Search NASA⌕ Search

SEARCH · Search NASA

Results for “Function Estimation”

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 73 records · Page 4
All sourcesNASA NTRSDOE OSTI
NASA NTRS1990

System Identification for Nonlinear Control Using Neural Networks

An approach to incorporating artificial neural networks in nonlinear, adaptive control systems is described. The controller contains three principal elements: a nonlinear inverse dynamic control law whose coefficients depend on a comprehensive model of the plant, a neural network that models system dynamics, and a state estimator whose outputs drive the control law and train the neural network. Attention is focused on the system identification task, which combines an extended Kalman filter with generalized spline function approximation. Continual learning is possible during normal operation, without taking the system off line for specialized training. Nonlinear inverse dynamic control requires smooth derivatives as well as function estimates, imposing stringent goals on the approximating technique.

Stengel, Robert F.↗
DOE OSTI2019

Simulation studies of R 2 (Δη,Δφ) and P 2 (Δη,Δφ) correlation functions in pp collisions with the PYTHIA and HERWIG models

We report studies of charge-independent (CI) and charge-dependent (CD) two-particle differential-number correlation functions, R 2 (Δη,Δφ), and transverse momentum (p T ) correlation functions, P 2 (Δη,Δφ), of charged particles in √s = 2.76 TeV pp collisions with the PYTHIA and HERWIG models. Model predictions are presented for inclusive charged hadrons (h ± ), as well as pions (π ± ), kaons (K ± ), and (anti-)protons (p¯/p) in the ranges 0.2 < p T ≤ 2.0 GeV/c, 2.0 < p T ≤ 5.0 GeV/c, and 5.0 < p T ≤ 30.0 GeV/c, with full azimuthal coverage in the range |η| < 1.0. We compare the model predictions for the strength and shape of the R 2 and P 2 correlators as these pertain to recent measurements by the ALICE collaboration. The R 2 and P 2 correlation functions estimated with PYTHIA and HERWIG exhibit qualitatively similar near-side and away-side correlation structures but feature important differences. Here, our analysis indicates that comparative studies of R 2 and P 2 correlation functions would provide valuable insight towards the understanding of particle production in pp collisions, and by extension, should also be useful in studies of heavy-ion collisions. Comparison of the Δη dependence of R 2 and P 2 could contribute, in particular, to a better understanding and modeling of the angular ordering of particles produced by hadronization in jets, as well as a better description of jet fragmentation functions of identified species at low momentum fraction (z).

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
NASA NTRS1971

Investigation of the Specht density estimator

The feasibility of using the Specht density estimator function on the IBM 360/44 computer is investigated. Factors such as storage, speed, amount of calculations, size of the smoothing parameter and sample size have an effect on the results. The reliability of the Specht estimator for normal and uniform distributions and the effects of the smoothing parameter and sample size are investigated.

Speed, F. M.↗
DOE OSTI2018

Snake River Plain Geothermal Play Fairway Analysis Project MT Data

MT is measured in the field by using induction coils to measure the time-varying magnetic source for frequencies between 1000-0.001~Hz, and electric dipoles to measure the Earth's electrical response. Because the magnetic source field is polarized, orthogonal directions of the fields need to be measured to get a complete description of the fields. In all measurements collected for this project, induction coils and electric dipoles were aligned with geomagnetic north and east. MT data were collected at 22 stations with a ZEN 32-bit data logger developed by Zonge International, magnetic fields were measured with ANT-4 induction coils, and electric fields where measured with Ag-AgCl reference electrodes from Borin on 50~m dipoles. The data was collected on a repeating schedule of 10~min at 4096~samples/s and 7 hours and 50 minutes at 256 samples/s over a 20-24 hour period. To convert time series data into the frequency domain and get estimations of the impedance tensor, the processing code BIRRP was used (Chave & Thompson 2004). Simultaneous measurements were used as remote references to reduce noise and bias in the data. Chave, A. D., & Thomson, D. J. 2004. Bounded inuence magnetotelluric response function estimation. Geophys. J. Int., 157, 988-1006.

