Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Models, Statistical”

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 595 records · Page 33
All sourcesNASA NTRSDOE OSTI
DOE OSTI2015

A NEW METHOD FOR DERIVING THE STELLAR BIRTH FUNCTION OF RESOLVED STELLAR POPULATIONS

We present a new method for deriving the stellar birth function (SBF) of resolved stellar populations. The SBF (stars born per unit mass, time, and metallicity) is the combination of the initial mass function (IMF), the star formation history (SFH), and the metallicity distribution function (MDF). The framework of our analysis is that of Poisson Point Processes (PPPs), a class of statistical models suitable when dealing with points (stars) in a multidimensional space (the measurement space of multiple photometric bands). The theory of PPPs easily accommodates the modeling of measurement errors as well as that of incompleteness. Our method avoids binning stars in the color–magnitude diagram and uses the whole likelihood function for each data point; combining the individual likelihoods allows the computation of the posterior probability for the population's SBF. Within the proposed framework it is possible to include nuisance parameters, such as distance and extinction, by specifying their prior distributions and marginalizing over them. The aim of this paper is to assess the validity of this new approach under a range of assumptions, using only simulated data. Forthcoming work will show applications to real data. Although it has a broad scope of possible applications, we have developed this method to study multi-band Hubble Space Telescope observations of the Milky Way Bulge. Therefore we will focus on simulations with characteristics similar to those of the Galactic Bulge.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2015

Luminescence of the (O{sub 2}(a{sup 1}Δ{sub g})){sub 2} collisional complex in the temperature range of 90-315 K: Experiment and theory

Experimental and theoretical studies of collision induced emission of singlet oxygen molecules O{sub 2}(a{sup 1}Δ{sub g}) in the visible range have been performed. The rate constants, half-widths, and position of peaks for the emission bands of the (O{sub 2}(a{sup 1}Δ{sub g})){sub 2} collisional complex centered around 634 nm (2) and 703 nm (3) have been measured in the temperature range of 90–315 K using a flow-tube apparatus that utilized a gas-liquid chemical singlet oxygen generator. The absolute values of the spontaneous emission rate constants k{sub 2} and k{sub 3} are found to be similar, with the k{sub 3}/k{sub 2} ratio monotonically decreasing from 1.1 at 300 K to 0.96 at 90 K. k{sub 2} slowly decreases with decreasing temperature but a sharp increase in its values is measured below 100 K. The experimental results were rationalized in terms of ab initio calculations of the ground and excited potential energy and transition dipole moment surfaces of singlet electronic states of the (O{sub 2}){sub 2} dimole, which were utilized to compute rate constants k{sub 2} and k{sub 3} within a statistical model. The best theoretical results reproduced experimental rate constants with the accuracy of under 40% and correctly described the observed temperature dependence. The main contribution to emission process (2), which does not involve vibrational excitation of O{sub 2} molecules at the ground electronic level, comes from the spin- and symmetry-allowed 1{sup 1}A{sub g}←{sup 1}B{sub 3u} transition in the rectangular H configuration of the dimole. Alternatively, emission process (3), in which one of the monomers becomes vibrationally excited in the ground electronic state, is found to be predominantly due to the vibronically allowed 1{sup 1}A{sub g}←2{sup 1}A{sub g} transition induced by the asymmetric O–O stretch vibration in the collisional complex. The strong vibronic coupling between nearly degenerate excited singlet states of the dimole makes the intensities of vibronically and symmetry-allowed transitions comparable and hence the rate constants k{sub 2} and k{sub 3} close to one another.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2015

High-Density and Very-Low-Density Lipoprotein Have Opposing Roles in Regulating Tumor-Initiating Cells and Sensitivity to Radiation in Inflammatory Breast Cancer

