Engineering Papers⌕ Search

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.

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
DOE OSTI2024

Metagenomic clustering links specific metabolic functions to globally relevant ecosystems

ABSTRACT Metagenomic sequencing has advanced our understanding of biogeochemical processes by providing an unprecedented view into the microbial composition of different ecosystems. While the amount of metagenomic data has grown rapidly, simple-to-use methods to analyze and compare across studies have lagged behind. Thus, tools expressing the metabolic traits of a community are needed to broaden the utility of existing data. Gene abundance profiles are a relatively low-dimensional embedding of a metagenome’s functional potential and are, thus, tractable for comparison across many samples. Here, we compare the abundance of KEGG Ortholog Groups (KOs) from 6,539 metagenomes from the Joint Genome Institute’s Integrated Microbial Genomes and Metagenomes (JGI IMG/M) database. We find that samples cluster into terrestrial, aquatic, and anaerobic ecosystems with marker KOs reflecting adaptations to these environments. For instance, functional clusters were differentiated by the metabolism of antibiotics, photosynthesis, methanogenesis, and surprisingly GC content. Using this functional gene approach, we reveal the broad-scale patterns shaping microbial communities and demonstrate the utility of ortholog abundance profiles for representing a rapidly expanding body of metagenomic data. IMPORTANCE Metagenomics, or the sequencing of DNA from complex microbiomes, provides a view into the microbial composition of different environments. Metagenome databases were created to compile sequencing data across studies, but it remains challenging to compare and gain insight from these large data sets. Consequently, there is a need to develop accessible approaches to extract knowledge across metagenomes. The abundance of different orthologs (i.e., genes that perform a similar function across species) provides a simplified representation of a metagenome’s metabolic potential that can easily be compared with others. In this study, we cluster the ortholog abundance profiles of thousands of metagenomes from diverse environments and uncover the traits that distinguish them. This work provides a simple to use framework for functional comparison and advances our understanding of how the environment shapes microbial communities.

54 ENVIRONMENTAL SCIENCES↗
NASA NTRS1995

An architecture and protocol for communications satellite constellations regarded as multi-agent systems

This paper presents an architecture for satellites regarded as intercommunicating agents. The architecture is based upon a postmodern paradigm of artificial intelligence in which represented knowledge is regarded as text, inference procedures are regarded as social discourse and decision making conventions and the semantics of representations are grounded in the situated behaviour and activity of agents. A particular protocol is described for agent participation in distributed search and retrieval operations conducted as joint activities.

Lindley, Craig A.↗
NASA NTRS1999

Unstructed Navier-Stokes Analysis of Wind-Tunnel Aeroelastic Effects on TCA Model 2

The aim of this work is to demonstrate a simple technique which accounts for aeroelastic deformations experienced by HSR wind-tunnel models within CFD computations. With improved correlations, CFD can become a more effective tool for augmenting the post-test understanding of experimental data. The present technique involves the loose coupling of a low-level structural representation within the ELAPS code, to an unstructured Navier-Stokes flow solver, USM3Dns. The ELAPS model is initially calibrated against bending characteristics of the wind-tunnel model. The strength of this method is that, with a single point calibration of a simple structural representation, the static aeroelastic effects can be accounted for in CFD calculations across a range of test conditions. No prior knowledge of the model deformation during the wind-on test is required. This approach has been successfully applied to the high aspect-ratio planforms of subsonic transports. The current challenge is to adapt the procedure to low aspect-ratio planforms typical of HSR configurations.

Frink, Neal T.↗
NASA NTRS2001

Modeling Nitrogen Oxides in the Lower Stratosphere

This talk will focus on the status of current understanding (not a historical review) as regards modeling nitrogen oxides (NOy) in the lower stratosphere (LS). The presentation will be organized around three major areas of process understanding: 1) NOy sources, sinks, and transport to the LS, 2) NOy species partitioning, and 3) polar multiphase processes. In each area, process topics will be identified with an estimate of the degree of confidence associated with their representation in numerical models. Several exotic and/or speculative processes will also be discussed. Those topics associated with low confidence or knowledge gaps, weighted by their prospective importance in stratospheric chemical modeling, will be collected into recommendations for further study. Suggested approaches to further study will be presented for discussion.

Kawa, S. Randy↗
NASA NTRS2004

Formal Consistency Verification of Deliberative Agents with Respect to Communication Protocols

The aim of this paper is to show a method that is able to detect inconsistencies in the reasoning carried out by a deliberative agent. The agent is supposed to be provided with a hybrid Knowledge Base expressed in a language called CCR-2, based on production rules and hierarchies of frames, which permits the representation of non-monotonic reasoning, uncertain reasoning and arithmetic constraints in the rules. The method can give a specification of the scenarios in which the agent would deduce an inconsistency. We define a scenario to be a description of the initial agent s state (in the agent life cycle), a deductive tree of rule firings, and a partially ordered set of messages and/or stimuli that the agent must receive from other agents and/or the environment. Moreover, the method will make sure that the scenarios will be valid w.r.t. the communication protocols in which the agent is involved.

