Search NASA⌕ Search

SEARCH · Search NASA

Results for “Statistical Testing Approach”

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 991 records · Page 55
All sourcesNASA NTRSDOE OSTI
DOE OSTI2019

Analysis of Localized Deformation Processes in Highly Irradiated Austenitic Stainless Steel through In Situ Techniques

This report describes new experimental results obtained during in situ mechanical tests with neutron-irradiated (10.7 dpa) austenitic 304L steel specimens. The in situ tensile tests inside a scanning electron microscope (SEM) were accompanied by modern high-resolution electron backscatter diffraction (HR-EBSD) analysis. The HR-EBSD approach allowed for investigating and quantifying internal stresses and dislocation structures in nondeformed and deformed steels, materials used in light water reactor internal components. The literature regarding the HR-EBSD analysis of metals and alloys is very limited, and practically no publications on the HR-EBSD analysis coupled with in situ tests are available. Experimental work, performed at Oak Ridge National Laboratory’s Low Activation Materials Development and Analysis facility, is detailed in this report. Section 1 briefly describes strain and stress localization processes and underlines their importance and role in irradiated materials, materials used in nuclear power plants. The section demonstrates that very limited quantitative data exist regarding localized stresses and strains in the irradiated materials at the meso-level (i.e., inside the grain, for a group of interacting grains, or for channel–grain boundary [GB] interaction) whereas local stresses at GBs are key contributors to the stress corrosion crack initiation.Section 2 documents the key methods used to conduct the experimental work: the SEM/EBSD system, miniature tensile frame, and irradiated material. Irradiated specimen manufacturing is described, as this option may be of special interest for other projects. Section 3 provides details of the HR-EBSD in situ tests (e.g., tensile curves, geometry, dimensions of regions of interest, load, and deformation steps). Section 4 discusses the results related to the evolution of geometrically necessary dislocations (GNDs) and their density, spatial, and statistical distributions.Section 5 focuses on the HR-EBSD analysis with respect to a channel–GB interaction. Two aspects are discussed in detail: mapping and analyzing the GND fields before and after loading, and the possibility of investigating local stresses and elastic strains at the channel–GB interaction. The present work results are beneficial for understanding strain and stress localization processes leading to crack initiation, for analyzing the material after long-term in-service life, and for developing predictive models. Section 6 summarizes the work performed and discusses future activities within this research direction.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
NASA NTRS2016

Automating Microbial Directed Evolution For Bioengineering Applications

From a micro-biology perspective, directed evolution is a technique that uses controlled environmental pressures to select for a desired phenotype. Directed evolution has the distinct advantage over rational design of not needing extensive knowledge of the genome or pathways associated with a microorganism to induce phenotypes. However, there are currently limitations to the applicability of this technique including being time-consuming, error-prone, and dependent on existing assays that may lack selectivity for the given phenotype. The AADEC (Autonomous Adaptive Directed Evolution Chamber) system is a proof-of-concept instrument to automate and improve the technique such that directed evolution can be used more effectively as a general bioengineering tool. A series of tests using the automated system and comparable by-hand survival assay measurements have been carried out using UV-C radiation and Escherichia coli cultures in order to demonstrate the advantages of the AADEC versus traditional implementations of directed evolution such as random mutagenesis. AADEC uses UV-C exposure as both a source of environmental stress and mutagenesis, so in order to evaluate the UV-C tolerance obtained from the cultures, a manual UV-C exposure survival assay was developed alongside the device to compare the survival fractions at a fixed dosage. This survival assay involves exposing E. coli to UV-C radiation using a custom-designed exposure hood to control the flux and dose. Surviving cells are counted then transferred to the next iteration and so on for several iterations to calculate the survival fractions for each exposure iteration.This survival assay primarily serves as a baseline for the AADEC device, allowing quantification of the differences between the AADEC system over the manual approach. The primary data of comparison is survival fractions; this is obtained by optical density and plate counts in the manual assay and by optical density growth curve fits pre- and post-exposure in the automated case. This data can then be compiled to calculate trends over the iterations to characterize increasing UV-C resistance of the E.coli strains. The observed trends are statistically indistinguishable through several iterations from both sources.

