SEARCH · Engineering Papers

Results for “uncertainty analysis”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,675 records · Page 93

All sources NASA NTRS DOE OSTI

DOE OSTI2020

Dark Energy Survey Year 3 results: cosmology with moments of weak lensing mass maps – validation on simulations

ABSTRACT We present a simulated cosmology analysis using the second and third moments of the weak lensing mass (convergence) maps. The second moment, or variances, of the convergence as a function of smoothing scale contains information similar to standard shear two-point statistics. The third moment, or the skewness, contains additional non-Gaussian information. The analysis is geared towards the third year (Y3) data from the Dark Energy Survey (DES), but the methodology can be applied to other weak lensing data sets. We present the formalism for obtaining the convergence maps from the measured shear and for obtaining the second and third moments of these maps given partial sky coverage. We estimate the covariance matrix from a large suite of numerical simulations. We test our pipeline through a simulated likelihood analyses varying 5 cosmological parameters and 10 nuisance parameters and identify the scales where systematic or modelling uncertainties are not expected to affect the cosmological analysis. Our simulated likelihood analysis shows that the combination of second and third moments provides a 1.5 per cent constraint on S8 ≡ σ8(Ωm/0.3)0.5 for DES Year 3 data. This is 20 per cent better than an analysis using a simulated DES Y3 shear two-point statistics, owing to the non-Gaussian information captured by the inclusion of higher order statistics. This paper validates our methodology for constraining cosmology with DES Year 3 data, which will be presented in a subsequent paper.

79 ASTRONOMY AND ASTROPHYSICS↗

NASA NTRS2005

Aerocapture Performance Analysis of A Venus Exploration Mission

A performance analysis of a Discovery Class Venus Exploration Mission in which aerocapture is used to capture a spacecraft into a 300km polar orbit for a two year science mission has been conducted to quantify its performance. A preliminary performance assessment determined that a high heritage 70 sphere-cone rigid aeroshell with a 0.25 lift to drag ratio has adequate control authority to provide an entry flight path angle corridor large enough for the mission s aerocapture maneuver. A 114 kilograms per square meter ballistic coefficient reference vehicle was developed from the science requirements and the preliminary assessment s heating indicators and deceleration loads. Performance analyses were conducted for the reference vehicle and for sensitivity studies on vehicle ballistic coefficient and maximum bank rate. The performance analyses used a high fidelity flight simulation within a Monte Carlo executive to define the aerocapture heating environment and deceleration loads and to determine mission success statistics. The simulation utilized the Program to Optimize Simulated Trajectories (POST) that was modified to include Venus specific atmospheric and planet models, aerodynamic characteristics, and interplanetary trajectory models. In addition to Venus specific models, an autonomous guidance system, HYPAS, and a pseudo flight controller were incorporated in the simulation. The Monte Carlo analyses incorporated a reference set of approach trajectory delivery errors, aerodynamic uncertainties, and atmospheric density variations. The reference performance analysis determined the reference vehicle achieves 100% successful capture and has a 99.87% probability of attaining the science orbit with a 90 meters per second delta V budget for post aerocapture orbital adjustments. A ballistic coefficient trade study conducted with reference uncertainties determined that the 0.25 L/D vehicle can achieve 100% successful capture with a ballistic coefficient of 228 kilograms per square meter and that the increased ballistic coefficient increases post aerocapture V budget to 134 meters per second for a 99.87% probability of attaining the science orbit. A trade study on vehicle bank rate determined that the 0.25 L/D vehicle can achieve 100% successful capture when the maximum bank rate is decreased from 30 deg/s to 20 deg/s. The decreased bank rate increases post aerocapture delta V budget to 102 meters per second for a 99.87% probability of attaining the science orbit.

Starr, Brett R.↗

DOE OSTI2023

Tank 27 Solids Disturbance Testing Using M-Star for the Impact of CSMP Usage on Blend Tank Mixing Effectiveness

