Search NASA⌕ Search

SEARCH · Search NASA

Results for “nonlinear 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 55 records · Page 3

Time-dependent oral absorption models

The plasma concentration-time profiles following oral administration of drugs are often irregular and cannot be interpreted easily with conventional models based on first- or zero-order absorption kinetics and lag time. Six new models were developed using a time-dependent absorption rate coefficient, ka(t), wherein the time dependency was varied to account for the dynamic processes such as changes in fluid absorption or secretion, in absorption surface area, and in motility with time, in the gastrointestinal tract. In the present study, the plasma concentration profiles of propranolol obtained in human subjects following oral dosing were analyzed using the newly derived models based on mass balance and compared with the conventional models. Nonlinear regression analysis indicated that the conventional compartment model including lag time (CLAG model) could not predict the rapid initial increase in plasma concentration after dosing and the predicted Cmax values were much lower than that observed. On the other hand, all models with the time-dependent absorption rate coefficient, ka(t), were superior to the CLAG model in predicting plasma concentration profiles. Based on Akaike's Information Criterion (AIC), the fluid absorption model without lag time (FA model) exhibited the best overall fit to the data. The two-phase model including lag time, TPLAG model was also found to be a good model judging from the values of sum of squares. This model also described the irregular profiles of plasma concentration with time and frequently predicted Cmax values satisfactorily. A comparison of the absorption rate profiles also suggested that the TPLAG model is better at prediction of irregular absorption kinetics than the FA model. In conclusion, the incorporation of a time-dependent absorption rate coefficient ka(t) allows the prediction of nonlinear absorption characteristics in a more reliable manner.

Non-NASA Center↗

Analysis of nanostructure and nanochemistry by ASAXS: Accessing phase composition of oxyfluoride glass ceramics doped with Er{sup 3+}/Yb{sup 3+}

Here, we describe the analysis of the nanostructure and average chemical compositions of each phase present in an oxyfluoride glass ceramic, which is composed of fluoride nanocrystals and an oxide glass matrix. The overall composition of the oxyfluoride glass ceramic as prepared is 21.1%SiO{sub 2} 6.5%B{sub 2}O{sub 3} 7.0%Al{sub 2}O{sub 3} 21.0%PbF{sub 2} 14.3%CdF{sub 2} 11.0%YbF{sub 3} 0.5%ErF{sub 3} 11.0%PbO 7.6%CdO(mole %). Nanocrystals begin to grow at temperatures above the glass transformation temperature at 678 K as observed by x-ray diffraction. We report results from anomalous small-angle x-ray scattering taken at energies of x-ray absorption edges of Er, Yb, Pb, and Cd. By nonlinear regression of the scattering curves obtained from different edges simultaneously, the nanocrystals were found to be describable as polydisperse spheroids. The length of the smaller axis was found to be 6.4+-1.4 nm while the larger axis was found to be 17.7+-3.9 nm. By analyzing the scattering contrast as a function of the x-ray energy we found cadmium only in the glass matrix.

36 MATERIALS SCIENCE↗

Growth and nonlinear saturation of electromagnetic ion cyclotron waves in multi-ion species magnetospheric plasma

Here, the growth and saturation of electromagnetic ion cyclotron (EMIC) waves is essential to the magnetospheric dynamics. Determining and isolating the effects of multiple ion parameters such as temperatures, anisotropies, and relative abundances is important for quantifying these processes in the magnetospheric plasma. In order to study these process, we utilize a 2.5–D hybrid model (where ions are modeled with the particle–in–cell (PIC) method, and electrons are modeled as background neutralizing fluid) to study the nonlinear electromagnetic wave–particle interactions of hot H + , cold H + , cold He + , and cold or hot O + ions for a broad range of typical magnetospheric parameters. The excitation of EMIC waves is driven by the temperature anisotropy of hot H + in our model. As a result, we quantify the parametric dependence of the linear growth, the nonlinear saturation level of perpendicular magnetic fluctuations, and the temporal evolution of the ion temperature anisotropies. We establish the relation between key plasma parameters and the saturated EMIC wave power, using either power law fits or a nonlinear regression method. We construct the dispersion relation of the waves using the results of the model and investigate the energy content in the various branches of the dispersion ( k ∥ –ω space), showing that the different modes can generate wave power in different regions of k space. We find that large O + concentration reduces the growth and saturated amplitude of the waves; but the waves are less sensitive to the temperature of the O + in the temperature range relevant to the magnetosphere.

