Search NASA⌕ Search

SEARCH · Search NASA

Results for “Distributed 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.

At least 919 records · Page 51

Halogenated Dibenzo[ f , h ]quinoxaline Units Constructed 2D‐Conjugated Guest Acceptors for 19% Efficiency Organic Solar Cells

Abstract Halogenation of Y‐series small‐molecule acceptors (Y‐SMAs) is identified as an effective strategy to optimize photoelectric properties for achieving improved power‐conversion‐efficiencies (PCEs) in binary organic solar cells (OSCs). However, the effect of different halogenation in the 2D‐structured large π‐fused core of guest Y‐SMAs on ternary OSCs has not yet been systematically studied. Herein, four 2D‐conjugated Y‐SMAs (X‐QTP‐4F, including halogen‐free H‐QTP‐4F, chlorinated Cl‐QTP‐4F, brominated Br‐QTP‐4F, and iodinated I‐QTP‐4F) by attaching different halogens into 2D‐conjugation extended dibenzo[ f , h ]quinoxaline core are developed. Among these X‐QTP‐4F, Cl‐QTP‐4F has a higher absorption coefficient, optimized molecular crystallinity and packing, suitable cascade energy levels, and complementary absorption with PM6:L8‐BO host. Moreover, among ternary PM6:L8‐BO:X‐QTP‐4F blends, PM6:L8‐BO:Cl‐QTP‐4F obtains a more uniform and size‐suitable fibrillary network morphology, improved molecular crystallinity and packing, as well as optimized vertical phase distribution, thus boosting charge generation, transport, extraction, and suppressing energy loss of OSCs. Consequently, the PM6:L8‐BO:Cl‐QTP‐4F‐based OSCs achieve a 19.0% efficiency, which is among the state‐of‐the‐art OSCs based on 2D‐conjugated Y‐SMAs and superior to these devices based on PM6:L8‐BO host (17.70%) and with guests of H‐QTP‐4F (18.23%), Br‐QTP‐4F (18.39%), and I‐QTP‐4F (17.62%). The work indicates that halogenation in 2D‐structured dibenzo[ f , h ]quinoxaline core of Y‐SMAs guests is a promising strategy to gain efficient ternary OSCs.

14 SOLAR ENERGY↗

Hydrogen Localization and Cluster Formation in α-Zr from First-Principles Investigations

We report alpha-zirconium (α-Zr) is a hydrogen atom host material used in nuclear energy and semiconducting applications due to Zr's excellent properties such as corrosion resistance, low thermal neutrons cross-section, and thermodynamically favorable interstitial sequestering of H atoms. Herein we use density functional theory simulations to systematically analyze the initial stages of H insertion in the Zr matrix, and determine optimal hydrogen localization, distribution, and clustering tendencies from a thermodynamics and kinetics perspective. Our results suggest that H occupation of tetrahedral sites is strongly favored, consistent with previous experimental and theoretical studies. Further, a cooperative effect in H localization is shown where H atoms favor clustering. The H pair is found to be the most stable by occupying tetrahedral sites with pairs at second nearest-neighbor positions (TTn2). Among the four H atom clusters, the most favorable configuration has two TTn2 pairs forming "box " or "zig-zag" shaped-cluster with the pairs separated by 3.03 or 3.25 Å, respectively. Nudged elastic band simulations suggest that the formation of TTn2 is facile and is limited by the small diffusion barrier of the monomer of less than 0.1 eV. Further, the formation routes for the 4H atom clusters are also presented. The current study can serve as a foundation to systematically model H atom clustering in metals and alloys.

