SEARCH · Search NASA

Results for “Model fitting”

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 1,225 records · Page 68

All sources NASA NTRS DOE OSTI

DOE OSTI2018

Measurement of the Se 80 ( γ , n ) reaction with linearly polarized γ rays

This paper reports the results of the first measurements of the differential cross section of the 80 Se(γ , n) 79 Se reaction with a linearly polarized gamma-ray (γ-ray) beam. The cross section was measured at three incident γ-ray beam energies: 15.6, 15.8, and 16.0 MeV, with a beam energy spread of 3.0% full width at half-maximum (FWHM). The differential cross section for the excitation spectrum in 79Se was measured at two scattering angles in the plane of the beam polarization: θ = 90° and 135°, and at one angle in the plane perpendicular to the plane of polarization: θ = 90°. The total photoneutron cross sections determined from these data are between 0.8 and 1.3 standard deviations smaller than previously published results. Here, the excitation spectra measured in this work were fit with a Hauser-Feshbach model. Better fits to the data were obtained with a constant-temperature formulation of the nuclear level density (NLD) than with a Fermi-gas NLD model. The parameters for the constant-temperature NLD model obtained in this work are consistent with those obtained for medium-mass nuclei in previous studies.

59 ≤ A ≤ 89↗

DOE OSTI2026

OmicsMLMentor: A Web Application for Guided Machine Learning Analysis of Omics Data

Expression-based omics technologies (e.g. proteomics, metabolomics, transcriptomics, etc.) increasingly rely on supervised and unsupervised machine learning (ML) models to find key biomolecules distinguishing conditions, identify natural groupings in biological data, or generate predictions for outcomes of interest. Fitting ML models to omics data presents several challenges, including handling missing data, selecting a normalization method, choosing a valid model, and optimizing hyperparameters, all requiring statistical programming skills to address these challenges. Thus, the open-source web application SLOPE was designed to lower the barrier to ML modeling for omics data. SLOPE supports the fitting of 15 ML models (10 supervised and 5 unsupervised) tailored to omics datasets, such as proteomics, metabolomics, lipidomics, and transcriptomics. SLOPE offers several omics-specific features, including methods for handling missingness (imputation, conversion, removal), normalization tests, ranking of models based on the structure of a user’s data and user input, and optimal hyperparameter selections using cross-validation splits. By streamlining ML workflows for omics analysis, SLOPE address critical gaps in existing online web tools, facilitating a broader adoption of these models for omics research. Here, SLOPE is applied to data from a lignin exposure study to highlight the workflow for fitting both supervised and unsupervised models to data.

lipidomics↗

NASA NTRS2012

Modeling and Maximum Likelihood Fitting of Gamma-Ray and Radio Light Curves of Millisecond Pulsars Detected with Fermi

Pulsed gamma rays have been detected with the Fermi Large Area Telescope (LAT) from more than 20 millisecond pulsars (MSPs), some of which were discovered in radio observations of bright, unassociated LAT sources. We have fit the radio and gamma-ray light curves of 19 LAT-detected MSPs in the context of geometric, outermagnetospheric emission models assuming the retarded vacuum dipole magnetic field using a Markov chain Monte Carlo maximum likelihood technique. We find that, in many cases, the models are able to reproduce the observed light curves well and provide constraints on the viewing geometries that are in agreement with those from radio polarization measurements. Additionally, for some MSPs we constrain the altitudes of both the gamma-ray and radio emission regions. The best-fit magnetic inclination angles are found to cover a broader range than those of non-recycled gamma-ray pulsars.

Johnson, T. J.↗

DOE OSTI2021

Understanding the extreme luminosity of DES14X2fna

We present DES14X2fna, a high-luminosity, fast-declining Type IIb supernova (SN IIb) at redshift z = 0.0453, detected by the Dark Energy Survey (DES). DES14X2fna is an unusual member of its class, with a light curve showing a broad, luminous peak reaching M r ≃ –19.3 mag 20 d after explosion. This object does not show a linear decline tail in the light curve until ≃60 d after explosion, after which it declines very rapidly (4.30 ± 0.10 mag 100 d –1 in the r band). By fitting semi-analytic models to the photometry of DES14X2fna, we find that its light curve cannot be explained by a standard 56 Ni decay model as this is unable to fit the peak and fast tail decline observed. Inclusion of either interaction with surrounding circumstellar material or a rapidly-rotating neutron star (magnetar) significantly increases the quality of the model fit. Furthermore, we also investigate the possibility for an object similar to DES14X2fna to act as a contaminant in photometric samples of SNe Ia for cosmology, finding that a similar simulated object is misclassified by a recurrent neural network (RNN)-based photometric classifier as an SN Ia in ~1.1–2.4 per cent of cases in DES, depending on the probability threshold used for a positive classification.

