SEARCH · Engineering Papers

Results for “knowledge representation”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 883 records · Page 49

All sources NASA NTRS DOE OSTI

DOE OSTI2024

Comparison of Machine Learning Approaches for Prediction of the Equivalent Alkane Carbon Number for Microemulsions Based on Molecular Properties

The chemical properties of oils are vital in the design of microemulsion systems. The hydrophilic–lipophilic difference equation used to predict microemulsions’ phase behavior expresses the oils’ physiochemical properties as the equivalent alkane carbon number (EACN). The experimental determination of EACN requires knowledge of the temperature dependence of the microemulsion system and the effects of different surfactant concentrations. Thus, the experimental determination is time-intensive and tedious, requiring days to months for proper separations. Furthermore, the experiments require high purity of chemicals because microemulsions are sensitive to impurities. Our work focuses on the quick and reliable predictions of the EACN with machine learning (ML) models. Due to the immaturity of ML chemical predictions, we compare three graph neural networks (GNNs) and a gradient-boosted tree algorithm, known as XGBoost. The GNNs use the molecular structures represented as simplified molecular-input line-entry system (SMILES) codes for the initial input, which allows us to assess whether geometry optimization is necessary for reliable results. The XGBoost model also begins with the SMILES representations of the molecules but uses molecular descriptors instead of geometry optimizations. As a result, the best model tested (crystal graph convolutional neural network with Merck molecular force field-94) has an error of 1.15 EACN units of the true EACN for unknown data with the errors skewed toward zero and an R² score of 0.9

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

NASA NTRS2019

Automatic Incorporation of Flaws into Cad and FEA Models Based On Measured NDE Data

Automatic incorporation of non-destructive evaluation (NDE) data into CAD and finite element (FE) models for structural analysis is essential for accurate residual-life and strength analyses of composite structures. Currently, two significant but independent challenges exist for incorporating NDE data in such analyses: interpretation of the NDE signals to locate and characterize the structural defects and the incorporation of that information into structural analysis tools. The first challenge will be addressed through linear inversion of calculated synthetic responses generated by a simulation engine to evaluate flaw size, location, and other characteristics. The second challenge will be addressed by building on prior and parallel defect modeling efforts to integrate suitable flaw models into the CAD/finite-element representation of a specimen. Ultrasonic NDE information is usually represented using C-Scans of time-of-flight or intensity information. Other modalities such as thermography and computed tomography (CT) could also be used to characterize the structure. This information will first be projected into material coordinates and then aligned with the specimen’s finite element mesh. The NDE data can then be incorporated by inserting flaw geometry into the CAD and FE analysis model and automatic modification of local material properties and boundary conditions. This process will enable the use of structural analysis tools even for practical modalities such as ultrasound and thermography which may not directly give accurate-enough geometric information to fully define the flaw. The full flaw characteristics will be inferred using prior expert and modeling knowledge to categorize flaw signatures. Outputs for this project will include demonstrated methods for creating structural analysis geometric models that incorporate flaw characterization information; these models can then be used directly in popular commercial FEA software. In addition, the primary deliverable for the project will include a software that will be able to automatically incorporate different kinds of flaws into an existing FE models. The models will be generated in standard file formats to enable direct import into the analysis software.

Adarsh Krishnamurthy↗

DOE OSTI2022

Projecting the Thermal Response in a HTGR-Type System during Conduction Cooldown Using Graph-Laplacian Based Machine Learning