Bioengineering↗
DOE OSTI2023

Probabilistic Failure Criterion of SiC/SiC Composites Under Multiaxial Loading

Owing to its excellent mechanical properties and stability under high temperature and neutron irradiation conditions, SiC/SiC composites have emerged as a promising material for light water reactors (LWRs) in the development of accident-tolerant fuel (ATF) systems. Structural integrity and retention of hermeticity are two crucial requirements for SiC/SiC claddings during normal operations, and both of them are closely related to the proportional limit stress (PLS) of the material. Understanding the behavior of SiC/SiC composites under multiaxial stress states and developing a probabilistic approach for evaluating the structural vulnerability are of paramount importance for reliability-based analysis and design of SiC/SiC composite claddings. So far, there has been very limited effort towards experimental and analytical investigations of probabilistic failure of SiC/SiC claddings. This critical knowledge gap motivates this research. A probabilistic failure criterion for SiC/SiC composites under multi-axial loading is developed, and this criterion is incorporated into reliability analysis of the structural integrity of SiC/SiC fuel cladding. The research consists of two parts: 1) experimental investigation of multiaxial failure behavior of SiC/SiC composites, and 2) theoretical modeling of time-dependent probabilistic failure of SiC/SiC cladding. In the experimental investigation, the PLS is determined through the examination of stress-strain response, the acoustic emission measurement, as well as the X-ray computed tomography. The theoretical framework is derived by combin- ing the finite weakest-link statistical model and the subcritical damage growth model. This theoretical model captures the time-dependent failure mechanism of the material, which has a major consequence for predicting the lifetime distribution of the cladding. Meanwhile, the model also predicts that the failure statistics of the cladding depends strongly on the cladding length. The results of the multiaxial experiments reveal the level of statistical variation of the PLS of SiC/SiC materials under different stress states. The theoretical model provides a robust analytical tool for extrapolation of small-scale laboratory test results to the behavior of full-scale claddings. These findings establish a scientific foundation for the development of reliability-based design of SiC/SiC fuel claddings, which will play an essential role in improving the structural safety and integrity of LWRs.

42 ENGINEERING↗
NASA NTRS2013

An Alternative Flight Software Trigger Paradigm: Applying Multivariate Logistic Regression to Sense Trigger Conditions Using Inaccurate or Scarce Information

In late 2014, NASA will fly the Orion capsule on a Delta IV-Heavy rocket for the Exploration Flight Test-1 (EFT-1) mission. For EFT-1, the Orion capsule will be flying with a new GPS receiver and new navigation software. Given the experimental nature of the flight, the flight software must be robust to the loss of GPS measurements. Once the high-speed entry is complete, the drogue parachutes must be deployed within the proper conditions to stabilize the vehicle prior to deploying the main parachutes. When GPS is available in nominal operations, the vehicle will deploy the drogue parachutes based on an altitude trigger. However, when GPS is unavailable, the navigated altitude errors become excessively large, driving the need for a backup barometric altimeter to improve altitude knowledge. In order to increase overall robustness, the vehicle also has an alternate method of triggering the parachute deployment sequence based on planet-relative velocity if both the GPS and the barometric altimeter fail. However, this backup trigger results in large altitude errors relative to the targeted altitude. Motivated by this challenge, this paper demonstrates how logistic regression may be employed to semi-automatically generate robust triggers based on statistical analysis. Logistic regression is used as a ground processor pre-flight to develop a statistical classifier. The classifier would then be implemented in flight software and executed in real-time. This technique offers improved performance even in the face of highly inaccurate measurements. Although the logistic regression-based trigger approach will not be implemented within EFT-1 flight software, the methodology can be carried forward for future missions and vehicles.

