Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “NETWORK ANALYSIS”

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 595 records · Page 33
All sourcesNASA NTRSDOE OSTI
NASA NTRSDate not supplied

Feature-Guided Analysis of Neural Networks

Applying standard software engineering practices to neural networks is challenging due to the lack of high-level abstractions describing a neural network’s behavior. To address this challenge, we propose to extract high-level task-specific features from the neural network internal representation, based on monitoring the neural network activations.The extracted feature representations can serve as a link to high-level requirements and can be leveraged to enable fundamental software engineering activities, such as automated testing, debugging, requirements analysis, and formal verification, leading to better engineering of neural networks. Using two case studies, we present initial empirical evidence demonstrating the feasibility of our ideas.

Features↗
DOE OSTI2016

Digital image analysis to quantify carbide networks in ultrahigh carbon steels

A method has been developed and demonstrated to quantify the degree of carbide network connectivity in ultrahigh carbon steels through digital image processing and analysis of experimental micrographs. It was shown that the network connectivity and carbon content can be correlated to toughness for various ultrahigh carbon steel specimens. The image analysis approach first involved segmenting the carbide network and pearlite matrix into binary contrast representations via a grayscale intensity thresholding operation. Next, the carbide network pixels were skeletonized and parceled into braches and nodes, allowing the determination of a connectivity index for the carbide network. Intermediate image processing steps to remove noise and fill voids in the network are also detailed. The connectivity indexes of scanning electron micrographs were consistent in both secondary and backscattered electron imaging modes, as well as across two different (50 × and 100 ×) magnifications. Results from ultrahigh carbon steels reported here along with other results from the literature generally showed lower connectivity indexes correlated with higher Charpy impact energy (toughness). A deviation from this trend was observed at higher connectivity indexes, consistent with a percolation threshold for crack propagation across the carbide network. - Highlights: • A method for carbide network analysis in steels is proposed and demonstrated. • ImageJ method extracts a network connectivity index from micrographs. • Connectivity index consistent in different imaging conditions and magnifications. • Impact energy may plateau when a critical network connectivity is exceeded.

36 MATERIALS SCIENCE↗
DOE OSTI2019

Slip Tendency Analysis of Fracture Networks to Determine Suitability of Candidate Testbeds for the EGS Collab Hydroshear Experiment

Experiment 2 of the EGS Collab project is aimed at testing stimulation by hydro-shearing of existing natural fractures, versus Experiment 1, which is focused on hydraulic fracturing a rock mass to enhance permeability. The main criterion for the testbed selection in Experiment 2 is the presence of an interconnected network of fractures, at least 10 meters in extent, and with orientations such that shear slip can be induced at injection pressures less than the minimum horizontal stress (Shmin). The feasibility analysis for this experiment requires a well-constrained stress state along with well-characterized fracture networks. The fracture systems at two candidate locations in the Sanford Underground Research Facility (SURF), the 4850 Level (number refers to depth below ground surface in feet) and the 4100 Level, have been characterized to different extents. The 4850 Level has a well-characterized Discrete Fracture Network (DFN) from borehole (drilled for an experimental test bed) and drift observations. In contrast, the 4100 level has a fracture network characterized only by observations from the drift wall, which provides little constraint on fracture extents. This paper will present assessments for the interpreted slip potential for the natural fractures at the two locations. Data uncertainties are addressed by performing a probabilistic analysis that takes into account the uncertainty in the stress state, uncertainty in the fracture properties, and preliminary borehole locations. There is strong evidence of shear stimulation of some natural fractures in Experiment 1 on the 4850 level. We use examples of natural fractures strongly linked to shear stimulation in Experiment 1 from multiple monitoring indicators to test the consistency of the stress model and guide the uncertainty interpretation. This analysis is intended to guide the site selection process for Experiment 2 by highlighting the fracture orientations that are likely to be shear-stimulated in a majority of modeled realizations.

