Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “distributed processing”

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 253 records · Page 14
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Overview and status of the Long-Baseline Neutrino Facility Far Site cryogenics system

The Sanford Underground Research Facility (SURF) will host the Far Detector of the Deep Underground Neutrino Experiment (DUNE), an international multi-kiloton Long-Baseline neutrino experiment that will be installed about one and a half kilometers underground in Lead, SD. Detectors will be located inside four cryostats filled with almost 70,000 metric tons of ultrapure liquid argon, with a level of impurities lower than 100 parts per trillion of oxygen equivalent contamination. The cryogenics infrastructure supporting this experiment is provided by the Long-Baseline Neutrino Facility (LBNF). This contribution presents modes of operation, layout, and main features of the LBNF Far Site cryogenic system, which is composed of three subsystems: Infrastructure, Proximity, and Internal cryogenics. The Infrastructure cryogenics supports the needs of the cryostat and Proximity cryogenics. It includes the equipment to receive the argon in liquid phase, vaporize it and transfer it underground as a gas, the nitrogen system (composed of the refrigeration system, liquid nitrogen buffer tanks and liquid and gaseous nitrogen distribution), liquid and gaseous argon distribution and process controls. The Proximity cryogenics receives fluids from the Infrastructure cryogenics and delivers them to the Internal cryogenics at the required temperature, pressure, purity and mass flow rate. It includes the argon condensers, liquid and gaseous argon purification and regeneration systems, nitrogen and argon phase separators, piping, valves, and instrumentation. The Internal cryogenics comprises the liquid and gaseous argon distribution inside the cryostats for the commissioning, cool down, fill, and steady state operations of the cryostats and detectors. An international engineering team is designing these systems and will manufacture, install, commission, and qualify them. This contribution describes the main features, performance, functional requirements, and modes of operation of the LBNF Far Site cryogenics system. It also presents the status of the design, along with present and future needs to support the DUNE experiment.

43 PARTICLE ACCELERATORS↗
DOE OSTI2015

Soft evolution of multi-jet final states

We present a new framework for computing resummed and matched distributions in processes with many hard QCD jets. The intricate color structure of soft gluon emission at large angles renders resummed calculations highly non-trivial in this case. We automate all ingredients necessary for the color evolution of the soft function at next-to-leading-logarithmic accuracy, namely the selection of the color bases and the projections of color operators and Born amplitudes onto those bases. Explicit results for all QCD processes with up to 2 → 5 partons are given. We also devise a new tree-level matching scheme for resummed calculations which exploits a quasi-local subtraction based on the Catani-Seymour dipole formalism. We implement both resummation and matching in the Sherpa event generator. As a proof of concept, we compute the resummed and matched transverse-thrust distribution for hadronic collisions.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2012

The r-PROCESS IN THE NEUTRINO-DRIVEN WIND FROM A BLACK-HOLE TORUS

We examine r-process nucleosynthesis in the neutrino-driven wind from the thick accretion disk (or 'torus') around a black hole. Such systems are expected as remnants of binary neutron star or neutron star-black hole mergers. We consider a simplified, analytic, time-dependent evolution model of a 3 M{sub Sun} central black hole surrounded by a neutrino emitting accretion torus with 90 km radius, which serves as basis for computing spherically symmetric neutrino-driven wind solutions. We find that ejecta with modest entropies ({approx}30 per nucleon in units of the Boltzmann constant) and moderate expansion timescales ({approx}100 ms) dominate in the mass outflow. The mass-integrated nucleosynthetic abundances are in good agreement with the solar system r-process abundance distribution if a minimal value of the electron fraction at the charged-particle freezeout, Y{sub e,min} {approx} 0.2, is achieved. In the case of Y{sub e,min} {approx} 0.3, the production of r-elements beyond A {approx} 130 does not reach to the third peak but could still be important for an explanation of the abundance signatures in r-process deficient stars in the early Galaxy. The total mass of the ejected r-process nuclei is estimated to be {approx}1 Multiplication-Sign 10{sup -3} M{sub Sun }. If our model was representative, this demands a Galactic event rate of {approx}2 Multiplication-Sign 10{sup -4} yr{sup -1} for black-hole-torus winds from merger remnants to be the dominant source of the r-process elements. Our result thus suggests that black-hole-torus winds from compact binary mergers have the potential to be a major, but probably not the dominant, production site of r-process elements.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2019

