Search NASA⌕ Search

SEARCH · Search NASA

Results for “functional regression”

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 325 records · Page 18

Rarefied aerodynamics and upper atmosphere density from multiple orbiter flight measurements

Flight data from six flights of the experimental High Resolution Accelerometer Package (HIRAP), a micro-g accelerometer system, have been analyzed to produce rarefied flow aerodynamic coefficients of the Orbiter and freestream density in the altitude range from 60 km to 160 km. The direct measurements of the lift to drag ratio (L/D) are used to obtain individual normalized body-axis force coefficients in a least-squares regression scheme. An aerodynamics model based on an exponential function of Knudsen number is used for the individual coefficients. The calculated L/D from the flight determined coefficients agree with the multiple flight data measurements to within about 5 percent. Simultaneously, an upper altitude density variation is also obtained from the data. Density altitude profiles exhibit a wave feature on all flights. The wave has an amplitude of about + or - 25 percent relative to the 1962 Standard atmosphere model.

Buck, G. M.↗

Durability of Ceramic Matrix Composites in Combustion Environments

SiO2 scale volatility and the resulting recession of SiC has been previously identified for CVD, as well as sintered, SiC and Si3N4 materials. Linear weight loss and surface recession of SiC was observed as a result of SiO2 volatility for both fuel-lean and fuel-rich gas mixtures. A strong Arrhenius temperature dependence exist and parametric studies using multiple linear regression were used to develop recession relationships as a function of e(sup -Q/RT) central dot P(sup X)V(sup y) for both lean and rich combustion conditions. Recession of 0.2 - 2.0 micron/hr is predicted for typical combustion conditions at 1200 - 1400 C, with a somewhat higher absolute recession rate and activation energy for rich combustion versus lean.

Robinson, Raymond C.↗

Digital Correlation Microwave Polarimetry: Analysis and Demonstration

The design, analysis, and demonstration of a digital-correlation microwave polarimeter for use in earth remote sensing is presented. We begin with an analysis of three-level digital correlation and develop the correlator transfer function and radiometric sensitivity. A fifth-order polynomial regression is derived for inverting the digital correlation coefficient into the analog statistic. In addition, the effects of quantizer threshold asymmetry and hysteresis are discussed. A two-look unpolarized calibration scheme is developed for identifying correlation offsets. The developed theory and calibration method are verified using a 10.7 GHz and a 37.0 GHz polarimeter. The polarimeters are based upon 1-GS/s three-level digital correlators and measure the first three Stokes parameters. Through experiment, the radiometric sensitivity is shown to approach the theoretical as derived earlier in the paper and the two-look unpolarized calibration method is successfully compared with results using a polarimetric scheme. Finally, sample data from an aircraft experiment demonstrates that the polarimeter is highly-useful for ocean wind-vector measurement.

Piepmeier, J. R.↗

Validation of TRMM Satellite Rainfall Products Over Oklahoma for a Three Year Period (1998-2000)

