Search NASA⌕ Search

SEARCH · Search NASA

Results for “Statistical modeling”

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 685 records · Page 38

Evidence for charged-particle decay of dipole-excited {sup 4}H clusters embedded in {sup 6}He and {sup 7}He

Charged-particle decay from the dipole resonances at E{sub x}=24 MeV in {sup 6}He and at E{sub x}=18 MeV in {sup 7}He was studied via the {sup 6,7}Li({sup 7}Li,{sup 7}Be x) coincidence experiment at 455 MeV and at 0 deg. These resonances are those that were interpreted as the analogs of the dipole resonances in the {alpha} clusters of {sup 6}Li and {sup 7}Li in our previous work. The dominant decay channels (branching ratios) of the dipole resonances are found to be d+t+n (26{approx}32%) and {alpha}+2n (47{approx}69%) for {sup 6}He and t+t+n (47{approx}55%) and {alpha}+3n (38{approx}47%) for {sup 7}He. The branching ratios measured for the d+t+n and t+t+n channels in {sup 6}He and {sup 7}He, respectively, are larger than those calculated in a simple statistical model by a factor of {approx}100. These results are consistent with the picture that the observed dipole resonances are analogs of the corresponding dipole resonances in the {alpha} clusters of {sup 6}Li and {sup 7}Li.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Dipole strength in {sup 144}Sm studied via (gamma,n), (gamma,p), and (gamma,alpha) reactions

Photoactivation measurements on {sup 144}Sm have been performed with bremsstrahlung endpoint energies from 10.0 to 15.5 MeV at the bremsstrahlung facility of the superconducting electron accelerator ELBE of Forschungszentrum Dresden-Rossendorf. The measured activation yield for the {sup 144}Sm(gamma,n) reaction is compared with the calculated yield using cross sections from previous photoneutron experiments. The activation yields measured for all disintegration channels {sup 144}Sm(gamma,n), (gamma,p), and (gamma,alpha) are compared to the yield calculated by using Hauser-Feshbach statistical models. A new parametrization of the photon strength function is presented and the yield simulated by using the modified photon strength parameters is compared to the experimental data.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Identification of fission-like events in the {sup 16}O + {sup 181}Ta system: Mass and isotopic yield distribution

In this paper, nuclear reaction cross sections for 24 fission-like fragments (30{<=}Z{<=}60) have been measured for the 6.5 MeV/A {sup 16}O + {sup 181}Ta system. The recoil-catcher activation technique was employed followed by off-line {gamma} spectroscopy. The isotopic yield distributions for yttrium and indium isotopes have been obtained from the experimental data. The variance of the presently measured isotopic yield distributions have been found to be in agreement with the literature values. However, the variance of the mass distribution of fission residues has found to be narrower as compared to other relatively heavier systems. A self-consistent approach to determining the isobaric charge dispersion parameters has been adopted. The measured fission cross sections at 97 and 100 MeV are satisfactorily described by a statistical model code. An attempt has been made to explain the production cross sections of intermediate mass residues in the fission of heavy residues populated via complete and/or incomplete fusion processes.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Modeling and Measurement Constraints in Fault Diagnostics for HVAC Systems

Many studies have shown that energy savings of five to fifteen percent are achievable in commercial buildings by detecting and correcting building faults, and optimizing building control systems. However, in spite of good progress in developing tools for determining HVAC diagnostics, methods to detect faults in HVAC systems are still generally undeveloped. Most approaches use numerical filtering or parameter estimation methods to compare data from energy meters and building sensors to predictions from mathematical or statistical models. They are effective when models are relatively accurate and data contain few errors. In this paper, we address the case where models are imperfect and data are variable, uncertain, and can contain error. We apply a Bayesian updating approach that is systematic in managing and accounting for most forms of model and data errors. The proposed method uses both knowledge of first principle modeling and empirical results to analyze the system performance within the boundaries defined by practical constraints. We demonstrate the approach by detecting faults in commercial building air handling units. We find that the limitations that exist in air handling unit diagnostics due to practical constraints can generally be effectively addressed through the proposed approach.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

{alpha}-particle optical potential tests below the Coulomb barrier

The results of two recent papers concerning ({alpha},{gamma}) and ({alpha},n) reaction cross sections close to the reaction thresholds are discussed with regard to predictions of a recent {alpha}-particle regional optical potential. It is found that the new measured cross sections are rather well described especially for the dominant reaction channels. Particular features of the {alpha}-particle optical potential at energies below the Coulomb barrier explain the failure of a former regional potential obtained by analysis of {alpha}-particle elastic scattering alone at higher energies. Additional limitations of statistical model calculations for minor reaction channels are also discussed.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Statistical spatio-temporal properties of the Laser MegaJoule speckle