15 GEOTHERMAL ENERGY↗
DOE OSTI2020

GIWAXS-SIIRkit: scattering intensity, indexing and refraction calculation toolkit for grazing-incidence wide-angle X-ray scattering of organic materials

Grazing-incidence wide-angle X-ray scattering (GIWAXS) has become an increasingly popular technique for quantitative structural characterization and comparison of thin films. For this purpose, accurate intensity normalization and peak position determination are crucial. At present, few tools exist to estimate the uncertainties of these measurements. Here, a simulation package is introduced called GIWAXS-SIIRkit , where SIIR stands for scattering intensity, indexing and refraction. Furthermore, the package contains several tools that are freely available for download and can be executed in MATLAB. The package includes three functionalities: estimation of the relative scattering intensity and the corresponding uncertainty based on experimental setup and sample dimensions; extraction and indexing of peak positions to approximate the crystal structure of organic materials starting from calibrated GIWAXS patterns; and analysis of the effects of refraction on peak positions. Additionally, each tool is based on a graphical user interface and designed to have a short learning curve. A user guide is provided with detailed usage instruction, tips for adding functionality and customization, and exemplary files.

36 MATERIALS SCIENCE↗
DOE OSTI2019

Anticipating global terrestrial ecosystem state change using FLUXNET

Ecosystems can be characterized as complex systems that traverse a variety of functional and structural states in response to changing bioclimatic forcings. A central challenge of global change biology is the robust empirical description of these states and state transitions. An ecosystem's functional state can be empirically described using Process Networks (PN) that use timeseries observations to determine the strength of process-level functional couplings between ecosystem components. A globally extensive source of in-situ observations of terrestrial ecosystem dynamics is the FLUXNET eddy-covariance network that provides standardized observations of micrometeorology and carbon, water, and energy flux dynamics. We employ the LaThuile FLUXNET synthesis dataset to delineate each month's functional state for 204 sites, yielding the LaThuile PN version 1.0 database that describes the strength of an ecosystem's functional couplings from air temperature and precipitation to carbon fluxes during each site-month. Then we calculate the elasticity of these couplings to seasonal scale forcings: air temperature, precipitation, solar radiation, and phenophase. Finally, we train artificial neural networks to extrapolate these elasticities from 204 sites to the globe, yielding maps of the estimated functional elasticity of every terrestrial ecosystem's functional states to changing seasonal bioclimatic forcings. These maps provide theoretically novel resource that can be used to anticipate ecological state transitions in response to climate change and to validate process-based models of ecological change. These elasticity maps show that each ecosystem can be expected to respond uniquely to changing forcings. Tropical forests, hot deserts, savannas, and high elevations are most elastic to climate change, and elasticity of ecosystems to seasonal air temperature is on average an order of magnitude higher than elasticity to other bioclimatic forcings. We also observe a reasonable amount of moderate relationships between functional elasticity and structural state change across different ecosystems.

59 BASIC BIOLOGICAL SCIENCES↗
NASA NTRS1995

The luminosity function of quasars

We propose a new evolutionary model for the optical luminosity function of quasars. Our analytical model is derived from fits to the empirical luminosity function estimated by Hartwick and Schade and Warren, Hewett, and Osmer on the basis of more than 1200 quasars over the range of redshifts 0 approximately less than z approximately less than 4.5. We find that the evolution of quasars over this entire redshift range can be well fitted by a Gaussian distribution, while the shape of the luminosity function can be well fitted by either a double power law or an exponential L(exp 1/4) law. The predicted number counts of quasars, as a function of either apparent magnitude or redshift, are fully consistent with the observed ones. Our model indicates that the evolution of quasars reaches its maximum at z approximately = 2.8 and declines at higher redshifts. An extrapolation of the evolution to z approximately greater than 4.5 implies that quasars may have started their cosmic fireworks at z(sub f) approximately = 5.2-5.5. Forthcoming surveys of quasars at these redshifts will be critical to constrain the epoch of quasar formation. All the results we derived are based on observed quasars and are therefore subject to the bias of obscuration by dust in damped Ly alpha systems. Future surveys of these absorption systems at z approximately greater than 3 will also be important if the formation epoch of quasars is to be known unambiguously.