Smith, Kelly M.↗
NASA NTRS2013

An Alternative Flight Software Paradigm: Applying Multivariate Logistic Regression to Sense Trigger Conditions using Inaccurate or Scarce Information

In late 2014, NASA will fly the Orion capsule on a Delta IV-Heavy rocket for the Exploration Flight Test-1 (EFT-1) mission. For EFT-1, the Orion capsule will be flying with a new GPS receiver and new navigation software. Given the experimental nature of the flight, the flight software must be robust to the loss of GPS measurements. Once the high-speed entry is complete, the drogue parachutes must be deployed within the proper conditions to stabilize the vehicle prior to deploying the main parachutes. When GPS is available in nominal operations, the vehicle will deploy the drogue parachutes based on an altitude trigger. However, when GPS is unavailable, the navigated altitude errors become excessively large, driving the need for a backup barometric altimeter to improve altitude knowledge. In order to increase overall robustness, the vehicle also has an alternate method of triggering the parachute deployment sequence based on planet-relative velocity if both the GPS and the barometric altimeter fail. However, this backup trigger results in large altitude errors relative to the targeted altitude. Motivated by this challenge, this paper demonstrates how logistic regression may be employed to semi-automatically generate robust triggers based on statistical analysis. Logistic regression is used as a ground processor pre-flight to develop a statistical classifier. The classifier would then be implemented in flight software and executed in real-time. This technique offers improved performance even in the face of highly inaccurate measurements. Although the logistic regression-based trigger approach will not be implemented within EFT-1 flight software, the methodology can be carried forward for future missions and vehicles

Smith, Kelly↗
DOE OSTI2016

flowVS: channel-specific variance stabilization in flow cytometry

Abstract Background Comparing phenotypes of heterogeneous cell populations from multiple biological conditions is at the heart of scientific discovery based on flow cytometry (FC). When the biological signal is measured by the average expression of a biomarker, standard statistical methods require that variance be approximately stabilized in populations to be compared. Since the mean and variance of a cell population are often correlated in fluorescence-based FC measurements, a preprocessing step is needed to stabilize the within-population variances. Results We present a variance-stabilization algorithm, called flowVS, that removes the mean-variance correlations from cell populations identified in each fluorescence channel. flowVS transforms each channel from all samples of a data set by the inverse hyperbolic sine (asinh) transformation. For each channel, the parameters of the transformation are optimally selected by Bartlett’s likelihood-ratio test so that the populations attain homogeneous variances. The optimum parameters are then used to transform the corresponding channels in every sample. flowVS is therefore an explicit variance-stabilization method that stabilizes within-population variances in each channel by evaluating the homoskedasticity of clusters with a likelihood-ratio test. With two publicly available datasets, we show that flowVS removes the mean-variance dependence from raw FC data and makes the within-population variance relatively homogeneous. We demonstrate that alternative transformation techniques such as flowTrans, flowScape, logicle, and FCSTrans might not stabilize variance. Besides flow cytometry, flowVS can also be applied to stabilize variance in microarray data. With a publicly available data set we demonstrate that flowVS performs as well as the VSN software, a state-of-the-art approach developed for microarrays. Conclusions The homogeneity of variance in cell populations across FC samples is desirable when extracting features uniformly and comparing cell populations with different levels of marker expressions. The newly developed flowVS algorithm solves the variance-stabilization problem in FC and microarrays by optimally transforming data with the help of Bartlett’s likelihood-ratio test. On two publicly available FC datasets, flowVS stabilizes within-population variances more evenly than the available transformation and normalization techniques. flowVS-based variance stabilization can help in performing comparison and alignment of phenotypically identical cell populations across different samples. flowVS and the datasets used in this paper are publicly available in Bioconductor.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2022

Realistic galaxy image simulation via score-based generative models

ABSTRACT We show that a denoising diffusion probabilistic model (DDPM), a class of score-based generative model, can be used to produce realistic mock images that mimic observations of galaxies. Our method is tested with Dark Energy Spectroscopic Instrument (DESI) grz imaging of galaxies from the Photometry and Rotation curve OBservations from Extragalactic Surveys (PROBES) sample and galaxies selected from the Sloan Digital Sky Survey. Subjectively, the generated galaxies are highly realistic when compared with samples from the real data set. We quantify the similarity by borrowing from the deep generative learning literature, using the ‘Fréchet inception distance’ to test for subjective and morphological similarity. We also introduce the ‘synthetic galaxy distance’ metric to compare the emergent physical properties (such as total magnitude, colour, and half-light radius) of a ground truth parent and synthesized child data set. We argue that the DDPM approach produces sharper and more realistic images than other generative methods such as adversarial networks (with the downside of more costly inference), and could be used to produce large samples of synthetic observations tailored to a specific imaging survey. We demonstrate two potential uses of the DDPM: (1) accurate inpainting of occluded data, such as satellite trails, and (2) domain transfer, where new input images can be processed to mimic the properties of the DDPM training set. Here we ‘DESI-fy’ cartoon images as a proof of concept for domain transfer. Finally, we suggest potential applications for score-based approaches that could motivate further research on this topic within the astronomical community.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2018