This paper investigates a statistical model to describe the spatial and temporal properties of hot spots generated by the superimposition of multiple laser beams. In the context of the Laser MegaJoule design, we introduce the formula for contrasts, trajectories and velocities of the speckle pattern. Single bundle of four beams, two-cones and three-cones configurations are considered. Statistical properties of the speckle in the zone where all the beams overlap are studied with different configurations of polarizations. These properties are shown to be very different from the case of one single bundle of four beams. The configuration of polarization has only a slight effect in the two-cones or three cones configuration. Indeed, the impact of the double polarization smoothing is reduced in the area in which all the beams overlap, while it is much more significant when they split. Moreover, the size of the hot-spots decreases as the number of laser beams increases, but we show that their velocity decreases. As a matter of fact, the maximal velocity of hot spots is found to be only about 10{sup -5} of the velocity of light and the integrated contrast is about 15% when the beams overlap.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

In-process monitoring and prediction of droplet quality in droplet-on-demand liquid metal jetting additive manufacturing using machine learning

Abstract In droplet-on-demand liquid metal jetting (DoD-LMJ) additive manufacturing, complex physical interactions govern the droplet characteristics, such as size, velocity, and shape. These droplet characteristics, in turn, determine the functional quality of the printed parts. Hence, to ensure repeatable and reliable part quality it is necessary to monitor and control the droplet characteristics. Existing approaches for in-situ monitoring of droplet behavior in DoD-LMJ rely on high-speed imaging sensors. The resulting high volume of droplet images acquired is computationally demanding to analyze and hinders real-time control of the process. To overcome this challenge, the objective of this work is to use time series data acquired from an in-process millimeter-wave sensor for predicting the size, velocity, and shape characteristics of droplets in DoD-LMJ process. As opposed to high-speed imaging, this sensor produces data-efficient time series signatures that allows rapid, real-time process monitoring. We devise machine learning models that use the millimeter-wave sensor data to predict the droplet characteristics. Specifically, we developed multilayer perceptron-based non-linear autoregressive models to predict the size and velocity of droplets. Likewise, a supervised machine learning model was trained to classify the droplet shape using the frequency spectrum information contained in the millimeter-wave sensor signatures. High-speed imaging data served as ground truth for model training and validation. These models captured the droplet characteristics with a statistical fidelity exceeding 90%, and vastly outperformed conventional statistical modeling approaches. Thus, this work achieves a practically viable sensing approach for real-time quality monitoring of the DoD-LMJ process, in lieu of the existing data-intensive image-based techniques.

Gaikwad, Aniruddha (ORCID:0000000285642621)↗

Deep learning-based predictive models for laser direct drive at the Omega Laser Facility

The rich and complex physics of inertial confinement fusion provides a unique and challenging space for high-fidelity first-principles modeling. Consequently, simulation codes that are used to design experiments are computationally expensive and lack the predictive capability required for extensive parameter exploration in search of a high-performing design for laser direct drive. In this article, we present two deep-learning-based predictive models intended to address these difficulties. The first model (TL DNN) acts as a fast emulator of simulations as well as experiments at the Omega Laser Facility. This model is trained on a simulation database and subsequently calibrated on experimental data using transfer learning. To facilitate the development of this model, an autoencoder is developed to reduce the dimensionality of the input space by compressing the laser pulse input. The model predicts key experimental scalar observables of Omega experiments with high accuracy and minimal computational cost. This deep neural net enables rapid exploration of a high-dimensional input parameter space for an optimal implosion design. The second model (DNN SM+) aims to extend the statistical modeling work of Lees et al. [Phys. Rev. Lett. 127, 105001 (2021)], by increasing the complexity of the model space and allowing for coupling between degradation terms. Since the model capacity of DNN SM+ is higher than the model of Lees et al., DNN SM+ can potentially provide an improvement in predictive capability, and we use this model to provide insight into complicated degradation dependencies.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

InClass nets: independent classifier networks for nonparametric estimation of conditional independence mixture models and unsupervised classification

Abstract Conditional independence mixture models (CIMMs) are an important class of statistical models used in many fields of science. We introduce a novel unsupervised machine learning technique called the independent classifier networks (InClass nets) technique for the nonparameteric estimation of CIMMs. InClass nets consist of multiple independent classifier neural networks (NNs), which are trained simultaneously using suitable cost functions. Leveraging the ability of NNs to handle high-dimensional data, the conditionally independent variates of the model are allowed to be individually high-dimensional, which is the main advantage of the proposed technique over existing non-machine-learning-based approaches. Two new theorems on the nonparametric identifiability of bivariate CIMMs are derived in the form of a necessary and a (different) sufficient condition for a bivariate CIMM to be identifiable. We use the InClass nets technique to perform CIMM estimation successfully for several examples. We provide a public implementation as a Python package called RainDancesVI.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Unified description of fission in fusion and spallation reactions