Optimization of process models for determining volatility distribution and viscosity of organic aerosols from isothermal particle evaporation data

The composition of organic aerosol under different ambient conditions as well as their phase state have been a subject of intense study in recent years. One way to study particle properties is to measure the particle size shrinkage in a diluted environment at isothermal conditions. From these measurements it is possible to separate the fraction of low-volatility compounds from high-volatility compounds. In this work, we analyse and evaluate a method for obtaining particle composition and viscosity from measurements using process models coupled with input optimization algorithms. Two optimization methods, the Monte Carlo genetic algorithm and Bayesian inference, are used together with process models describing the dynamics of particle evaporation. The process model optimization scheme in inferring particle composition in a volatility-basis-set sense and composition-dependent particle viscosity is tested with artificially generated data sets and real experimental data. Optimizing model input so that the output matches these data yields a good match for the estimated quantities. Both optimization methods give equally good results when they are used to estimate particle composition to artificially test data. Finally, the timescale of the experiments and the initial particle size are found to be important in defining the range of values that can be identified for the properties from the optimization.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

VDiSC: An Open Source Framework for Distributed Smart City Vision and Biometric Surveillance Networks

Recent global growth in the interest of smart cities has led to trillions of dollars of investment toward research and development. These connected cities have the potential to create a symbiosis of technology and society and revolutionize the cost of living, safety, ecological sustainability, and quality of life of societies on a world-wide scale. Some key components of the smart city construct are connected smart grids, self-driving cars, federated learning systems, smart utilities, large-scale public transit, and proactive surveillance systems. While exciting in prospect, these technologies and their subsequent integration cannot be attempted without addressing the potential societal impacts of such a high degree of automation and data sharing. Additionally, the feasibility of coordinating so many disparate tasks will require a fast, extensible, unifying framework. To that end, we propose the Distributed Smart City framework for Vision, or VDiSC. VDiSC serves as a unified biometric API harness that allows for seamless evaluation, deployment, and simple pipeline creation for heterogeneous biometric software. VDiSC additionally provides a fully declarative capability for defining and coordinating custom machine learning and sensor pipelines, allowing the distribution of processes across otherwise incompatible hardware and networks. VDiSC ultimately provides a way to quickly configure, hot-swap, and expand large coordinated or federated systems online without interruptions for maintenance. Because much of the data collected in a smart city contains Personally Identifying Information (PII), VDiSC also provides built-in tools and layers to ensure secure and encrypted streaming, storage, and access of PII data across distributed systems.

Brogan, Joel↗
DOE OSTI2017

Trace: a high-throughput tomographic reconstruction engine for large-scale datasets

Abstract Background Modern synchrotron light sources and detectors produce data at such scale and complexity that large-scale computation is required to unleash their full power. One of the widely used imaging techniques that generates data at tens of gigabytes per second is computed tomography (CT). Although CT experiments result in rapid data generation, the analysis and reconstruction of the collected data may require hours or even days of computation time with a medium-sized workstation, which hinders the scientific progress that relies on the results of analysis. Methods We present Trace, a data-intensive computing engine that we have developed to enable high-performance implementation of iterative tomographic reconstruction algorithms for parallel computers. Trace provides fine-grained reconstruction of tomography datasets using both (thread-level) shared memory and (process-level) distributed memory parallelization. Trace utilizes a special data structure called replicated reconstruction object to maximize application performance. We also present the optimizations that we apply to the replicated reconstruction objects and evaluate them using tomography datasets collected at the Advanced Photon Source. Results Our experimental evaluations show that our optimizations and parallelization techniques can provide 158× speedup using 32 compute nodes (384 cores) over a single-core configuration and decrease the end-to-end processing time of a large sinogram (with 4501 × 1 × 22,400 dimensions) from 12.5 h to <5 min per iteration. Conclusion The proposed tomographic reconstruction engine can efficiently process large-scale tomographic data using many compute nodes and minimize reconstruction times.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2011