Savannah River Mission Completion (SRMC) plans to operate Tank 27as a blend tank for future waste disposal operations. It is desirable to avoid additional worker exposure by using the planned commercial submersible mixing pump (CSMP) installations in Risers B2 and B4 to function as the blend pumps for prolonged blend tank operations. The purpose of this task is to evaluate the impacts to mixing across a range of pump installation heights from 1 to 201 inches above the tank floor. The CSMPs shall maintain adequate blending within a 24 hour maximum timeframe and an adequate nozzle discharge velocity as outlined in U-ESR-G-00030, using the specifications of the planned CSMPs listed in PO SRRA149971 and M-DS-G-00081. Specifications include: maximum pump diameter of D=22.5 in., dual nozzles placed tangentially opposed with nominal inside diameter of 2 in., total flow range (both nozzles combined) of 500 to 1950 GPM and maximum RPM of 1800. Tank 27 is expected to have a fill height maximum of 360". This analysis began by using results from the 2010 testing and applying them to the Tank 27 design to calculate a mixing time as a function of pump flow rate. These results were modified to account for the higher liquid level in Tank 27. Once the liquid level was increased to 360 inches, the results were modified to account for a higher viscosity in Tank 27 than in the testing. Once the viscosity was increased, the effect of changing pump elevation was applied to calculations of the mixing time in Tank 27. Finally, the uncertainty from the testing was included in the analysis to recommend a blend time as a function of pump flow rate and elevation. The conclusion from this study is that the particles in tank 27 will be suspended at both the 600 and 900 gpm flow rates. Only particles above 90 micron will not be suspended in significant amounts when the flow rate is 600 gpm. At 900 gpm even 100 micron particles were thoroughly suspended in the tank. From the series of simulations it has been concluded that it will not be possible to mix tank 27 at the proposed fluid flow rates and not suspend the solid particles at the bottom of the tank.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

NASA NTRS2010

Finite Element Model Calibration Approach for Ares I-X

Ares I-X is a pathfinder vehicle concept under development by NASA to demonstrate a new class of launch vehicles. Although this vehicle is essentially a shell of what the Ares I vehicle will be, efforts are underway to model and calibrate the analytical models before its maiden flight. Work reported in this document will summarize the model calibration approach used including uncertainty quantification of vehicle responses and the use of nonconventional boundary conditions during component testing. Since finite element modeling is the primary modeling tool, the calibration process uses these models, often developed by different groups, to assess model deficiencies and to update parameters to reconcile test with predictions. Data for two major component tests and the flight vehicle are presented along with the calibration results. For calibration, sensitivity analysis is conducted using Analysis of Variance (ANOVA). To reduce the computational burden associated with ANOVA calculations, response surface models are used in lieu of computationally intensive finite element solutions. From the sensitivity studies, parameter importance is assessed as a function of frequency. In addition, the work presents an approach to evaluate the probability that a parameter set exists to reconcile test with analysis. Comparisons of pre-test predictions of frequency response uncertainty bounds with measured data, results from the variance-based sensitivity analysis, and results from component test models with calibrated boundary stiffness models are all presented.

Horta, Lucas G.↗

NASA NTRS1986

Predictive momentum management for a space station measurement and computation requirements

An analysis is made of the effects of errors and uncertainties in the predicting of disturbance torques on the peak momentum buildup on a space station. Models of the disturbance torques acting on a space station in low Earth orbit are presented, to estimate how accurately they can be predicted. An analysis of the torque and momentum buildup about the pitch axis of the Dual Keel space station configuration is formulated, and a derivation of the Average Torque Equilibrium Attitude (ATEA) is presented, for the case of no MRMS (Mobile Remote Manipulation System) motion, Y vehicle axis MRMS motion, and Z vehicle axis MRMS motion. Results showed the peak momentum buildup to be approximately 20000 N-m-s and to be relatively insensitive to errors in the predicting torque models, for Z axis motion of the MRMS was found to vary significantly with model errors, but not exceed a value of approximately 15000 N-m-s for the Y axis MRMS motion with 1 deg attitude hold error. Minimum peak disturbance momentum was found not to occur at the ATEA angle, but at a slightly smaller angle. However, this minimum peak momentum attitude was found to produce significant disturbance momentum at the end of the predicting time interval.

Adams, John Carl↗

NASA NTRS1992

Iron abundance in the hot DA white dwarfs Feige 24 and G191 B2B

Attention is given to model calculations of the far- and extreme-UV line spectra of highly ionized Fe species (Fe IV, Fe V, and Fe VI) for hot high-gravity H-rich stars. A spectral analysis of 31 hr of exposure of the DA white dwarf Feige 24 with IUE in the echelle mode reveals the presence of Fe with an abundance relative to H by number of (5-10) x 10 exp -6 with an uncertainty dominated by the determination of stellar parameters. An analysis of IUE data from the white dwarf G191 B2B results in a similar Fe abundance if this star shares similar atmospheric parameters (Teff, g) with Feige 24. Fe is thus the second most abundant photospheric element in hot DA white dwarfs.

