Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Distribution networks”

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 1,621 records · Page 90

Issues in ATM Support of High-Performance, Geographically Distributed Computing

This report experimentally assesses the effect of the underlying network in a cluster-based computing environment. The assessment is quantified by application-level benchmarking, process-level communication, and network file input/output. Two testbeds were considered, one small cluster of Sun workstations and another large cluster composed of 32 high-end IBM RS/6000 platforms. The clusters had Ethernet, fiber distributed data interface (FDDI), Fibre Channel, and asynchronous transfer mode (ATM) network interface cards installed, providing the same processors and operating system for the entire suite of experiments. The primary goal of this report is to assess the suitability of an ATM-based, local-area network to support interprocess communication and remote file input/output systems for distributed computing.

Claus, Russell W.↗

Solving high-dimensional inverse problems using amortized likelihood-free inference with noisy and incomplete data

Here, we present a likelihood-free probabilistic inversion method based on normalizing flows for high-dimensional inverse problems. The proposed method is composed of two complementary networks: a summary network for data compression and an inference network for parameter estimation. The summary network encodes raw observations into a fixed-size vector of summary features, while the inference network generates samples of the approximate posterior distribution of the model parameters based on these summary features. The posterior samples are produced in a deep generative fashion by sampling from a latent Gaussian distribution and passing these samples through an invertible transformation. We construct this invertible transformation by sequentially alternating conditional invertible neural network and conditional neural spline flow layers. The summary and inference networks are trained simultaneously. We apply the proposed method to an inversion problem in groundwater hydrology to estimate the posterior distribution of the log-conductivity field conditioned on spatially sparse time-series observations of the system’s hydraulic head responses. The conductivity field is represented with 706 degrees of freedom in the considered problem. Comparison with the likelihood-based iterative ensemble smoother PEST-IES method demonstrates that the proposed method accurately estimates the parameter posterior distribution and the observations’ predictive posterior distribution at a fraction of the inference time of PEST-IES.

conditional invertible neural network↗

Solving inverse problems in stochastic models using deep neural networks and adversarial training

Inverse problems associated with stochastic models constitute a significant portion of scientific and engineering applications. In such cases the unknown quantities are distributions. The applicability of traditional methods is limited because of their demanding assumptions or prohibitive computational consumption; for example, maximum likelihood methods require closed-form density functions, and Markov Chain Monte Carlo needs a large number of simulations. We propose a new method that estimates the unknown distribution by matching the statistical properties between observed and simulated random processes. We leverage the expressive power of neural networks to approximate the unknown distribution and use a discriminative neural network for computing the statistical discrepancies between the observed and simulated random processes. Here we demonstrated numerically that the proposed methods can estimate both the model parameters and learn complicated unknown distributions.

42 ENGINEERING↗

Photospheric models of solar active regions and the network based on the Mg II h and k line wings

From a comparison between observed and computed wings of the Mg II resonance lines, distributions of temperature versus mass column density for solar photospheric layers in plages and in the chromospheric network are derived. The theoretical profiles are computed assuming partial coherent scattering. In the active regions, temperatures exceed those in the quiet sun by up to 200 K near the temperature minimum and up to 400 K in deeper layers. In the observed network structure, the temperature is enhanced by 200 K at the temperature minimum but is the same as that in the quiet sun at greater depths. The difference in the slope of the temperature distribution between the network and plages is real, but may refer only to long elements of the network rather than to the brightest portions. Adjacent to the network is a region in which the temperatures are similar to those in the quiet sun, except immediately below the temperature minimum, where the temperatures are depressed by 150 K.

Morrison, N. D.↗

BIO-Plex Information System Concept

This paper describes a suggested design for an integrated information system for the proposed BIO-Plex (Bioregenerative Planetary Life Support Systems Test Complex) at Johnson Space Center (JSC), including distributed control systems, central control, networks, database servers, personal computers and workstations, applications software, and external communications. The system will have an open commercial computing and networking, architecture. The network will provide automatic real-time transfer of information to database server computers which perform data collection and validation. This information system will support integrated, data sharing applications for everything, from system alarms to management summaries. Most existing complex process control systems have information gaps between the different real time subsystems, between these subsystems and central controller, between the central controller and system level planning and analysis application software, and between the system level applications and management overview reporting. An integrated information system is vitally necessary as the basis for the integration of planning, scheduling, modeling, monitoring, and control, which will allow improved monitoring and control based on timely, accurate and complete data. Data describing the system configuration and the real time processes can be collected, checked and reconciled, analyzed and stored in database servers that can be accessed by all applications. The required technology is available. The only opportunity to design a distributed, nonredundant, integrated system is before it is built. Retrofit is extremely difficult and costly.