15 GEOTHERMAL ENERGY↗
DOE OSTI2020

Slip Tendency Analysis of Fracture Networks to Determine Suitability of Candidate Testbeds for the EGS Collab Hydroshear Experiment

Experiment 2 of the EGS Collab project is aimed at testing stimulation by hydro-shearing of existing natural fractures, versus Experiment 1, which is focused on hydraulic fracturing a rock mass to enhance permeability. The main criterion for the testbed selection in Experiment 2 is the presence of an interconnected network of fractures, at least 10 meters in extent, and with orientations such that shear slip can be induced at injection pressures less than the minimum horizontal stress (Shmin). The feasibility analysis for this experiment requires a well-constrained stress state along with well-characterized fracture networks. The fracture systems at two candidate locations in the Sanford Underground Research Facility (SURF), the 4850 Level (number refers to depth below ground surface in feet) and the 4100 Level, have been characterized to different extents. The 4850 Level has a well-characterized Discrete Fracture Network (DFN) from borehole (drilled for an experimental test bed) and drift observations. In contrast, the 4100 level has a fracture network characterized only by observations from the drift wall, which provides little constraint on fracture extents. This paper will present assessments for the interpreted slip potential for the natural fractures at the two locations. Data uncertainties are addressed by performing a probabilistic analysis that takes into account the uncertainty in the stress state, uncertainty in the fracture properties, and preliminary borehole locations. There is strong evidence of shear stimulation of some natural fractures in Experiment 1 on the 4850 level. We use examples of natural fractures strongly linked to shear stimulation in Experiment 1 from multiple monitoring indicators to test the consistency of the stress model and guide the uncertainty interpretation. This analysis is intended to guide the site selection process for Experiment 2 by highlighting the fracture orientations that are likely to be shear-stimulated in a majority of modeled realizations.

40 EE - Geothermal Technologies Office (EE-4G)↗
NASA NTRS2024

High-Rate Delay Tolerant Networking (HDTN) Software Requirements Analysis

This document serves as a detailed analysis of the main networking protocols implemented by HDTN. Sources of the protocol specifications include Internet Engineering Task Force (IETF) Request for Comments (RFC) and Consultative Committee for Space Data Systems (CCSDS) standards. The focus of this report is to derive software requirements suitable for NASA Procedural Requirements 7150.2D Class B compliance, including requirements traceability and software verification and validation, from the source specifications. This analysis will be incorporated into the finalized HDTN Software Requirements Specification (SRS) but does not encompass the full scope of the HDTN SRS. Requirements in this document are considered draft. The complete requirements will include bundle application requirements, interface requirements, computer resource requirements, software quality factors, and additional requirements as determined by the project. This document is publicly released to the greater community to receive feedback and foster collaboration opportunities.

Rachel Dudukovich↗
DOE OSTI2021

Deep Cellular Recurrent Network for Efficient Analysis of Time-Series Data With Spatial Information

Efficient processing of large-scale time series data is an intricate problem in machine learning. Conventional sensor signal processing pipelines with hand engineered feature extraction often involve huge computational cost with high dimensional data. Deep recurrent neural networks have shown promise in automated feature learning for improved time-series processing. However, generic deep recurrent models grow in scale and depth with increased complexity of the data. This is particularly challenging in presence of high dimensional data with temporal and spatial characteristics. Consequently, this work proposes a novel deep cellular recurrent neural network (DCRNN) architecture to efficiently process complex multi-dimensional time series data with spatial information. Here, the cellular recurrent architecture in the proposed model allows for location-aware synchronous processing of time series data from spatially distributed sensor signal sources. Extensive trainable parameter sharing due to cellularity in the proposed architecture ensures efficiency in the use of recurrent processing units with high-dimensional inputs. This study also investigates the versatility of the proposed DCRNN model for classification of multi-class time series data from different application domains. Consequently, the proposed DCRNN architecture is evaluated using two time-series datasets: a multichannel scalp EEG dataset for seizure detection, and a machine fault detection dataset obtained in-house. The results suggest that the proposed architecture achieves state-of-the-art performance while utilizing substantially less trainable parameters when compared to comparable methods in the literature.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2015