Purpose: We previously demonstrated that cholesterol-lowering agents regulate radiation sensitivity of inflammatory breast cancer (IBC) cell lines in vitro and are associated with less radiation resistance among IBC patients who undergo postmastectomy radiation. We hypothesized that decreasing IBC cellular cholesterol induced by treatment with lipoproteins would increase radiation sensitivity. Here, we examined the impact of specific transporters of cholesterol (ie lipoproteins) on the responses of IBC cells to self-renewal and to radiation in vitro and on clinical outcomes in IBC patients. Methods and Materials: Two patient-derived IBC cell lines, SUM 149 and KPL4, were incubated with low-density lipoproteins (LDL), very-low-density lipoproteins (VLDL), or high-density lipoproteins (HDL) for 24 hours prior to irradiation (0-6 Gy) and mammosphere formation assay. Cholesterol panels were examined in a cohort of patients with primary IBC diagnosed between 1995 and 2011 at MD Anderson Cancer Center. Lipoprotein levels were then correlated to patient outcome, using the log rank statistical model, and examined in multivariate analysis using Cox regression. Results: VLDL increased and HDL decreased mammosphere formation compared to untreated SUM 149 and KPL4 cells. Survival curves showed enhancement of survival in both of the IBC cell lines when pretreated with VLDL and, conversely, radiation sensitization in all cell lines when pretreated with HDL. In IBC patients, higher VLDL values (>30 mg/dL) predicted a lower 5-year overall survival rate than normal values (hazard ratio [HR] = 1.9 [95% confidence interval [CI]: 1.05-3.45], P=.035). Lower-than-normal patient HDL values (<60 mg/dL) predicted a lower 5-year overall survival rate than values higher than 60 mg/dL (HR = 3.21 [95% CI: 1.25-8.27], P=.015). Conclusions: This study discovered a relationship among the plasma levels of lipoproteins, overall patient response, and radiation resistance in IBC patients and IBC patient-derived cell lines. A more expansive study is needed to verify these observations.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2020

Artificial Intelligence Application to D and D - 20492

As aging facilities across the DOE complex await decommissioning, there is an ongoing need to understand any changes in the structural conditions. Many of these facilities were built over 50 years ago and, in some cases, these facilities have gone beyond the expected operational lifetime. Many facilities have been placed in a state of 'cold and dark,' sitting unused and awaiting decommissioning. Especially challenging are the aging facilities that provide unique operational/production capabilities to support critical DOE missions and cannot be shut down. In any of these scenarios, the structural integrity of these facilities may become compromised as time passes. It is critical that adequate inspections be performed on a continual basis and that the data collected undergoes sufficient analysis to support timely identification of any new or worsening structural issues as well as prompt needed maintenance and repairs to maintain the facilities in a safe condition. In recent days, Artificial Intelligence (AI) [1] and its application to various domains are growing at fast speed. FIU is performing research in this area and exploring the associated technologies to solve nuclear decommissioning problems. Artificial intelligence refers to the capability of a program to autonomously act, react and adapt to the working environment. AI enables the machine to behave like humans and perform the cognitive functions such as 'learning' and 'problem solving'. AI systems gradually moving from traditional approaches (algorithms and expert systems) towards more efficient and advanced technologies (machine learning [1] and deep learning [2] [3]). AI is the study of algorithms and statistical models that is being used by computers to perform specific tasks without using explicit instructions. FIU is working to develop a pilot-scale infrastructure to implement structural health monitoring using AI technologies with focus on machine learning, deep learning. This research is focused on Computer Vision/Image Classification area of AI applications. This can also be expanded to other areas of AI related to Object Recognition and Character Recognition in images. In addition to utilizing existing data sets, FIU will collect and investigate image and video data using FIU test-bed mockups to monitor structural health of the facility. Resulting data will be processed and analyzed using machine learning/deep learning technologies. The proposed pilot system is intended to serve as a starting point to engage the DOE field sites on related data sets and their decision making needs. It is anticipated that proposed machine learning/deep learning technologies can be effectively employed using anomaly detection to solve EM challenges in surveillance and maintenance of the D and D facilities. FIU will work with research stakeholders to identify applications at various sites and other DOE facilities. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2018

Long-term exposure to ambient air pollution (including PM1) and metabolic syndrome: The 33 Communities Chinese Health Study (33CCHS)