Accurate prediction of an off-normal event in a nuclear reactor is dependent upon the availability of sensory data, reactor core physical condition, and understanding of the underlying phenomenon. This work presents a method to project the data from some discrete sensory locations to the overall reactor domain during conduction cooldown scenarios similar to High Temperature Gas-cooled Reactors (HTGRs). The existing models for conductive cooldown in a heterogeneous multi-body system, such as an assembly of prismatic blocks or pebble beds relies on knowledge of the thermal contact conductance, requiring significant knowledge of local thermal contacts and heat transport possibilities across those contacts. With a priori knowledge of bulk geometry features and some discrete sensors, a machine learning approach was devised. The presented work uses an experimental facility to mimic conduction cooldown with an assembly of 68 cylindrical rods initially heated to 1200 K. High-fidelity temperature data were collected using an infrared (IR) camera to provide training data to the model and validate the predicted temperature data. The machine learning approach used here first converts the macroscopic bulk geometry information into Graph-Laplacian, and then uses the eigenvectors of the Graph-Laplacian to develop Kernel functions. Support vector regression (SVR) was implemented on the obtained Kernels and used to predict the thermal response in a packed rod assembly during a conduction cooldown experiment. The usage of SVR modeling differs from most models today because of its representation of thermal coupling between rods in the core. When trained with thermographic data, the average normalized error is less than 2% over 400 s, during which temperatures of the assembly have dropped by more than 500 K. The rod temperature prediction performance was significantly better for rods in the interior of the assembly compared to those near the exterior, likely due to the model simplification of the surroundings.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

NASA NTRS1993

Designers' models of the human-computer interface

Understanding design models of the human-computer interface (HCI) may produce two types of benefits. First, interface development often requires input from two different types of experts: human factors specialists and software developers. Given the differences in their backgrounds and roles, human factors specialists and software developers may have different cognitive models of the HCI. Yet, they have to communicate about the interface as part of the design process. If they have different models, their interactions are likely to involve a certain amount of miscommunication. Second, the design process in general is likely to be guided by designers' cognitive models of the HCI, as well as by their knowledge of the user, tasks, and system. Designers do not start with a blank slate; rather they begin with a general model of the object they are designing. The author's approach to a design model of the HCI was to have three groups make judgments of categorical similarity about the components of an interface: human factors specialists with HCI design experience, software developers with HCI design experience, and a baseline group of computer users with no experience in HCI design. The components of the user interface included both display components such as windows, text, and graphics, and user interaction concepts, such as command language, editing, and help. The judgments of the three groups were analyzed using hierarchical cluster analysis and Pathfinder. These methods indicated, respectively, how the groups categorized the concepts, and network representations of the concepts for each group. The Pathfinder analysis provides greater information about local, pairwise relations among concepts, whereas the cluster analysis shows global, categorical relations to a greater extent.

Gillan, Douglas J.↗

DOE OSTI2021

An efficient reconstruction algorithm for diffusion on triangular grids using the nodal discontinuous Galerkin method

High-energy-density (HED) hydrodynamics studies such as those relevant to inertial confinement fusion and astrophysics require highly disparate densities, temperatures, viscosities, and other diffusion parameters over relatively short spatial scales. This presents a challenge for high-order accurate methods to effectively resolve the hydrodynamics at these scales, particularly in the presence of highly disparate diffusion. A significant volume of engineering and physics applications use an unstructured discontinuous Galerkin (DG) method developed based on the finite element mesh generation and algorithmic framework. This work discusses the application of an affine reconstructed nodal DG method for unstructured grids of triangles. Solving the diffusion terms in the DG method is non-trivial due to the solution representations being piecewise continuous. Hence, the diffusive flux is not defined on the interface of elements. The proposed numerical approach reconstructs a smooth solution in a parallelogram that is enclosed by the quadrilateral formed by two adjacent triangle elements. The interface between these two triangles is the diagonal of the enclosed parallelogram. Similar to triangles, the mapping of parallelograms from a physical domain to a reference domain is an affine mapping, which is necessary for an accurate and efficient implementation of the numerical algorithm. Thus, all computations can still be performed on the reference domain, which promotes efficiency in computation and storage. This reconstruction does not make assumptions on choice of polynomial basis. Reconstructed DG algorithms have previously been developed for modal implementations of the convection–diffusion equations. However, to the best of the authors’ knowledge, this is the first practical guideline that has been proposed for applying the reconstructed algorithm on a nodal discontinuous Galerkin method with a focus on accuracy and efficiency. As a result, the algorithm is demonstrated on a number of benchmark cases as well as a challenging substantive problem in HED hydrodynamics with highly disparate diffusion parameters.