Accelerated Sensitivity Analysis in High-Dimensional Stochastic Reaction Networks

Existing sensitivity analysis approaches are not able to handle efficiently stochastic reaction networks with a large number of parameters and species, which are typical in the modeling and simulation of complex biochemical phenomena. In this paper, a two-step strategy for parametric sensitivity analysis for such systems is proposed, exploiting advantages and synergies between two recently proposed sensitivity analysis methodologies for stochastic dynamics. The first method performs sensitivity analysis of the stochastic dynamics by means of the Fisher Information Matrix on the underlying distribution of the trajectories; the second method is a reduced-variance, finite-difference, gradient-type sensitivity approach relying on stochastic coupling techniques for variance reduction. Here in this paper we demonstrate that these two methods can be combined and deployed together by means of a new sensitivity bound which incorporates the variance of the quantity of interest as well as the Fisher Information Matrix estimated from the first method. The first step of the proposed strategy labels sensitivities using the bound and screens out the insensitive parameters in a controlled manner. In the second step of the proposed strategy, a finite-difference method is applied only for the sensitivity estimation of the (potentially) sensitive parameters that have not been screened out in the first step. Results on an epidermal growth factor network with fifty parameters and on a protein homeostasis with eighty parameters demonstrate that the proposed strategy is able to quickly discover and discard the insensitive parameters and in the remaining potentially sensitive parameters it accurately estimates the sensitivities. The new sensitivity strategy can be several times faster than current state-of-the-art approaches that test all parameters, especially in “sloppy” systems. In particular, the computational acceleration is quantified by the ratio between the total number of parameters over the number of the sensitive parameters.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

“Understanding Robustness Lottery”: A Geometric Visual Comparative Analysis of Neural Network Pruning Approaches

Deep learning approaches have provided state-of-the-art performance in many applications by relying on large and overparameterized neural networks. However, such networks are very brittle and are difficult to deploy on resource-limited platforms. Model pruning, i.e., reducing the size of the network, is a widely adopted strategy that can lead to a more robust and compact model. Many heuristics exist for model pruning, but our understanding of the pruning process remains limited due to the black-box nature of a neural network model. Empirical studies show that some heuristics improve performance whereas others can make models more brittle. Here, this work aims to shed light on how different pruning methods alter the network’s internal feature representation and the corresponding impact on model performance. To facilitate a comprehensive comparison and characterization of the high-dimensional model feature space, we introduce a visual geometric analysis of feature representations. We evaluated a set of critical geometric concepts decomposed from the commonly adopted classification loss and used them to design a visualization system to compare and highlight the impact of pruning on model performance and feature representation. The proposed tool provides an environment for an in-depth comparison of pruning methods and a comprehensive understanding of how the model responds to common data corruption. By leveraging the proposed visualization, machine learning researchers can reveal the similarities between pruning methods and redundancy in robustness evaluation benchmarks, obtain geometric insights about the differences between pruned models that achieve superior robustness performance, and identify samples that are robust or fragile to model pruning and common data corruption.

Li, Zhimin [Univ. of Utah, Salt Lake City, UT (Uni↗
DOE OSTI2017

Synthesis and analysis of separation networks for the recovery of intracellular chemicals generated from microbial-based conversions

