SEARCH · Search NASA

Results for “algorithm selection”

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 649 records · Page 36

All sources NASA NTRS DOE OSTI

NASA NTRS2022

MLtool Python Code

Machine Learning (ML) is a subfield of Artificial Intelligence that gives computers the ability to learn from past data without being explicitly programmed. The predictive capabilities of ML models have already been used to facilitate several scientific breakthroughs. However, the practical application of ML is often limited due to the gaps in technical knowledge of its users. The common issue faced by many scientific researchers is the inability to choose the appropriate ML pipelines that are needed to treat real-world data, which is often sparse and noisy. To solve this problem, we have developed an automated Machine Learning tool (MLtool) that includes a set of ML algorithms and approaches to aid scientific researchers. The current version of MLtool is implemented as an object-oriented Python code that is easily extensible. It includes 44 different regression algorithms used to model data. MLtool helps users select the best model for their data, based on the scoring metrics used. Besides regression algorithms, MLtool also includes a suite of pre- and post-processing techniques such as missing value imputation, categorical variable encoding, input feature normalization, uncertainty quantification, exploratory data analysis (EDA), etc. MLtool was tested on several publicly available multi-dimensional data sets and was found capable of making accurate predictions.

Machine Learning↗

NASA NTRS1973

Application of modern control theory to the design of optimum aircraft controllers

The synthesis procedure presented is based on the solution of the output regulator problem of linear optimal control theory for time-invariant systems. By this technique, solution of the matrix Riccati equation leads to a constant linear feedback control law for an output regulator which will maintain a plant in a particular equilibrium condition in the presence of impulse disturbances. Two simple algorithms are presented that can be used in an automatic synthesis procedure for the design of maneuverable output regulators requiring only selected state variables for feedback. The first algorithm is for the construction of optimal feedforward control laws that can be superimposed upon a Kalman output regulator and that will drive the output of a plant to a desired constant value on command. The second algorithm is for the construction of optimal Luenberger observers that can be used to obtain feedback control laws for the output regulator requiring measurement of only part of the state vector. This algorithm constructs observers which have minimum response time under the constraint that the magnitude of the gains in the observer filter be less than some arbitrary limit.

Power, L. J.↗

DOE OSTI2023

Autonomous and dynamic precursor selection for solid-state materials synthesis

Abstract Solid-state synthesis plays an important role in the development of new materials and technologies. While in situ characterization and ab-initio computations have advanced our understanding of materials synthesis, experiments targeting new compounds often still require many different precursors and conditions to be tested. Here we introduce an algorithm (ARROWS 3 ) designed to automate the selection of optimal precursors for solid-state materials synthesis. This algorithm actively learns from experimental outcomes to determine which precursors lead to unfavorable reactions that form highly stable intermediates, preventing the target material’s formation. Based on this information, ARROWS 3 proposes new experiments using precursors it predicts to avoid such intermediates, thereby retaining a larger thermodynamic driving force to form the target. We validate this approach on three experimental datasets, containing results from over 200 synthesis procedures. In comparison to black-box optimization, ARROWS 3 identifies effective precursor sets for each target while requiring substantially fewer experimental iterations. These findings highlight the importance of domain knowledge in optimization algorithms for materials synthesis, which are critical for the development of fully autonomous research platforms.

36 MATERIALS SCIENCE↗

DOE OSTI2021

Novel usage of deep learning and high-performance computing in long-baseline neutrino oscillation experiments