79 ASTRONOMY AND ASTROPHYSICS↗

Deep Learning for Rapid Analysis of Spectroscopic Ellipsometry Data

High‐throughput experimental approaches to rapidly develop new materials require high‐throughput data analysis methods to match. Spectroscopic ellipsometry is a powerful method of optical properties characterization, but for unknown materials and/or layer structures the data analysis using traditional methods of nonlinear regression is too slow for autonomous, closed‐loop, high‐throughput experimentation. Herein, three methods (termed spectral, piecewise, and pointwise) of spectroscopic ellipsometry data analysis based on deep learning are introduced and studied. After initial training, the incremental time for inferring optical properties can be a thousand times faster than traditional methods. Results for multilayer sample structures with optically isotropic materials are presented, appropriate for high‐throughput studies of thin films of phase‐change materials such as GeSbTe (GST) alloys. Results for studies on highly birefringent layered materials are also presented, exemplified by the transition metal dichalcogenide MoS 2 . How the materials under test and the experimental objectives may guide the choice of analysis methods are discussed. The utility of our approach is demonstrated by analyzing data measured on a composition spread of GeSbTe phase‐change alloys containing 177 distinct compositions, and identifying the composition with optimal phase‐change figure of merit in only 1.4 s of analysis time.

Li, Yifei↗

Energy modeling and efficiency analysis of aluminum die-casting processes

Energy modeling and efficiency analysis are considered the foundation of manufacturing process optimization to improve quality and efficiency and reduce energy consumption and carbon emissions during aluminum die-casting processes. Our report introduced an energy modeling method to connect gas and electric energy consumption with production rate of aluminum die-casting processes based on data collected at workshops with various combination of machines and products. The detailed modeling process involved the development of a data-acquiring system and the comparison of various kinds of nonlinear regression methods. The resulting models were confirmed with actual production data and were also used to improve production scheduling. Furthermore, it was found that if the modeling results are reasonably used and production is accordingly well-scheduled, 10 to 15% of energy savings could be realized without sacrificing profits.

42 ENGINEERING↗

Transient analysis of advancing contact angle measurements on polished rock surfaces

We report that contact angle measurements for gas-liquid-rock systems are important for modeling multi-phase flow and transport in the subsurface. These data are needed in applications such as the extraction of oil and gas resources, geologic sequestration of carbon dioxide, contaminant fate and transport, and aquifer recharge through the vadose zone. Contact angles are frequently measured with the sessile drop method. Previous research has largely ignored the dynamic behavior of sessile droplets on geologic materials. This study investigates the dynamic behavior of sessile water droplets on prepared rock surfaces in the presence of air. Droplet diameter and advancing contact angle were determined at 0.5 s intervals for ~90 s on flat polished disks of Burlington limestone, Crossville sandstone, Mancos shale, Sierra White granite, Vermilion Bay granite, and Westerly granite using a Krüss DSA 30 Drop Shape Analyzer. The droplet diameter and advancing contact angle data sets were nonlinearly regressed against time using two different two-parameter models. The median coefficients of determination for the fits were 0.85 and 0.96, respectively. The resulting parameter estimates were used to compute the apparent equilibrium contact angle, θ e , for each disk following droplet diameter stabilization. Estimates of θ e ranged from 37.2° for Mancos shale to 75.6° for Burlington limestone. Analysis of variance indicated statistically significant differences in θ e between the rock types at the 95% confidence level. The variability of θe on the polished rock surfaces, as quantified by the coefficient of variation (CV) for θ e , varied between ~3 and ~9%; there were no significant differences in CV between the rock types. Neutron radiography indicated changes droplet morphology over time were due to the spontaneous imbibition of water into the rock matrix. Finally, the transient analysis employed in this study permits a more meaningful estimate of the equilibrium contact angle for rocks than taking the initial value or averaging over time as is frequently done.