Highlights: • Associations of ambient air pollution with metabolic syndrome were scarcely studied. • Data from 15,477 Chinese adults were analyzed. • Long-term exposure to air pollution was associated with metabolic syndrome. • Demographic and lifestyle factors modified the association between metabolic syndrome and air pollution. Little evidence exists about the effects of long-term exposure to ambient air pollution on metabolic syndrome (MetS). This study aimed to determine the association between long-term ambient air pollution and MetS in China. A total of 15,477 adults who participated in the 33 Communities Chinese Health Study (33CCHS) in 2009 were evaluated. MetS was defined based on the recommendation by the Joint Interim Societies. Exposure to air pollutants was assessed using data from monitoring stations and a spatial statistical model (including particles with diameters ≤ 1.0 µm (PM{sub 1}), ≤ 2.5 µm (PM{sub 2.5}), and ≤ 10 µm (PM{sub 10}), sulfur dioxide (SO{sub 2}), nitrogen dioxide (NO{sub 2}), and ozone (O{sub 3})). Two-level logistic regression analyses were utilized to assess the associations between air pollutants and MetS. The prevalence of MetS was 30.37%. The adjusted odds ratio of MetS per 10 µg/m{sup 3} increase in PM{sub 1}, PM{sub 2.5}, PM{sub 10}, SO{sub 2}, NO{sub 2}, and O{sub 3} were 1.12 (95% CI = 1.00–1.24), 1.09 (95% CI = 1.00–1.18), 1.13 (95% CI = 1.08–1.19), 1.10 (95% CI = 1.02–1.18), 1.33 (95% CI = 1.12–1.57), and 1.10 (95% CI = 1.01–1.18), respectively. Stratified analyses indicated that the above associations were stronger in participants with the demographic variables of males, < 50 years of age, and higher income, as well as with the behavioral characteristics of smoking, drinking, and consuming sugar-sweetened soft drinks frequently. This study indicates that long-term exposure to ambient air pollutants may increase the risk of MetS, especially among males, the young to middle aged, those of low income, and those with unhealthy lifestyles.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2015

Online adaptation and verification of VMAT

Purpose: This work presents a method for fast volumetric modulated arc therapy (VMAT) adaptation in response to interfraction anatomical variations. Additionally, plan parameters extracted from the adapted plans are used to verify the quality of these plans. The methods were tested as a prostate class solution and compared to replanning and to their current clinical practice. Methods: The proposed VMAT adaptation is an extension of their previous intensity modulated radiotherapy (IMRT) adaptation. It follows a direct (forward) planning approach: the multileaf collimator (MLC) apertures are corrected in the beam’s eye view (BEV) and the monitor units (MUs) are corrected using point dose calculations. All MLC and MU corrections are driven by the positions of four fiducial points only, without need for a full contour set. Quality assurance (QA) of the adapted plans is performed using plan parameters that can be calculated online and that have a relation to the delivered dose or the plan quality. Five potential parameters are studied for this purpose: the number of MU, the equivalent field size (EqFS), the modulation complexity score (MCS), and the components of the MCS: the aperture area variability (AAV) and the leaf sequence variability (LSV). The full adaptation and its separate steps were evaluated in simulation experiments involving a prostate phantom subjected to various interfraction transformations. The efficacy of the current VMAT adaptation was scored by target mean dose (CTV{sub mean}), conformity (CI{sub 95%}), tumor control probability (TCP), and normal tissue complication probability (NTCP). The impact of the adaptation on the plan parameters (QA) was assessed by comparison with prediction intervals (PI) derived from a statistical model of the typical variation of these parameters in a population of VMAT prostate plans (n = 63). These prediction intervals are the adaptation equivalent of the tolerance tables for couch shifts in the current clinical practice. Results: The proposed adaptation of a two-arc VMAT plan resulted in the intended CTV{sub mean} (Δ ≤ 3%) and TCP (ΔTCP ≤ 0.001). Moreover, the method assures the intended CI{sub 95%} (Δ ≤ 11%) resulting in lowered rectal NTCP for all cases. Compared to replanning, their adaptation is faster (13 s vs 10 min) and more intuitive. Compared to the current clinical practice, it has a better protection of the healthy tissue. Compared to IMRT, VMAT is more robust to anatomical variations, but it is also less sensitive to the different correction steps. The observed variations of the plan parameters in their database included a linear dependence on the date of treatment planning and on the target radius. The MCS is not retained as QA metric due to a contrasting behavior of its components (LSV and AAV). If three out of four plan parameters (MU, EqFS, AAV, and LSV) need to lie inside a 50% prediction interval (3/4—50%PI), all adapted plans will be accepted. In contrast, all replanned plans do not meet this loose criterion, mainly because they have no connection to the initially optimized and verified plan. Conclusions: A direct (forward) VMAT adaptation performs equally well as (inverse) replanning but is faster and can be extended to real-time adaptation. The prediction intervals for the machine parameters are equivalent to the tolerance tables for couch shifts in the current clinical practice. A 3/4—50%PI QA criterion accepts all the adapted plans but rejects all the replanned plans.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2017