Pei, Yichuan C.↗
NASA NTRS1987

Time-dependent approximation schemes for some problems of parameter estimation in distributed systems

A parameter estimation method that can be used to estimate functional parameters in delay differential equations and moving boundary problems is discussed. In either problem, the original model equation (which is infinite-dimensional) is approximated by a system of ordinary differential equations that can be solved numerically in an efficient way. The approximation scheme is based on time-dependent spline elements. For the delay equation with time-varying delay, convergence results are presented that indicate the estimates obtained using the approximating system. Numerical test examples converge in some sense to a best-fit parameter for the original system are included by means of which time-varying and state-dependent delays and a time-varying diffusion coefficient in a one-phase, one-dimensional Stefan problem are estimated.

Murphy, K. A.↗
NASA NTRS1992

Using neural networks and Dyna algorithm for integrated planning, reacting and learning in systems

The traditional AI answer to the decision making problem for a robot is planning. However, planning is usually CPU-time consuming, depending on the availability and accuracy of a world model. The Dyna system generally described in earlier work, uses trial and error to learn a world model which is simultaneously used to plan reactions resulting in optimal action sequences. It is an attempt to integrate planning, reactive, and learning systems. The architecture of Dyna is presented. The different blocks are described. There are three main components of the system. The first is the world model used by the robot for internal world representation. The input of the world model is the current state and the action taken in the current state. The output is the corresponding reward and resulting state. The second module in the system is the policy. The policy observes the current state and outputs the action to be executed by the robot. At the beginning of program execution, the policy is stochastic and through learning progressively becomes deterministic. The policy decides upon an action according to the output of an evaluation function, which is the third module of the system. The evaluation function takes the following as input: the current state of the system, the action taken in that state, the resulting state, and a reward generated by the world which is proportional to the current distance from the goal state. Originally, the work proposed was as follows: (1) to implement a simple 2-D world where a 'robot' is navigating around obstacles, to learn the path to a goal, by using lookup tables; (2) to substitute the world model and Q estimate function Q by neural networks; and (3) to apply the algorithm to a more complex world where the use of a neural network would be fully justified. In this paper, the system design and achieved results will be described. First we implement the world model with a neural network and leave Q implemented as a look up table. Next, we use a lookup table for the world model and implement the Q function with a neural net. Time limitations prevented the combination of these two approaches. The final section discusses the results and gives clues for future work.

Lima, Pedro↗
DOE OSTI2019

Computing three-point correlation function randoms counts without the randoms catalogue

ABSTRACT As we move towards future galaxy surveys, the three-point statistics will be increasingly leveraged to enhance the constraining power of the data on cosmological parameters. An essential part of the three-point function estimation is performing triplet counts of synthetic data points in random catalogues. Since triplet counting algorithms scale at best as $\mathcal {O}(N^2\log N)$ with the number of particles and the random catalogues are typically at least 50 times denser than the data; this tends to be by far the most time-consuming part of the measurements. Here, we present a simple method of computing the necessary triplet counts involving uniform random distributions through simple one-dimensional integrals. The method speeds up the computation of the three-point function by orders of magnitude, eliminating the need for random catalogues, with the simultaneous pair and triplet counting of the data points alone being sufficient.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS1972

Estimation of correlation functions by stochastic approximation.

Consideration of the autocorrelation function of a zero-mean stationary random process. The techniques are applicable to processes with nonzero mean provided the mean is estimated first and subtracted. Two recursive techniques are proposed, both of which are based on the method of stochastic approximation and assume a functional form for the correlation function that depends on a number of parameters that are recursively estimated from successive records. One technique uses a standard point estimator of the correlation function to provide estimates of the parameters that minimize the mean-square error between the point estimates and the parametric function. The other technique provides estimates of the parameters that maximize a likelihood function relating the parameters of the function to the random process. Examples are presented.