Jones, Harry↗

Reducing Communication in Graph Neural Network Training

Graph Neural Networks (GNNs) are powerful and flexible neural networks that use the naturally sparse connectivity information of the data. GNNs represent this connectivity as sparse matrices, which have lower arithmetic intensity and thus higher communication costs compared to dense matrices, making GNNs harder to scale to high concurrencies than convolutional or fully-connected neural networks. Here, we introduce a family of parallel algorithms for training GNNs and show that they can asymptotically reduce communication compared to previous parallel GNN training methods. We implement these algorithms, which are based on 1D, 1. 5D, 2D, and 3D sparse-dense matrix multiplication, using torch.distributed on GPU-equipped clusters. Our algorithms optimize communication across the full GNN training pipeline. We train GNNs on over a hundred GPUs on multiple datasets, including a protein network with over a billion edges.

97 MATHEMATICS AND COMPUTING↗

NASA Near Earth Network (NEN) Support for Lunar and L1/L2 CubeSats

The NASA Near Earth Network (NEN) consists of globally distributed tracking stations, including NASA, commercial, and partner ground stations, that are strategically located to maximize the coverage provided to a variety of orbital and suborbital missions, including those in LEO, GEO, HEO, lunar and L1/L2 orbits. The NENs future mission set includes and will continue to include CubeSat missions. The majority of the CubeSat missions destined to fly on EM-1, launching in late 2018, many in a lunar orbit, will communicate with ground based stations via X-band and will utilize the NASA Jet Propulsion Laboratory (JPL) developed IRIS radio. The NEN recognizes the important role CubeSats are beginning to play in carrying out NASAs mission and is therefore investigating the modifications needed to provide IRIS radio compatibility. With modification, the NEN could potentially expand support to the EM-1 lunar CubeSats.The NEN could begin providing significant coverage to lunar CubeSat missions utilizing three to four of the NENs mid-latitude sites. This coverage would supplement coverage provided by the JPL Deep Space Network (DSN). The NEN, with smaller apertures than DSN, provides the benefit of a larger beamwidth that could be beneficial in the event of uncertain ephemeris data. In order to realize these benefits the NEN would need to upgrade stations targeted based on coverage ability and current configuration/ease of upgrade, to ensure compatibility with the IRIS radio. In addition, the NEN is working with CubeSat radio developers to ensure NEN compatibility with alternative CubeSat radios for Lunar and L1/L2 CubeSats. The NEN has provided NEN compatibility requirements to several radio developers who are developing radios that offer lower cost and, in some cases, more capabilities with fewer constraints. The NEN is ready to begin supporting CubeSat missions. The NEN is considering network upgrades to broaden the types of CubeSat missions that can be supported and is supporting both the CubeSat community and radio developers to ensure future CubeSat missions have multiple options when choosing a network for their communications support.

Schaire, Scott↗

Science Rationale for a Micro-Met Mission to Augment InterMarsNet

The 2003 opportunity has the potential to carry out for the first time in Mars exploration history, coordinated measurements from the surface and from orbit that can address fundamental issues associated with the Martian global circulation and climate system. Coordinated measurements are defined here to mean collecting meteorological data from a network of 12-16 globally distributed surface stations simultaneously with an orbiter carrying an atmospheric sounder. With such measurements it is possible to define the horizontally varying (barotropic) and vertically varying (baroclinic) components of the global circulation from which the full 3-dimensional horizontal wind field can be reconstructed. It is also possible to precisely define the CO2 cycle, the main component of the current climate system. InterMarsNet, as currently envisioned, consists of 3-4 landers with a supporting communications orbiter that may carry some instrumentation. The landers are likely to touch down in low latitudes and will probably be configured to optimize seismological objectives. We propose to augment the InterMarsNet meteorological objectives by flying an additional 10-15 "MicroMet" landers equipped to measure surface pressure and nothing else. Surface pressure is the most fundamental meteorological parameter and it is the easiest to measure. The sensors are light, operate with minimal power, and do not require orientation or deployment. Consequently the landers can be very small (< 10 kg) and 10-15 of them can be delivered by a Med-lite launcher. This would enable global network science for meteorology and, when combined wAh the more sophisticated measurements from the InterMarsNet landers and the simultaneous temperature and dust profiling measurements from an atmospheric sounder aboard the orbiter, it would enable us to reconstruct global scale circulation patterns.