Brain Metastasis Velocity: A Novel Prognostic Metric Predictive of Overall Survival and Freedom From Whole-Brain Radiation Therapy After Distant Brain Failure Following Upfront Radiosurgery Alone

Purpose: Prior statistical models attempted to identify risk factors for time to distant brain failure (DBF) or time to salvage whole-brain radiation therapy (WBRT) to predict the benefit of early WBRT versus stereotactic radiosurgery (SRS) alone. We introduce a novel clinical metric, brain metastasis velocity (BMV), for predicting clinical outcomes after initial DBF following upfront SRS alone. Methods and Materials: BMV was defined as the cumulative number of new brain metastases that developed over time since first SRS in years. Patients were classified by BMV into low-, intermediate-, and high-risk groups, consisting of <4, 4 to 13, and >13 new metastases per year, respectively. Histology, number of metastases at the time of first SRS, and systemic disease status were assessed for effect on BMV. Results: Of 737 patients treated at our institution with upfront SRS without WBRT, 286 had ≥1 DBF event. A lower BMV predicted for improved overall survival (OS) following initial DBF (log-rank P<.0001). Median OS for the low, intermediate, and high BMV groups was 12.4 months (95% confidence interval [CI], 10.4-16.9 months), 8.2 months (95% CI, 5.0-9.7 months), and 4.3 months (95% CI, 2.6-6.7 months), respectively. Multivariate analysis showed that BMV remained the dominant predictor of OS, with a hazard ratio of 2.75 for the high BMV group (95% CI, 1.94-3.89; P<.0001) and a hazard ratio of 1.65 for the intermediate BMV group (95% CI, 1.18-2.30; P<.004). A lower BMV was associated with decreased rates of salvage WBRT (P=.02) and neurologic death (P=.008). Factors predictive for a higher BMV included ≥2 initial brain metastases (P=.004) and melanoma histology (P=.008). Conclusions: BMV is a novel metric associated with OS, neurologic death, and need for salvage WBRT after initial DBF following upfront SRS alone.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
NASA NTRS2014

Peak Oxygen Uptake during and after Long-duration Space Flight

Aerobic capacity (VO2peak) previously has not been measured during or after long-duration spaceflight. PURPOSE: To measure VO2peak and submaximal exercise responses during and after International Space Station (ISS) missions. METHODS: Astronauts (9 M, 5 F: 49 +/- 5 yr, 175 +/- 7 cm, 77.2 +/- 15.1 kg, 40.6 +/- 6.4 mL/kg/min [mean +/-SD]) performed graded peak cycle tests ~90 days before spaceflight, 15 d (FD15) after launch and every ~30 d thereafter during flight, and 1 (R+1), 10 (R+10), and 30 d (R+30) after landing. Oxygen consumption (VO2) and heart rate (HR) were measured from rest to peak exercise, while cardiac output (Q), stroke volume (SV), and arterial-venous oxygen difference (a-vO2diff) were measured only during rest and submaximal exercise. Data were analyzed using mixed-model linear regression. Body mass contributed significantly to statistical models, and thus results are reported as modeled estimates for an average subject. RESULTS: Early inflight (FD15) VO2peak was 17% lower (95% CI = - 22%, -13%) than preflight. VO2peak increased during spaceflight (0.001 L/min/d, P = 0.02) but did not return to preflight levels. On R+1 VO2peak was 15% (95% CI = -19%, -10%) lower than preflight but recovered to within 2% of preflight by R+30 (95% CI = -6%, +3%). Peak HR was not significantly different from preflight at any time. Inflight submaximal VO2 and a-vO2diff were generally lower than preflight, but the Q vs. VO2 slope was unchanged. In contrast, the SV vs. VO2 slope was lower (P < 0.001), primarily due to elevated SV at rest, and the HR vs. VO2 slope was greater (P < 0.001), largely due to elevated HR during more intense exercise. On R+1 although the relationships between VO2 and Q, SV, and HR were not statistically different than preflight, resting and submaximal exercise SV was lower (P < 0.001), resting and submaximal exercise HR was higher (P < 0.002), and a-vO2diff was unchanged. HR and SV returned to preflight levels by R+30. CONCLUSION: In the average astronaut VO2peak was reduced during spaceflight and immediately after landing but factors contributing to lower VO2peak may be different during spaceflight and recovery. Maintaining Q while VO2 is reduced inflight may be suggestive of an elevated blood flow to vascular beds other than exercising muscles, but decreased SV after flight likely reduces Q at peak exertion.