This study evaluates the level 2 and level 3 rainfall products from the Tropical Rainfall Measuring Mission gridded distribution of 118 high resolution tipping buckets from the Oklahoma Mesonet. The Tropical Rainfall Measuring Mission, a joint satellite mission between NASA and the National Space Development Agency (NASDA) of Japan, was designed to estimate global precipitation between 40 S and 40 N latitude. The TRMM satellite, consisting of two main precipitation sensors, a passive microwave (TMI) and precipitation radar (PR) sensors, was launched in 1997. Although the great advantage of TRMM has been its ability to sample precipitation globally over the tropical oceans in places where ground sensors do not exist, regional points over land still offer the best opportunity for validating these rain estimates. For this study, the gauge data was gridded at I x I degree resolution between 34 and 36 N and -95 and -100 W. The location of Oklahoma is somewhat unique, in that, it is located near the turning point for the satellite. This study investigates effects of temporal sampling on the satellite measurements and the resulting rainfall bias observed between space and ground based sensors. The first part of the study analyzes the monthly rainfall statistics of the 3AI2 (TMI), the 3A25 (PR) and the 31142 (combined sensors) with the rain gauges. It uses regression techniques and a computation of the probability density function (PDF) and cumulative probability density function (CDF) for each sensor to evaluate the effects of temporal sampling on the satellite products and the sufficiency of using two overpasses a day to estimate rainfall on a monthly scale. In further probing this issue, the study also looked at the direct coincidence of ground and satellite measurements by comparing instantaneous comparisons obtained from the level 2 orbital track data, e.g. 2A12, 2A25. Lastly, the effects of temporal sampling were further studied by sub-sampling the gauge data only at overpass times. This methodology provides a quantitative way of separately inferring the fraction of bias due to temporal sampling and the rainfall algorithm. In addition to showing the salient effects of temporal sampling, the study also revisits the historical problem of correlating point gauge estimates with area estimates, though, the problem is ameliorated by averaging and gridding the data to the area resolution scale of the satellite.

Fisher, Brad↗

Instrumentation and methodology for quantifying GFP fluorescence in intact plant organs

The General Fluorescence Plant Meter (GFP-Meter) is a portable spectrofluorometer that utilizes a fiber-optic cable and a leaf clip to gather spectrofluorescence data. In contrast to traditional analytical systems, this instrument allows for the rapid detection and fluorescence measurement of proteins under field conditions with no damage to plant tissue. Here we discuss the methodology of gathering and standardizing spectrofluorescence data from tobacco and canola plants expressing GFP. Furthermore, we demonstrate the accuracy and effectiveness of the GFP-Meter. We first compared GFP fluorescence measurements taken by the GFP-Meter to those taken by a standard laboratory-based spectrofluorometer, the FluoroMax-2. Spectrofluorescence measurements were taken from the same location on intact leaves. When these measurements were tested by simple linear regression analysis, we found that there was a positive functional relationship between instruments. Finally, to exhibit that the GFP-Meter recorded accurate measurements over a span of time, we completed a time-course analysis of GFP fluorescence measurements. We found that only initial measurements were accurate; however, subsequent measurements could be used for qualitative purposes.

Evaluation Studies↗

Analysis of Wind Tunnel Lateral Oscillatory Data of the F-16XL Aircraft

Static and dynamic wind tunnel tests were performed on an 18% scale model of the F-16XL aircraft. These tests were performed over a wide range of angles of attack and sideslip with oscillation amplitudes from 5 deg. to 30 deg. and reduced frequencies from 0.073 to 0.269. Harmonic analysis was used to estimate Fourier coefficients and in-phase and out-of-phase components. For frequency dependent data from rolling oscillations, a two-step regression method was used to obtain unsteady models (indicial functions), and derivatives due to sideslip angle, roll rate and yaw rate from in-phase and out-of-phase components. Frequency dependence was found for angles of attack between 20 deg. and 50 deg. Reduced values of coefficient of determination and increased values of fit error were found for angles of attack between 35 deg. and 45 deg. An attempt to estimate model parameters from yaw oscillations failed, probably due to the low number of test cases at different frequencies.