Haberle, Robert M.↗

A hybrid architecture for volt-var control in active distribution grids

Modern active distribution grids are characterized by the increasing penetration of distributed energy resources (DERs). The proper coordination and scheduling of a large numbers of these small-scale and spatially distributed DERs is necessary, and warrants the use of novel distributed approaches. In this paper, we propose a hybrid volt-var control architecture for the distribution grid, which leverages existing centralized and local approaches to planning, decision making, and control, and augments it with distributed optimization and distributed control for DER management. First, we propose a convex model to describe the power physics of distribution grids of meshed topology and unbalanced structure, based on current injection and McCormick Envelopes. Second, we employ the distributed proximal atomic coordination (PAC) algorithm to coordinate DERs to provide voltage support. We implement volt-var optimization by optimally coordinating DERs including PV smart inverters and demand response. We present results using the IEEE-34 bus network, using real data from a distribution feeder in Hawaii, to model load and PV generation. Different levels of DER penetration and objective functions are simulated. Finally, our results show the need for the coordination of DERs to improve voltage profiles, even in networks with existing voltage control devices. Further, we show the need for flexible reactive power capabilities to achieve desired grid performance.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Distributed Multi-GPU Community Detection on Exascale Computing Platforms

Community detection is a fundamental operation in graph mining, and by uncovering hidden structures and patterns within complex systems it helps solve fundamental problems pertaining to social networks, such as information diffusion, epidemics, and recommender systems. Scaling graph algorithms for massive networks becomes challenging on modern distributed-memory multi-GPU (Graphics Processing Unit) systems due to limitations such as irregular memory access patterns, load imbalances, higher communication-computation ratios, and cross-platform support. We present a novel algorithm HiPDPL-GPU (distributed parallel Louvain) to address these challenges. We conduct experiments involving different partitioning techniques to achieve optimized performance of HiPDPL-GPU on the two largest supercomputers: Frontier and Summit. Remarkably, HiPDPL-GPU processes a graph with 4.2 billion edges in less than 3 minutes using 1024 GPUs. Qualitatively performance of HiPDPL-GPU is similar or better compared to other state-of-the-art CPU- and GPU-based implementations. While prior GPU implementations have predominantly employed CUDA, our first-of-its-kind implementation for community detection is cross-platform, accommodating both AMD and NVIDIA GPUs.

graph algorithms, high performance comptuing↗

Distributed Multi-GPU Community Detection on Exascale Computing Platforms

Community detection is a fundamental operation in graph mining, and by uncovering hidden structures and patterns within complex systems it helps solve fundamental problems pertaining to social networks, such as information diffusion, epidemics, and recommender systems. Scaling graph algorithms for massive networks becomes challenging on modern distributed-memory multi-GPU (Graphics Processing Unit) systems due to limitations such as irregular memory access patterns, load imbalances, higher communication-computation ratios, and cross-platform support. We present a novel algorithm HiPDPL-GPU (Distributed Parallel Louvain) to address these challenges. We conduct experiments involving different partitioning techniques to achieve an optimized performance of HiPDPL-GPU on the two largest supercomputers: Frontier and Summit. Remarkably, HiPDPL-GPU processes a graph with 4.2 billion edges in less than 3 minutes using 1024 GPUs. Qualitatively, the performance of HiPDPL-GPU is similar or better compared to other state-of-the-art CPU- and GPU-based implementations. While prior GPU implementations have predominantly employed CUDA, our first-of-its-kind implementation for community detection is cross-platform, accommodating both AMD and NVIDIA GPUs.

Sattar, Naw Safrin↗

NASA Near Earth Network (NEN) Support for Lunar and L1/L2 CubeSats

The NASA Near Earth Network (NEN) consists of globally distributed tracking stations, including NASA, commercial, and partner ground stations, that are strategically located to maximize the coverage provided to a variety of orbital and suborbital missions, including those in LEO, GEO, HEO, lunar and L1/L2 orbits. The NENs future mission set includes and will continue to include CubeSat missions. The first NEN supported CubeSat mission will be the Cubesat Proximity Operations Demonstration (CPOD) launching into low earth orbit (LEO) in early 2017. The majority of the CubeSat missions destined to fly on EM-1, launching in late 2018, many in a lunar orbit, will communicate with ground based stations via X-band and will utilize the NASA Jet Propulsion Laboratory (JPL) developed IRIS radio. The NEN recognizes the important role CubeSats are beginning to play in carrying out NASAs mission and is therefore investigating the modifications needed to provide IRIS radio compatibility. With modification, the NEN could potentially expand support to the EM-1 lunar CubeSats. The NEN could begin providing significant coverage to lunar CubeSat missions utilizing three to four of the NENs mid-latitude sites. This coverage would supplement coverage provided by the JPL Deep Space Network (DSN). The NEN, with smaller apertures than DSN, provides the benefit of a larger beamwidth that could be beneficial in the event of uncertain ephemeris data. In order to realize these benefits the NEN would need to upgrade stations targeted based on coverage ability and current configurationease of upgrade, to ensure compatibility with the IRIS radio.In addition, the NEN is working with CubeSat radio developers to ensure NEN compatibility with alternative CubeSat radios for Lunar and L1/L2 CubeSats. The NEN has provided NEN compatibility requirements to several radio developers who are developing radios that offer lower cost and, in some cases, more capabilities with fewer constraints. The NEN is ready to begin supporting CubeSat missions. The NEN is considering network upgrades to broaden the types of CubeSat missions that can be supported and is supporting both the CubeSat community and radio developers to ensure future CubeSat missions have multiple options when choosing a network for their communications support.