58 GEOSCIENCES↗

State space parameterization of explosive eigenvalues during autoignition

Explosive modes such as ignition and extinction are characterized by an eigenvalue of the chemical Jacobian matrix with positive real part, representing the transient instability of chain-branching chemistry and thermal feedback. Formation and eigen-decomposition of the Jacobian matrix are expensive operations whose cost increases cubically with chemical mechanism size. As an alternative to directly computing the eigenvalues of the Jacobian, we explore principal component analysis (PCA) along with nonlinear regression as a methodology to parameterize the eigenvalues by state variables (or linear combinations thereof). We evaluate this modeling strategy using homogeneous autoignition data on two different applications: pseudotransient continuation (Ψtc)-based ODE solvers and chemical explosive mode analysis (CEMA). Results indicate that the PCA-based parameterization of the eigenvalues appears feasible for Ψtc solvers in autoignition calculations over a range of temperatures and pressures. Our results also show that eigenvalue models are capable of tracking sharp discontinuities (such as ignition or extinction) in the eigenvalue for computational flame diagnostics such as CEMA.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Capillary pressure – saturation relationships for gas shales measured using a water activity meter

Hydraulic fracturing of gas shale formations involves pumping a large volume of fracking fluid into a hydrocarbon reservoir to fracture the rock and thus increase its permeability. The majority of the fracking fluid introduced is never recovered and the fate of this lost fluid, often called “leak off,” has become the source of much debate. Information on the capillary pressure – saturation relationship for each wetting phase is needed to simulate leak off using numerical reservoir models. The petroleum industry commonly employs air – water capillary pressure – saturation curves to predict these relationships for mixed wet reservoirs. Traditional methods of measuring this curve are unsuitable for gas shales due to high capillary pressures associated with the small pores present. Still, a possible alternative method is the water activity meter which is used widely in the soil sciences for such measurements. However, its application to lithified material has been limited. Here, this study utilized a water activity meter to measure air – water capillary pressures (ranging from 1.3 to 219.6 MPa) at several water saturation levels in both the wetting and drying directions. Water contents were measured gravimetrically. Seven types of gas producing shale with different porosities (2.5–13.6%) and total organic carbon contents (0.4–13.5%) were investigated. Nonlinear regression was used to fit the resulting capillary pressure – water saturation data pairs for each shale type to the Brooks and Corey equation. Data for six of the seven shale types investigated were successfully fitted (median R 2 = 0.93), indicating this may be a viable method for parameterizing capillary pressure – saturation relationships for inclusion in numerical reservoir models. As expected, the different shale types had statistically different Brooks and Corey parameters. However, there were no significant differences between the Brooks and Corey parameters for the wetting and drying measurements, suggesting that hysteresis may not need to be taken into account in leak off simulations.

54 ENVIRONMENTAL SCIENCES↗

Adsorption of arsenic(III) into modified lamellar Na-magadiite in aqueous medium-Thermodynamic of adsorption process

Synthetic Na-magadiite sample was used for organofunctionalization process with N-propyldiethylenetrimethoxysilane and bis[3-(triethoxysilyl)propyl]tetrasulfide, after expanding the interlayer distance with polar organic solvents such as dimethylsulfoxide (DMSO). The resulted materials were submitted to process of adsorption with arsenic solution at pH 2.0 and 298{+-}1 K. The adsorption isotherms were adjusted using a modified Langmuir equation with regression nonlinear; the net thermal effects obtained from calorimetric titration measurements were adjusted to a modified Langmuir equation. The adsorption process was exothermic ({delta}{sub int}H=-4.15-5.98 kJ mol{sup -1}) accompanied by increase in entropy ({delta}{sub int}S=41.32-62.20 J k{sup -1} mol{sup -1}) and Gibbs energy ({delta}{sub int}G=-22.44-24.56 kJ mol{sup -1}). The favorable values corroborate with the arsenic (III)/basic reactive centers interaction at the solid-liquid interface in the spontaneous process. - Grapical Abstract: The results suggest that the adsorption capacities increased with an increase of reactive basic centers in the pendant organic chains of the intercalated agent.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Temperature-dependent structural studies of mullite-type Bi{sub 2}Fe{sub 4}O{sub 9}