Ramirez, Jaime↗
NASA NTRS1998

Some Formal Aspects of Human-Machine Interaction

While automated control systems such as autopilots and medical devices are introduced at a rapid pace, it is widely recognized that user interaction with these machines is problematic (Abbott, Slotte, & Stimson, 1996). One factor commonly cited in the literature is the discrepancy between the machine's behavior and the user's expectations. Design guidelines to reduce this discrepancy focus on two elements: (1) improvement of the "feedback" about what the automation is actually doing, and (2) improvement of the user's "mental model" of the automation (Norman, 1990; Sarter and Woods, 1995). This presentation describes a methodology for investigating these two elements via a formal (Le., mathematical) approach. The method involves two representations: (1) a finite state model of the machine's behavior (2) a finite state model of the user's knowledge and expectations about the machine's behavior. In the analysis phase we compare these two models and identify discrepancies. Such discrepancies can be compensated by augmenting the display and/or the user's model. A taxonomy of these discrepancies will be discussed using examples from automated Eight control systems of modern "glass cockpit" jetliners.

Degani, Asaf↗
DOE OSTI2020

Evaporation Submodel Development for Volume of Fluid (eVOF) Method Applicable to Spray-Wall Interaction Including Film Characteristics with Validation at High Pressure and Temperature Conditions

Internal combustion engines have seen a great evolution over the last several decades through application of high pressure direct injection, multiple injections, and other technologies to reduced fuel consumption, NOx, and PM. Although combustion systems with advanced injection strategies have been studied extensively, there exists a significant fundamental knowledge gap on the fuel-spray interactions with the piston surface and chamber walls. Advanced computational codes validated with experimental techniques have to be developed for accurate representation of the drop impingement, fuel film formation, and vaporization. Current engine CFD (Computational Fluid Dynamics) spray models utilize a Lagrangian framework for modeling which lacks critical considerations of the physics pertaining to these interactions and thus requiring extensive parameterization, tuning and validation. The team from Michigan Technological University, University of Massachusetts Dartmouth, and Argonne National Laboratory is composed of experts in sprays, combustion, engines and CFD with a wide spectrum of knowledge including specific expertise in the area under consideration. In the proposed work, a VOF (Volume of Fluid) modeling approach has been adopted for the spray-wall interaction, film formation and spreading, and vaporization. With the inclusion of a vaporization submodel, a more predictive and accurate simulation of the spray-film was performed without extensive need of parameterization and tuning. Extensive experimentation of the spray-wall interaction under the range of conditions matching the thermodynamic and surface temperatures that occur in diesel and gasoline engines were conducted to validate the SWI submodels and for development of the evaporation submodel, which has been implemented in the flow solver.

42 ENGINEERING↗
DOE OSTI2022

The Influence of Visual Provenance Representations on Strategies in a Collaborative Hand-off Data Analysis Scenario

Conducting data analysis tasks rarely occur in isolation. Especially in intelligence analysis scenarios where different experts contribute knowledge to a shared understanding, members must communicate how insights develop to establish common ground among collaborators. The use of provenance to communicate analytic sensemaking carries promise by describing the interactions and summarizing the steps taken to reach insights. Yet, no universal guidelines exist for communicating provenance in different settings. Our work here focuses on the presentation of provenance information and the resulting conclusions reached and strategies used by new analysts. In an open-ended, 30-minute, textual exploration scenario, we qualitatively compare how adding different types of provenance information (specifically data coverage and interaction history) affects analysts' confidence in conclusions developed, propensity to repeat work, filtering of data, identification of relevant information, and typical investigation strategies. We see that data coverage (i.e., what was interacted with) provides provenance information without limiting individual investigation freedom. On the other hand, while interaction history (i.e., when something was interacted with) does not significantly encourage more mimicry, it does take more time to comfortably understand, as represented by less confident conclusions and less relevant information-gathering behaviors. In conclusion, our results contribute empirical data towards understanding how provenance summarizations can influence analysis behaviors.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