We present a statistical-model description of fission, in the framework of compound-nucleus decay, which is found to simultaneously reproduce data from both heavy-ion-induced fusion reactions and proton-induced spallation reactions at around 1 GeV. For the spallation reactions, the initial compound-nucleus population is predicted by the Liege intranuclear cascade model. We are able to reproduce experimental fission probabilities and fission-fragment mass distributions in both reactions types with the same parameter sets. However, no unique parameter set was obtained for the fission probability. The introduction of fission transients can be offset by an increase of the ratio of level-density parameters for the saddle-point and ground-state configurations. Changes to the finite-range fission barriers could be offset by a scaling of the Bohr-Wheeler decay width as predicted by Kramers. The parameter sets presented allow accurate prediction of fission probabilities for excitation energies up to 300 MeV and spins up to 60 ({h_bar}/2{pi}).

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Total reaction cross sections from elastic {alpha}-nucleus scattering angular distributions around the Coulomb barrier

The total reaction cross section {sigma}{sub reac} is a valuable measure for the prediction of {alpha}-induced reaction cross sections within the statistical model and for the comparison of scattering of tightly bound projectiles to weakly bound and exotic projectiles. Here we provide the total reaction cross sections {sigma}{sub reac} derived from our previously published angular distributions of elastic {alpha}-nucleus scattering on {sup 89}Y, {sup 92}Mo, {sup 112,124}Sn, and {sup 144}Sm at energies around the Coulomb barrier.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Nucleon-content effect on the fission lifetime of excited nuclei

It was shown previously that the fission lifetime of a nucleus excited to about 100 MeV depends strongly and nonmonotonically on the initial value of its angular momentum L{sub 0}. This result was obtained on the basis of a refined version of the combined dynamical and statistical model. The present study is devoted to a theoretical analysis of the dependence of the fission time on the nucleonic composition of the nucleus involved. The respective calculations were performed within the same model. The dependence of the average fission time <t{sub f}> on the initial fissility parameter (Z{sup 2}/A){sub 0} appears to be of a resonance type and is similar to its dependence on L{sub 0}. This dependence of the average fission time on (Z{sup 2}/A){sub 0} stems both from statistical calculations and from a dynamical simulation of the fission mode with allowance for friction. The conditions under which the average fission time reaches a maximum are specified. The dependence of the average fission time on (Z{sup 2}/A){sub 0} remains nonmonotonic in the fusion-fission reaction as well, in which case the distribution of compound nuclei with respect to the initial angular momentum is broad.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Validation and Calibration of Nuclear Thermal Hydraulics Multiscale Multiphysics Models - Subcooled Flow Boiling Study

In addition to validation data plan, development of advanced techniques for calibration and validation of complex multiscale, multiphysics nuclear reactor simulation codes are a main objective of the CASL VUQ plan. Advanced modeling of LWR systems normally involves a range of physico-chemical models describing multiple interacting phenomena, such as thermal hydraulics, reactor physics, coolant chemistry, etc., which occur over a wide range of spatial and temporal scales. To a large extent, the accuracy of (and uncertainty in) overall model predictions is determined by the correctness of various sub-models, which are not conservation-laws based, but empirically derived from measurement data. Such sub-models normally require extensive calibration before the models can be applied to analysis of real reactor problems. This work demonstrates a case study of calibration of a common model of subcooled flow boiling, which is an important multiscale, multiphysics phenomenon in LWR thermal hydraulics. The calibration process is based on a new strategy of model-data integration, in which, all sub-models are simultaneously analyzed and calibrated using multiple sets of data of different types. Specifically, both data on large-scale distributions of void fraction and fluid temperature and data on small-scale physics of wall evaporation were simultaneously used in this work’s calibration. In a departure from traditional (or common-sense) practice of tuning/calibrating complex models, a modern calibration technique based on statistical modeling and Bayesian inference was employed, which allowed simultaneous calibration of multiple sub-models (and related parameters) using different datasets. Quality of data (relevancy, scalability, and uncertainty) could be taken into consideration in the calibration process. This work presents a step forward in the development and realization of the “CIPS Validation Data Plan” at the Consortium for Advanced Simulation of LWRs to enable quantitative assessment of the CASL modeling of Crud-Induced Power Shift (CIPS) phenomenon, in particular, and the CASL advanced predictive capabilities, in general. This report is prepared for the Department of Energy’s Consortium for Advanced Simulation of LWRs program’s VUQ Focus Area.