Graphite Characterization: Baseline Variability Analysis Report

The Idaho National INL’s Baseline Material Property program provides many measurements of different material properties on a variety of nuclear graphite grades being investigated for irradiation-induced changes in material properties. The baseline material properties provide an unirradiated reference for comparison with measurements of those properties after irradiation. This study describes how uncertainty, or variance, in estimates of important material properties, like compressive strength, may depend on sample size, sample distribution and parameter estimation methods. Compressive strength data from two large collections of specimens from a billet of PCEA graphite (sample size = 230) and a billet of IG-110 graphite (sample size = 48) are used to illustrate these dependencies. The grades of graphite used represent approximate end members in terms of relative variability in strength measurements. Monte Carlo simulation and bootstrapping methods were used to examine variance in Weibull-parameter estimates as a function of sample size. Results were consistent with what previous statistical analyses have shown about the relationship between sample size and width of the confidence interval for maximum likelihood (ML) estimated parameters when samples are repeatedly drawn from a Weibull distribution. As a function of sample size, the precision of the modulus estimate increases faster than that of the characteristic strength. The modulus maximum likelihood estimates (MLEs) are biased for low sample sizes. The American Society for Testing and Materials document ASTM D7846-16 provides equations that describe the dependence of the variance of parameter MLEs on sample size. Using the same methodology, we show that the polynomials provided for variance dependence on sample size are inaccurate for sample sizes greater than ~120. We provide approximations to that relationship that illustrate relative constancy for sample size > ~10, in the logspace relationship of the modulus variance and the log transformed characteristic strength variance. Bootstrapping provided a means of comparing Weibull parameters obtained using ASME’s guidance for specimen collection to those obtained using different weighting in a case where strength is related to position in a billet and to grain orientation and a single distribution is used to characterize the billet. ASME specifies that samples be collected with equal representation in all locations and grain orientations. Parameter values obtained using ASME’s method were different from those obtained using random sampling of the complete data sets, and this was particularly true for the PCEA data. In addition, while the ASME approach yielded more conservative parameters for the PCEA data, the opposite was true for the IG-110 data. The difference observed for the PCEA data set likely reflects the larger number of against-grain specimens collected, which had generally higher strength than the with-grain specimens. This emphasizes the importance of the dependence of strength on location or grain orientation and suggests that it may be prudent to weight specimen collection toward the weaker subgroups if a conservative estimate as desired, as the equal weighting method recommended by ASME would will not always provide the most conservative Weibull parameters. Because spatial correlations commonly exist in certain grades of graphite, future work may involve developing a methodology for parameter inference that relaxes the assumption of independent specimens. ASTM and ASME codes recommend different parameter estimation methods and models of the Weibull distribution in different situations. Descriptions of the effects on parameter estimates for the example billets used in this study provide an example of how the method may affect the parameter estimate, and how—in this case—that compares to differences due to other factors, such as sample size and dependence on grain orientation or sample location.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
NASA NTRS1992