Background. Bioseparations can contribute to more than 70% in the total production cost of a bio-based chemical, and if the desired chemical is localized intracellularly, there can be additional challenges associated with its recovery. Based on the properties of the desired chemical and other components in the stream, there can be multiple feasible options for product recovery. These options are composed of several alternative technologies, performing similar tasks. The suitability of a technology for a particular chemical depends on (1) its performance parameters, such as separation efficiency; (2) cost or amount of added separating agent; (3) properties of the bioreactor effluent (e.g., biomass titer, product content); and (4) final product specifications. Our goal is to first synthesize alternative separation options and then analyze how technology selection affects the overall process economics. To achieve this, we propose an optimization-based framework that helps in identifying the critical technologies and parameters. Results. We study the separation networks for two representative classes of chemicals based on their properties. The separation network is divided into three stages: cell and product isolation (stage I), product concentration (II), and product purification and refining (III). Each stage exploits differences in specific product properties for achieving the desired product quality. The cost contribution analysis for the two cases (intracellular insoluble and intracellular soluble) reveals that stage I is the key cost contributor (>70% of the overall cost). Further analysis suggests that changes in input conditions and technology performance parameters lead to new designs primarily in stage I. Conclusions. The proposed framework provides significant insights for technology selection and assists in making informed decisions regarding technologies that should be used in combination for a given set of stream/product properties and final output specifications. Additionally, the parametric sensitivity provides an opportunity to make crucial design and selection decisions in a comprehensive and rational manner. This will prove valuable in the selection of chemicals to be produced using bioconversions (bioproducts) as well as in creating better bioseparation flow sheets for detailed economic assessment and process implementation on the commercial scale.

09 BIOMASS FUELS↗
DOE OSTI2021

Analysis of mobility data to build contact networks for COVID-19

As social distancing policies and recommendations went into effect in response to COVID-19, people made rapid changes to the places they visit. These changes are clearly seen in mobility data, which records foot traffic using location trackers in cell phones. While mobility data is often used to extract the number of customers that visit a particular business or business type, it is the frequency and duration of concurrent occupancy at those sites that governs transmission. Understanding the way people interact at different locations can help target policies and inform contact tracing and prevention strategies. This paper outlines methods to extract interactions from mobility data and build networks that can be used in epidemiological models. Several measures of interaction are extracted: interactions between people, the cumulative interactions for a single person, and cumulative interactions that occur at particular businesses. Network metrics are computed to identify structural trends which show clear changes based on the timing of stay-at-home orders. Measures of interaction and structural trends in the resulting networks can be used to better understand potential spreading events, the percent of interactions that can be classified as close contacts, and the impact of policy choices to control transmission.

60 APPLIED LIFE SCIENCES↗
NASA NTRS1979

Telecommunications network optimization

Analysis discusses STACOM (state criminal justic communication) network topology program used to design and evaluate digital telecommunications networks STACOM employs ESAU-WILLIAMS technique to search for direct links between system terminations and regional switching center. Inputs include traffic data, terminal locations, and functional requirements.

Lee, J.↗
DOE OSTI2017

Protein Mutation Stability Ternary Classification using Neural Networks and Rigidity Analysis

Discerning how a mutation affects the stability of a protein is central to the study of a wide range of diseases. Machine learning and statistical analysis techniques can inform how to allocate limited resources to the considerable time and cost associated with wet lab mutagenesis experiments. In this work we explore the effectiveness of predicting the change in the stability of a protein due to a mutation using a neural network classfier. Assessing the accuracy of our approach is dependent on the use of experimental data about the effects of mutations performed in vitro. Because the experimental data is prone to discrepancies when similar experiments have been performed by multiple laboratories, the use of the data near the juncture of stabilizing and destabilizing mutations is question- able. We address this later problem via a systematic approach in which we explore the use of a three-way classification scheme with stabilizing, destabilizing, and inconclusive labels. For a systematic search of potential classification cutoff values our classfier achieved 68 percent accuracy on ternary classification for cutoff values of -0.6 and 0.7 with a low rate of classifying stabilizing as destabilizing and vice versa.

Neural Network, Classification, protein stability,↗
DOE OSTI2020

Artificial neural network based isotopic analysis of airborne radioactivity measurement for radiological incident detection