36 MATERIALS SCIENCE↗

Estimating the impact of tariff-driven behind-the-meter storage operation on distribution grid investments

Increasing growth of distributed solar photovoltaics (PV) and electric vehicles (EV) can strain local distribution networks and require costly upgrades. Distributed battery storage, often deployed alongside PV, can be used to mitigate those costs, depending on how batteries are operated. This study evaluates the potential deferral value of distributed battery storage across a range of tariff structures, focusing on the rate structures most commonly available to residential customers today and related variants. Deferrals are evaluated with a least-cost distribution grid expansion optimization model to identify requirements on line reconductoring, transformer upgrades, and voltage regulator installations under each tariff. Results show that TOU rates and net billing tariffs can yield meaningful deferral value, depending on specific tariff structure features. Under the best performing tariff structure tested, storage produced a median annualized deferral value of $7.18 per kW of storage capacity ( kW S ) across all feeders in the sample, though deferral values were considerably larger for feeders with peak loads that coincide with utility system peak, i.e., timing of TOU peak period. In contrast, under an unrestricted TOU design with no restrictions on grid charging or discharging, the median deferral value was $0/ kW S illustrating the critical importance of tariff structure details.

Rodriguez-Garcia, Luis↗

Efficient Heterogeneous Execution on Large Multicore and Accelerator Platforms: Case Study Using a Block Tridiagonal Solver

The algorithmic and implementation principles are explored in gainfully exploiting GPU accelerators in conjunction with multicore processors on high-end systems with large numbers of compute nodes, and evaluated in an implementation of a scalable block tridiagonal solver. The accelerator of each compute node is exploited in combination with multicore processors of that node in performing block-level linear algebra operations in the overall, distributed solver algorithm. Optimizations incorporated include: (1) an efficient memory mapping and synchronization interface to minimize data movement, (2) multi-process sharing of the accelerator within a node to obtain balanced load with multicore processors, and (3) an automatic memory management system to efficiently utilize accelerator memory when sub-matrices spill over the limits of device memory. Results are reported from our novel implementation that uses MAGMA and CUBLAS accelerator software systems simultaneously with ACML for multithreaded execution on processors. Overall, using 940 nVidia Tesla X2090 accelerators and 15,040 cores, the best heterogeneous execution delivers a 10.9-fold reduction in run time relative to an already efficient parallel multicore-only baseline implementation that is highly optimized with intra-node and inter-node concurrency and computation-communication overlap. Detailed quantitative results are presented to explain all critical runtime components contributing to hybrid performance.

Park, Alfred J↗

Fast neural network training on a cluster of GPUs for action recognition with high accuracy

In this work, we propose algorithms and techniques to accelerate training of deep neural networks for action recognition on a cluster of GPUs. The convergence analysis of our algorithm shows it is possible to reduce communication cost and at the same time minimize the number of iterations needed for convergence. We customize the Adam optimizer for our distributed algorithm to improve efficiency. In addition, we employ transfer-learning to further reduce training time while improving validation accuracy. For the UCF101 and HMDB51 datasets, the validation accuracies achieved are 93.1% and 67.9% respectively. With an additional end-to-end trained temporal stream, the validation accuracies achieved for UCF101 and HMDB51 are 93.47% and 81.24% respectively. As far as we know, these are the highest accuracies achieved with the two-stream approach using ResNet that does not involve computationally expensive 3D convolutions or pretraining on much larger datasets.

97 MATHEMATICS AND COMPUTING↗

The Baghdad Atlas: A relational database of inelastic neutron-scattering (n,n ' γ) data

A relational database has been developed based on the original (n,n'γ) work carried out by A. M. Demidov et al., at the Nuclear Research Institute in Baghdad, Iraq (Demidov et al., 1978) for 105 independent measurements comprising 76 elemental samples of natural composition and 29 isotopically-enriched samples. The information from this Atlas includes: γ-ray energies and relative intensities; nuclide and level data corresponding to the residual nucleus and meta data associated with the target sample that allows for the extraction of the flux-weighted (n,n'γ) cross sections for a given transition relative to a defined value. The optimized angular-distribution-corrected fast-neutron flux-weighted partial γ-ray cross section for the production of the 846.8-keV 21+→0gs+γ-ray transition in 56Fe, determined to be $\langle$σγ$\rangle$=143(29) mb, is used for this purpose. However, different values for the adopted cross section can be readily implemented to accommodate user preference based on revised determinations of this quantity. The Atlas (n,n'γ) data has been compiled into a series of CSV-style ASCII data sets and a suite of Python scripts have been developed to build and install the database locally. The database can then be accessed directly through the SQLite engine, or using alternative methods such as the Jupyter Notebook Python-browser interface. Several examples exploiting different interaction methodologies are distributed with the complete software package.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Bidirectional threshold switching in engineered multilayer (Cu{sub 2}O/Ag:Cu{sub 2}O/Cu{sub 2}O) stack for cross-point selector application