The 1991 AVIRIS/POLDER experiment in Camargue, France

Airborne campaigns during the eighties provided high spectral resolution data, collected with imaging instruments such as AIS, AVIRIS, FLI, CAESI, and ISM, in order to investigate the relationship with canopy biophysical characteristics. The statistical approaches used to analyze these data do not allow investigation of the causality and the applicability of the observed correlations. Further, statistical studies demonstrated the high degree of redundancy of the spectral information amongst many others. And for retrieving vegetation biophysical characteristics, few results demonstrate the real information gain attributable to the high spectral resolution capability as compared to the use of a few wide wavelength bands. With several new imaging spectrophotometers scheduled for launch during the next 10 years (MERIS, MODIS, HIRIS), progress in the description and understanding of the mechanisms that drive the spectral variation of canopy reflectance is required. Most of these new sensor systems will also have the capability to observe the target under differing view directions. The problem of the combination and the use of the synergy between both the spectral and the directional sources of canopy reflectance variations has to be addressed. Apart from the atmospheric effects, the spectral variation of the light reflected by canopies originates from the leaves, the soil or the other vegetation elements such as branches and fruits. At leaf level, both diffuse reflectance and transmittance may be simulated by simple models, although no accurate information exists on the absorption features of the biochemicals (except water) in the 900-2500nm wavelength range. Many models mimic the directional variability of canopy reflectance at a given wavelength. Combining a leaf spectral model with a canopy directional model provides a powerful tool to analyze this problem. Some of us have initiated such a study, but our approach and theory remain to be tested using canopy data with their complexity, associated experimental error, and atmospheric effects. The main objective of the 1990 POLDER/AVIRIS experiment in Camargue was to provide a consistent data set over various canopies in order to test the applicability of the theory. The experiment, part of the 1991 MAC Europe experiment, involved simultaneous data collection using two sensors: AVIRIS and POLDER which measures the bidirectional and polarization properties of the targets at 670 and 880nm wavebands.

Baret, F.↗
DOE OSTI2020

A generalized Boltzmann kinetic theory for strongly magnetized plasmas with application to friction

Coulomb collisions in plasmas are typically modeled using the Boltzmann collision operator, or its variants, which apply to weakly magnetized plasmas in which the typical gyroradius of particles significantly exceeds the Debye length. Conversely, O’Neil has developed a kinetic theory to treat plasmas that are so strongly magnetized that the typical gyroradius of particles is much smaller than the distance of the closest approach in a binary collision. Here, we develop a generalized collision operator that applies across the full range of magnetization strength. Since there is no closed-form solution for the scattering cross section when plasma is strongly magnetized, the input to the collision operator is obtained by numerically calculating particle trajectories. To demonstrate novel physics associated with strong magnetization, it is used to compute the friction force on a massive test charge. In addition to the traditional stopping power component, this is found to exhibit a transverse component that is perpendicular to both the velocity and Lorentz force vectors in the strongly magnetized regime, as was predicted recently using linear response theory. As a result, good agreement is found between the collision theory and linear response theory in the regime in which both apply, but the new collision theory also applies to stronger magnetization strength regimes than the linear response theory is expected to apply in.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
NASA NTRS2018

Considerations for Implementing Voice-Controlled Spacecraft Systems Through a Human-Centered Design Approach

As computational power and speech recognition algorithms improve, the consumer market will see better-performing speech recognition applications. The cell phone and Internet-related service industry have further enhanced speech recognition applications using artificial intelligence and statistical data-mining techniques. These improvements to speech recognition technology (SRT) may one day help astronauts on future deep space human missions that require control of complex spacecraft systems or spacesuit applications by voice. Though SRT and more advanced speech recognition techniques show promise, use of this technology for a space application such as vehicle/habitat/spacesuit requires careful considerations. There are still recognition challenges to overcome such as background noise, human speech variability, and task loading. However, implemented correctly, a voice-controlled spacecraft system (VCSS) can provide a useful, natural and efficient form of human-machine communications during complex tasks such as when the hands and eyes are busy or as an aid for vehicle situational awareness inquiries or collaborative human-robot tasks. This document provides considerations and guidance for the use of SRT in VCSS applications for space missions, specifically in command-and-control (C2) applications where the commanding is user-initiated. First, current SRT limitations as known at the time of this report are given. Then, highlights of SRT used in the space program provide the reader with a history of some of the human spaceflight applications and research. Next, an overview of the speech production process and the intrinsic variations of speech are provided. Finally, general guidance and considerations are given for development of a VCSS using a human-centered design approach for space applications that includes vocabulary selection and performance testing, as well as VCSS considerations for C2 dialogue management design, feedback, error handling, and evaluation/usability testing. Available from

Salazar, George A.↗
DOE OSTI2022

Structure of Offshore Low-Level Jet Turbulence and Implications to Mesoscale-to-Microscale Coupling