The power of event-driven analytics in Large Scale Data Processing

FeedZai is a software company specialized in creating high-­‐throughput low-­‐latency data processing solutions. FeedZai develops a product called "FeedZai Pulse" for continuous event-­‐driven analytics that makes application development easier for end users. It automatically calculates key performance indicators and baselines, showing how current performance differ from previous history, creating timely business intelligence updated to the second. The tool does predictive analytics and trend analysis, displaying data on real-­‐time web-­‐based graphics. In 2010 FeedZai won the European EBN Smart Entrepreneurship Competition, in the Digital Models category, being considered one of the "top-­‐20 smart companies in Europe". The main objective of this seminar/workshop is to explore the topic for large-­‐scale data processing using Complex Event Processing and, in particular, the possible uses of Pulse in the scope of the data processing needs of CERN. Pulse is available as open-­‐source and can be licensed both for non-­‐commercial and commercial applications. FeedZai is interested in exploring possible synergies with CERN in high-­‐volume low-­‐latency data processing applications. The seminar will be structured in two sessions, the first one being aimed to expose the general scope of FeedZai's activities, and the second focused on Pulse itself: 10:00-11:00 FeedZai and Large Scale Data Processing; Introduction to FeedZai; FeedZai Pulse and Complex Event Processing; Demonstration; Use-Cases and Applications; Conclusion and Q&A. 11:00-11:15 Coffee break 11:15-12:30 FeedZai Pulse Under the Hood; A First FeedZai Pulse Application; PulseQL overview; Defining KPIs and Baselines; Conclusion and Q&A. About the speakers Nuno Sebastião is the CEO of FeedZai. Having worked for many years for the European Space Agency (ESA), he was responsible the overall design and development of Satellite Simulation Infrastructure of the agency. Having left ESA to found FeedZai, Nuno is currently responsible for the whole operations of the company. Nuno holds an M.Eng. in Informatics Engineering for the University of Coimbra, and an MBA from the London Business School. Paulo Marques is the CTO of FeedZai, being responsible for product development. Paulo is an Assistant Professor at the University of Coimbra, in the area of Distributed Data Processing, and an Adjunct Associated Professor at Carnegie Mellon, in the US. In the past Paulo lead a large number of projects for institutions like the ESA, Microsoft Research, SciSys, Siemens, among others, being now fully dedicated to FeedZai. Paulo holds a Ph.D. in Distributed Systems from the University of Coimbra.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Hybrid Quantum Networks with Discrete Polarizations and Continuous Quadrature Variables

Final Scientific/Technical Report for DOE Project DE-SC0022069. Quantum communication and computation systems have evolved along two largely independent paradigms: discrete-variable (DV) systems that encode information in qubits such as photon polarizations or photon-number states, and continuous-variable (CV) systems that encode information in optical field quadratures. Each approach offers distinct advantages—DVs provide low error rates and compatibility with single-photon platforms, while CVs support deterministic operations and efficient quantum state manipulation. A fundamental challenge for building a scalable Quantum Internet lies in interfacing these two regimes into a unified hybrid architecture that can coherently distribute and process quantum information across heterogeneous quantum nodes. This project aims to develop and demonstrate a hybrid optical quantum network that seamlessly integrates DV and CV systems. Specifically, we investigate a new class of hybrid entanglement between the discrete polarizations of single photons and the continuous quadrature variables of optical cat states, overcoming incompatibilities in existing DV–CV demonstrations. Using this new entanglement resource, the team investigates a multi-node hybrid quantum local area network (Q-LAN) testbed capable of DV–CV entanglement generation, swapping, and distribution across fiber links.

74 ATOMIC AND MOLECULAR PHYSICS↗
DOE OSTI2011

Residential Real-time Price Response Simulation

The electric industry is gaining experience with innovative price responsive demand pilots and limited roll-outs to customers. One of these pilots is investigating real-time pricing signals to engage end-use systems and local distributed generation and storage in a distributed optimization process. Attractive aspects about the approach include strong scalability characteristics, simplified interfaces between automation devices, and the adaptability to integrate a wide variety of devices and systems. Experience in this nascent field is revealing a rich array of for engineering decisions and the application of complexity theory. To test the decisions, computer simulations are used to reveal insights about design, demand elasticity, and the limits of response (including consumer fatigue). Agent-based approaches lend themselves well in the simulation to modeling the participation and interaction of each piece of equipment on a distribution feeder. This paper discusses rate design and simulation experiences at the distribution feeder level where consumers and their HVAC systems and water heaters on a feeder receive real-time pricing signals.