We report on the temperature-dependent structural studies on the mullite-type Bi{sub 2}Fe{sub 4}O{sub 9} compound. The crystal structures were determined using both powder X-ray diffraction and single crystal neutron diffraction. The thermal expansion of the cell parameters from smallest to largest occurred in the order a<c<b; their expansion behavior was fit using either linear or nonlinear regression according to the region of choice. The anisotropic thermal expansion was monitored with respect to the expansion coefficients, anisotropy factor, and polyhedral geometry. At 900 K an abrupt change of the anisotropy factor is observed which correlates with a slope in the BiO{sub 6} polyhedral volume. At the same temperature the FeO{sub 4} tetrahedral distortion starts to increase. The temperature-dependent stereochemically activity of the Bi{sup 3+} 6s{sup 2} lone electron pair was probed by the absolute value of eccentricity parameter. This parameter shows a significant drop at around 773 K. Additionally, the average crystal size and micro-strain of the powder samples were studied. A healing of intrinsic defects is observed with respect to a decreasing micro-strain followed by crystal growth above about 900 K. The temperature-dependent infrared spectra were described in three distinct zones. Deconvolution of the absorption features between 370 cm{sup -1} and 1000 cm{sup -1} required nine fitted bands, and the decomposed Pseudo-Voigt elements were assigned to Fe-O stretching, Fe-O-Fe bending and O-Fe-O bending vibrations. The shift of the observed modes was described with respect to temperature. An effective linewidth parameter {Delta}corr was determined by autocorrelation analysis between 740 cm{sup -1} and 900 cm{sup -1} for each temperature-dependent spectrum. The break of the slope of {Delta}corr at about 773 K was interpreted as the change of some vibrational modes of the FeO{sub 4} tetrahedra. The intrinsic effects observed around 773 K are responsible for extrinsic parameter changes found around 900 K. - Graphical abstract: Temperature-dependent stereochemical activity (eccentricity |{Phi}{sub i}|) of the Bi{sup 3+} 6s{sup 2} lone electron pair directing in the non-bonded region of the Bi-O coordination in the Bi{sub 2}Fe{sub 4}O{sub 9} iron framework (given as polyhedra). Highlights: Black-Right-Pointing-Pointer Complex expansion-contraction mechanism explains anisotropic thermal expansion. Black-Right-Pointing-Pointer Bi 6s{sup 2} lone electron pair sphericity influences analyses by FTIR autocorrelation. Black-Right-Pointing-Pointer Intrinsic effects at 773 K are responsible for extrinsic parameter changes at 900 K.

36 MATERIALS SCIENCE↗

Single toxin dose-response models revisited

The goal of this paper is to offer a rigorous analysis of the sigmoid shape single toxin dose-response relationship. The toxin efficacy function is introduced and four special points, including maximum toxin efficacy and inflection points, on the dose-response curve are defined. The special points define three phases of the toxin effect on mortality: (1) toxin concentrations smaller than the first inflection point or (2) larger then the second inflection point imply low mortality rate, and (3) concentrations between the first and the second inflection points imply high mortality rate. Probabilistic interpretation and mathematical analysis for each of the four models, Hill, logit, probit, and Weibull is provided. Two general model extensions are introduced: (1) the multi-target hit model that accounts for the existence of several vital receptors affected by the toxin, and (2) model with a nonzero mortality at zero concentration to account for natural mortality. Special attention is given to statistical estimation in the framework of the generalized linear model with the binomial dependent variable as the mortality count in each experiment, contrary to the widespread nonlinear regression treating the mortality rate as continuous variable. The models are illustrated using standard EPA Daphnia acute (48 h) toxicity tests with mortality as a function of NiCl or CuSO{sub 4} toxin. - Highlights: • The paper offers a rigorous study of a sigmoid dose-response relationship. • The concentration with highest mortality rate is rigorously defined. • A table with four special points for five morality curves is presented. • Two new sigmoid dose-response models have been introduced. • The generalized linear model is advocated for estimation of sigmoid dose-response relationship.