Vennes, Stephane↗

DOE OSTI2026

Prioritizing Uncertainties in Hydrogen Contribution to Risk in Post-Crash Outcomes for Rail

This report presents analysis from Sandia National Laboratories predicting contributions to risk associated with the use of hydrogen technology for rail. Event sequence diagrams are used to describe possible accident scenarios and progressions. Initiating event frequencies and branch event probabilities for each scenario are quantified with uncertainty using distributions fit to Federal Railroad Administration and U.S. Department of Transportation Pipeline and Hazardous Materials Safety Administration data on applicable accidents from 2000 to 2020. Uncertainty is propagated through the event sequence diagram to estimate the frequency and conditional probability of accident end states. The analysis identifies four scenarios with significant contributions to risk from hydrogen that are predicted to occur relatively frequently, which may inform priorities for reducing uncertainty. These scenarios are 1) overpressure events resulting from collisions with hydrogen release due to mechanical damage and delayed ignition, 2) jet fire events resulting from collisions with hydrogen release due to mechanical damage and immediate ignition, 3) jet fires resulting from fire or explosion initiating events involving the hydrogen tank and correct operation of the thermally-activated pressure relief device (TPRD) subsequent to the thermal insult, and 4) pressure burst resulting from fire or explosion initiating events involving the hydrogen tank and failure of the TPRD. Delayed and immediate hydrogen ignition probabilities are identified as being highly uncertain and potential candidates for reducing conservatism in the predicted frequencies for these two scenarios.

08 HYDROGEN↗

DOE OSTI2021

Uniformly Ordered Binary Decision Algorithm for Benchmark Experiment Correlations in Whisper Validation

When performing a validation exercise for determining the upper subcritical limit of a nuclear criticality safety application, an analyst should select and perform a statistical analysis on a population of benchmark experiments that are neutronically similar to the application. The size of this population should be sufficiently large such that the statistical analysis has a high degree of confidence that the bias plus bias uncertainty (calculational margin) has been accurately quantified. A complication arises because many benchmark experiments share common components, leading to correlations in their measured effective multiplication factors. Correlations between benchmark experiments within the population reduces its predictive power. This motivates the need for methods that consider benchmark experiment correlations and ensure adequate statistical significance of results. The Whisper code is a statistical analysis pack- age that incorporates nuclear data sensitivity coefficients from MCNP to assess benchmark experiment similarity and then performs an extreme-value analysis to estimate the bias plus bias uncertainty. The original methodology in Whisper does not consider the effect of benchmark experiment correlations when making this estimation, and this summary proposes the uniformly ordered binary decision algorithm to address this shortcoming. The original methodology in Whisper computes similarity coefficients ck for an application compared to all benchmark experiments in its library and develops weighting factors for a selected population proportional to the ck values. The methodology can be interpreted as statistically emulating a validation exercise for a particular application where the weighting factors may be viewed as the likelihood that an analyst would include a particular benchmark experiment within the population. The effective sample size of the population is the expected or mean number of benchmark experiments in the population. The uniformly ordered binary decision algorithm identifies clusters of correlated benchmark experiments within the population and then computes adjusted weighting factors based on the magnitude of the correlation coefficients within the cluster to compute a reduced effective sample size accounting for the lower information content because of correlations. Benchmark experiments within the cluster are ordered randomly with equal probability and probabilistic decisions are made as to whether a benchmark. experiment within the cluster should treated as redundant with a previous one; if two redundant benchmark experiments are included, then the conservative worst case bias plus bias uncertainty is used and the pair is counted as a single benchmark experiment in the population. Results are provided for HEU solutions in a research version of the Whisper software using benchmark experiment correlations provided by DICE, the Database for the International Criticality Safety Benchmark Evaluation Project (ICSBEP). These show that there can be a significant increase in the bias plus bias uncertainty because the effective sample size is reduced, and therefore the algorithm, needing to meet sample size requirements, expands the benchmark experiment population by accepting less similar benchmark experiments that would have otherwise not been included.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