99 GENERAL AND MISCELLANEOUS↗
DOE OSTI2018

Cost Recovery for Standby and Ancillary Services for Distributed Generation: Considerations for New Mexico

A New Mexico statute allows investor-owned utilities to recover costs of “standby and ancillary services” resulting from distributed generation (DG). In attempting to do so, it is important to recognize both the physical and economic impacts, positive and negative, that distributed generation creates. Distributed generation has different impacts on ancillary and standby services depending on the location, capacity, and production patterns of the systems. Low penetrations of small distributed generation systems will not create significant contingencies for utilities. System impact of fluctuations in DG production may be no larger than normal fluctuations caused by customer demand, such as an air conditioner cycling. Higher penetrations and large DG system may produce noticeable impacts; so magnitude is critical when determining potential additional recoverable utility costs. Understanding DG impact on ancillary services is a relatively new and emerging area of study and as such there are not well-established formulas or protocols. Discretion is required. A hosting capacity analysis can help determine how much DG can be implemented before the impacts become substantive. Standby rates are traditionally used for large industrial generators or combined heat and power systems and threshold system size levels are often used in determining applicability of standby rates. System size thresholds may also be appropriate for standby rates and DG. Relatively small, geographically diverse DG systems are unlikely to create a noticeable coincident outage, so developing a standby charge based on assumed coincidence may not be appropriate. Individual lines and feeders may have considerable excess capacity during coincident outages, or they may be fully utilized and facing upgrades in the near future, which could vary over time. Understanding specific events that create concern for utilities is critical to developing an appropriate rate mechanism (e.g., single residential or community system outage, cascading event, clouds, eclipse). Any modeling or studies of the grid should also recognize the unique situation created when DG is coupled with a smart inverter or storage. These devices change the value proposition and can help mitigate some of the voltage and ramping issues DG can pose for a utility. Utilities can recover their fixed costs in a variety of ways, and some are better suited for DG than others. Beyond charging for standby and ancillary services, other options include fixed customer charges, demand charges, minimum bills, time-of-use rates and updated net energy-metering policies. Studies and analyses of the specific impacts that DG has on the grid should inform utilities’ decision-making process because distributed generation creates both costs and benefits for the grid. Any charge specific to DG customers should be backed up with analysis. At the heart of this conversation is utilities attempting to recover fixed costs. Allowing utilities to indiscriminately recover all stranded fixed costs can negatively affect long-term market efficiency by blunting utility incentives to lower costs and acting as a barrier to entry and exit. The current status of New Mexico’s grid, impending infrastructure updates, and likely customer adoption patterns for DG (backed up by transparent analysis) should remain at the forefront when considering charges for standby and ancillary services. Recognizing that utility regulation is largely a reactive process, it is recommended that applications for DG related standby or ancillary services cost recovery mechanisms be thoroughly reviewed by an experienced economist. If the commission has no such skill set on staff, one might be hired.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2011

Dosimetric Impact of Interfraction Catheter Movement in High-Dose Rate Prostate Brachytherapy

Purpose: To evaluate the impact of interfraction catheter movement on dosimetry in prostate high-dose-rate (HDR) brachytherapy. Methods and Materials: Fifteen patients were treated with fractionated HDR brachytherapy. Implants were performed on day 1 under transrectal ultrasound guidance. A computed tomography (CT) scan was performed. Inverse planning simulated annealing was used for treatment planning. The first fraction was delivered on day 1. A cone beam CT (CBCT) was performed on day 2 before the second fraction was given. A fusion of the CBCT and CT was performed using intraprostatic gold markers as landmarks. Initial prostate and urethra contours were transferred to the CBCT images. Bladder and rectum contours were drawn, and catheters were digitized on the CBCT. The planned treatment was applied to the CBCT dataset, and dosimetry was analyzed and compared to the initial dose distribution. This process was repeated after a reoptimization was performed, using the same constraints used on day 1. Results: Mean interfraction catheter displacement was 5.1 mm. When we used the initial plan on day 2, the mean prostate V100 (volume receiving 100 Gy or more) decreased from 93.8% to 76.2% (p < 0.01). Rectal V75 went from 0.75 cm{sup 3} to 1.49 cm{sup 3} (p < 0.01). A reoptimization resulted in a mean prostate V100 of 88.1%, closer to the initial plan (p = 0.05). Mean rectal V75 was also improved with a value of 0.59 cm{sup 3}. There was no significant change in bladder and urethra dose on day 2. Conclusions: A mean interfraction catheter displacement of 5.1 mm results in a significant decrease in prostate V100 and an increase in rectum dose. A reoptimization before the second treatment improves dose distribution.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2018