Global Tracking Network Ground Station↗

Bridging the gap between NMR measured mean silicate chain length and nano-scale silicate polymorphism of calcium silicate hydrates

Variable stoichiometry and silicate polymorphism in calcium-silicate-hydrates (C-S-H) has impeded the revelation of point defects distribution in the silicate tetrahedral network of C-S-H, which resembles tobermorite crystal structure with some bridging tetrahedra (BT) and paired tetrahedra (PT) vacancies in dreierketten chains. Here we use a computational approach to characterize silicate polymorphism by introducing the vacancy ratio of BT to PT (α) and establishing a three-term empirical mean chain length (MCL) formula for different calcium-to‑silicon ratios (Ca/Si) ranging from 1.2 to 2.3. The formula identifies BT and PT controls at low and high α respectively and allows an inverse mapping of polymorphic range based on NMR experiments. The proposed computational framework quantitatively describes silicate polymorphs and links NMR-measured MCL to C-S-H atomistic configurations at the molecular level.

36 MATERIALS SCIENCE↗

Generating Massive Scale-free Networks: Novel Parallel Algorithms using the Preferential Attachment Model

Recently, there has been substantial interest in the study of various random networks as mathematical models of complex systems. As real-life complex systems grow larger, the ability to generate progressively large random networks becomes all the more important. This motivates the need for efficient parallel algorithms for generating such networks. Naïve parallelization of sequential algorithms for generating random networks is inefficient due to inherent dependencies among the edges and the possibility of creating duplicate (parallel) edges. In this article, we present message passing interface-based distributed memory parallel algorithms for generating random scale-free networks using the preferential-attachment model. Our algorithms are experimentally verified to scale very well to a large number of processing elements (PEs), providing near-linear speedups. The algorithms have been exercised with regard to scale and speed to generate scale-free networks with one trillion edges in 6 minutes using 1,000 PEs.

97 MATHEMATICS AND COMPUTING↗

Near Earth Network (NEN) CubeSat Communications

The NASA Near Earth Network (NEN) consists of globally distributed tracking stations, including NASA, commercial, and partner ground stations, that are strategically located to maximize the coverage provided to a variety of orbital and suborbital missions, including those in LEO (Low Earth Orbit), GEO (Geosynchronous Earth Orbit), HEO (Highly Elliptical Orbit), lunar and L1-L2 orbits. The NEN's future mission set includes and will continue to include CubeSat missions. The first NEN-supported CubeSat mission will be the Cubesat Proximity Operations Demonstration (CPOD) launching into LEO in 2017. The majority of the CubeSat missions destined to fly on EM-1, launching in late 2018, many in a lunar orbit, will communicate with ground-based stations via X-band and will utilize the NASA Jet Propulsion Laboratory (JPL)-developed IRIS (Satellite Communication for Air Traffic Management) radio. The NEN recognizes the important role CubeSats are beginning to play in carrying out NASAs mission and is therefore investigating the modifications needed to provide IRIS radio compatibility. With modification, the NEN could potentially expand support to the EM-1 (Exploration Mission-1) lunar CubeSats. The NEN could begin providing significant coverage to lunar CubeSat missions utilizing three to four of the NEN's mid-latitude sites. This coverage would supplement coverage provided by the JPL Deep Space Network (DSN). The NEN, with smaller apertures than DSN, provides the benefit of a larger beamwidth that could be beneficial in the event of uncertain ephemeris data. In order to realize these benefits the NEN would need to upgrade stations targeted based on coverage ability and current configuration ease of upgrade, to ensure compatibility with the IRIS radio. In addition, the NEN is working with CubeSat radio developers to ensure NEN compatibility with alternative CubeSat radios for Lunar and L1-L2 CubeSats. The NEN has provided NEN compatibility requirements to several radio developers who are developing radios that offer lower cost and, in some cases, more capabilities with fewer constraints. The NEN is ready to begin supporting CubeSat missions. The NEN is considering network upgrades to broaden the types of CubeSat missions that can be supported and is supporting both the CubeSat community and radio developers to ensure future CubeSat missions have multiple options when choosing a network for their communications support.