This paper explores realistic nonstationary atmospheric boundary layer (ABL) turbulence arising from nonstationarity at the mesoscale, particularly within offshore low-level jets with implications to offshore wind farms, using high-fidelity multiscale large-eddy simulations (LES). To this end, we analyzed the single-point turbulence statistical structure of a North-Atlantic offshore LLJ event simulated using high-resolution LES (AMR-Wind). The nonstationary LLJ is simulated using a mesoscale-to-microscale coupled (MMC) simulation procedure involving data assimilation of mesoscale velocity and temperature data from the Weather Research and Forecasting (WRF) model. Unlike the assimilation of mesoscale velocity data into the LES, the direct assimilation of temperature profiles had a strong impact on turbulence stratification, thereby causing erroneous predictions of turbulence both above and within the jet layer. Various approaches to mitigate this effect have resulted in multiple (four) variants of this MMC strategy. Outcomes from this work clearly show that the turbulence within the low-level jet is a strong function of the MMC approach as the turbulence structure within the low-level jet is dependent on the flux of residual turbulence from outside the jet, which in turn depends on the temperature forcing history. Additionally, the turbulence predicted by all these different methods (as well as the observation data) show similar deviations from equilibrium as evidenced by comparisons with idealized atmospheric turbulence structure obtained using the same numerical method. In general, we observe that the predicted LLJ turbulence tends to differ from canonical ABL turbulence with comparable shear. Particularly, the combination of shear and turbulence observed in such nonstationary low-level turbulence cannot be matched using equilibrium settings and therefore, represents a critical use-case for both testing and leveraging meso–micro coupling strategies.

17 WIND ENERGY↗
DOE OSTI2019

Model reduction for fractured porous media: a machine learning approach for identifying main flow pathways

Discrete fracture networks (DFN) are often used to model flow and transport in fractured porous media. The accurate resolution of flow and transport behavior on a large DFN involving thousands of fractures is computationally expensive. This makes uncertainty quantification studies of quantities of interest such as travel time through the network computationally intractable, since hundreds to thousands of runs of the DFN model are required to get bounds on the uncertainty of the predictions. Prior works on the subject demonstrated that the complexity of a DFN could be reduced by considering a sub-network of it (often termed a “backbone” sub-network), one whose flow and transport properties were then shown to be similar to that of the full network. The technique is tantamount to partitioning the complete set of fractures of a network into two disjoint sets, one of which is the backbone sub-network while the other its complement. It is in this context that we present a system-reduction technique for DFNs using supervised machine learning via a Random Forest Classifier that selects a backbone sub-network from the full set of fractures. The in-sample errors (in terms of precision and recall scores) of the trained classifier are found to be very accurate indicators of the out-of-sample errors, thus exhibiting that the classifier generalizes well to test data. Moreover, this system-reduction technique yields sub-networks as small as 12% of the full DFN that still recover transport characteristics of the full network such as the peak dosage and tailing behavior for late times. Most importantly, the sub-networks remain connected, and their size can be controlled by a single dimensionless parameter. Furthermore, measures of KL-divergence and KS-statistic for the breakthrough curves of the sub-networks with respect to the full network show physically realistic trends in that the measures decrease monotonically as the size of the sub-networks increase. In conclusion, the computational efficiency gained by this technique depends on the size of the sub-network, but large reductions in computational time can be expected for small sub-networks, yielding as much as 90% computational savings for sub-networks that are as small as 10-12% of the full network.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2018

Development of ARCHER towards clinical use: modeling and simulation of Varian Linac for radiation therapy dose calculations - 25449

Monte Carlo particle transport methods are well recognized as the most accurate approach to dose calculations for radiation therapy treatment planning. However, the lengthy calculation time has prevented Monte Carlo methods from being applied to dose calculations in the planning of radiation therapy. Recently, the emerging parallel computing hardware accelerators, such as general-purpose Graphics Processing Unit (GPU), have significantly boosted the performance of Monte Carlo codes. The accuracy of a clinical dose engine depends greatly on the accuracy of source term modeling and the methodology of dose calculations. Hence, in this study, an efficient and accurate source model of the Varian TrueBeam{sup R} Linac is developed and integrated with a fast photon-electron transport module of the recently developed high-performance Monte Carlo code, ARCHER. The validity of the source model and the particle transport physics are verified with established dose calculation codes such as EGSnrc and validated with experimental measurements. To demonstrate the capability of the platform in clinical applications, three clinical treatment plans (the lung, breast, and prostate) are simulated and analyzed. From the perspective of performance, the entire simulation of each clinical plan, starting from the source modeling to the end of particle transport within the patient phantom, can be finished in less than one minute using a single Titan X GPU card with 1% statistical error. In comparison, the same cases take the EGSnrc code hundreds of CPU hours to finish. The results suggest that this new version of ARCHER code has been successfully developed and tested as a clinically deployable dose engine for external photon beam IMRT plans in the clinical setting. (authors)