79 ASTRONOMY AND ASTROPHYSICS↗

NASA NTRS1995

Lithospheric bending at subduction zones based on depth soundings and satellite gravity

A global study of trench flexure was performed by simultaneously modeling 117 bathymetric profiles (original depth soundings) and satellite-derived gravity profiles. A thin, elastic plate flexure model was fit to each bathymetry/gravity profile by minimization of the L(sub 1) norm. The six model parameters were regional depth, regional gravity, trench axis location, flexural wavelength, flexural amplitude, and lithospheric density. A regional tilt parameter was not required after correcting for age-related trend using a new high-resolution age map. Estimates of the density parameter confirm that most outer rises are uncompensated. We find that flexural wavelength is not an accurate estimate of plate thickness because of the high curvatures observed at a majority of trenches. As in previous studies, we find that the gravity data favor a longer-wavelength flexure than the bathymetry data. A joint topography-gravity modeling scheme and fit criteria are used to limit acceptable parameter values to models for which topography and gravity yield consistent results. Even after the elastic thicknesses are converted to mechanical thicknesses using the yield strength envelope model, residual scatter obscures the systematic increase of mechanical thickness with age; perhaps this reflects the combination of uncertainties inherent in estimating flexural wavelength, such as extreme inelastic bending and accumulated thermoelastic stress. The bending moment needed to support the trench and outer rise topography increases by a factor of 10 as lithospheric age increases from 20 to 150 Ma; this reflects the increase in saturation bending moment that the lithosphere can maintain. Using a stiff, dry-olivine rheology, we find that the lithosphere of the GDH1 thermal model (Stein and Stein, 1992) is too hot and thin to maintain the observed bending moments. Moreover, the regional depth seaward of the oldest trenches (approximately 150 Ma) exceeds the GDH1 model depths by about 400 m.

Levitt, Daniel A.↗

NASA NTRS1961

A Study of the Free Oscillations of the Earth

Published observations on the toroidal oscillations of the earth are critically reviewed. A supplementary analysis of the record obtained by the Lamont strain seismometer is presented. Eleven toroidal modes are identified, and it is concluded that the periods are known to within 1 per cent. A perturbation scheme involving the ratio of the angular velocity of the earth to the resonant frequency is used in calculating the effects due to the rotation of the earth on the resonant frequency. The free oscillations are viewed as a superposition of traveling waves. In a nonrotating system two traveling waves combine to produce a stationary standing wave. In a rotating system, the rotation distinguishes between waves that travel in the direction of rotation and those that travel in the opposite direction. Rotation removes a degeneracy and results in a splitting of a spectral peak of order l into 2 times l plus 1 peaks. The fractional displacement in frequency for the lowest-order toroidal oscillations is 1/206 and of the same order as the Q of the peak, so that splitting will probably not be observed in the toroidal oscillations. Viewed locally, rotation causes a particle to precess about a direction parallel to the axis of rotation. This precession will cause a variation of amplitude with time if the motion is recorded by an instrument with an anisotropic response function. Care is therefore needed in studying the time decay of a given spectral peak. Rotation also couples the normal coordinates so that a motion that is initially purely horizontal will develop a vertical component. It is expected that vertical seismometers should record particle motion with the toroidal frequencies. The perturbations of the toroidal oscillations due to core-mantle interaction are treated in detail. An exact expression is obtained for the rate of energy dissipated by a finitely conducting plate oscillating across a magnetic field. The energy dissipated at the core-mantle boundary due to viscous and hydromagnetic coupling is shown to be insignificant as compared with the energy dissipated within the mantle. The toroidal magnetic field leaking into the lower mantle combines with the dipole field, resulting in a stress on the mantle, tending to stiffen the lower boundary. The stress is of sufficient magnitude to produce a displacement toward higher frequency in the lower-order toroidal oscillations. Observations on the (sub 0) T (sub 2) oscillations lead to an estimate of the toroidal magnetic field in the lower mantle. A calculation of elastic energy in the low-order oscillations suggests a value of 10 (sup 18) ergs per cycles per hour for the energy density at low frequencies in the Chilean earthquake. Each mode of oscillation has a characteristic radial distribution of elastic energy associated with it. This distribution determines which parts of the earth contribute most heavily in determining a particular resonant frequency. The distribution of energy for the lower 17 modes for a homogeneous and a Gutenberg model earth is calculated. The resonant frequencies for models of the earth based on the Gutenberg and Lehmann distribution of elastic properties are presented. It is shown that the Gutenberg model earth fits the observations more closely than the Lehmann model and that a slight alteration of the Gutenberg model gives a significantly better fit to the observations. The alteration involves a lower shear-wave velocity in the lower mantle while the Gutenberg velocity distribution is maintained in the upper mantle. Various studies of the earth's oscillations coupled with surface-wave investigations substantiate Gutenberg's hypothesis of a layer of low velocity in the upper mantle. The physical conditions required for the formation of a region of low velocity are examined in detail. The results confirm Birch's earlier statement that a temperature gradient in excess of 6 degrees to 7 degrees per kilometer is needed to produce a decrease in velocity. The low-velocity layer does not require that the temperature approach or exceed the melting temperature. If tile upper mantle is homogeneous, the region of lower velocity should commence at the base of the crust and extend to 150 kilometers under the oceans and about 100 kilometers under continental regions. The distribution of thermal conductivity and radioactivity consistent with the low-velocity layer is also considered.