Explicitly Accounting for Multiple Spatial Scales in Data Informing Kd Distribution Development - 18414

This work provides an example of incorporating information about spatial scale into input parameter distribution development for probabilistic performance assessment (PA) models and highlights the importance of having an interdisciplinary team working together in the distribution development process. The novel approach is demonstrated for contaminant partitioning coefficients for water/rock (Kds), providing a specific context for discussing the approach and challenges, but the strategy is general and applies to many other PA model inputs. In this example, a PA model of a radioactive waste disposal site is assumed to have large cells representing spatial scales of up to thousands of square meters in area and several to tens of meters in depth. Spatial variability in Kds is expected to occur at scales much smaller than the units in the model domain over which values are applied, and therefore values are intended to represent possible averages over the modeled volume (i.e., spatial scale represented by a PA model cell). Information about possible Kd values for each material category comes from a variety of literature sources, and the reported results used to inform the distribution are available at a wide range of scales (e.g., individual batch experiment results from small-scale samples, averages based on samples collected over site-scales, averages based on samples from large regions, etc.). Effectively addressing the mismatch in spatial scale between available information and the PA Model is a challenge within PA distribution development and it takes effort from database managers, subject matter experts, modelers, and statisticians. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2015

Performance analysis of data intensive cloud systems based on data management and replication: a survey

As we delve deeper into the ‘Digital Age’, we witness an explosive growth in the volume, velocity, and variety of the data available on the Internet. For example, in 2012 about 2.5 quintillion bytes of data was created on a daily basis that originated from myriad of sources and applications including mobiledevices, sensors, individual archives, social networks, Internet of Things, enterprises, cameras, software logs, etc. Such ‘Data Explosions’ has led to one of the most challenging research issues of the current Information and Communication Technology era: how to optimally manage (e.g., store, replicated, filter, and the like) such large amount of data and identify new ways to analyze large amounts of data for unlocking information. It is clear that such large data streams cannot be managed by setting up on-premises enterprise database systems as it leads to a large up-front cost in buying and administering the hardware and software systems. Therefore, next generation data management systems must be deployed on cloud. The cloud computing paradigm provides scalable and elastic resources, such as data and services accessible over the Internet Every Cloud Service Provider must assure that data is efficiently processed and distributed in a way that does not compromise end-users’ Quality of Service (QoS) in terms of data availability, data search delay, data analysis delay, and the like. In the aforementioned perspective, data replication is used in the cloud for improving the performance (e.g., read and write delay) of applications that access data. Through replication a data intensive application or system can achieve high availability, better fault tolerance, and data recovery. In this paper, we survey data management and replication approaches (from 2007 to 2011) that are developed by both industrial and research communities. The focus of the survey is to discuss and characterize the existing approaches of data replication and management that tackle the resource usage and QoS provisioning with different levels of efficiencies. Moreover, the breakdown of both influential expressions (data replication and management) to provide different QoS attributes is deliberated. Furthermore, the performance advantages and disadvantages of data replication and management approaches in the cloud computing environments are analyzed. Open issues and future challenges related to data consistency, scalability, load balancing, processing and placement are also reported.

Malik, Saif Ur Rehman↗
DOE OSTI2015

A Novel Energy-Aware Distributed Clustering Algorithm for Heterogeneous Wireless Sensor Networks in the Mobile Environment

In order to prolong the network lifetime, energy-efficient protocols adapted to the features of wireless sensor networks should be used. This paper explores in depth the nature of heterogeneous wireless sensor networks, and finally proposes an algorithm to address the problem of finding an effective pathway for heterogeneous clustering energy. The proposed algorithm implements cluster head selection according to the degree of energy attenuation during the network’s running and the degree of candidate nodes’ effective coverage on the whole network, so as to obtain an even energy consumption over the whole network for the situation with high degree of coverage. Simulation results show that the proposed clustering protocol has better adaptability to heterogeneous environments than existing clustering algorithms in prolonging the network lifetime.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2015