Moore, Alan D., Jr.↗
DOE OSTI2025

First 𝛽-Delayed Two-Neutron Spectroscopy of the 𝑟-Process Nucleus 134 In and Observation of the 𝑖 13/2 Single-Particle Neutron State in 133 Sn

This manuscript reports on the direct observation of a 𝛽-delayed two-neutron emission in a study of 134 In at the ISOLDE Decay Station using neutron spectroscopy. We also report on the first measurement in 𝛽 − decay of the long-sought 13/2 + excited state in 133 Sn, attributed to be the neutron single-particle 𝑖 13/2 orbital. The observation of sequential neutron emission is used to extract the relative population of the 𝑖 13/2 state, which was found to be much smaller than the predictions of the statistical model. The experiment was possible because of the innovative use of a neutron array with neutron discrimination and interaction tracking capabilities. This is the first study of the details of the two-neutron emission for a nucleus, which belongs to the 𝑟-process path. Understanding 𝛽-delayed two-neutron emission probabilities is essential to validate models used in astrophysical 𝑟-process nucleosynthesis calculations. Observing two-neutron emissions in 𝛽 − decay paves the way for new experiments to study energy and angular correlations for 𝛽-delayed multineutron emitters.

Beta decay↗
DOE OSTI2019

Improving Precipitation Estimation Using Convolutional Neural Network

Precipitation process is generally considered to be poorly represented in numerical weather/climate models. Statistical downscaling (SD) methods, which relate precipitation with model resolved dynamics, often provide more accurate precipitation estimates compared to model's raw precipitation products. Herein we introduce the convolutional neural network model to foster this aspect of SD for daily precipitation prediction. Specifically, we restrict the predictors to the variables that are directly resolved by discretizing the atmospheric dynamics equations. In this sense, our model works as an alternative to the existing precipitation-related parameterization schemes for numerical precipitation estimation. We train the model to learn precipitation-related dynamical features from the surrounding dynamical fields by optimizing a hierarchical set of spatial convolution kernels. We test the model at 14 geogrid points across the contiguous United States. Results show that provided with enough data, precipitation estimates from the convolutional neural network model outperform the reanalysis precipitation products, as well as SD products using linear regression, nearest neighbor, random forest, or fully connected deep neural network. Evaluation for the test set suggests that the improvements can be seamlessly transferred to numerical weather modeling for improving precipitation prediction. Based on the default network, we examine the impact of the network architectures on model performance. Also, we offer simple visualization and analyzing approaches to interpret the models and their results. Our study contributes to the following two aspects: First, we offer a novel approach to enhance numerical precipitation estimation; second, the proposed model provides important implications for improving precipitation-related parameterization schemes using a data-driven approach.

54 ENVIRONMENTAL SCIENCES↗
NASA NTRS1993

Mental Models of Software Forecasting

The majority of software engineers resist the use of the currently available cost models. One problem is that the mathematical and statistical models that are currently available do not correspond with the mental models of the software engineers. In an earlier JPL funded study (Hihn and Habib-agahi, 1991) it was found that software engineers prefer to use analogical or analogy-like techniques to derive size and cost estimates, whereas curren CER's hide any analogy in the regression equations. In addition, the currently available models depend upon information which is not available during early planning when the most important forecasts must be made.