MacDonald, Gordon J. F.↗

DOE OSTI2021

Identifiability and predictability of integer- and fractional-order epidemiological models using physics-informed neural networks

Here we analyze a plurality of epidemiological models through the lens of physics-informed neural networks (PINNs) that enable us to identify time-dependent parameters and data-driven fractional differential operators. In particular, we consider several variations of the classical susceptible-infectious-removed (SIR) model by introducing more compartments and fractional-order and time-delay models. We report the results for the spread of COVID-19 in New York City, Rhode Island and Michigan states and Italy, by simultaneously inferring the unknown parameters and the unobserved dynamics. For integer-order and time-delay models, we fit the available data by identifying time-dependent parameters, which are represented by neural networks. In contrast, for fractional differential models, we fit the data by determining different time-dependent derivative orders for each compartment, which we represent by neural networks. We investigate the structural and practical identifiability of these unknown functions for different datasets, and quantify the uncertainty associated with neural networks and with control measures in forecasting the pandemic.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2015

Determining neutron star properties by fitting oblate-star waveform models to X-ray burst oscillations

We describe sophisticated new Bayesian analysis methods that make it possible to estimate quickly the masses and radii of rapidly rotating, oblate neutron stars by fitting oblate-star waveform models to energy-resolved observations of the waveforms of X-ray burst oscillations produced by such stars. We find that a 25% variation of the temperature of the hot spot in the north-south direction does not significantly bias estimates of the mass M and equatorial radius R{sub eq} derived by fitting a model that assumes a uniform-temperature spot. Our results show that fits of oblate-star waveform models to waveform data can simultaneously determine M and R{sub eq} with 1σ uncertainties ≲7% if (1) the star's rotation rate is ≳ 600 Hz; (2) the spot center and observer's sightline are both within 30{sup ∘} of the star's rotational equator; and (3) ∼10{sup 7} counts are collected during burst oscillations that have a fractional rms amplitude ≳10%. A fractional amplitude of ∼10% is realistic, and the accepted NICER and proposed LOFT and AXTAR space missions could collect 10{sup 7} counts from a single star by combining data from many X-ray bursts from the star. Uncertainties ≲7% are small enough to improve substantially our understanding of cold, ultradense matter.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2023

Long-term sinking of nuclear waste canisters in salt formations by low-stress creep at high temperature

Abstract Rock salt has a self-sealing capacity, low permeability, and high thermal conductivity, making it a potential host for heat-generating nuclear waste. The feasibility of nuclear waste disposal within salt formations has been investigated mostly for small-sized canisters. Geologic disposal of larger-sized canisters originally designed for spent fuel storage and transportation has lately been examined as a cost-effective alternative. This raises questions about their long-term vertical movement due to their weight and high decay heat. Low-stress creep governs this movement; however, most salt constitutive models do not incorporate it. In this paper, the Norton and the WIPP creep models are compared with the Lux/Wolters/Lerche (LWL) model and a simpler model that combines linear and Norton creep laws (named combined creep model). The LWL and combined creep models consider pressure solution creep, though all incorporate dislocation creep. The models are first applied to creep tests under various stress levels. The LWL and the combined creep models results fit the experimental data well in both high and low stress ranges, whereas the Norton and WIPP models results only fit in higher-stress ranges. The different models are further applied for analyzing long-term canister movement. A sinking rate of $$-4.4\times 10^{-7}$$ - 4.4 × 10 - 7 mm/year was predicted using the Norton and WIPP models versus $$-2.1\times 10^{-2}$$ - 2.1 × 10 - 2 mm/year and $$-3.1\times 10^{-2}$$ - 3.1 × 10 - 2 mm/year using the LWL and the combined creep models, respectively. This comparative study confirms that creep models calibrated exclusively against high-deviatoric stress data might result in an inaccurate estimation of waste packages sinking rate in salt formations.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