61 RADIATION PROTECTION AND DOSIMETRY↗
DOE OSTI2019

Turbulent nonpremixed cool flames: Experimental measurements, Direct Numerical Simulation, and manifold-based combustion modeling

Turbulence, low-temperature chemistry, and their interactions in the form of turbulent cool flames are critical to understanding and improving advanced engines. Design of such engines requires tractable simulations which in turn necessitate turbulent combustion models that can account for cool flames. While manifold-based turbulent combustion models are an attractive option for hot flames, their applicability to cool flames is not yet fully understood. This is partially due to the lack of turbulent cool flame experiments, which has made model validation difficult. This work addresses these points with a combined experimental and computational investigation of turbulent nonpremixed dimethyl ether cool flames. First, a Co-flow Axisymmetric Reactor-Assisted Turbulent (CARAT) burner is developed and tested. Turbulent cool flames are studied using the formaldehyde planar laser-induced fluorescence (PLIF), acetone PLIF, and planar Rayleigh scattering techniques. The acetone PLIF signals are converted into mixture fraction values, and measurements of time-averaged temperature are derived from the Rayleigh scattering signals by taking advantage of the similarities between cool flames and unburned mixtures. Second, a Direct Numerical Simulation (DNS) of the CARAT burner is conducted. The flame is shown to be sensitive to thermal boundary conditions such that the stabilization method is dependent on a 10 K change in inlet temperature. Comparisons of first- and second-order statistics of temperature, mixture fraction, and formaldehyde between the DNS and experiment show good agreement. The validity of manifold-based turbulent combustion models in turbulent cool flames is then explored by first studying the effective Lewis numbers in the flame using a local differential diffusion parameter analysis. Turbulent cool flames are shown to have effective unity Lewis number transport even at a much lower Reynolds number than typically required for hot flames. Good agreement is shown for temperature and formaldehyde between DNS conditional means and one-dimensional nonpremixed flame solutions. Additionally, the motion of the cool flame in mixture fraction space is shown to be well described by one-dimensional nonpremixed flame solutions. These points indicate that the fundamental physics of turbulent cool flames are analogous to turbulent hot flames, implying that the same reduced-order manifold-based modeling approaches utilized for turbulent hot flames can be utilized for turbulent cool flames.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
NASA NTRSDate not supplied

Expedited Model Building Method for Rocket Nozzle Side Loading

Statistical analysis of test data is currently the preferred method of creating models for uncharacterized side loading in rocket nozzles. However, statistical analysis requires multiple data sets to be reliable; getting these sets requires building and testing multiple rocket motors, which is not feasible given the incredible time and cost required to do so. By using a combination of geometry, simplified finite element analysis (FEA) analogs, and algebra we can create a limited predictive model from only one test data set. Once created, the model requires only the input from one strain gauge located at each yaw and pitch attachment bracket for the actuation system. This process is not computationally intensive, therefore saving an abundance of time and budget over using full 3D FEA models to develop a similar predictive model. To develop the model, conversion factors between the strain at the gauge locations and corresponding actuator load is derived from the data. These conversion factors are used to create a “conversion ellipse” that bounds the constitutive properties of the nozzle as one lumped factor. Using a 2D cross-sectional analog of the nozzle exit plane, we create a function that allows us to determine the azimuthal location of the unexpected load based on the ratio between yaw and pitch strain readings. Combining this with the “conversion ellipse” yields a function that given the strain reading outputs an induced actuator load, regardless of the azimuthal location of the unexpected load. As previously mentioned, this predictive model is limited in its application as there is inaccuracy introduced by the use of data by only one test as well as simplifying assumptions. This approach makes a useful tool to quickly determine whether uncharacterized actuator loading events can damage actuation systems or induce loading that exceeds design limits.