Responders need tools to rapidly detect and identify airborne alpha radioactivity during consequence management scenarios. Traditional continuous air monitor systems used for this purpose compute the net counts in various energy windows to determine the presence of specified isotopes, such as 235U, 239Pu, and 241Am. These calculations rely on having a well-calibrated detector, which is challenging in low-background environments. Here an alternative approach of using artificial neural networks to classify alpha spectra is presented. Two network architectures, fully connected and convolutional networks, were trained to classify alpha spectra into four categories: background and background plus the three isotopes above. Sources were injected into measured background at various fractions of the derived response level (DRL) corresponding to early-phase Protective Action Guides. The convolutional network identifies all sources at 1% of the DRL with average probability of detection of 95% and false alarm probability of 1%. Further, the network identifies sources ranging between 0.25% and 1% of the DRL with higher than 80% probability of detection and lower than 7% false alarm probability. Most significantly, the network performance improves in low-count background conditions, increasing its minimum probability of detection to 93% and reducing the false alarm probabilities to lower than 0.25%. These results show that, once trained on datasets representing a range of detection scenarios, artificial neural networks can accurately identify alpha isotopes of interest without the need for detector calibration.

Woldegiorgis, Surafel F.↗
NASA NTRSDate not supplied

Human Systems Risk Network - A Ranking Analysis of Risks

INTRODUCTION The Human Systems Risk Board (HSRB) is responsible for understanding, managing, and mitigating the risks associated with spaceflight. For a particular mission, the HSRB assigns each human system risk a rating on a 5x5 grid assessing its likelihood and consequence, which is ultimately used to compare and rank the risks. The HSRB approaches risk management by primarily establishing the context of each human system risk individually with the understanding that mitigating one risk might affect the likelihood, consequence, and mitigation approaches of another. To support this effort the HSRB, subject matter experts, and risk custodian teams created directed acyclic graphs (DAG), often called a causal graph, for the twenty-nine risks. In this presentation, we propose a new ranking algorithm for the risks which includes the downstream influence of each risk according to the information in the DAGs and provide an application of graph theoretic tools. METHODS In 2014, Mindock and Klaus proposed a taxonomy for human system risk influences which we have adopted to categorize the nodes in each DAG. Analyzing the nodes that correspond to the risks in this taxonomy allows us to analyze and understand how each risk influences the others. We construct an auxiliary network, which we call the Primary Risk Network (PRN), where the nodes are the twenty-nine space flight risks and, a directed edge connects Risk A to Risk B if Risk A has some influence on the likelihood or consequence of Risk B as described in the DAGS. We perform a variety of graph theoretic ranking methods on the nodes (or risks) in the PRN, including Katz centrality. RESULTS We rank the nodes in the PRN using the Katz centrality score. The ten risks with the highest score are pictured in Figure 1, colored (light to dark) according to their score. We analyze other centrality measures like betweenness centrality, eigenvector centrality, and the Estrada index, and provide the meaning of the corresponding rankings in terms of the risks. Future work includes analyzing the other categories in the taxonomy defined by Mindock and Klaus [1]. For example, we are interested in analyzing the nodes that are labeled as countermeasures or capabilities and perform similar analysis to measure their effect on certain medical conditions.

dag↗
DOE OSTI2021

Utilizing a Dynamic Segmentation Convolutional Neural Network for Microstructure Analysis of Additively Manufactured Superalloy 718

Additive manufacturing (AM) is revolutionizing almost all industries through the production of intricate geometries previously prohibited by cost or machinability. Ni-based superalloys form a primary alloy class for high temperature applications in the petrochemical, aerospace, and nuclear industries because of their intrinsic resistance to creep and the ability to heat treat the superalloy for high strength. Despite these attractive properties, the extreme work hardening of Ni-based superalloys makes traditional manufacturing of complex shapes difficult and thus, these alloys are an attractive target for AM. Superalloy 718 was chosen as an example superalloy because of the wide variety of precipitates that can form within its composition space from the repetitive heating and cooling cycles of the AM process. The precipitates and other microstructure features, such as grain boundaries and dislocations, will dictate the mechanical properties and thus, there is an extensive challenge to characterize the size, number density, composition, and volume fraction of each microstructural feature from AM fabrication using analytical electron microscopy. This work focused on the application of a pixel-wise classification machine learning (ML) model called a dynamic segmentation convolutional neural network (DSCNN) to identify the microstructural features of an as-fabricated additively manufactured superalloy 718.

Taller, Stephen↗