Mención Internacional en el título de doctorDeep-learning methods are playing a crucial role in numerous scientific and industrialapplications. Over the past two decades, these techniques have helped in the collection,reconstruction, and analysis of large data samples in particle physics experiments. Themain topic of this PhD research is the study of deep-learning techniques in long-baselineneutrino oscillation experiments. Neutrinos are mysterious light elementary particles,and their investigation is essential to shed light on some of the remaining open questionsin physics. The work presented here describes an algorithm based on a convolutionalneural network developed to provide highly accurate and efficient selections of electronneutrino and muon neutrino interactions in the Deep Underground Neutrino Experiment(DUNE). With this algorithm, the electron neutrino (antineutrino) selection efficiencypeaks at 90% (94%) and exceeds 85% (90%) for reconstructed neutrino energies between2-5 GeV. The selection efficiency for muon neutrino (antineutrino) interactions is foundto have a maximum of 96% (97%) and exceeds 90% (95%) efficiency for reconstructedneutrino energies above 2 GeV. When considering all electron neutrino and antineutrinointeractions as signal (both those appearing from oscillations and those intrinsic tothe beam), a selection purity of 90% is achieved. These event selections are criticalto maximise the sensitivity of the experiment to CP-violating effects, key to furtherunderstand the matter-antimatter asymmetry of the Universe.In high-energy physics experiments, deep learning has also been explored for producingfast simulations and physically-motivated manipulations of simulated images. Some ofthose simulations, such as the light production and detection, are very computationallyexpensive and require novel methods to produce the necessary samples while controllingthe varied underlying physics model parameters. To do so, we invented the model-assistedgenerative adversarial network (MAGAN), first validated on simple generic case studiesand then successfully applied to the DUNE photon-detector simulation.Moreover, we also developed graph neural networks for 3D-voxel classification ofambiguities and optical crosstalk for a different particle physics experiment, most preciselyfor the proposed SuperFGD. This novel 3D-granular plastic-scintillator neutrino detectorwill be used to upgrade the near detector of the T2K neutrino oscillation experiment, and our method reports efficiencies and purities of 94-96% per event in the classificationof particle track voxels.Due to the growth and complexity of deep neural networks, researchers have beeninvestigating techniques to train those networks in a more computationally-efficient way.Many efforts have been made by the community to optimise deep-learning models byparallelising or distributing their training computation across multiple devices. In thisthesis, we study an approach based on data locality for those neural networks that cannotbenefit from scaling their computation due to a significant bottleneck in the data I/O.The research also includes a detailed study on the performance of deep neural networkson hardware accelerator boards.Los métodos de aprendizaje profundo son cada vez más utilizados en numerosas aplicacionescientíficas e industriales hoy en día. Durante las dos últimas décadas, estastécnicas se han empleado en la recolección, reconstrucción y análisis de la gran cantidadde datos generados por experimentos de física de partículas. El tema principal de estatesis doctoral es el uso de estos modelos de aprendizaje profundo en experimentos defísica de neutrinos, en concreto en los experimentos de larga distancia DUNE y T2K. Losneutrinos, partículas fundamentales neutras, de las más ligeras del Universo, pueden serclave para explicar algunas de las cuestiones todavía sin resolver en física fundamental.Entre las diferentes contribuciones que esta tesis ha hecho a su estudio, cabe destacar eldesarrollo de un algoritmo basado en una red de neuronas convolucional para seleccionarcon gran eficiencia y precisión las interacciones de neutrinos electrónicos y muónicos enel Deep Underground Neutrino Experiment (DUNE). La eficiencia de selección obtenidapara neutrinos (antineutrinos) electrónicos alcanza un máximo del 90% (94%) y supera el85% (90%) para neutrinos con energías reconstruidas en el rango 2-5 GeV. La selección deneutrinos (antineutrinos) muónicos tiene una eficiencia máxima del 96% (97%) y excedeel 90% (95%) para neutrinos con energías reconstruidas de más de 2 GeV. Considerandocomo señal todas las interacciones de neutrinos y antineutrinos electrónicos (procedentestanto de oscilaciones como intrínsecos en el haz inicial), se logra una pureza en la seleccióndel 90%. Dichas selecciones de eventos son fundamentales para maximizar la sensibilidaddel experimento a los efectos de violació...