Seetha A Kolli↗
DOE OSTI2024

Deep Learning and Likelihood Approaches for Viral Phylogeography Converge on the Same Answers Whether the Inference Model Is Right or Wrong

Abstract Analysis of phylogenetic trees has become an essential tool in epidemiology. Likelihood-based methods fit models to phylogenies to draw inferences about the phylodynamics and history of viral transmission. However, these methods are often computationally expensive, which limits the complexity and realism of phylodynamic models and makes them ill-suited for informing policy decisions in real-time during rapidly developing outbreaks. Likelihood-free methods using deep learning are pushing the boundaries of inference beyond these constraints. In this paper, we extend, compare, and contrast a recently developed deep learning method for likelihood-free inference from trees. We trained multiple deep neural networks using phylogenies from simulated outbreaks that spread among 5 locations and found they achieve close to the same levels of accuracy as Bayesian inference under the true simulation model. We compared robustness to model misspecification of a trained neural network to that of a Bayesian method. We found that both models had comparable performance, converging on similar biases. We also implemented a method of uncertainty quantification called conformalized quantile regression that we demonstrate has similar patterns of sensitivity to model misspecification as Bayesian highest posterior density (HPD) and greatly overlap with HPDs, but have lower precision (more conservative). Finally, we trained and tested a neural network against phylogeographic data from a recent study of the SARS-Cov-2 pandemic in Europe and obtained similar estimates of region-specific epidemiological parameters and the location of the common ancestor in Europe. Along with being as accurate and robust as likelihood-based methods, our trained neural networks are on average over 3 orders of magnitude faster after training. Our results support the notion that neural networks can be trained with simulated data to accurately mimic the good and bad statistical properties of the likelihood functions of generative phylogenetic models.

Evolutionary Biology↗
DOE OSTI2022

Do Machine Learning Approaches Offer Skill Improvement for Short-Term Forecasting of Wind Gust Occurrence and Magnitude?

Abstract Wind gusts, and in particular intense gusts, are societally relevant but extremely challenging to forecast. This study systematically assesses the skill enhancement that can be achieved using artificial neural networks (ANNs) for forecasting of wind gust occurrence and magnitude. Geophysical predictors from the ERA5 reanalysis are used in conjunction with an autoregressive term in regression and ANN models with different predictors, and varying model complexity. Models are derived and assessed for the warm (April–September) and cold (October–March) seasons for three high passenger volume airports in the United States. Model uncertainty is assessed by deriving models for 1000 different randomly selected training (70%) and testing (30%) subsets. Gust prediction fidelity in independent test samples is critically dependent on inclusion of an autoregressive term. Gust occurrence probabilities derived using five-layer ANNs exhibit consistently higher fidelity than those from regression models and shallower ANNs. Inclusion of the autoregressive term and increasing the number of hidden layers in ANNs from 1 to 5 also improve the model performance for gust magnitudes (lower RMSE, increased correlation, and model standard deviations that more closely approximate observed values). Deeper ANNs (e.g., 20 hidden layers) exhibit higher skill in forecasting strong (17–25.7 m s −1 ) and damaging (≥25.7 m s −1 ) wind gusts. However, such deep networks exhibit evidence of overfitting and still substantially underestimate (by 50%) the frequency of strong and damaging wind gusts at the three airports considered herein. Significance Statement Improved short-term forecasting of wind gusts will enhance aviation safety and logistics and may offer other societal benefits. Here we present a rigorous investigation of the relative skill of models of wind gust occurrence and magnitude that employ different statistical methods. It is shown that artificial neural networks (ANNs) offer considerable skill enhancement over regression methods, particularly for strong and damaging wind gusts. For wind gust magnitudes in particular, application of deeper learning networks (e.g., five or more hidden layers) offers tangible improvements in forecast accuracy. However, deeper networks are vulnerable to overfitting and exhibit substantial variability with the specific training and testing data subset used. Also, even deep ANNs reproduce only half of strong and damaging wind gusts. These results indicate the need for future work to elucidate the dynamical mechanisms of intense wind gusts and advance solutions to their prediction.

54 ENVIRONMENTAL SCIENCES↗