CLASSIFYING X-RAY BINARIES: A PROBABILISTIC APPROACH

In X-ray binary star systems consisting of a compact object that accretes material from an orbiting secondary star, there is no straightforward means to decide whether the compact object is a black hole or a neutron star. To assist in this process, we develop a Bayesian statistical model that makes use of the fact that X-ray binary systems appear to cluster based on their compact object type when viewed from a three-dimensional coordinate system derived from X-ray spectral data where the first coordinate is the ratio of counts in the mid- to low-energy band (color 1), the second coordinate is the ratio of counts in the high- to low-energy band (color 2), and the third coordinate is the sum of counts in all three bands. We use this model to estimate the probabilities of an X-ray binary system containing a black hole, non-pulsing neutron star, or pulsing neutron star. In particular, we utilize a latent variable model in which the latent variables follow a Gaussian process prior distribution, and hence we are able to induce the spatial correlation which we believe exists between systems of the same type. The utility of this approach is demonstrated by the accurate prediction of system types using Rossi X-ray Timing Explorer All Sky Monitor data, but it is not flawless. In particular, non-pulsing neutron systems containing “bursters” that are close to the boundary demarcating systems containing black holes tend to be classified as black hole systems. As a byproduct of our analyses, we provide the astronomer with the public R code which can be used to predict the compact object type of XRBs given training data.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2011

Expected Performance of the ATLAS Experiment - Detector, Trigger and Physics

The Large Hadron Collider (LHC) at CERN promises a major step forward in the understanding of the fundamental nature of matter. The ATLAS experiment is a general-purpose detector for the LHC, whose design was guided by the need to accommodate the wide spectrum of possible physics signatures. The major remit of the ATLAS experiment is the exploration of the TeV mass scale where groundbreaking discoveries are expected. In the focus are the investigation of the electroweak symmetry breaking and linked to this the search for the Higgs boson as well as the search for Physics beyond the Standard Model. In this report a detailed examination of the expected performance of the ATLAS detector is provided, with a major aim being to investigate the experimental sensitivity to a wide range of measurements and potential observations of new physical processes. An earlier summary of the expected capabilities of ATLAS was compiled in 1999 [1]. A survey of physics capabilities of the CMS detector was published in [2]. The design of the ATLAS detector has now been finalised, and its construction and installation have been completed [3]. An extensive test-beam programme was undertaken. Furthermore, the simulation and reconstruction software code and frameworks have been completely rewritten. Revisions incorporated reflect improved detector modelling as well as major technical changes to the software technology. Greatly improved understanding of calibration and alignment techniques, and their practical impact on performance, is now in place. The studies reported here are based on full simulations of the ATLAS detector response. A variety of event generators were employed. The simulation and reconstruction of these large event samples thus provided an important operational test of the new ATLAS software system. In addition, the processing was distributed world-wide over the ATLAS Grid facilities and hence provided an important test of the ATLAS computing system - this is the origin of the expression 'CSC studies' ('computing system commissioning'), which is occasionally referred to in these volumes. The work reported does generally assume that the detector is fully operational, and in this sense represents an idealised detector: establishing the best performance of the ATLAS detector with LHC proton-proton collisions is a challenging task for the future. The results summarised here therefore represent the best estimate of ATLAS capabilities before real operational experience of the full detector with beam. Unless otherwise stated, simulations also do not include the effect of additional interactions in the same or other bunch-crossings, and the effect of neutron background is neglected. Thus simulations correspond to the low-luminosity performance of the ATLAS detector. This report is broadly divided into two parts: firstly the performance for identification of physics objects is examined in detail, followed by a detailed assessment of the performance of the trigger system. This part is subdivided into chapters surveying the capabilities for charged particle tracking, each of electron/photon, muon and tau identification, jet and missing transverse energy reconstruction, b-tagging algorithms and performance, and finally the trigger system performance. In each chapter of the report, there is a further subdivision into shorter notes describing different aspects studied. The second major subdivision of the report addresses physics measurement capabilities, and new physics search sensitivities. Individual chapters in this part discuss ATLAS physics capabilities in Standard Model QCD and electroweak processes, in the top quark sector, in b-physics, in searches for Higgs bosons, supersymmetry searches, and finally searches for other new particles predicted in more exotic models.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2025

Harnessing distributed GPU computing for generalizable graph convolutional networks in power grid reliability assessments