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

DOE OSTI2022

A Reinforcement Learning Approach to Parameter Selection for Distributed Optimal Power Flow

With the increasing penetration of distributed energy resources, distributed optimization algorithms have attracted significant attention for power systems applications due to their potential for superior scalability, privacy, and robustness to a single point-of-failure. The Alternating Direction Method of Multipliers (ADMM) is a popular distributed optimization algorithm; however, its convergence performance is highly dependent on the selection of penalty parameters, which are usually chosen heuristically. In this work, we use reinforcement learning (RL) to develop an adaptive penalty parameter selection policy for alternating current optimal power flow (ACOPF) problem solved via ADMM with the goal of minimizing the number of iterations until convergence. We train our RL policy using deep Q-learning and show that this policy can result in significantly accelerated convergence (up to a 59% reduction in the number of iterations compared to existing, curvatureinformed penalty parameter selection methods). Furthermore, we show that our RL policy demonstrates promise for generalizability, performing well under unseen loading schemes as well as under unseen losses of lines and generators (up to a 50% reduction in iterations). This work thus provides a proof-of-concept for using RL for parameter selection in ADMM for power systems applications.

alternating current optimal power flow↗

NASA NTRS2011

Introduction of Parallel GPGPU Acceleration Algorithms for the Solution of Radiative Transfer

General-purpose computing on graphics processing units (GPGPU) is a recent technique that allows the parallel graphics processing unit (GPU) to accelerate calculations performed sequentially by the central processing unit (CPU). To introduce GPGPU to radiative transfer, the Gauss-Seidel solution of the well-known expressions for 1-D and 3-D homogeneous, isotropic media is selected as a test case. Different algorithms are introduced to balance memory and GPU-CPU communication, critical aspects of GPGPU. Results show that speed-ups of one to two orders of magnitude are obtained when compared to sequential solutions. The underlying value of GPGPU is its potential extension in radiative solvers (e.g., Monte Carlo, discrete ordinates) at a minimal learning curve.

Godoy, William F.↗

DOE OSTI2022

A Hybrid DC Fault Primary Protection Algorithm for Multi-Terminal HVdc Systems

Protection against dc faults is one of the main technical hurdles faced when operating converter-based HVdc systems. Protection becomes even more challenging for multi-terminal dc (MTdc) systems with more than two terminals/converter stations. In this paper, a hybrid primary fault detection algorithm for MTdc systems is proposed to detect a broad range of failures. Sensor measurements, i.e., line currents and dc reactor voltages measured at local terminals, are first processed by a top-level context clustering algorithm. For each cluster, the best fault detector is selected among a detector pool according to a rule resulting from a learning algorithm. The detector pool consists of several existing detection algorithms, each performing differently across fault scenarios. The proposed hybrid primary detection algorithm: i) offers superior performance compared to an individual detector through a data-driven approach; ii) detects all major fault types including pole-to-pole (P2P), pole-to-ground (P2G), and external dc faults; iii) identifies faults with various fault locations and impedances; iv) is more robust to noisy sensor measurements compared to existing methods; v) does not require exhaustive simulation and sampling for training the model. Performance and effectiveness of the proposed algorithm are evaluated and verified based on time-domain simulations in the PSCAD/EMTDC software environment. The results confirm satisfactory operation, accuracy, and detection speed of the proposed algorithm under various fault scenarios.

24 POWER TRANSMISSION AND DISTRIBUTION↗

NASA NTRS1977

Low cost Ku-band earth terminals for voice/data/facsimile

A Ku-band satellite earth terminal capable of providing two way voice/facsimile teleconferencing, 128 Kbps data, telephone, and high-speed imagery services is proposed. Optimized terminal cost and configuration are presented as a function of FDMA and TDMA approaches to multiple access. The entire terminal from the antenna to microphones, speakers and facsimile equipment is considered. Component cost versus performance has been projected as a function of size of the procurement and predicted hardware innovations and production techniques through 1985. The lowest cost combinations of components has been determined in a computer optimization algorithm. The system requirements including terminal EIRP and G/T, satellite size, power per spacecraft transponder, satellite antenna characteristics, and link propagation outage were selected using a computerized system cost/performance optimization algorithm. System cost and terminal cost and performance requirements are presented as a function of the size of a nationwide U.S. network. Service costs are compared with typical conference travel costs to show the viability of the proposed terminal.