In this study, we achieved bidirectional threshold switching (TS) for selector applications in a Ag-Cu{sub 2}O-based programmable-metallization-cell device by engineering the stack wherein Ag was intentionally incorporated in the oxide (Cu{sub 2}O) layer by a simple approach comprising co-sputtering and subsequent optimized annealing. The distribution of the Ag was directly confirmed by transmission electron microscopy and energy dispersive spectroscopy line profiling. The observed TS occurred because of the spontaneous self-rupturing of the unstable Ag filament that formed in the oxide layer.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Ordering nanoparticles with polymer brushes

Ordering nanoparticles into a desired super-structure is often crucial for their technological applications. We use molecular dynamics simulations to study the assembly of nanoparticles in a polymer brush randomly grafted to a planar surface as the solvent evaporates. Initially, the nanoparticles are dispersed in a solvent that wets the polymer brush. After the solvent evaporates, the nanoparticles are either inside the brush or adsorbed at the surface of the brush, depending on the strength of the nanoparticle-polymer interaction. For strong nanoparticle-polymer interactions, a 2-dimensional ordered array is only formed when the brush density is finely tuned to accommodate a single layer of nanoparticles. When the brush density is higher or lower than this optimal value, the distribution of nanoparticles shows large fluctuations in space and the packing order diminishes. For weak nanoparticle-polymer interactions, the nanoparticles order into a hexagonal array on top of the polymer brush as long as the grafting density is high enough to yield a dense brush. As a result, an interesting healing effect is observed for a low-grafting-density polymer brush that can become more uniform in the presence of weakly adsorbed nanoparticles.

74 ATOMIC AND MOLECULAR PHYSICS↗

Initial position optimization in molecular dynamics simulations for a Coulomb system

A new algorithm for molecular dynamics (MD) simulations is developed to optimize plasma particle distributions at given initial temperatures. By combining velocity scaling and reassignment, the method effectively eliminates the initial rise and oscillation in temperatures observed with randomly distributed positions. These rises and oscillations are undesired numerical artifacts observed in conventional plasma MD simulations, arising from unoptimized particle positions. The algorithm demonstrates temperature relaxation without initial rises or oscillations, as well as precise flow velocity relaxation, enabling accurate measurement of relaxation times. The code is accelerated using graphics processing units for parallel processing, enhancing the study of plasma dynamics. The proposed method for distributing physically valid particles in MD simulations enables accurate studies of intrinsic collision processes in plasmas, including the dynamics of strongly coupled plasmas, plasma–wave interactions, and transport phenomena in magnetized plasmas. The paper concludes with a discussion of potential applications and future enhancements to the algorithm.

Jo, Jawon (ORCID:0009000924193285)↗

Efficient Extraction of High Centrality Vertices in Distributed Graphs

Betweenness centrality (BC) is an important measure for identifying high value or critical vertices in graphs, in variety of domains such as communication networks, road networks, and social graphs. However, calculating betweenness values is prohibitively expensive and, more often, domain experts are interested only in the vertices with the highest centrality values. In this paper, we first propose a partition-centric algorithm (MS-BC) to calculate BC for a large distributed graph that optimizes resource utilization and improves overall performance. Further, we extend the notion of approximate BC by pruning the graph and removing a subset of edges and vertices that contribute the least to the betweenness values of other vertices (MSL-BC), which further improves the runtime performance. We evaluate the proposed algorithms using a mix of real-world and synthetic graphs on an HPC cluster and analyze its strengths and weaknesses. The experimental results show an improvement in performance of upto 12x for large sparse graphs as compared to the state-of-the-art, and at the same time highlights the need for better partitioning methods to enable a balanced workload across partitions for unbalanced graphs such as small-world or power-law graphs.