Computational efficiency↗

DOE OSTI2022

Uncertainties in Local Intense Precipitation Flood Modeling

In the Task 1 report for the U.S. Nuclear Regulatory Commission (NRC) funded Local Intense Precipitation (LIP) Probabilistic Flood Hazard Assessment (PFHA) Pilot Study, the Pacific Northwest National Laboratory (PNNL) reviewed hydrologic and hydraulic modeling approaches for estimation of LIP flooding and their implementation in readily available simulation software packages (Prasad and Yuan 2020). NRC’s interest is in developing guidance for performing LIP PFHA at nuclear power plant (NPP) sites. This report describes the work done under Task 2 of the study that focused on describing uncertainties associated with LIP flood modeling. Two main types of uncertainties exist in LIP flood modeling: (1) aleatory uncertainties that arise from the inherent natural variability of the hydrometeorological system and (2) epistemic uncertainties that arise from the analysts’ incomplete knowledge of the hydrometeorological and site configuration. Principal sources for aleatory variability in LIP flood simulations at NPP sites include precipitation (i.e., magnitude, duration, and temporal distribution of LIP events), initial conditions (e.g., soil moisture content, stormwater drainage discharge, surface storage), boundary conditions (e.g., upstream discharge, downstream water surface elevations), and long-term temporal trends (e.g., climate change). This report describes approaches used for estimating the aleatory variability in precipitation including precipitation-frequency analysis, numerical weather prediction, and stochastic weather generation. Data sources that can be used to estimate aleatory variability in initial and boundary conditions are also described. Approaches to include effects of long-term trends like those from climate change into estimation of aleatory variability are summarized. Sources of epistemic uncertainty in LIP flood simulations at NPP sites include process representation (e.g., multiple approaches to represent runoff generation, stormwater drainage, and hydraulic routing), site configurations (e.g., site layout, flow features, status of temporary flood protection, blockage of drains), model resolution, and long-term temporal trends (e.g., known/planned site alterations, land-use changes at and in the vicinity of the site). The report describes alternative process representations (methods and models), particularly those implemented in the LIP flood simulation software packages reviewed in Task 1 report, and lists the associated model parameters. Approaches for estimating model parameters when surface and subsurface water exchanges occur are also described.

42 ENGINEERING↗

NASA NTRSDate not supplied

Merging Analytic Collaborative Frameworks with New Observing Strategies Toward a Digital Twin: Earth – Episodic Pulse Event Impacts on Ocean Carbon Cycle as an Example

Virtual representations of the Earth will allow us to address some of the most critical environmental issues of our time. Here, we show the first steps toward representation of riverine, estuarine, and coastal carbon processes to enable scenario driven “what-if” analyses of the carbon system and human footprint. Excess sediment and nutrient runoff from land-based human activities impact water quality and can pose serious threats to coastal and marine ecosystems. Episodic pulse events, such as extreme precipitation events, can increase the amount of nutrients entering estuaries and coastal regions, potentially leading to large phytoplankton blooms followed by anoxic conditions. Consequences of coastal runoff are predicted to increase with the higher intensity and frequency of extreme events. Beyond the threat to coastal ecosystems, recent findings suggest these episodic pulses might play a significant role for biological production influencing regional and global carbon fluxes and budgets. An improved understanding of these events through optimal, dynamic observing strategies will increase our knowledge of the land-ocean continuum and how regional events and nutrient fluxes affect the carbon cycle and ocean ecosystem. This conceptual framework enables focused science investigations by pairing data analytics and artificial intelligence tools (otherwise termed an Analytic Center Framework, ACF) with targeted measurement acquisition through distributed sensing and intelligent asset tasking (or New Observing Strategies, NOS). This NOS and ACF iterative approach acquires and integrates complementary and coincident satellite, in-situ and model data to build a more complete and in-depth picture of science phenomena. Specifically, Apache Science Data Analytic Platform (SDAP) is extended to incorporate relevant datasets for data access, harmonized analysis, and anomaly detection. When conditions are met for a likely pulse event, NASA’s D-SHIELD (Distributed Spacecraft with Heuristic Intelligence to Enable Logistical Decisions) tool is triggered to optimize asset overpass frequency and schedule observations for persistent monitoring. Targeted data is ingested by SDAP for enhanced investigation via iterative analysis until the trigger criteria is no longer met - steps toward a digital twin.