Cost Models↗
NASA NTRS2014

Bayesian Statistics and Uncertainty Quantification for Safety Boundary Analysis in Complex Systems

The analysis of a safety-critical system often requires detailed knowledge of safe regions and their highdimensional non-linear boundaries. We present a statistical approach to iteratively detect and characterize the boundaries, which are provided as parameterized shape candidates. Using methods from uncertainty quantification and active learning, we incrementally construct a statistical model from only few simulation runs and obtain statistically sound estimates of the shape parameters for safety boundaries.

Active Learning↗
NASA NTRSDate not supplied

The effect of modeling dose uncertainty on low-boom community noise dose-response curves

In logistic dose-response modeling, failing to account for uncertainty in estimated doses can cause an artificial flattening or attenuation of the slope of the summary curve. In Lee et al. [J. Acoust. Soc. Am. 147(4), pp. 2222-2234 (2020)], data from two NASA low-amplitude sonic boom community noise survey tests were modeled using a Bayesian multilevel logistic regression (MLR) statistical model that assumed there was no uncertainty in the noise dose estimates. However, in these community tests, the noise dose uncertainty was estimated by Page et al. [NASA/CR-2014-218180 and NASA/CR-2020-220589/Volume I] using a leave-one-out method. In the current work, a term was added to extend the Bayesian MLR model to account for the estimated noise dose uncertainty quantified in the Page et al. analyses. This uncertainty term was included in two ways, either as classical or as Berkson uncertainty, and yield similar results. When the uncertainty is accounted for in the Bayesian MLR model, the dose-response curves become 5-10% steeper, but the difference in the noise dose that elicits a 5% highly annoyed response is small (less than 1 dB). This result is encouraging for future X-59 community tests whose survey area will be sparsely populated with noise monitors.

X-59↗
NASA NTRS2023

Geometric Analysis and Visualization of Maxillofacial Anthropometry

As part of head borne systems, the maxillofacial and mandibular components are critical for Soldier protection in extreme and dynamic conditions. Often this equipment must not only cover the lower face of the wearer, but also accommodate motion of the lower jaw and resultant deformation of the face shape. Although the application is different, head borne equipmentdesigned for astronauts has similarly critical design requirements including comfort and protection from other environmental hazards while allowing for maximum performance. To improve the design of such systems and to ensure adequate accommodation of the user population, NASA Johnson Space Center is collaborating with the U.S. Army’s Development Command Soldier Center (DEVCOM SC) to statistically analyze facial landmark and measurement data from large Soldier anthropometric databases ( ANSUR 1988 and ANSUR II). Anthropometric measurements were incorporated from 5383 males and 3594 females for analysis, including individuals aged 17-58 across a range of races. Facial landmarks were analyzed, specifically the 3D coordinates of the Menton, Prementon, Gonion, Tragion, Stomion, Zygion, Pronasale, and Subnasale. These data were statistically correlated with demographiccharacteristics, such as sex and population origin. A statistical model of the lower head-face surface geometry is currently in development to describe the variance of head and face shape across the population. The model will be able to visually show the statistical trends of the different shapes and sizes of the head, along with the synthetically represented boundary cases. The specific analytical methods and outcomes will be detailed in the final proceeding. This study will provide a more comprehensive assessment of how face shape varies across the military population compared to traditional linear measurements. The outcome of this study is expected to provide critical insight and recommendations for improved protective equipment design and development, not only with Soldiers but also astronauts in future exploration missions, as well as first responders and others that require critical equipment fitted to the maxillofacial and or mandibular region.