A Study on Efficient Reinforcement Learning Through Knowledge Transfer

Although Reinforcement Learning (RL) algorithms have made impressive progress in learning complex tasks over the past years, there are still prevailing short-comings and challenges. Specifically, the sample-inefficiency and limited adaptation across tasks often make classic RL techniques impractical for real-world applications despite the gained representational power when combining deep neural networks with RL, known as Deep Reinforcement Learning (DRL). Recently, a number of approaches to address those issues have emerged. Many of those solutions are based on smart DRL architectures that enhance single task algorithms with the capability to share knowledge between agents and across tasks by introducing Transfer Learning (TL) capabilities. Here this survey addresses strategies of knowledge transfer from simple parameter sharing to privacy preserving federated learning and aims at providing a general overview of the field of TL in the DRL domain, establishes a classification framework, and briefly describes representative works in the area.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Machine learning on neutron and x-ray scattering and spectroscopies

Neutron and x-ray scattering represent two classes of state-of-the-art materials characterization techniques that measure materials structural and dynamical properties with high precision. These techniques play critical roles in understanding a wide variety of materials systems from catalysts to polymers, nanomaterials to macromolecules, and energy materials to quantum materials. In recent years, neutron and x-ray scattering have received a significant boost due to the development and increased application of machine learning to materials problems. This article reviews the recent progress in applying machine learning techniques to augment various neutron and x-ray techniques, including neutron scattering, x-ray absorption, x-ray scattering, and photoemission. We highlight the integration of machine learning methods into the typical workflow of scattering experiments, focusing on problems that challenge traditional analysis approaches but are addressable through machine learning, including leveraging the knowledge of simple materials to model more complicated systems, learning with limited data or incomplete labels, identifying meaningful spectra and materials representations, mitigating spectral noise, and others. We present an outlook on a few emerging roles machine learning may play in broad types of scattering and spectroscopic problems in the foreseeable future.

Chen, Zhantao↗
NASA NTRS1987

SWITCH user's manual

The planning program, SWITCH, and its surrounding changed-goal-replanning program, Runaround, are described. The evolution of SWITCH and Runaround from an earlier planner, DEVISER, is recounted. SWITCH's plan representation, and its process of building a plan by backward chaining with strict chronological backtracking, are described. A guide for writing knowledge base files is provided, as are narrative guides for installing the program, running it, and interacting with it while it is running. Some utility functions are documented. For the sake of completeness, a narrative guide to the experimental discrepancy-replanning feature is provided. Appendices contain knowledge base files for a blocksworld domain, and a DRIBBLE file illustrating the output from, and user interaction with, the program in that domain.

Source record↗
NASA NTRS1988

Requirements for a tool to aid in the development of expert systems for space station applications

The Space Station Program is currently in the early stages of design. The extended life of the program and the wide range of planned capabilities present NASA with a major challenge to make effective use of all available resources. Expert systems may represent one of the most important techniques NASA can use to improve performance and functionality of the Space Station. The recent growth in expert system applications has been spurred by the availability of commercial expert system development tools. These state-of-the-art tools provide multiple representation techniques, sophisticated user interfaces, and powerful development tools; allowing programmers to develop expert systems without requiring years of training in knowledge engineering. The requirements are described for a software tool to aid in the development and delivery of expert systems intended for use in the Space Station Program. The intent is to define a tool that would be useful for a very broad range of expert system applications.

Culbert, Chris↗
NASA NTRS1991

Classification-based reasoning

A representation formalism for N-ary relations, quantification, and definition of concepts is described. Three types of conditions are associated with the concepts: (1) necessary and sufficient properties, (2) contingent properties, and (3) necessary properties. Also explained is how complex chains of inferences can be accomplished by representing existentially quantified sentences, and concepts denoted by restrictive relative clauses as classification hierarchies. The representation structures that make possible the inferences are explained first, followed by the reasoning algorithms that draw the inferences from the knowledge structures. All the ideas explained have been implemented and are part of the information retrieval component of a program called Snowy. An appendix contains a brief session with the program.

Gomez, Fernando↗
NASA NTRSDate not supplied

Coupling of Dust Mineralogy to Radiation in NASA GISS ModelE: Impact Upon Dust Optical Properties and Direct Radiative Effect

Estimates of dust direct radiative effect (DRE) are still largely uncertain mainly due to the poor constraining of regional variations of dust absorption, whose known connection with mineral composition is still largely neglected by most Earth System Models. The inaccurate knowledge of soil mineralogy and insufficient measurements of dust physical and optical properties still present the main challenges for a proper representation of the dust-radiation interaction in climate models. We recently upgraded the coupling of dust minerals to shortwave radiation in the NASA GISS ModelE. In particular, we extended the interaction to online-calculated minerals and implemented two different mixing configurations: i) external mixing of three main mineral components and ii) internal mixing of all minerals, allowing a dynamic fraction of iron oxides. We used a semi-empirical approach for the optical calculations, mostly based on aerosol chamber measurements from natural soils, aiming to reduce the influence of uncertain assumptions on the optical properties and mixing state of individual minerals. In this contribution, we present the results of our first calculations of dust optical properties and DREs based on different mixing configurations of interactive minerals. We show the changes to dust optical depth and single scattering albedo, with respect to the control case using homogeneous dust composition, and the consequent global and regional variations in dust DREs at the top of atmosphere and at surface. We also compare the model optical properties with AERONET data, filtered to identify dust scenes. We find that accounting for the diverse properties of dust minerals modifies the pattern of regional dust absorption and therefore its impact upon climate. We also show that the mineralogy effect is sensitive to the assumed mixing state of minerals.