Laura Rogers↗

NASA NTRS2015

Convective Enhancement of Icing Roughness Elements in Stagnation Region Flows

To improve existing ice accretion simulation codes, more data regarding ice roughness and its effects on convective heat transfer are required. To build on existing research on this topic, this study used the Vertical Icing Studies Tunnel (VIST) at NASA Glenn Research to model realistic ice roughness in the stagnation region of a NACA 0012 airfoil. Using the VIST, a test plate representing the leading 2% chord of the airfoil was subjected to flows of 7.62 m/s (25 ft/s), 12.19 m/s (40 ft/s), and 16.76 m/s (55 ft/s). The test plate was fitted with 3 surfaces, each with a different representation of ice roughness: 1) a control surface with no ice roughness, 2) a surface with ice roughness with element height scaled by 10x and streamwise rough zone width from the stagnation point scaled by 10x, and 3) a surface with ice roughness with element height scaled by 10x and streamwise rough zone width from the stagnation point scaled by 25x. Temperature data from the tests were recorded using an infrared camera and thermocouples imbedded in the test plate. From the temperature data, a convective heat transfer coefficient map was created for each case. Additional testing was also performed to validate the VIST's flow quality. These tests included five-hole probe and hot-wire probe velocity traces to provide flow visualization and to study boundary layer formation on the various test surfaces. The knowledge gained during the experiments will help improve ice accretion codes by providing heat transfer coefficient validation data and by providing flow visualization data helping understand current and future experiments performed in the VIST.

Icing↗

NASA NTRS1992

Inverse problems: Fuzzy representation of uncertainty generates a regularization

In many applied problems (geophysics, medicine, and astronomy) we cannot directly measure the values x(t) of the desired physical quantity x in different moments of time, so we measure some related quantity y(t), and then we try to reconstruct the desired values x(t). This problem is often ill-posed in the sense that two essentially different functions x(t) are consistent with the same measurement results. So, in order to get a reasonable reconstruction, we must have some additional prior information about the desired function x(t). Methods that use this information to choose x(t) from the set of all possible solutions are called regularization methods. In some cases, we know the statistical characteristics both of x(t) and of the measurement errors, so we can apply statistical filtering methods (well-developed since the invention of a Wiener filter). In some situations, we know the properties of the desired process, e.g., we know that the derivative of x(t) is limited by some number delta, etc. In this case, we can apply standard regularization techniques (e.g., Tikhonov's regularization). In many cases, however, we have only uncertain knowledge about the values of x(t), about the rate with which the values of x(t) can change, and about the measurement errors. In these cases, usually one of the existing regularization methods is applied. There exist several heuristics that choose such a method. The problem with these heuristics is that they often lead to choosing different methods, and these methods lead to different functions x(t). Therefore, the results x(t) of applying these heuristic methods are often unreliable. We show that if we use fuzzy logic to describe this uncertainty, then we automatically arrive at a unique regularization method, whose parameters are uniquely determined by the experts knowledge. Although we start with the fuzzy description, but the resulting regularization turns out to be quite crisp.

Kreinovich, V.↗

NASA NTRS2014

Temperature Knowledge and Model Correlation for the Soil Moisture Active and Passive (SMAP) Reflector Mesh

NASA's Earth observing Soil Moisture Active & Passive (SMAP) Mission is scheduled to launch in November 2014 into a 685 km near-polar, sun synchronous orbit. SMAP will provide comprehensive global mapping measurements of soil moisture and freeze/thaw state in order to enhance understanding of the processes that link the water, energy, and carbon cycles. The primary objectives of SMAP are to improve worldwide weather and flood forecasting, enhance climate prediction, and refine drought and agriculture monitoring during its 3 year mission. The SMAP instrument architecture incorporates an L-band radar and an L-band radiometer which share a common feed horn and parabolic mesh reflector. The instrument rotates about the nadir axis at approximately 15 rpm, thereby providing a conically scanning wide swath antenna beam that is capable of achieving global coverage within 3 days. In order to make the necessary precise surface emission measurements from space, a temperature knowledge of 60 deg C for the mesh reflector is required. In order to show compliance, a thermal vacuum test was conducted using a portable solar simulator to illuminate a non flight, but flight-like test article through the quartz window of the vacuum chamber. The molybdenum wire of the antenna mesh is too fine to accommodate thermal sensors for direct temperature measurements. Instead, the mesh temperature was inferred from resistance measurements made during the test. The test article was rotated to five separate angles between 10 deg and 90 deg via chamber breaks to simulate the maximum expected on-orbit solar loading during the mission. The resistance measurements were converted to temperature via a resistance versus temperature calibration plot that was constructed from data collected in a separate calibration test. A simple thermal model of two different representations of the mesh (plate and torus) was created to correlate the mesh temperature predictions to within 60 deg C. The on-orbit mesh temperature will be predicted using the correlated analytical thermal model since direct measurements from in-situ flight thermal sensors are not possible.