97 MATHEMATICS AND COMPUTING↗

Validation of MCNP Critical Benchmark Models of PU-MET-FAST-016 [Slides]

The validation of PU-MET-FAST-016 contributed to the centralized LANL benchmark repository currently under development. The revisions made the MCNP models statistically, significantly more similar to the benchmark models. The overall impact on the USL is negligible. The revisions to the PU-MET-FAST-016 models provide value to the Los Alamos Benchmark Suite without invalidating past Whisper results.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Predictive Indicators of the Performance of Large Language Models

In several mission contexts, it is desirable to estimate the performance of large language models (LLMs) on tasks that we cannot run directly. In light of published “scaling laws” our hypothesis is that some tasks should be consistently more challenging than others based on characteristics of the task. The goal of this project was to begin quantifying how much information about LLM performance can be gained from the features of a model and a task. Two of our statistical models struggled to converge. Pass/fail test results may provide limited information for inference beyond model quality and task difficulty, but we see no evidence at this time for significant feature interaction effect sizes, arguing for simple models. Future work extending the models to capitalize on perplexity of ground truth answers is suggested. This project also introduces “Depth of Knowledge Variant Testing” as a strategy for more finely assessing language models on open domain question and answer tasks. We developed sets of questions that ask a language model to produce similar information while demonstrating increasing depth of knowledge, and also relabeled existing Q&A test questions with their depth of knowledge. Our results suggest further consideration of Bloom’s taxonomy and further refinement of prompts to properly elicit information at varying depths. In the course of this work, we set up a basic infrastructure for standardizing tasks and testing many language models on these tasks. In addition to testing the predictive quality of model features and performance across test suites, with this project we have introduced two new task features to contextualize each test question: the Dewey Classification main category of information covered, and the Bloom’s taxonomy level that corresponds to the depth of knowledge probed by the question. Splits across these and other features produced over five hundred task subtypes with distinct feature vectors, which we tested on half a dozen models.

97 MATHEMATICS AND COMPUTING↗

Bayesian chain graph models to characterize microbe-environment dynamics

Microbiome data require statistical models that can simultaneously decode microbes' reaction to the environment and interactions among microbes. While a multiresponse linear regression model seems like a straight-forward solution, we argue that treating it as a graphical model is problematic given that the regression coefficient matrix does not encode the conditional dependence structure between response and predictor nodes. This observation is especially important in biological settings when we have prior knowledge on the edges from specific experimental interventions that can only be properly encoded under a conditional dependence model. Here, we propose a chain graph model with two sets of nodes (predictors and responses) whose solution yields a graph with edges that indeed represent conditional dependence, thus agreeing with the experimenter's intuition on the average behavior of nodes under treatment. The solution to our model is sparse via the Bayesian linear regression (LASSO). In addition, we propose an adaptive extension so that different shrinkages can be applied to different edges to incorporate edge-specific prior knowledge. Our model is computationally inexpensive through an efficient Gibbs sampling algorithm and can account for binary, counting, and compositional responses via an appropriate hierarchical structure. We test the performance of our model in a variety of simulated datasets, thereby showing superior performance to state-of-the-art approaches. We further apply our model to human gut and soil microbial compositional datasets, and we highlight that CG-LASSO can estimate biologically meaningful network structures in the data.

compositional data↗

Coordination procedure for radio relay and communication satellite services

A global rain rate statistic model is used to link microwave propagation statistics to measurable rain statistics in order to develop international telecommunication site criteria for radio relay and communication satellite services that minimize interference between receivers and transmitters. This rain coordination procedure utilizes a rain storm cell size, a statistical description of the rainfall rate within the cell valid for most of the earth's surface, approximations between Raleigh scatter and constancy of precipitation with altitude, and an analytic relation between radar reflectivity and rain rate.

Eckerman, J.↗

Magnetospheric plasma modeling (0-100 keV)

Spacecraft surface charging, which is primarily a current balance phenomenon, is in general a function of the dominant currents to and from the vehicle's surface. Within the near-earth magnetosphere the dominant currents to the surface are the ambient space plasma fluxes between approximately 0 and 100 keV. A major effort to understand the near-earth environment was initiated when spacecraft charging became a major issue. The present paper has the objective to summarize the basic features of the models which have resulted from this effort. A description is given of four categories of models, based primarily on the degree of empirical and theoretical input. Types of quantitative models are discussed, taking into account definitions, statistical models, analytic models, static models, and time-dependent models. Engineering models are also considered, giving attention to baseline models and 'worst-case' models.

Garrett, H. B.↗