dust direct radiative effect↗
DOE OSTI2025

Enhancing occupant behavior representation for interoperability between building information modeling and building energy modeling

Building Performance Simulation (BPS) has been adopted as an essential tool for designing, operating, and retrofitting buildings to optimize energy efficiency throughout the building life cycle. The Green Building XML (gbXML) schema facilitates seamless data exchange between Building Information Modeling (BIM) and Building Energy Modeling (BEM) software tools. However, limited occupant behavior (OB) representation in BIM often leads to inconsistent and inaccurate energy simulation in BEM software. This paper presents 154 systematic enhancements to the existing occupant behavior XML (obXML) schema v1.3.4, initially developed for standardizing OB representation for BEM, to address existing limitations and improve interoperability with BIM models. The enhancements encompass improved integration with BIM models through extended building representations and system operations, expanded support for advanced OB models with additional environmental parameters and mathematical capabilities, and implementation of a standardized model documentation framework. To facilitate seamless data transformation between gbXML and obXML schemas, we developed a publicly available gb-obXML Schema Converter. Three case studies demonstrate the enhanced schema’s capabilities: representation of building information using a two-story office building model, documentation of a window operation behavior model, and validation of the schema converter’s functionality. The enhanced obXML schema v1.4 enables sophisticated modeling of occupant-building interactions while maintaining consistency with industry-standard BIM schemas. The standardized documentation framework facilitates reproducibility and knowledge sharing in the OB research community, while the schema converter automates the integration of building information into OB simulation workflows. These enhancements establish a foundation for more accurate building performance simulation by supporting sophisticated representation of occupant behavior within the BIM-to-BEM simulation workflows.

Chung, Jihoon↗
DOE OSTI2022

Hypothesis testing via AI: Generating physically interpretable models of scientific data with machine learning (Full Technical Report)

Deep learning has demonstrated an exceptional ability to solve complex tasks (an engineering success); however, it has done so at the expense of the ability to generate new knowledge (a scientific failure). We propose an alternative framework—entitled Deep Symbolic Regression (DSR)—in which artificial neural networks (NNs) rapidly generate hypotheses about physical relationships among inputs. This framework bypasses the need to interpret an NN altogether, while still leveraging the representational power of deep learning. The resulting models are tractable mathematical expressions, which are inherently and readily human interpretable and can provide insights into underlying physical phenomena. Further, we fold this methodology into the scientific process by allowing the scientist to directly integrate a priori knowledge and beliefs to accelerate learning. We demonstrate this methodology on symbolic regression—the problem of rediscovering underlying expressions describing a dataset—and achieve state-of-the-art performance across a wide variety of symbolic regression problems. Further, we generalize our DSR framework to apply to the more general class of symbolic optimization problems, in which one seeks to optimize a sequence of symbols or “tokens” under a black-box reward function. Examples of other symbolic optimization problems include neural architecture search and computational antibody design. Our generalized tool, Deep Symbolic Optimization (DSO), has been demonstrated on the task of learning symbolic control policies for reinforcement learning environments, and has been adopted as an enabling capability for computational antibody design.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1992

Steady and transient performance calculation method for prediction, analysis, and identification

The detailed design and development of turbofans Involves the prediction and identification, by means of test analysis, of the performance of the engine and its components. The thermodynamic simulation and analysis codes integrate existing knowledge and Interpretations of the detailed operating procedure of the components of the engine being developed. The relevance of the predicted performance depends on the quality of the representation of the various physical phenomena affecting the characteristics of the components and, consequently, on the incorporation of experimental correlation in the modelling. In this context, the representation of compressor and turbine characteristics is particularly important. Firstly, we will analyze the ability of corrected parameters to represent MACH similitude at the component inlet under various conditions. The various measurements achieved on the engine during development are used to enhance the modelling. The methods of identifying the thermodynamic calculation code with the various measurements, considered here with their uncertainties, are then presented and described. The analysis of the tests performed on the powerplant, designed to Identify the real characteristics of the engine components, can be undertaken, considering one or more engine test points and incorporating knowledge acquired through experimentation or on the component test bench. There are many possible fields of use for these identification methods ranging from the rematching of components to the optimization of the control system.

Jean Pierre Duponchel↗