methods used to formally derive macroscale equations from pore-scale (mass, momentum and energy) conservation laws. This review includes the volume averaging method, mixture theory, thermodynamically constrained averaging, homogenization, and renormalization group techniques. We apply these methods to a number of specific problems ranging from food processing to human bronchial system, and from diffusion to multiphase flow, to demonstrate the methods generality and flexibility in handling different applications. The primary intent of such an overview is not to provide a thorough review of all currently available upscaling techniques, nor a complete mathematical treatment of the ones presented, but rather a primer on some of the tools available for upscaling, the basic principles they are based upon, and their specific advantages and drawbacks, so to guide the reader in the choice of the most appropriate method for particular applications and of the most relevant technical literature.

42 ENGINEERING↗