DOE OSTI2015

Automatic Segmentation of the Eye in 3D Magnetic Resonance Imaging: A Novel Statistical Shape Model for Treatment Planning of Retinoblastoma

Purpose: Proper delineation of ocular anatomy in 3-dimensional (3D) imaging is a big challenge, particularly when developing treatment plans for ocular diseases. Magnetic resonance imaging (MRI) is presently used in clinical practice for diagnosis confirmation and treatment planning for treatment of retinoblastoma in infants, where it serves as a source of information, complementary to the fundus or ultrasonographic imaging. Here we present a framework to fully automatically segment the eye anatomy for MRI based on 3D active shape models (ASM), and we validate the results and present a proof of concept to automatically segment pathological eyes. Methods and Materials: Manual and automatic segmentation were performed in 24 images of healthy children's eyes (3.29 ± 2.15 years of age). Imaging was performed using a 3-T MRI scanner. The ASM consists of the lens, the vitreous humor, the sclera, and the cornea. The model was fitted by first automatically detecting the position of the eye center, the lens, and the optic nerve, and then aligning the model and fitting it to the patient. We validated our segmentation method by using a leave-one-out cross-validation. The segmentation results were evaluated by measuring the overlap, using the Dice similarity coefficient (DSC) and the mean distance error. Results: We obtained a DSC of 94.90 ± 2.12% for the sclera and the cornea, 94.72 ± 1.89% for the vitreous humor, and 85.16 ± 4.91% for the lens. The mean distance error was 0.26 ± 0.09 mm. The entire process took 14 seconds on average per eye. Conclusion: We provide a reliable and accurate tool that enables clinicians to automatically segment the sclera, the cornea, the vitreous humor, and the lens, using MRI. We additionally present a proof of concept for fully automatically segmenting eye pathology. This tool reduces the time needed for eye shape delineation and thus can help clinicians when planning eye treatment and confirming the extent of the tumor.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

DOE OSTI2016

Foreground bias from parametric models of far-IR dust emission

We use simple toy models of far-IR dust emission to estimate the accuracy to which the polarization of the cosmic microwave background can be recovered using multi-frequency fits, if the parametric form chosen for the fitted dust model differs from the actual dust emission. Commonly used approximations to the far-IR dust spectrum yield CMB residuals comparable to or larger than the sensitivities expected for the next generation of CMB missions, despite fitting the combined CMB + foreground emission to precision 0.1% or better. The Rayleigh–Jeans approximation to the dust spectrum biases the fitted dust spectral index by Δβ{sub d}=0.2 and the inflationary B-mode amplitude by Δr=0.03. Fitting the dust to a modified blackbody at a single temperature biases the best-fit CMB by Δr>0.003 if the true dust spectrum contains multiple temperature components. A 13-parameter model fitting two temperature components reduces this bias by an order of magnitude if the true dust spectrum is in fact a simple superposition of emission at different temperatures, but fails at the level Δr=0.006 for dust whose spectral index varies with frequency. Restricting the observing frequencies to a narrow region near the foreground minimum reduces these biases for some dust spectra but can increase the bias for others. Data at THz frequencies surrounding the peak of the dust emission can mitigate these biases while providing a direct determination of the dust temperature profile.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2021