Although machine learning (ML) has emerged as a powerful tool for rapidly assessing grid contingencies, prior studies have largely considered a static grid topology in their analyses. This limits their application, since they need to be re-trained for every new topology. Here, this paper explores the development of generalizable graph convolutional network (GCN) models by pre-training them across a range of grid topologies and contingency types. We found that a GCN model with auto-regressive moving average (ARMA) layers with a line graph representation of the grid offered the best predictive performance in predicting voltage magnitudes (VM) and voltage angles (VA). We introduced the concept of phantom nodes to consider disparate grid topologies with a varying number of nodes and lines. For pre-training the GCN ARMA model across a variety of topologies, distributed graphics processing unit (GPU) computing afforded us significant training scalability. The predictive performance of this model on grid topologies that were part of the training data is substantially better than the direct current (DC) approximation. Although direct application of the pre-trained model to topologies that are not part of the grid is not particularly satisfactory, fine-tuning with small amounts of data from a specific topology of interest significantly improves predictive performance. In general, this paper highlights the feasibility of training large-scale GNN models to assess the reliability of power grids by considering a wide variety of grid topologies and contingency types. With the advent of foundational models in ML and the exponential increase in GPU computing clusters, generalizable ML models will significantly enhance how utilities manage power systems and make decisions in real-time or near-real-time.

24 - POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2012

Foundations of Feature Selection and Classification for Non-Gaussian Distributed Targets

This report describes the author’s research toward creating algorithms for selecting useful statistical features in pattern/target classification problems in which the features are non-Gaussian distributed. In engineering practice, it is common to either (1) Not perform any feature selection procedure, or (2) Use a feature selection algorithm that assumes the features are Gaussian distributed. This process can be far from optimal if the data are actually non-Gaussian, as they frequently are in practice. This research has the goal of mitigating that problem by creating a useful algorithm that can be used in practice. The approach is to focus on the performance indices used in common feature selection algorithms, such as Branch and Bound algorithms and the Sequential Forward Selection Algorithm. Ordinarily, the performance index used to measure the class separation in feature space involves assuming the data are Gaussian and deriving tractable performance indices that use closed form equations that represent the probability density functions of the class data. The advantage of this approach is that it produces feature selection algorithms that have low computational complexity and do not require knowledge of the actual data densities. The disadvantage is that these algorithms may not perform reasonably when the data are non-Gaussian, as they commonly are in practice. This research examines the use of information-theoretic class separability measures that can deal with the non-Gaussian case. In particular, this work shows that the Hellinger Distance (a divergence measure), has very desirable mathematical properties and can be useful for feature selection. The work in this paper employs a multivariate kernel density estimator along with the Hellinger Distance and an optimal subset selection algorithm to do feature selection. This report (1) Defines the classification and feature selection problems, (2) Describes the continuous variable Hellinger distance in one dimension and discusses its mathematical properties in comparison with other distance measures, (3) Develops the discrete variable Hellinger distance expression for one and two dimensions, and shows that the discrete Hellinger distance for multiple dimensions is a straightforward extension, and (4) Presents experimental density estimation results using a new MATLAB implementation of a multivariate kernel density estimator. The draft of this report served as a starting point for Ensign Matthew Wilder’s MSEE thesis work at the Naval Postgraduate School (NPS) in 2010 and 2011. Ensign Wilder completed his thesis and graduated from NPS in June 2011 (M. J. Wilder, “Automatic Target Recognition: Statistical Feature Selection of Non-Gaussian Distributed Target Classes,” MSEE Thesis, Naval Postgraduate School, Thesis Advisor: Grace A. Clark, Second Reader: Monique P. Fargues, June 2011). In the thesis, Ensign Wilder created implementations of the Branch and Bound and Sequential Feature Selection algorithms, extended the multivariate kernel density estimator code, and evaluated the performance of Hellinger Distance-based feature selection algorithms using both simulated data and data from a real-world benchmark data set. The new algorithm was shown to be very effective for both Gaussian and non-Gaussian data sets. The advantages gained for non-Gaussian data come at the cost of increased computational complexity. Using a laptop computer for practical problems, we generally use initial feature vectors containing no more than about ten or twelve features. Proposed future research lies in creating faster kernel density estimator algorithms that would allow for the use of higher dimensional data.

42 ENGINEERING↗