Mikhaylov, Rebecca↗

DOE OSTI2021

Machine learning of Kondo physics using variational autoencoders and symbolic regression

In this work, we employ variational autoencoders to extract physical insight from a dataset of one-particle Anderson impurity model spectral functions. Autoencoders are trained to find a low-dimensional, latent space representation that faithfully characterizes each element of the training set, as measured by a reconstruction error. Variational autoencoders, a probabilistic generalization of standard autoencoders, further condition the learned latent space to promote highly interpretable features. In our study, we find that the learned latent variables strongly correlate with well known, but nontrivial, parameters that characterize emergent behaviors in the Anderson impurity model. In particular, one latent variable correlates with particle-hole asymmetry, while another is in near one-to-one correspondence with the Kondo temperature, a dynamically generated low-energy scale in the impurity model. Using symbolic regression, we model this variable as a function of the known bare physical input parameters and "rediscover" the nonperturbative formula for the Kondo temperature. The machine learning pipeline we develop suggests a general purpose approach, which opens opportunities to discover new domain knowledge in other physical systems.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

DOE OSTI2023

Potential bioavailability of representative pyrogenic organic matter compounds in comparison to natural dissolved organic matter pools

Abstract. Pyrogenic organic matter (PyOM) from wildfires impacts river corridors globally and is widely regarded as resistant to biological degradation. Though recent work suggests PyOM may be more bioavailable than historically perceived, estimating bioavailability across its chemical spectrum remains elusive. To address this knowledge gap, we assessed potential bioavailability of representative PyOM compounds relative to ubiquitous dissolved organic matter (DOM) with a substrate-explicit model. The range of potential bioavailability of PyOM was greater than natural DOM; however, the predicted thermodynamics, metabolic rates, and carbon use efficiencies (CUEs) overlapped significantly between all OM pools. Compound type (e.g., natural versus PyOM) had approximately 6-fold less impact on predicted respiration rates than simulated carbon and oxygen limitations. Within PyOM, the metabolism of specific chemistries differed strongly between unlimited and oxygen-limited conditions – degradations of anhydrosugars, phenols, and polycyclic aromatic hydrocarbons (PAHs) were more favorable under oxygen limitation than other molecules. Notably, amino sugar-like, protein-like, and lignin-like PyOM had lower carbon use efficiencies relative to natural DOM of the same classes, indicating potential impacts in process-based model representations. Overall, our work illustrates how similar PyOM bioavailability may be to that of natural DOM in the river corridor, furthering our understanding of how PyOM may influence riverine biogeochemical cycling.