Klein, Vladislav↗

DEVELOPMENT OF A POPULATION BALANCE MODEL TO SIMULATE FRACTIONATION OF GROUND SWITCHGRASS

The population balance model represents a time-dependent formulation of mass conservation for a ground biomass that flows through a set of sieves. The model is suitable for predicting the change in size and distribution of ground biomass while taking into account the flow rate processes of particles through a grinder. This article describes the development and application of this model to a switchgrass grinding operation. The mass conservation formulation of the model contains two parameters: breakage rate and breakage ratio. A laboratory knife mill was modified to act as a batch or flow-through grinder. The ground switchgrass was analyzed over a set of six Tyler sieves with apertures ranging from 5.66 mm (top sieve) to 1 mm (bottom sieve). The breakage rate was estimated from the sieving tests. For estimating the breakage ratio, each of the six fractions was further ground and sieved to 11 fractions on a set of sieves with apertures ranging from 5.66 to 0.25 mm (and pan). These data formed a matrix of values for determining the breakage ratio. Using the two estimated parameters, the transient population balance model was solved numerically. Results indicated that the population balance model generally underpredicted the fractions remaining on sieves with 5.66, 4.00, and 2.83 mm apertures and overpredicted fractions remaining on sieves with 2.00, 1.41, and 1.00 mm apertures. These trends were similar for both the batch and flow-through grinder configurations. The root mean square of residuals (RSE), representing the difference between experimental and simulated mass of fractions, was 0.32 g for batch grinding and 0.1 g for flow-through grinding. The breakage rate exhibited a linear function of the logarithm of particle size, with a regression coefficient of 0.99.

09 BIOMASS FUELS↗

Hybrid NN/SVM Computational System for Optimizing Designs

A computational method and system based on a hybrid of an artificial neural network (NN) and a support vector machine (SVM) (see figure) has been conceived as a means of maximizing or minimizing an objective function, optionally subject to one or more constraints. Such maximization or minimization could be performed, for example, to optimize solve a data-regression or data-classification problem or to optimize a design associated with a response function. A response function can be considered as a subset of a response surface, which is a surface in a vector space of design and performance parameters. A typical example of a design problem that the method and system can be used to solve is that of an airfoil, for which a response function could be the spatial distribution of pressure over the airfoil. In this example, the response surface would describe the pressure distribution as a function of the operating conditions and the geometric parameters of the airfoil. The use of NNs to analyze physical objects in order to optimize their responses under specified physical conditions is well known. NN analysis is suitable for multidimensional interpolation of data that lack structure and enables the representation and optimization of a succession of numerical solutions of increasing complexity or increasing fidelity to the real world. NN analysis is especially useful in helping to satisfy multiple design objectives. Feedforward NNs can be used to make estimates based on nonlinear mathematical models. One difficulty associated with use of a feedforward NN arises from the need for nonlinear optimization to determine connection weights among input, intermediate, and output variables. It can be very expensive to train an NN in cases in which it is necessary to model large amounts of information. Less widely known (in comparison with NNs) are support vector machines (SVMs), which were originally applied in statistical learning theory. In terms that are necessarily oversimplified to fit the scope of this article, an SVM can be characterized as an algorithm that (1) effects a nonlinear mapping of input vectors into a higher-dimensional feature space and (2) involves a dual formulation of governing equations and constraints. One advantageous feature of the SVM approach is that an objective function (which one seeks to minimize to obtain coefficients that define an SVM mathematical model) is convex, so that unlike in the cases of many NN models, any local minimum of an SVM model is also a global minimum.

Rai, Man Mohan↗

Logistic Regression in Clinical Studies

• A logistic regression model is used when the outcome of interest is binary. The term “logistic” refers to the underlying “logit” (log odds) function that is used to model the binary outcome. • Odds ratios are produced from a logistic regression model and have a useful interpretation. • Tips, tricks and concepts used to fit logistic regression models are similar to those used in linear regression models. • Modeling building that is knowledge-based rather than automatic is preferred in most applications of logistic regression. • A logistic regression model that is overparameterized (ie, too many variables for too few events) can result in odds ratios that are implausibly large and confidence intervals that are wide and uninterpretable. These types of “overfitted” models should be avoided. • Logistic regression models can be fit using most standard statistical software.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Sparse Bayesian Regression with Integrated Feature Selection for Nuclear Reactor Analysis

High-dimensional, nonlinear function estimation using large datasets is a current area of interest in the machine learning community, and applications may be found throughout the analytical sciences, where ever-growing datasets are making more information available to the analyst. In this paper, we leverage the existing relevance vector machine, a sparse Bayesian version of the well-studied support vector machine, and expand the method to include integrated feature selection and automatic function shaping. These innovations produce an algorithm that is able to distinguish variables that are useful for making predictions of a response from variables that are unrelated or confusing. We test the technology using synthetic data, conduct initial performance studies, and develop a model capable of making position-independent predictions of the coreaveraged burnup using a single specimen drawn randomly from a nuclear reactor core.