97 MATHEMATICS AND COMPUTING↗

On the Need for Communication for Voltage Regulation of Power Distribution Grids

We consider the problem of regulating the voltage profile of a power distribution grid by controlling the reactive power injection of distributed microgenerators. We define a very general class of purely local feedback controllers in which reactive power injection is adjusted based on the local voltage measurements. This class includes most of the solutions proposed in the literature and in the latest grid codes. We show that these strategies do not guarantee the desired regulation, as each of them can have an equilibrium that is not feasible with respect to the desired voltage and power constraints. We then propose a networked feedback law to show that, by adding short-range communication between microgenerators, it is possible to design control strategies that provably converge to the feasible set. This fundamental performance gap between local and networked strategies is finally illustrated via simulations.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Thermal Management System for an Electric Machine With Additively Manufactured Hollow Conductors With Integrated Heat Pipes

Here, this paper discusses steps taken to develop a novel thermal management system for an aircraft propulsion electric machine containing additively manufactured coils integrated with heat pipes aimed at boosting its specific power. Experimental setups are used to size and characterize heat pipes for the application and 3D thermal finite element analysis is used to determine optimum heat transfer coefficient of the convective boundaries. For some of the convective boundaries, fin-based surface area enhancement is required to reach a target combined overall heat transfer coefficient and surface area performance (UA). This enhancement is worked out using a combination of the Engineering Equation Solver tool and 3D thermal FEA. The thermal management system's UA, and by extension its specific power, sensitivity to coolant temperature is explored. Temperature distribution plots of optimized machine components are also presented and discussed. Lastly, additional heat pipe testing is carried out to study its maximum heat transfer capability's sensitivity to condenser coolant temperature and configuration.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Deep Reinforcement Learning Enabled Physical-Model-Free Two-Timescale Voltage Control Method for Active Distribution Systems

Active distribution networks are being challenged by frequent and rapid voltage violations due to renewable energy integration. Conventional model-based voltage control methods rely on accurate parameters of the distribution networks, which are difficult to achieve in practice. This paper proposes a novel physical-model-free two-timescale voltage control framework for active distribution systems. To achieve fast control of PV inverters, the whole network is first partitioned into several subnetworks using voltage-reactive power sensitivity. Then, the scheduling of PV inverters in the multiple sub-networks is formulated as Markov games and solved by a multi-agent soft actor-critic (MASAC) algorithm, where each subnetwork is modeled as an intelligent agent. All agents are trained in a centralized manner to learn a coordinated strategy while being executed based on only local information for fast response. For the slower time-scale control, OLTCs and switched capacitors are coordinated by a single agent-based SAC algorithm using the global information with considering control behaviors of the inverters. Particularly, the two-level agents are trained concurrently with information exchange according to the reward signal calculated from the data-driven surrogate model. Comparative tests with different benchmark methods on IEEE 33-and 123-bus systems and 342-node low voltage distribution system demonstrate that the proposed method can effectively mitigate the fast voltage violations and achieve systematical coordination of different voltage regulation assets without the knowledge of accurate system model.

24 POWER TRANSMISSION AND DISTRIBUTION↗

SU-E-T-579: Impact of Cylinder Size in High-Dose Rate Brachytherapy (HDRBT) for Primary Cancer in the Vagina