CubeSat↗

Cloud-Based Demodulation and Data Distribution of a Satellite Downlink

Ground station networks connected to the cloud allow space missions to have global communications coverage without operating their own infrastructure. In this work, we describe the communications architecture for the TechEdSat-13 mission, which performed the first in-space characterization of a neuromorphic processor. The mission utilizes a commercial provider for S-band downlinks. A suite of cloud services and open-source software such as GNU Radio are leveraged to demodulate signals received by an AWS ground station during passes with TechEdSat-13 and store recovered data. Once a pass is scheduled, the entire process takes place without human intervention. On-orbit results the past year of operations are presented, demonstrating the advantages of this approach over traditional operator-owned ground stations. Use of software-defined radio makes possible custom signal processing. The homogeneity of apertures and their interfaces to the cloud simplifies scaling across many sites. This abundance of candidate links lays the groundwork for intelligent scheduling agents to optimize pass selection across several factors, automatically recover from failed contacts, and gather metrics to learn from past performance.

cloud demodulation↗

The NASA Integrated Information Technology Architecture

This document defines an Information Technology Architecture for the National Aeronautics and Space Administration (NASA), where Information Technology (IT) refers to the hardware, software, standards, protocols and processes that enable the creation, manipulation, storage, organization and sharing of information. An architecture provides an itemization and definition of these IT structures, a view of the relationship of the structures to each other and, most importantly, an accessible view of the whole. It is a fundamental assumption of this document that a useful, interoperable and affordable IT environment is key to the execution of the core NASA scientific and project competencies and business practices. This Architecture represents the highest level system design and guideline for NASA IT related activities and has been created on the authority of the NASA Chief Information Officer (CIO) and will be maintained under the auspices of that office. It addresses all aspects of general purpose, research, administrative and scientific computing and networking throughout the NASA Agency and is applicable to all NASA administrative offices, projects, field centers and remote sites. Through the establishment of five Objectives and six Principles this Architecture provides a blueprint for all NASA IT service providers: civil service, contractor and outsourcer. The most significant of the Objectives and Principles are the commitment to customer-driven IT implementations and the commitment to a simpler, cost-efficient, standards-based, modular IT infrastructure. In order to ensure that the Architecture is presented and defined in the context of the mission, project and business goals of NASA, this Architecture consists of four layers in which each subsequent layer builds on the previous layer. They are: 1) the Business Architecture: the operational functions of the business, or Enterprise, 2) the Systems Architecture: the specific Enterprise activities within the context of IT systems, 3) the Technical Architecture: a common, vendor-independent framework for design, integration and implementation of IT systems and 4) the Product Architecture: vendor=specific IT solutions. The Systems Architecture is effectively a description of the end-user "requirements". Generalized end-user requirements are discussed and subsequently organized into specific mission and project functions. The Technical Architecture depicts the framework, and relationship, of the specific IT components that enable the end-user functionality as described in the Systems Architecture. The primary components as described in the Technical Architecture are: 1) Applications: Basic Client Component, Object Creation Applications, Collaborative Applications, Object Analysis Applications, 2) Services: Messaging, Information Broker, Collaboration, Distributed Processing, and 3) Infrastructure: Network, Security, Directory, Certificate Management, Enterprise Management and File System. This Architecture also provides specific Implementation Recommendations, the most significant of which is the recognition of IT as core to NASA activities and defines a plan, which is aligned with the NASA strategic planning processes, for keeping the Architecture alive and useful.

Baldridge, Tim↗

Pseudo Aircraft Systems - A multi-aircraft simulation system for air traffic control research

Pseudo Aircraft Systems (PAS) is a computerized flight dynamics and piloting system designed to provide a high fidelity multi-aircraft real-time simulation environment to support Air Traffic Control research. PAS is composed of three major software components that run on a network of computer workstations. Functionality is distributed among these components to allow the system to execute fast enough to support real-time operation. PAS workstations are linked by an Ethernet Local Area Network, and standard UNIX socket protocol is used for data transfer. Each component of PAS is controlled and operated using a custom designed Graphical User Interface. Each of these is composed of multiple windows, and many of the windows and sub-windows are used in several of the components. Aircraft models and piloting logic are sophisticated and realistic and provide complex maneuvering and navigational capabilities. PAS will continually be enhanced with new features and improved capabilities to support ongoing and future Air Traffic Control system development.

Weske, Reid A.↗