Dayman, Ken J↗

Physics-based dimension reduction in uncertainty quantification for radiative transfer

We present a physics-based methodology for quantifying the uncertainty in a given quantity of interest (QOI) that is contributed by uncertainties in opacities in radiation transport problems. Typically, opacities are tabulated as a function of density, temperature, and photon energy group. The size of this table makes a study of uncertainties at this level challenging because of the well-known 'curse of dimensionality.' We address this by studying uncertain parameters in the underlying physical model that generates the opacity tables. At this level, there are fewer uncertain parameters but still too many to analyze directly through computationally expensive radiation transport simulations. In order to explore this large uncertain parameter space, we develop two simplified radiation transport problems that are much less computationally demanding than the target problem of interest. An emulator is created for each QOI for each simplified problem using Bayesian Multivariate Adaptive Regression Splines (BMARS). This emulator is used to create a functional relationship between the QOIs and the uncertain parameters. Sensitivity analysis is performed using the emulator to determine which parameters contribute significantly to the uncertainty. This physics-based screening process reduces the dimension of the parameter space that is then studied via the computationally expensive radiation transport calculation to generate distributions of quantities of interest. Results of this research demonstrate that the QOIs for the target problem agree for varying screening criteria determined by the sensitivity analysis, and the QOIs agree well for varying Latin Hypercube Design (LHD) sample sizes for the uncertain space. (authors)

97 MATHEMATICS AND COMPUTING↗

Adaptive surrogate modeling by ANOVA and sparse polynomial dimensional decomposition for global sensitivity analysis in fluid simulation

The Polynomial Dimensional Decomposition (PDD) is employed in this work for the global sensitivity analysis and uncertainty quantification (UQ) of stochastic systems subject to a moderate to large number of input random variables. Due to the intimate connection between the PDD and the Analysis of Variance (ANOVA) approaches, PDD is able to provide a simpler and more direct evaluation of the Sobol' sensitivity indices, when compared to the Polynomial Chaos expansion (PC). Unfortunately, the number of PDD terms grows exponentially with respect to the size of the input random vector, which makes the computational cost of standard methods unaffordable for real engineering applications. In order to address the problem of the curse of dimensionality, this work proposes essentially variance-based adaptive strategies aiming to build a cheap meta-model (i.e. surrogate model) by employing the sparse PDD approach with its coefficients computed by regression. Three levels of adaptivity are carried out in this paper: 1) the truncated dimensionality for ANOVA component functions, 2) the active dimension technique especially for second- and higher-order parameter interactions, and 3) the stepwise regression approach designed to retain only the most influential polynomials in the PDD expansion. During this adaptive procedure featuring stepwise regressions, the surrogate model representation keeps containing few terms, so that the cost to resolve repeatedly the linear systems of the least-squares regression problem is negligible. The size of the finally obtained sparse PDD representation is much smaller than the one of the full expansion, since only significant terms are eventually retained. Consequently, a much smaller number of calls to the deterministic model is required to compute the final PDD coefficients.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

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↗

A Multi-Objective Bayesian Optimization Approach Using the Weighted Tchebycheff Method