Purpose: To evaluate the dosimetric impact of cylinder size in high dose rate Brachytherapy for primary vaginal cancers. Methods: Patients treated with HDR vaginal vault radiation in a list of cylinders ranging from 2.5 to 4 cm in diameter at 0.5 cm increment were analyzed. All patients’ doses were prescribed at the 0.5 cm from the vaginal surface with different treatment lengths. A series of reference points were created to optimize the dose distribution. The fraction dose was 5.5 Gy, the treatment was repeated for 4 times in two weeks. A cylinder volume was contoured in each case according to the prescribed treatment length, and then expanded to 5 mm to get a volume Cylinder-5mm-exp. A volume of PTV-Eval was obtained by subtracting the cylinder volume from the Cylinder-5mm-exp. The shell volume, PTV-Eval serves as the target volume for dosimetric evaluation. Results: DVH curves and average doses of PTV-Eval were obtained. Our results indicated that the DVH curves shifted toward higher dose side when larger cylinder was used instead of smaller ones. When 3.0 cm cylinder was used instead of 2.5 cm, for 3.0 cm treatment length, the average dose only increased 1%, from 790 to 799 cGy. However, the average doses for 3.5 and 4 cm cylinders respectively are 932 and 1137 cGy at the same treatment length. For 5.0 cm treatment length, the average dose is 741 cGy for 2.5 cm cylinder, and 859 cGy for 3 cm cylinder. Conclusion: Our data analysis suggests that for the vaginal intracavitary HDRBT, the average dose is at least 35% larger than the prescribed dose in the studied cases; the size of the cylinder will impact the dose delivered to the target volume. The cylinder with bigger diameter tends to deliver larger average dose to the PTV-Eval.

60 APPLIED LIFE SCIENCES↗

WE-D-17A-06: Optically Stimulated Luminescence Detectors as ‘LET-Meters’ in Proton Beams

Purpose: To demonstrate and evaluate the potential of optically stimulated luminescence (OSL) detectors (OSLDs) for measurements of linear energy transfer (LET) in therapeutic proton beams. Methods: Batches of Al2O2:C OSLDs were irradiated with an absorbed dose of 0.2 Gy in un-modulated proton beams of varying LET (0.67 keV/μm to 2.58 keV/μm). The OSLDs were read using continuous wave (CW-OSL) and pulsed (P-OSL) stimulation modes. We parameterized and calibrated three characteristics of the OSL signals as functions of LET: CW-OSL curve shape, P-OSL curve shape and the ratio of the two OSL emission band intensities (ultraviolet/blue ratio). Calibration curves were created for each of these characteristics to describe their behaviors as functions of LET. The true LET values were determined using a validated Monte Carlo model of the proton therapy nozzle used to irradiate the OSLDs. We then irradiated batches of OSLDs with an absorbed dose of 0.2 Gy at various depths in two modulated proton beams (140 MeV, 4 cm wide spread-out Bragg peak (SOBP) and 250 MeV, 10 cm wide SOBP). The LET values were calculated using the OSL response and the calibration curves. Finally, measured LET values were compared to the true values determined using Monte Carlo simulations. Results: The CW-OSL curve shape, P-OSL curve shape and the ultraviolet/blue-ratio provided proton LET estimates within 12.4%, 5.7% and 30.9% of the true values, respectively. Conclusion: We have demonstrated that LET can be measured within 5.7% using Al2O3:C OSLDs in the therapeutic proton beams used in this investigation. From a single OSLD readout, it is possible to measure both the absorbed dose and LET. This has potential future applications in proton therapy quality assurance, particularly for treatment plans based on optimization of LET distributions. This research was partially supported by the Natural Sciences and Engineering Research Council of Canada.

07 ISOTOPE AND RADIATION SOURCES↗

Avoiding Communication in Primal and Dual Block Coordinate Descent Methods

Primal and dual block coordinate descent methods are iterative methods for solving regularized and unregularized optimization problems. Distributed-memory parallel implementations of these methods have become popular in analyzing large machine learning datasets. However, existing implementations communicate at every iteration, which, on modern data center and supercomputing architectures, often dominates the cost of floating-point computation. Recent results on communication-avoiding Krylov subspace methods suggest that large speedups are possible by re-organizing iterative algorithms to avoid communication. We show how applying similar algorithmic transformations can lead to primal and dual block coordinate descent methods that only communicate every $s$ iterations---where $s$ is a tuning parameter---instead of every iteration for the regularized least-squares problem. We show that the communication-avoiding variants reduce the number of synchronizations by a factor of $s$ on distributed-memory parallel machines without altering the convergence rate and attain strong scaling speedups of up to $6.1\times$ over the “standard algorithm" on a Cray XC30 supercomputer.

97 MATHEMATICS AND COMPUTING↗