59 BASIC BIOLOGICAL SCIENCES↗

NASA NTRS2007

Intelligent Entity Behavior Within Synthetic Environments

This paper describes some elements in the development of realistic performance and behavior in the synthetic entities (players) which support Modeling and Simulation (M&S) applications, particularly military training. Modern human-in-the-loop (virtual) training systems incorporate sophisticated synthetic environments, which provide: 1. The operational environment, including, for example, terrain databases; 2. Physical entity parameters which define performance in engineered systems, such as aircraft aerodynamics; 3. Platform/system characteristics such as acoustic, IR and radar signatures; 4. Behavioral entity parameters which define interactive performance, including knowledge/reasoning about terrain, tactics; and, 5. Doctrine, which combines knowledge and tactics into behavior rule sets. The resolution and fidelity of these model/database elements can vary substantially, but as synthetic environments are designed to be compose able, attributes may easily be added (e.g., adding a new radar to an aircraft) or enhanced (e.g. Amending or replacing missile seeker head/ Electronic Counter Measures (ECM) models to improve the realism of their interaction). To a human in the loop with synthetic entities, their observed veridicality is assessed via engagement responses (e.g. effect of countermeasures upon a closing missile), as seen on systems displays, and visual (image) behavior. The realism of visual models in a simulation (level of detail as well as motion fidelity) remains a challenge in realistic articulation of elements such as vehicle antennae and turrets, or, with human figures; posture, joint articulation, response to uneven ground. Currently the adequacy of visual representation is more dependant upon the quality and resolution of the physical models driving those entities than graphics processing power per Se. Synthetic entities in M&S applications traditionally have represented engineered systems (e.g. aircraft) with human-in-the-loop performance characteristics (e.g. visual acuity) included in the system behavioral specification. As well, performance affecting human parameters such as experience level, fatigue and stress are coming into wider use (via AI approaches) to incorporate more uncertainty as to response type as well as performance (e.g. Where an opposing entity might go and what it might do, as well as how well it might perform).

Kruk, R. V.↗

DOE OSTI2024

A Mass‐Conserving‐Perceptron for Machine‐Learning‐Based Modeling of Geoscientific Systems

Although decades of effort have been devoted to building Physical-Conceptual (PC) models for predicting the time-series evolution of geoscientific systems, recent work shows that Machine Learning (ML) based Gated Recurrent Neural Network technology can be used to develop models that are much more accurate. However, the difficulty of extracting physical understanding from ML-based models complicates their utility for enhancing scientific knowledge regarding system structure and function. Here, we propose a physically interpretable Mass-Conserving-Perceptron (MCP) as a way to bridge the gap between PC-based and ML-based modeling approaches. The MCP exploits the inherent isomorphism between the directed graph structures underlying both PC models and GRNNs to explicitly represent the mass-conserving nature of physical processes while enabling the functional nature of such processes to be directly learned (in an interpretable manner) from available data using off-the-shelf ML technology. As a proof of concept, we investigate the functional expressivity (capacity) of the MCP, explore its ability to parsimoniously represent the rainfall-runoff (RR) dynamics of the Leaf River Basin, and demonstrate its utility for scientific hypothesis testing. To conclude, we discuss extensions of the concept to enable ML-based physical-conceptual representation of the coupled nature of mass-energy-information flows through geoscientific systems.

58 GEOSCIENCES↗

NASA NTRSDate not supplied

Identifying Chemical and Optical Smoke Signatures in Suborbital Observations to Improve Model Projections

Biomass Burning (BB) fires are expected to occur more frequently in the future due to climate change, the build‐up of fuels due to fire suppression, and the expansion of the wildland‐urban interface [Schoennagel et al., 2017; Shivdenko & Schepaschenko, 2013; Stevens et al., 2014; Turner et al., 2019; Yue et al., 2015; Tilman et al., 2001]. Accurate knowledge of the abundance, distribution, and optical properties of absorbing aerosols such as BB smoke is essential to understanding regional and global aerosol radiative effects [Boucher et al., 2013]. It is recognized that differing fuel types (e.g., rice straw, ponderosa pine), combustion modes (i.e., flaming, smoldering or both) and temperature/ moisture content produce smoke with differing optical characteristics (e.g., aerosol light scattering and absorption) and component mixing ratios [e.g., BC, OC and inorganic species such as sulfate, potassium, chloride; Streets et al. 2003; Reid et al., 2005a, b; Janhäll et al., 2010; Sayer et al., 2014]. However, climate models often ascribe a fixed set of emission factors [i.e., amount of particulate for a given species released in the atmosphere per unit dry fuel burned; Darmenov and Dasilva, 2013] and optical properties [Hess et al., 1998] to all BB aerosol plumes, inherently limiting accurate representation of the global and regional heterogeneity of BB. The overarching goal of the proposed work is to improve the microphysical and optical representation of different types of Biomass Burning in Chemistry Transport Models. Our strategy is to capitalize on multiple airborne campaigns (e.g., ARCTAS, SEAC4RS, ORACLES, FIREX-AQ), which sampled different types of BB plumes at different aging stages. Models, when improved and combined with satellite, airborne, and ground observations will help address many aerosol science objectives from the NASA Decadal survey [National Academies of Sciences, Engineering, and Medicine. 2018].