GLEAM: Galaxy Line Emission & Absorption Modeling

We present Galaxy Line Emission & Absorption Modeling (gleam), a Python tool for fitting Gaussian models to emission and absorption lines in large samples of 1D extragalactic spectra. gleam is tailored to work well in batch mode without much human interaction. With gleam, users can uniformly process a variety of spectra, including galaxies and active galactic nuclei, in a wide range of instrument setups and signal-to-noise regimes. gleam also takes advantage of multiprocessing capabilities to process spectra in parallel. With the goal of enabling reproducible workflows for its users, gleam employs a small number of input files, including a central, user-friendly configuration in which fitting constraints can be defined for groups of spectra and overrides can be specified for edge cases. For each spectrum, gleam produces a table containing measurements and error bars for the detected spectral lines and continuum and upper limits for nondetections. For visual inspection and publishing, gleam can also produce plots of the data with fitted lines overlaid. In the present paper, we describe gleam’s main features, the necessary inputs, expected outputs, and some example applications, including thorough tests on a large sample of optical/infrared multi-object spectroscopic observations and integral field spectroscopic data. gleam is developed as an open-source project hosted at https://github.com/multiwavelength/gleam and welcomes community contributions.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2014

Free kick instead of cross-validation in maximum-likelihood refinement of macromolecular crystal structures

The maximum-likelihood free-kick target, which calculates model error estimates from the work set and a randomly displaced model, proved superior in the accuracy and consistency of refinement of crystal structures compared with the maximum-likelihood cross-validation target, which calculates error estimates from the test set and the unperturbed model. The refinement of a molecular model is a computational procedure by which the atomic model is fitted to the diffraction data. The commonly used target in the refinement of macromolecular structures is the maximum-likelihood (ML) function, which relies on the assessment of model errors. The current ML functions rely on cross-validation. They utilize phase-error estimates that are calculated from a small fraction of diffraction data, called the test set, that are not used to fit the model. An approach has been developed that uses the work set to calculate the phase-error estimates in the ML refinement from simulating the model errors via the random displacement of atomic coordinates. It is called ML free-kick refinement as it uses the ML formulation of the target function and is based on the idea of freeing the model from the model bias imposed by the chemical energy restraints used in refinement. This approach for the calculation of error estimates is superior to the cross-validation approach: it reduces the phase error and increases the accuracy of molecular models, is more robust, provides clearer maps and may use a smaller portion of data for the test set for the calculation of R{sub free} or may leave it out completely.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

NASA NTRS2016

Foreground Bias from Parametric Models of Far-IR Dust Emission

We use simple toy models of far-IR dust emission to estimate the accuracy to which the polarization of the cosmic microwave background can be recovered using multi-frequency fits, if the parametric form chosen for the fitted dust model differs from the actual dust emission. Commonly used approximations to the far-IR dust spectrum yield CMB residuals comparable to or larger than the sensitivities expected for the next generation of CMB missions, despite fitting the combined CMB plus foreground emission to precision 0.1 percent or better. The Rayleigh-Jeans approximation to the dust spectrum biases the fitted dust spectral index by (Delta)(Beta)(sub d) = 0.2 and the inflationary B-mode amplitude by (Delta)(r) = 0.03. Fitting the dust to a modified blackbody at a single temperature biases the best-fit CMB by (Delta)(r) greater than 0.003 if the true dust spectrum contains multiple temperature components. A 13-parameter model fitting two temperature components reduces this bias by an order of magnitude if the true dust spectrum is in fact a simple superposition of emission at different temperatures, but fails at the level (Delta)(r) = 0.006 for dust whose spectral index varies with frequency. Restricting the observing frequencies to a narrow region near the foreground minimum reduces these biases for some dust spectra but can increase the bias for others. Data at THz frequencies surrounding the peak of the dust emission can mitigate these biases while providing a direct determination of the dust temperature profile.

cosmic microwave background↗

DOE OSTI2019

Kinetic Modeling of the Silicothermic Reduction of Manganese Oxide from Slag

A model was developed to describe the mixed rate control kinetics in systems with MnO-SiO{sub 2}-CaO-Al{sub 2}O{sub 3} slags and Fe-Si metal droplets. During the reaction of Fe-Si droplets and slag, Mn{sup 2+} transport was found to be part of controlling the system but could not sufficiently describe all initial conditions (chiefly, changes to initial silicon content). The current model describes the kinetics of the stated system and offers answers to the question of rate control; the model has been fitted to nine datasets of varying initial conditions including initial [Si], initial (MnO), initial droplet size, and reaction temperature. The fitted mass transfer coefficients for metal and slag were 2.3 × 10{sup −4} and 6.7 × 10{sup −4} m/s, respectively; these values are constant across the nine datasets. Previous claims about the efficiency of silicon usage in reducing manganese have been modified; it appears that the formation of silica is favored throughout the reaction, but that the formation of a silicon monoxide gas layer on the metal surface dramatically impacts the rate of reaction. As a measure of overall fit of the model, the average of the root-mean-square errors for all datasets is 14 pct. Mass transport in slag is twice as influential on rate control as the metal phase. The simultaneous transport of both Mn{sup 2+} and silicate dimers controls mass transport in the slag. Both [Si] and [Mn] can control from the metal side but are dependent on the initial conditions.