Kelley, R. L.↗

NASA NTRS1993

Modeling and analysis of selected space station communications and tracking subsystems

The Communications and Tracking System on board Space Station Freedom (SSF) provides space-to-ground, space-to-space, audio, and video communications, as well as tracking data reception and processing services. Each major category of service is provided by a communications subsystem which is controlled and monitored by software. Among these subsystems, the Assembly/Contingency Subsystem (ACS) and the Space-to-Ground Subsystem (SGS) provide communications with the ground via the Tracking and Data Relay Satellite (TDRS) System. The ACS is effectively SSF's command link, while the SGS is primarily intended as the data link for SSF payloads. The research activities of this project focused on the ACS and SGS antenna management algorithms identified in the Flight System Software Requirements (FSSR) documentation, including: (1) software modeling and evaluation of antenna management (positioning) algorithms; and (2) analysis and investigation of selected variables and parameters of these antenna management algorithms i.e., descriptions and definitions of ranges, scopes, and dimensions. In a related activity, to assist those responsible for monitoring the development of this flight system software, a brief summary of software metrics concepts, terms, measures, and uses was prepared.

Richmond, Elmer Raydean↗

DOE OSTI2026

Exploring small-angle emissions in charm quark jets in proton-proton collisions at $\sqrt{s}=5.02$ TeV

A measurement of the angular structure of inclusive jets and those containing a prompt D 0 meson in proton-proton collisions at the LHC at a center-of-mass energy of 5.02 TeV is presented. The data corresponding to an integrated luminosity of 301 pb −1 were collected by the CMS experiment in 2017. Two jet grooming algorithms, late-k T and soft drop, are used to study the intrajet radiation pattern using iterative Cambridge-Aachen declustering. The splitting-angle distributions of jets with transverse momentum (p T ) of around 100 GeV, obtained with these two algorithms, show that there is a shift of the distribution for jets containing a prompt D 0 meson with respect to inclusive jets. The suppression of emissions at small angles observed in the late-k T grooming approach is consistent with the dead-cone effect, whereas the similar suppression for splittings selected with the soft-drop algorithm appears to be induced by gluon splitting to charm quark-antiquark pairs at large angles. The measured distributions are corrected to the particle level and can be used to constrain model predictions for the substructure of high-p T charm quark jets.

Charm Physics↗

NASA NTRS1991

Evaluation of a transfinite element numerical solution method for nonlinear heat transfer problems

Laplace transform techniques have been widely used to solve linear, transient field problems. A transform-based algorithm enables calculation of the response at selected times of interest without the need for stepping in time as required by conventional time integration schemes. The elimination of time stepping can substantially reduce computer time when transform techniques are implemented in a numerical finite element program. The coupling of transform techniques with spatial discretization techniques such as the finite element method has resulted in what are known as transfinite element methods. Recently attempts have been made to extend the transfinite element method to solve nonlinear, transient field problems. This paper examines the theoretical basis and numerical implementation of one such algorithm, applied to nonlinear heat transfer problems. The problem is linearized and solved by requiring a numerical iteration at selected times of interest. While shown to be acceptable for weakly nonlinear problems, this algorithm is ineffective as a general nonlinear solution method.

Cerro, J. A.↗

NASA NTRS2000

Active Structural Acoustic Control of Interior Noise on a Raytheon 1900D

An active structural acoustic control system has been demonstrated on a Raytheon Aircraft Company 1900D turboprop airliner. Both single frequency and multi-frequency control of the blade passage frequency and its harmonics was accomplished. The control algorithm was a variant of the popular filtered-x LMS implemented in the principal component domain. The control system consisted of 21 inertial actuators and 32 microphones. The actuators were mounted to the aircraft's ring frames. The microphones were distributed uniformly throughout the interior at head height, both seated and standing. Actuator locations were selected using a combinatorial search optimization algorithm. The control system achieved a 14 dB noise reduction of the blade passage frequency during single frequency tests. Multi-frequency control of the first 1st, 2nd and 3rd harmonics resulted in 10.2 dB, 3.3 dB and 1.6 dB noise reductions respectively. These results fall short of the predictions which were produced by the optimization algorithm (13.5 dB, 8.6 dB and 6.3 dB). The optimization was based on actuator transfer functions taken on the ground and it is postulated that cabin pressurization at flight altitude was a factor in this discrepancy.