W Green↗
DOE OSTI2015

Adaptive Bayesian Inference for Prediction

Performing rigorous predictive simulation necessitates that all available data be incorporated into the modeling process in order to accurately quantify uncertainties in predictions of the model. Statistical Bayesian inference is the most commonly used approach for incorporating data into uncertainty estimates. Statistical Bayesian inference uses data and an assumed error model to inform posterior distributions of model input variables and model discrepancies. An explicit characterization of the posterior distribution is not formed, but rather methods such as Markov chain Monte-Carlo (MCMC) are used to draw samples from the posterior. These samples are then used to evaluate statistics, such as integrals, of predicted quantities of interest (QoI). Recently the authors have developed an alternative reformulation of Bayesian inference based upon a measure-theoretic inversion technique. Unlike traditional statistical Bayesian inference, this reformulation allows us to directly interrogate the posterior distribution and, consequently, to generate samples from the posterior without MCMC. In this work we use forward propagation uncertainty quantification techniques to reduce the cost of approximating the posterior distribution. Specifically we investigate goal-oriented adaptive sparse grid approximation methods. These adaptive methods concentrate high-fidelity model evaluations in regions of the parameter space that both significantly contribute to the uncertainty in specified QoI and are informed by the available data. Numerical results are presented to compare our approach with the standard statistical Bayesian approach and to demonstrate the effectiveness of the goal-oriented adaptivity.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2016

TOUGH2-seed: A coupled fluid flow and mechanical-stochastic approach to model injection-induced seismicity

Understanding the injection-induced triggering mechanism is a fundamental step towards controlling the seismicity generated by deep underground exploitation. Here we propose a modeling approach based on coupling the TOUGH2 simulator with a geomechanical-stochastic model. Additionally, the hydro-mechanical-stochastic model provides a good representation of different mechanisms influencing each other during and after the injection phase. Each mechanism affects the induced seismicity in a different way and at different times during the reservoir stimulation, confirming that a complex interaction is in place, and that more sophisticated and physics-based approaches coupled with statistical model are required to explain such a complex interaction. In addition to previous statistical and hybrid models, our approach accounts for a full 3D formulation of both stresses and fluid flow, further including all the TOUGH2 capabilities. Furthermore, it includes interactions between triggered seismic events through calculation of static stress transfer. Here, in this work, we present the main capabilities of TOUGH2-SEED and apply the model to the Basel EGS case, successfully reproducing the injection pressure as well as the evolution of the seismicity.

58 GEOSCIENCES↗
DOE OSTI2021

Comparing Emulation Methods for Computer Models with High Dimensional Output

This Master’s Capstone project will present a comparison of statistical models for computer simulation studies. The four models included in this comparison study were chosen for both their proven and diverse methodologies. We will present a case study on hurricane flood data in the Delaware bay which highlights the strengths and weaknesses of each model when applied to a very large spatial field. As computers have gotten faster, we have become interested in modeling increasingly large spatial fields in both size and resolution. Statistical algorithms that are able to efficiently handle these fields have never been more important. We therefore find this comparison to be extremely topical.

42 ENGINEERING↗
DOE OSTI2017

Spatio-Temporal Data Analysis at Scale Using Models Based on Gaussian Processes

Gaussian processes are the most commonly used statistical model for spatial and spatio-temporal processes that vary continuously. They are broadly applicable in the physical sciences and engineering and are also frequently used to approximate the output of complex computer models, deterministic or stochastic. We undertook research related to theory, computation, and applications of Gaussian processes as well as some work on estimating extremes of distributions for which a Gaussian process assumption might be inappropriate. Our theoretical contributions include the development of new classes of spatial-temporal covariance functions with desirable properties and new results showing that certain covariance models lead to predictions with undesirable properties. To understand how Gaussian process models behave when applied to deterministic computer models, we derived what we believe to be the first significant results on the large sample properties of estimators of parameters of Gaussian processes when the actual process is a simple deterministic function. Finally, we investigated some theoretical issues related to maxima of observations with varying upper bounds and found that, depending on the circumstances, standard large sample results for maxima may or may not hold. Our computational innovations include methods for analyzing large spatial datasets when observations fall on a partially observed grid and methods for estimating parameters of a Gaussian process model from observations taken by a polar-orbiting satellite. In our application of Gaussian process models to deterministic computer experiments, we carried out some matrix computations that would have been infeasible using even extended precision arithmetic by focusing on special cases in which all elements of the matrices under study are rational and using exact arithmetic. The applications we studied include total column ozone as measured from a polar-orbiting satellite, sea surface temperatures over the Pacific Ocean, and annual temperature extremes at a site in New York City. In each of these applications, our theoretical and computational innovations were directly motivated by the challenges posed by analyzing these and similar types of data.

97 MATHEMATICS AND COMPUTING↗