60 APPLIED LIFE SCIENCES↗

Machine Learning for Quantum Mechanical Properties of Atoms in Molecules

Here we introduce machine learning models of quantum mechanical observables of atoms in molecules. Instant out-of-sample predictions for proton and carbon nuclear chemical shifts, atomic core level excitations, and forces on atoms reach accuracies on par with density functional theory reference. Locality is exploited within nonlinear regression via local atom-centered coordinate systems. The approach is validated on a diverse set of 9 k small organic molecules. Linear scaling of computational cost in system size is demonstrated for saturated polymers with up to submesoscale lengths.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Data-Efficient Methods for Determining Flory–Huggins χ Parameters in Multicomponent Polymer Formulations

Polymer formulations are essential in diverse applications including personal care products, coatings, paints, adhesives, and plastic materials. Designing these formulations requires navigating large, complex design spaces, where phase and self-assembly behavior critically impact performance. The Flory–Huggins χ parameter, which quantifies segmental miscibility, is widely used to parametrize the excess free energy of mixing in formulation models. In this work, we introduce two data-efficient, top-down methods for estimating χ parameters using the Random Phase Approximation (RPA): (i) Boundary Nonlinear Regression (Boundary-NLR), which fits theoretical spinodal boundaries to experimental phase boundaries, and (ii) Surrogate Model Inverse Parameter Estimation (SMIPE), which uses a Gaussian Process Classifier to fit sparse phase maps via a surrogate model. Both methods allow rapid parametrization of polymer field-theoretic models without the need for additional experiments. We evaluate these approaches on data sets involving polymer–solvent–nonsolvent ternary mixtures and block copolymer–solvent systems, demonstrating their robustness to experimental noise and their relevance for real-world formulation design.

copolymers↗

A Kinetic and Mass Transfer Model for Glycerol Hydrogenolysis in a Trickle-Bed Reactor

A detailed model of glycerol hydrogenolysis in a trickle-bed reactor is presented that includes a mechanistically based kinetic rate expression, energy transport, mass transport across the gas-liquid and liquid-solid interfaces, intraparticle catalyst mass transfer, and partial wetting of the bed. Optimal kinetic parameters for the glycerol hydrogenolysis rate expression were determined via nonlinear regression analysis on the basis of experiments conducted in a laboratory-scale trickle-bed reactor over a broad range of operating conditions. Model predictions agree well with experimental data and accurately predict trends in reactor performance with liquid flow rate, temperature, hydrogen pressure, and base promoter concentration. The model is thus a useful tool for predicting laboratory reactor performance and for design of commercial-scale trickle-bed systems.

08 HYDROGEN↗

New Collector Efficiency Equation for Colloid Filtration in Both Natural and Engineered Flow Conditions

new equation for the collector efficiency (η) of the colloid filtration theory (CFT) is developed via nonlinear regression on the numerical data generated by a large number of Lagrangian simulations conducted in Happel's sphere-in-cell porous media model over a wide range of environmentally relevant conditions. The new equation expands the range of CFT's applicability in the natural subsurface primarily by accommodating departures from power law dependence of η on the Peclet and gravity numbers, a necessary but as of yet unavailable feature for applying CFT to large-scale field transport (e.g., of nanoparticles, radionuclides, or genetically modified organisms) under low groundwater velocity conditions. The new equation also departs from prior equations for colloids in the nanoparticle size range at all fluid velocities. These departures are particularly relevant to subsurface colloid and colloid-facilitated transport where low permeabilities and/or hydraulic gradients lead to low groundwater velocities and/or to nanoparticle fate and transport in porous media in general. We also note the importance of consistency in the conceptualization of particle flux through the single collector model on which most η equations are based for the purpose of attaining a mechanistic understanding of the transport and attachment steps of deposition. A lack of sufficient data for small particles and low velocities warrants further experiments to draw more definitive and comprehensive conclusions regarding the most significant discrepancies between the available equations.