Palumbo, Dan↗

DOE OSTI2020

Leveraging Real-World Driving Data for Design and Impact Evaluation of Energy Efficient Control Strategies

Modeling and simulation are crucial in the development of advanced energy efficient control strategies. Utilizing real-world driving data as the underlying basis for control design and simulation lends veracity to projected real-world energy savings. Standardized drive cycles are limited in their utility for evaluating advanced driving strategies that utilize connectivity and on-vehicle sensing, primarily because they are typically intended for evaluating emissions and fuel economy under controlled conditions. Real-world driving data, because of its scale, is a useful representation of various road types, driving styles, and driving environments. The scale of real-world data also presents challenges in effectively using it in simulations. A fast and efficient simulation methodology is necessary to handle the large number of simulations performed for design analysis and impact evaluation of control strategies. In this study, two methods are presented of leveraging real-world data in both design optimization of energy efficient control strategies and in evaluating the real-world impact of those control strategies upon large-scale deployment. Through these methodologies, strategies with highest impact on energy savings were selected to be implemented as control algorithms. The developed algorithms were incorporated into a vehicle dynamics and powertrain control architecture implemented on a Cadillac CT6 demonstration vehicle. The control algorithms were then exercised on real-world driving scenarios to determine their impact on collective energy savings. The methodology utilizes the large-scale driving data sets maintained by the National Renewable Energy Laboratory to extract real-world driving scenarios and efficient simulation software tools. The insights obtained through this research help in guiding technology selection for energy efficient driving controls.

ARPA-E↗

DOE OSTI2024

Advanced Method Optimization for Sampling and Analysis Instrumentation

This work presents a generalized approach for analytical method optimization that branches the gap between techniques historically employed and accurate modern optimization techniques suitable for various applications. The novelty of the described strategy is the utilization of multivariate, multiobjective optimization with Karush-Kuhn-Tucker conditions to bound the optimization space to solutions within the physical limitations of instrumentation. Briefly, the basic steps outlined in this paper are to (1) determine the objective(s) that should be maximized or minimized based on the goals of the analytical application, (2) conduct a screening experiment, (3) perform ANOVA to determine the parameters which have a statistically significant effect on the objective, (4) conduct an experiment (e.g., Box-Behnken design) to collect data for fitting the objective equation, and (5) determine the physical constraints of the parameters and solve the Lagrangian to determine the optimal method parameters. A broad approach to optimization target selection allows for robust method tuning to develop improved data sets amenable for chemometrics and machine learning algorithm development. Gas chromatography-mass spectrometry was selected as a use case due to its broad use across scientific fields and time-consuming method development involving numerous parameters. In conclusion, this strategy can reduce the cost of research, improve data quality, and enable the rapid development of new analytical technique.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

NASA NTRS1993

High-speed data duplication/data distribution: An adjunct to the mass storage equation

The term 'mass storage' invokes the image of large on-site disk and tape farms which contain huge quantities of low- to medium-access data. Although the cost of such bulk storage is recognized, the cost of the bulk distribution of this data rarely is given much attention. Mass data distribution becomes an even more acute problem if the bulk data is part of a national or international system. If the bulk data distribution is to travel from one large data center to another large data center then fiber-optic cables or the use of satellite channels is feasible. However, if the distribution must be disseminated from a central site to a number of much smaller, and, perhaps varying sites, then cost prohibits the use of fiber-optic cable or satellite communication. Given these cost constraints much of the bulk distribution of data will continue to be disseminated via inexpensive magnetic tape using the various next day postal service options. For non-transmitted bulk data, our working hypotheses are that the desired duplication efficiency of the total bulk data should be established before selecting any particular data duplication system; and, that the data duplication algorithm should be determined before any bulk data duplication method is selected.