NASA NTRS2008

Assessment of the LC-2 Prelaunch Fatigue Spectra of the CM-to-SM Flange Weld

The pad stay and rollout components of the Ares I-X life cycle can generate cyclic stress oscillations to the vehicle that could initiate and grow fatigue cracks from weld defects. The Ares I-X Project requested that a study be performed to determine if stabilization of the vehicle is required to reduce the stresses that could initiate and grow fatigue cracks at the flange-to-skin weld of the Command Module (CM) and Service Module (SM) interface. A fatigue crack growth analysis was conducted that used loads (LC-2) and stress analyses developed by the Ares I-X Project and utilized material data and analysis methods developed by a critical initial flaw size (CIFS) analysis conducted by NASA Engineering and Safety Center (NESC) for the Upper Stage Simulator (USS) of the Ares I-X vehicle. A full CIFS analysis for the CM-to-SM flange-to-skin weld was not performed because the full flight spectrum was not provided and was not necessary to answer the question posed by the Ares I-X Project. Instead, an approach was developed to determine if the crack growth due to the pad stay and rollout components of the flight spectrum would adversely influence the CIFS. The approach taken used a number of conservative assumptions that eliminated the need for high-fidelity analyses and additional material testing, but still provided a bounding solution for the uncertainties of the problem. The results from this analysis indicate that the LC-2 pad stay and rollout spectrum components would not produce significant fatigue crack growth on the CM-to-SM flange-to-skin weld. Thus, from a fatigue crack growth standpoint, no stabilization is required to reduce the LC-2 pad stay and rollout cyclic stresses on the CM-to-SM flange-to-skin weld.

Dawicke, David S.↗

DOE OSTI2024

Uncertainty characterization in a coupled human-natural system: Modeling agricultural adaptation in the Great Lakes Region

The Great Lakes Region's water quality and ecological health are threatened by the export of nutrients from agricultural lands, which causes eutrophication, hypoxia, and destructive algal blooms. The intensification of hydrologic cycles brought about by climate change is expected to exacerbate nutrient loading in the region, and, at the same time, agricultural adaptation to changing conditions is also expected to affect loading through shifting amounts and timing of fertilization. Quantifying these future effects and their interactions necessitates modeling both the human and natural processes as a coupled system, by pairing land use and agricultural management with hydrologic modeling. At the same time, compounding uncertainties arising from the complex interactions in both systems significantly limit our predictive understanding of the region's impacts. This study utilizes the Soil and Water Assessment Tool (SWAT), developed for simulating the impact of various farmer decisions on watershed functions in Western Lake Erie watersheds, and an under-development agent-based model (ABM) for agricultural management decisions. The aim of this study is to use global sensitivity analysis on the coupled ABM and SWAT models to quantify how uncertainty in both models interactively affects nutrient loading. To do so, we will conduct Sobol sensitivity analysis experiments at different levels of coupling assumptions to quantify how various uncertain factors (e.g., soil moisture and crop choice) and their interactions affect our estimates of nutrient loading. The results of this analysis will allow us to quantify how complex interactions and dependencies between both systems amplify the effect of uncertainties. Insights gained from this study will have broader implications for modeling the adaptive co-evolution of human and natural systems under climate change and can inform effective management of nutrient loading in the Great Lakes Region.

Climate Change↗

NASA NTRS2010

Probabilistic Structural Analysis Program

NASA/NESSUS 6.2c is a general-purpose, probabilistic analysis program that computes probability of failure and probabilistic sensitivity measures of engineered systems. Because NASA/NESSUS uses highly computationally efficient and accurate analysis techniques, probabilistic solutions can be obtained even for extremely large and complex models. Once the probabilistic response is quantified, the results can be used to support risk-informed decisions regarding reliability for safety-critical and one-of-a-kind systems, as well as for maintaining a level of quality while reducing manufacturing costs for larger-quantity products. NASA/NESSUS has been successfully applied to a diverse range of problems in aerospace, gas turbine engines, biomechanics, pipelines, defense, weaponry, and infrastructure. This program combines state-of-the-art probabilistic algorithms with general-purpose structural analysis and lifting methods to compute the probabilistic response and reliability of engineered structures. Uncertainties in load, material properties, geometry, boundary conditions, and initial conditions can be simulated. The structural analysis methods include non-linear finite-element methods, heat-transfer analysis, polymer/ceramic matrix composite analysis, monolithic (conventional metallic) materials life-prediction methodologies, boundary element methods, and user-written subroutines. Several probabilistic algorithms are available such as the advanced mean value method and the adaptive importance sampling method. NASA/NESSUS 6.2c is structured in a modular format with 15 elements.