Environmental Molecular Sciences Laboratory↗

Synchrotron radiation-based far-infrared spectroscopic ellipsometer with full Mueller-matrix capability

We developed far-IR spectroscopic ellipsometer at the U4IR beamline of the National Synchrotron Light Source in Brookhaven National Laboratory. This ellipsometer is able to measure both, rotating analyzer and full-Mueller matrix spectra using rotating retarders, and wire-grid linear polarizers. We utilize exceptional brightness of synchrotron radiation in the broad spectral range between about 20 and 4000 cm{sup -1}. Fourier-transform infrared (FT-IR) spectrometer is used for multi-wavelength data acquisition. The sample stage has temperature variation between 4.2 and 450 K, wide range of {theta}-2{theta} angular rotation, {chi} tilt angle adjustment, and X-Y-Z translation. A LabVIEW-based software controls the motors, sample temperature, and FT-IR spectrometer and also allows to run fully automated experiments with pre-programmed measurement schedules. Data analysis is based on Berreman's 4 Multiplication-Sign 4 propagation matrix formalism to calculate the Mueller matrix parameters of anisotropic samples with magnetic permeability {mu}{ne} 1. A nonlinear regression of the rotating analyzer ellipsometry and/or Mueller matrix (MM) spectra, which are usually acquired at variable angles of incidence and sample crystallographic orientations, allows extraction of dielectric constant and magnetic permeability tensors for bulk and thin-film samples. Applications of this ellipsometer setup for multiferroic and ferrimagnetic materials with {mu}{ne} 1 are illustrated with experimental results and simulations for TbMnO{sub 3} and Dy{sub 3}Fe{sub 5}O{sub 12} single crystals. We demonstrate how magnetic and electric dipoles, such as magnons and phonons, can be distinguished from a single MM measurement without adducing any modeling arguments. The parameters of magnetoelectric components of electromagnon excitations are determined using MM spectra of TbMnO{sub 3}.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Multi-species temperature and number density analysis of a laser-produced plasma using dual-comb spectroscopy

Dual-comb spectroscopy (DCS) represents a novel method of using absorption spectroscopy as a diagnostic tool for multispecies analysis of excitation temperatures and column densities in laser-produced plasmas (LPPs). DCS was performed on a LPP generated by ablating a multielement alloy containing Nd, Gd, and Fe. Transitions from all three elements were observed in absorption spectra measured from 530.08 to 535.19 nm at seven time-delays from 31 to 250 μs after ablation. The spectra were fit using a nonlinear regression algorithm to determine peak areas, and excitation temperatures and column densities were determined for the three atomic species separately using Boltzmann plots. The measured excitation temperatures of Nd I and Gd I showed good agreement at all time-delays, whereas the Fe I temperature was found to be higher, and the ratios between the column densities varied with delay. The observations are understood via effects of LPP spatial averaging, elemental fractionation, and molecular formation and are compared and contextualized with previous work studying LPPs using other spectroscopic techniques. A brief discussion of the precision and accuracy of the determined excitation temperatures and column densities is also presented.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Machine learning without a processor: Emergent learning in a nonlinear analog network

Standard deep learning algorithms require differentiating large nonlinear networks, a process that is slow and power-hungry. Electronic contrastive local learning networks (CLLNs) offer potentially fast, efficient, and fault-tolerant hardware for analog machine learning, but existing implementations are linear, severely limiting their capabilities. These systems differ significantly from artificial neural networks as well as the brain, so the feasibility and utility of incorporating nonlinear elements have not been explored. Here, we introduce a nonlinear CLLN—an analog electronic network made of self-adjusting nonlinear resistive elements based on transistors. We demonstrate that the system learns tasks unachievable in linear systems, including XOR (exclusive or) and nonlinear regression, without a computer. We find our decentralized system reduces modes of training error in order (mean, slope, curvature), similar to spectral bias in artificial neural networks. The circuitry is robust to damage, retrainable in seconds, and performs learned tasks in microseconds while dissipating only picojoules of energy across each transistor. This suggests enormous potential for fast, low-power computing in edge systems like sensors, robotic controllers, and medical devices, as well as manufacturability at scale for performing and studying emergent learning.

Science & Technology - Other Topics↗