Habibi, A.↗
DOE OSTI2020

Enhanced dynamic equivalent identification method of large-scale power systems using multiple events

The increasing complexity of the interconnected power system makes high-fidelity dynamic simulation models computationally more intensive. To improve computation efficiency, model reduction techniques have been investigated to only preserve the dynamics in a limited area of interest (study area), while deriving equivalent representation for the external area. For this purpose, a measurement-based reduction approach using system identification techniques has been previously proposed. Accordingly, external areas are represented by dynamic equivalent loads using transfer function estimation. In this paper, in order to enhance the accuracy of the reduced model in preserving dynamics of the study area, multiple “grid” events of different types and at different locations are considered for identifying the parameters of the equivalent loads. Additionally, case studies are carried out in the NPCC 140-bus system. Comparisons are made between multiple events training and single event training, highlighting the advantages of the proposed method in providing a better representation of the grid dynamics under different operating conditions.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2026

Adaptive Sampling-Based Bi-Fidelity Stochastic Trust Region Method for Stochastic Derivative-Free Optimization

Bi-fidelity stochastic optimization has gained increasing attention as an efficient approach to reduce computational costs by leveraging a low-fidelity (LF) model to optimize an expensive high-fidelity (HF) objective. In this paper, we propose ASTRO-BFDF, an adaptive sampling trust-region method specifically designed for unconstrained bi-fidelity stochastic derivative-free optimization problems. In ASTRO-BFDF, the LF function serves two purposes: (i) to identify better iterates for the HF function when the optimization process indicates a high correlation between them and (ii) to reduce the variance of the HF function estimates using bi-fidelity Monte Carlo (BFMC). The algorithm dynamically determines sample sizes while adaptively choosing between crude Monte Carlo and BFMC to balance the trade-off between optimization and sampling errors. We prove that the iterates generated by ASTRO-BFDF converge to a first-order stationary point almost surely. Additionally, we demonstrate the effectiveness of the proposed algorithm through numerical experiments on synthetic benchmarks and simulation optimization problems involving discrete event systems.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Microgrid energy scheduling under uncertain extreme weather: Adaptation from parallelized reinforcement learning agents

Microgrids are useful solutions for integrating renewable energy resources and providing seamless green electricity to minimize carbon footprint. In recent years, extreme weather events happened often worldwide and caused significant economic and societal losses. Such events bring uncertainties to the microgrid energy scheduling problems and increase the challenges of microgrid operation. Traditional optimization approaches suffer from the inaccuracy of the uncertain microgrid model and the unseen events. Existing reinforcement learning (RL) - based approaches are also hampered by the limited generalization and the increasing computational burden when stochastic formulations are required to accommodate the uncertainties. This paper proposes a new parallelized reinforcement learning (PRL) method based on the probabilistic events to handle the microgrid energy uncertainties. Specifically, several local learning agents are employed to interact with pertinent microgrid environments in a distributed manner and report outcomes to the global agent, which will optimize microgrid energy resources online during extreme events. The stochastic microgrid energy optimization problem is reformulated to include all possible scenarios with probabilities. The advantage estimate functions of learning agents are designed with a backward sweep to transfer the outcomes to the value function updating process. Two simulation studies, stochastic optimization and online testing, are performed to compare with several existing RL approaches. Results substantiate that the proposed PRL method can achieve up to 20% optimization performance improvement with 4 and 28 times less computation cost than Q-learning with experience replay and multi-agent Q-learning approaches, respectively.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2022

Structures of Vanadium-Containing Silicate and Borosilicate Glasses: Vanadium Potential Development and MD Simulations

Vanadium-containing multi-component glasses have been found with widely interests in various applications. However, the complex compositions, thus structures, lead to the challenges in experiments to understand the structural origin of the property change of these glasses. In this work, we developed compatible vanadium-related parameters for a widely used pair wise potential set to enable the simulations of the vanadium-containing multi-component glasses. Various crystal structures and glass compositions (with vanadium in different oxidation states) have been tested using the newly developed parameters, and structural information of cell parameters, pair distribution function, estimated bond distance, and coordination number have been obtained. The results are in good agreement with available experimental data, which indicates the new vanadium parameters can be used to simulate the vanadium-containing multi-component glasses and thus help study those applications with wide interests.