Abstract Bayesian optimization (BO) is a low-cost global optimization tool for expensive black-box objective functions, where we learn from prior evaluated designs, update a posterior surrogate Gaussian process model, and select new designs for future evaluation using an acquisition function. This research focuses upon developing a BO model with multiple black-box objective functions. In the standard multi-objective (MO) optimization problem, the weighted Tchebycheff method is efficiently used to find both convex and non-convex Pareto frontiers. This approach requires knowledge of utopia values before we start optimization. However, in the BO framework, since the functions are expensive to evaluate, it is very expensive to obtain the utopia values as a prior knowledge. Therefore, in this paper, we develop a MO-BO framework where we calibrate with multiple linear regression (MLR) models to estimate the utopia value for each objective as a function of design input variables; the models are updated iteratively with sampled training data from the proposed MO-BO. These iteratively estimated mean utopia values are used to formulate the weighted Tchebycheff MO acquisition function. The proposed approach is implemented in two numerical test examples and one engineering design problem of optimizing thin tube geometries under constant loading of temperature and pressure, with minimizing the risk of creep-fatigue failure and design cost, along with risk-based and manufacturing constraints. Finally, the model accuracy with frequentist, Bayesian and without MLR-based calibration are compared to true Pareto solutions.

Engineering↗

Accurate parameterization of the kinetic energy functional for calculations using exact-exchange

In this report, electronic structure calculations based on Kohn-Sham density functional theory (KSDFT) that incorporate exact exchange or hybrid functionals are associated with large computational expense, a consequence of the inherent cubic scaling bottleneck and large associated prefactor, which limits the length and time scales that can be accessed. Though orbital-free density functional theory (OFDFT) calculations scale linearly with system size and are associated with significantly smaller prefactor, they are limited by the absence of accurate density-dependent kinetic energy functionals. Therefore, the development of accurate density-dependent kinetic energy functionals is important for OFDFT calculations of large realistic systems. To this end, we propose a method to train kinetic energy functional models at the exact-exchange level of theory by using a dictionary of physically relevant terms that have been proposed in the literature in conjunction with linear or nonlinear regression methods to obtain the fitting coefficients. For our dictionary, we use gradient expansion of the kinetic energy, nonlocal models proposed in the literature and their nonlinear combinations, such as a model that incorporates spatial correlations between higher order derivatives of electron density at two points. The predictive capabilities of these models are assessed by using a variety of model one-dimensional systems that exhibit diverse bonding characteristics, such as a chain of eight hydrogens, LiF, LiH, C 4 H 2 , C 4 N 2 and C 3 O 2 . We show that by using data from model one-dimensional KSDFT calculations performed using the exact-exchange functional for only a few neutral structures, it is possible to generate models with high accuracy for charged systems as well as electron and kinetic energy densities during self-consistent field iterations. In addition, we show that it possible to learn both the orbital dependent terms, i.e., the kinetic energy and the exact-exchange energy, and models that incorporate additional nonlinearities in spatial correlations, such as a quadratic model, are needed to capture subtle features of the kinetic energy density that are present in exact-exchange-based KSDFT calculations.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Scalable learning of potentials to predict time-dependent Hartree–Fock dynamics

We propose a framework to learn the time-dependent Hartree–Fock (TDHF) inter-electronic potential of a molecule from its electron density dynamics. Although the entire TDHF Hamiltonian, including the inter-electronic potential, can be computed from first principles, we use this problem as a testbed to develop strategies that can be applied to learn a priori unknown terms that arise in other methods/approaches to quantum dynamics, e.g., emerging problems such as learning exchange–correlation potentials for time-dependent density functional theory. We develop, train, and test three models of the TDHF inter-electronic potential, each parameterized by a four-index tensor of size up to 60 × 60 × 60 × 60. Two of the models preserve Hermitian symmetry, while one model preserves an eight-fold permutation symmetry that implies Hermitian symmetry. Across seven different molecular systems, we find that accounting for the deeper eight-fold symmetry leads to the best-performing model across three metrics: training efficiency, test set predictive power, and direct comparison of true and learned inter-electronic potentials. All three models, when trained on ensembles of field-free trajectories, generate accurate electron dynamics predictions even in a field-on regime that lies outside the training set. To enable our models to scale to large molecular systems, we derive expressions for Jacobian-vector products that enable iterative, matrix-free training.