36 MATERIALS SCIENCE↗

DOE OSTI2015

A thermodynamic model to predict the aqueous solubility of hydrocarbon mixtures at two-phase hydrate-liquid water equilibrium

In order to understand the fate and transport of hydrocarbons and hydrocarbon mixtures in the deep sea and underlying sediments it requires an accurate determination of the two-phase hydrate (H)-liquid water (L w ) thermodynamic equilibrium in the absence of a free gas phase. In addition to controlling hydrate formation directly from the aqueous phase, the H-L w equilibrium provides the aqueous solubility of dissolving hydrate. The aqueous solubility of hydrocarbon mixture at H-L w equilibrium was calculated based on the van der Waals and Platteeuw statistical thermodynamic model and the Holder model. Langmuir constants were calculated from cell potential parameters obtained from ab initio intermolecular potentials and thus the model contains no fitting parameters. The model accurately predicts pure methane, ethane, and propane experimental H-L w equilibrium solubility data, including new data provided here for methane. Here, we present hydrocarbon solubilities in water and seawater at H-L w equilibrium at deep sea conditions for both methane and a thermogenic mixture based on the methane–ethane–propane ratio from the 2010 Macondo oil spill. We also present model predictions of the proportion of occupied large and small cages at H-L w equilibrium for both systems, as well as hydrocarbon ratios in the hydrate phase for the thermogenic mixture.

58 GEOSCIENCES↗

DOE OSTI2014

Random Forests to Predict Rectal Toxicity Following Prostate Cancer Radiation Therapy

Purpose: To propose a random forest normal tissue complication probability (RF-NTCP) model to predict late rectal toxicity following prostate cancer radiation therapy, and to compare its performance to that of classic NTCP models. Methods and Materials: Clinical data and dose-volume histograms (DVH) were collected from 261 patients who received 3-dimensional conformal radiation therapy for prostate cancer with at least 5 years of follow-up. The series was split 1000 times into training and validation cohorts. A RF was trained to predict the risk of 5-year overall rectal toxicity and bleeding. Parameters of the Lyman-Kutcher-Burman (LKB) model were identified and a logistic regression model was fit. The performance of all the models was assessed by computing the area under the receiving operating characteristic curve (AUC). Results: The 5-year grade ≥2 overall rectal toxicity and grade ≥1 and grade ≥2 rectal bleeding rates were 16%, 25%, and 10%, respectively. Predictive capabilities were obtained using the RF-NTCP model for all 3 toxicity endpoints, including both the training and validation cohorts. The age and use of anticoagulants were found to be predictors of rectal bleeding. The AUC for RF-NTCP ranged from 0.66 to 0.76, depending on the toxicity endpoint. The AUC values for the LKB-NTCP were statistically significantly inferior, ranging from 0.62 to 0.69. Conclusions: The RF-NTCP model may be a useful new tool in predicting late rectal toxicity, including variables other than DVH, and thus appears as a strong competitor to classic NTCP models.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

DOE OSTI2018

Impact of hydrate saturation on water permeability in hydrate-bearing sediments

Permeability of the hydrate-bearing sediments vitally affects the hydrate dissociation process as well as the rate and efficiency of gas production. Reported permeability values are observed to be widely are very scattered owing to the dependence on multiple factors such as experimental conditions and test procedures. It is critical that the permeability is measured accurately to enable prediction of long-term gas production using numerical simulation for an economic development of hydrate-bearing reservoirs. In this study, the tetrahydrofuran (THF) hydrates that exhibit pore habits of the pore-filling pattern are formed in sediments, and the water permeability is quantified as a function of hydrate saturation. The results reveal that a decrease in the permeability for an increase in the hydrate saturation. Based on the measured values that are bounded by Kozeny grain-coating and pore-filling models, the fitting parameters of the empirical permeability models are indicated for use in numerical simulation of long-term hydrate dissociation. Wave velocity measurements reveal pore habits of pore-filling pattern of THF hydrates in the sediments with S h < 0.5, and a gradual transition to patchy and load-bearing pore-habits when S h > 0.6. The numerical simulation results obtained using the complementary pore-network model suggest that the THF hydrates in the sediments may form in small clusters with an average patch size of ~4 pores.

58 GEOSCIENCES↗