Deng, Lu↗
DOE OSTI2022

Accurate predictions from small boxes: variance suppression via the Zel'dovich approximation

Simulations have become an indispensable tool for accurate modelling of observables measured in galaxy surveys, but can be expensive if very large dynamic range in scale is required. We describe how to combine Lagrangian perturbation theory models with N-body simulations to reduce the effects of finite computational volume in the prediction of ensemble average properties in the simulations within the context of control variates. In particular we use the fact that Zel'dovich displacements, computed during initial condition generation for any simulation, correlate strongly with the final density field. Since all the correlators of biased tracers can be computed with arbitrary precision for these displacements, pairing the Zel'dovich `simulation' with the N-body realization allows hundredfold reductions in sample variance for power spectrum or correlation function estimation. Additionally, Zel'dovich control variates can accurately extend matter or tracer field emulators to larger scales than previously possible, as well as improving measurements of statistics in simulations which are inherently limited to small volumes, such as hydrodynamical simulations of galaxy formation and reionization.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2019

Investigation of fast particle redistribution induced by sawtooth instability in NSTX-U

The effects of sawtooth on fast ion transport have been examined in reproducible, 2s long sawtoothing L-mode discharges during the 2016 experimental campaign on National Spherical Torus Experiment Upgrade (NSTX-U) [J.E. Menard et al. Nucl. Fusion (2012) 083015]. Analysis of the discharges showed that standard sawtooth models (full/partial reconnection models) in the TRANSP code were not capable to fully reproduce the fast ion redistribution induced by sawtooth crashes. Some global parameters such as neutron rate can be recovered while detailed features, e.g. distribution functions, estimated using the models were different from the experimental observation. The standard sawtooth models in TRANSP do not take into account the different effect of sawtooth crashes depending on fast ion energy and orbit type and that may cause the disagreement between experiments and simulations. In this work, the newly developed kick model has been applied to replace the standard sawtooth models for the fast ion transport. TRANSP simulation results using the kick model, taking into account the characteristics of fast ion such as energy and pitch angle, can reproduce experimental neutron rates within 10\% difference. The qualitative comparison of the measurements and synthetic diagnostics of FIDA and SSNPA using the TRANSP simulation results with kick model demonstrates good agreements.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2018

Aviation Transportation, Cyber Threats, and Network-of-Networks: Modeling Perspectives for Translating Theory to Practice

Understanding aviation transportation infrastructure system behavior and coupling with communication networks is essential for securing and restoring functionality against cyber-enabled threats. While significant progress has been made in the past decade on developing infrastructure resilience theories based on network structure and operations, translating and generalizing them to real-world practice has often been challenging due to imperfect data and inapplicability of modeling assumptions. These typically include: 1) stylized network structures without uncertainty, 2) node homogeneity, 3) static criticality measures, and 4) unrealistic cascade models originating from single points of failure. This paper presents modeling perspectives that aim to address these theory-to-practice challenges using a well-grounded network-of-networks (NoN) construct. Real-world modeling challenges are identified and a network theory-guided conceptual NoN model is developed that may be operationalized with the U.S. national airspace system airport network (NASAN) and Federal Aviation Administration (FAA) communication network as application domain. Improving accuracy in loss of functionality estimates due to cyber events in communication networks and cascading air travel delay impacts is key for informing consequence-based decision support and risk/contingency analyses.

Bhatia, Udit↗