97 MATHEMATICS AND COMPUTING↗

STANDARDIZING TYPE Ia SUPERNOVA ABSOLUTE MAGNITUDES USING GAUSSIAN PROCESS DATA REGRESSION

We present a novel class of models for Type Ia supernova time-evolving spectral energy distributions (SEDs) and absolute magnitudes: they are each modeled as stochastic functions described by Gaussian processes. The values of the SED and absolute magnitudes are defined through well-defined regression prescriptions, so that data directly inform the models. As a proof of concept, we implement a model for synthetic photometry built from the spectrophotometric time series from the Nearby Supernova Factory. Absolute magnitudes at peak B brightness are calibrated to 0.13 mag in the g band and to as low as 0.09 mag in the z = 0.25 blueshifted i band, where the dispersion includes contributions from measurement uncertainties and peculiar velocities. The methodology can be applied to spectrophotometric time series of supernovae that span a range of redshifts to simultaneously standardize supernovae together with fitting cosmological parameters.

79 ASTRONOMY AND ASTROPHYSICS↗

Multivariate analysis for the estimation of target localization errors in fiducial marker-based radiotherapy

Purpose: To assess the target localization error (TLE) in terms of the distance between the target and the localization point estimated from the surrogates (|TMD|), the average of respiratory motion for the surrogates and the target (|aRM|), and the number of fiducial markers used for estimating the target (n). Methods: This study enrolled 17 lung cancer patients who subsequently underwent four fractions of real-time tumor tracking irradiation. Four or five fiducial markers were implanted around the lung tumor. The three-dimensional (3D) distance between the tumor and markers was at maximum 58.7 mm. One of the markers was used as the target (P{sub t}), and those markers with a 3D |TMD{sub n}| ≤ 58.7 mm at end-exhalation were then selected. The estimated target position (P{sub e}) was calculated from a localization point consisting of one to three markers except P{sub t}. Respiratory motion for P{sub t} and P{sub e} was defined as the root mean square of each displacement, and |aRM| was calculated from the mean value. TLE was defined as the root mean square of each difference between P{sub t} and P{sub e} during the monitoring of each fraction. These procedures were performed repeatedly using the remaining markers. To provide the best guidance on the answer with n and |TMD|, fiducial markers with a 3D |aRM ≥ 10 mm were selected. Finally, a total of 205, 282, and 76 TLEs that fulfilled the 3D |TMD| and 3D |aRM| criteria were obtained for n = 1, 2, and 3, respectively. Multiple regression analysis (MRA) was used to evaluate TLE as a function of |TMD| and |aRM| in each n. Results: |TMD| for n = 1 was larger than that for n = 3. Moreover, |aRM| was almost constant for all n, indicating a similar scale for the marker’s motion near the lung tumor. MRA showed that |aRM| in the left–right direction was the major cause of TLE; however, the contribution made little difference to the 3D TLE because of the small amount of motion in the left–right direction. The TLE calculated from the MRA ({sup MRA}TLE) increased as |TMD| and |aRM| increased and adversely decreased with each increment of n. The median 3D {sup MRA}TLE was 2.0 mm (range, 0.6–4.3 mm) for n = 1, 1.8 mm (range, 0.4–4.0 mm) for n = 2, and 1.6 mm (range, 0.3–3.7 mm) for n = 3. Although statistical significance between n = 1 and n = 3 was observed in all directions, the absolute average difference and the standard deviation of the {sup MRA}TLE between n = 1 and n = 3 were 0.5 and 0.2 mm, respectively. Conclusions: A large |TMD| and |aRM| increased the differences in TLE between each n; however, the difference in 3D {sup MRA}TLEs was, at most, 0.6 mm. Thus, the authors conclude that it is acceptable to continue fiducial marker-based radiotherapy as long as |TMD| is maintained at ≤58.7 mm for a 3D |aRM| ≥ 10 mm.

60 APPLIED LIFE SCIENCES↗