Howard, Kevin↗

NASA NTRS2021

Using Intelligent Targeting to increase the science return of a Smart Ice Storm Hunting Radar

Smart Ice Cloud Sensing (SMICES) is a small-sat concept in which a radar intelligently targets ice storms based on information collected by a lookahead radiometer. Often space observations are performed by continuously collecting data from an instrument aimed at nadir (e.g. directly below the space platform). However, if the platform has the ability to assess science utility of features being overflown, an intelligent measurement scheme can improve science return. This can be achieved by controlling the on/off state of the instrument if it is not able to continuously operate (e.g. due to energy or thermal constraints), and by allowing the instrument to view off nadir if it has pointing capabilities.In the case of SMICES, power constraints and the rarity of storms means that with blind nadir targeting SMICES would collect a limited amount of ice storm radar data. The algorithms proposed acquire measurements to maximize acquired high interest storms while concurrently collecting a background sampling of all features. We use a cloud classification system to identify five different cloud types. Six algorithms ranging from “blind” to more selective are described and results from evaluation on a dataset of 13 ground swaths covering 72,399,600 km2 of data are presented. This data is from high quality science simulations that contain all five cloud types and multiple storms. When utilizing the radiometer’s lookahead and the full range of the radar the results show a 23.7x and 1.9x increase over the base algorithm in the most and second most important cloud types respectively.

Cooke, Caitlyn↗

NASA NTRS1989

An experimental investigation of fault tolerant software structures in an avionics application

The objective of this experimental investigation is to compare the functional performance and software reliability of competing fault tolerant software structures utilizing software diversity. In this experiment, three versions of the redundancy management software for a skewed sensor array have been developed using three diverse failure detection and isolation algorithms and incorporated into various N-version, recovery block and hybrid software structures. The empirical results show that, for maximum functional performance improvement in the selected application domain, the results of diverse algorithms should be voted before being processed by multiple versions without enforced diversity. Results also suggest that when the reliability gain with an N-version structure is modest, recovery block structures are more feasible since higher reliability can be obtained using an acceptance check with a modest reliability.

Caglayan, Alper K.↗

NASA NTRS2003

Enhancing and Archiving the APS Catalog of the POSS I

We have worked on two different projects: 1) Archiving the APS Catalog of the POSS I for distribution to NASA's NED at IPAC, SIMBAD in France, and individual astronomers and 2) The automated morphological classification of galaxies. We have completed archiving the Catalog into easily readable binary files. The database together with the software to read it has been distributed on DVD's to the national and international data centers and to individual astronomers. The archived Catalog contains more than 89 million objects in 632 fields in the first epoch Palomar Observatory Sky Survey. Additional image parameters not available in the original on-line version are also included in the archived version. The archived Catalog is also available and can be queried at the APS web site (URL: http://aps.umn.edu) which has been improved with a much faster and more efficient querying system. The Catalog can be downloaded as binary datafiles with the source code for reading it. It is also being integrated into the SkyQuery system which includes the Sloan Digital Sky Survey, 2MASS, and the FIRST radio sky survey. We experimented with different classification algorithms to automate the morphological classification of galaxies. This is an especially difficult problem because there are not only a large number of attributes or parameters and measurement uncertainties, but also the added complication of human disagreement about the adopted types. To solve this problem we used 837 galaxy images from nine POSS I fields at the North Galactic Pole classified by two independent astronomers for which they agree on the morphological types. The initial goal was to separate the galaxies into the three broad classes relevant to issues of large scale structure and galaxy formation and evolution: early (ellipticals and lenticulars), spirals, and late (irregulars) with an accuracy or success rate that rivals the best astronomer classifiers. We also needed to identify a set of parameters derived from the digitized images that separate the galaxies by type. The human eye can easily recognize complicated patterns in images such as spiral arms which can be spotty, blotchy affairs that are difficult for automated techniques. A galaxy image can potentially be described by hundreds of parameters, all of which may have some relation to the morphological type. In the set of initial experiments we used 624 such parameters, in two colors, blue and red. These parameters include the surface brightness and color measured at different radii, ratios of these parameters at different radii, concentration indices, Fourier transforms and wavelet decomposition coefficients. We experimented with three different classes of classification algorithms; decision trees, k-nearest neighbors, and support vector machines (SVM). A range of experiments were conducted and we eventually narrowed the parameters to 23 selected parameters. SVM consistently outperformed the other algorithms with both sets of features. By combining the results from the different algorithms in a weighted scheme we achieved an overall classification success of 86%.

Humphreys, Roberta M.↗