Meloe Shenandoah F Kacenelenbogen↗

DOE OSTI2025

Bimodal Visualization of Industrial X-Ray and Neutron Computed Tomography Data

Advanced manufacturing creates increasingly complex objects with material compositions that are often difficult to characterize by a single modality. Our collaborating domain scientists are going beyond traditional methods by employing both X-ray and neutron computed tomography to obtain complementary representations expected to better resolve material boundaries. However, the use of two modalities creates its own challenges for visualization, requiring either complex adjustments of bimodal transfer functions or the need for multiple views. Together with experts in nondestructive evaluation, we designed a novel interactive bimodal visualization approach to create a combined view of the co-registered X-ray and neutron acquisitions of industrial objects. Using an automatic topological segmentation of the bivariate histogram of X-ray and neutron values as a starting point, the system provides a simple yet effective interface to easily create, explore, and adjust a bimodal visualization. Here, we propose a widget with simple brushing interactions that enables the user to quickly correct the segmented histogram results. Our semiautomated system enables domain experts to intuitively explore large bimodal datasets without the need for either advanced segmentation algorithms or knowledge of visualization techniques. We demonstrate our approach using synthetic examples, industrial phantom objects created to stress bimodal scanning techniques, and real-world objects, and we discuss expert feedback.

image segmentation↗

NASA NTRS2014

Overview of the Contributing Factor Map and How it can Support your Research

The Contributing Factor Map (CFM) is a visual representation of a taxonomy of factors influencing human health and performance in space. This presentation will give an overview of its development and its structure. It will describe various uses of the CFM that can support researchers working within the Human Research Program (HRP) Architecture of Evidence-Risk-Gap-Task-Deliverable. For example, during the Risk phase, the CFM can be used as a "menu" to help formulate a qualitative model of the factors contributing to specific consequences of concern. It provides a reference set of factors from across the operational, vehicle design, and human domains that otherwise might not be considered if approaching a risk from a specific domain perspective. Using the CFM as a reference can increase awareness of potential cross-disciplinary collaborations for overall risk mitigation. The CFM can also be used as a framework for identifying gaps in knowledge about a risk. This identification can support the subsequent development of gaps and tasks comprising the research plan aimed at risk mitigation. Examples of these types of applications of the CFM will be discussed and information on the support available to researchers in using it will be provided.

Mindock, Jennifer A.↗

NASA NTRS1997

Knowledge-Based Aircraft Automation: Managers Guide on the use of Artificial Intelligence for Aircraft Automation and Verification and Validation Approach for a Neural-Based Flight Controller

The ultimate goal of this report was to integrate the powerful tools of artificial intelligence into the traditional process of software development. To maintain the US aerospace competitive advantage, traditional aerospace and software engineers need to more easily incorporate the technology of artificial intelligence into the advanced aerospace systems being designed today. The future goal was to transition artificial intelligence from an emerging technology to a standard technology that is considered early in the life cycle process to develop state-of-the-art aircraft automation systems. This report addressed the future goal in two ways. First, it provided a matrix that identified typical aircraft automation applications conducive to various artificial intelligence methods. The purpose of this matrix was to provide top-level guidance to managers contemplating the possible use of artificial intelligence in the development of aircraft automation. Second, the report provided a methodology to formally evaluate neural networks as part of the traditional process of software development. The matrix was developed by organizing the discipline of artificial intelligence into the following six methods: logical, object representation-based, distributed, uncertainty management, temporal and neurocomputing. Next, a study of existing aircraft automation applications that have been conducive to artificial intelligence implementation resulted in the following five categories: pilot-vehicle interface, system status and diagnosis, situation assessment, automatic flight planning, and aircraft flight control. The resulting matrix provided management guidance to understand artificial intelligence as it applied to aircraft automation. The approach taken to develop a methodology to formally evaluate neural networks as part of the software engineering life cycle was to start with the existing software quality assurance standards and to change these standards to include neural network development. The changes were to include evaluation tools that can be applied to neural networks at each phase of the software engineering life cycle. The result was a formal evaluation approach to increase the product quality of systems that use neural networks for their implementation.

Broderick, Ron↗