Pai, Shantaram S.↗

DOE OSTI2025

MOOSE ProbML: Parallelizable Probabilistic Machine Learning and Uncertainty Quantification Capabilities

The Multiphysics Object Oriented Simulation Environment (MOOSE) is a widely used open- source finite element software for performing multiphysics multiscale simulations in a massively parallel fashion. Recently, the computational team at Idaho National Laboratory (INL) has implemented Probabilistic Machine Learning (ProbML) capabilities in MOOSE—in a parallelized fashion—and enable active learning with large-scale computational models for tasks such as surrogate model development, scale bridging, forward/inverse uncertainty quantification (UQ), Bayesian optimization, etc. This presentation summarizes these developments in MOOSE along with demonstrations on several real applications relevant to nuclear energy. At the fundamental level, samplers like Monte Carlo/Latin Hypercube, variance reduction, parallelized Markov Chain Monte Carlo (MCMC) support uncertainty propagation in both forward and inverse settings. These samplers can be integrated with the Gaussian processes (GP) suite in MOOSE, which offer several variants like scalar GPs, multi-output GPs, and deep GPs, to enable active learning. These GPs can be tuned using gradient-based optimization methods like Adam and its variants or gradient-free methods like the elliptical slice sampler (a variant of MCMC adept under Gaussian settings) for more complex covariance kernels or likelihoods whose gradient computations can be cumbersome. A variety of batch acquisition functions permit parallelized evaluation of the computational model and support different learning objectives with high efficiency like Bayesian inference, global surrogate development, optimization, etc. Furthermore, libtorch integration supports training, evaluation, and re-training of neural networks and other complex machine learning models in active learning settings. The impacts of these developments are shown on several real applications: (1) nuclear fuel inverse UQ and model inadequacy assessment using the Kennedy O’Hagan framework; (2) uncertainty aware surrogate modeling for additive manufacturing to predict field quantities; (3) nuclear reactor rare events analysis; and (4) complex fluid flow prediction using a global surrogate with quantified prediction uncertainty. Finally, the outlook of MOOSE ProbML is discussed for both outer-loop and inner-loop computations in the broad view to accelerate fuels and materials qualification, address gaps in knowledge and data, and assess new reactor/fuel systems.

11 - NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

DOE OSTI2021

Climate-Relevant Gas Absorption Properties from AWARE and Other ARM Spectral Measurements (Final Report)

The objective of this project has been to use measurements from the Atmospheric Emitted Radiance Interferometer (AERI) deployed at ARM sites to improve our knowledge of uncertain infrared spectroscopic parameters of importance to climate, remote sensing, and data assimilation. A primary focus has been the water vapor continuum in the infrared atmospheric window. Radiation codes used to predict climate and weather base their representation of the water vapor continuum in this window on the MT_CKD model, but in this spectral region the MT_CKD water vapor continuum absorption coefficients were derived almost two decades ago by an analysis of AERI measurements for conditions with a limited range of precipitable water vapor (PWV) and temperature values, leading to uncertainty in the derived coefficients. This project has included a new, comprehensive analysis of all aspects of the water vapor continuum in the atmospheric window -- the self continuum, the foreign continuum, and the self continuum temperature dependence, all resolved spectrally. Having a wide range of PWV amounts and temperatures from deployments of the ARM Mobile Facility for GoAmazon as well as the more than a decade of AERI measurements at SGP has provided an appropriate foundation for such a comprehensive analysis.

54 ENVIRONMENTAL SCIENCES↗

NASA NTRS1988

A Design Study of Onboard Navigation and Guidance During Aerocapture at Mars

The navigation and guidance of a high lift-to-drag ratio sample return vehicle during aerocapture at Mars are investigated. Emphasis is placed on integrated systems design, with guidance algorithm synthesis and analysis based on vehicle state and atmospheric density uncertainty estimates provided by the navigation system. The latter utilizes a Kalman filter for state vector estimation, with useful update information obtained through radar altimeter measurements and density altitude measurements based on IMU-measured drag acceleration. A three-phase guidance algorithm, featuring constant bank numeric predictor/corrector atmospheric capture and exit phases and an extended constant altitude cruise phase, is developed to provide controlled capture and depletion of orbital energy, orbital plane control, and exit apoapsis control. Integrated navigation and guidance systems performance are analyzed using a four degree-of-freedom computer simulation. The simulation environment includes an atmospheric density model with spatially correlated perturbations to provide realistic variations over the vehicle trajectory. Navigation filter initial conditions for the analysis are based on planetary approach optical navigation results. Results from a selection of test cases are presented to give insight into systems performance.

Fuhry, Douglas Paul↗

NASA NTRS2001

On the Application of Hilbert Spectral Analysis for Climate Studies

The Hilbert spectral analysis (Huang et al, 1998, Proceedings of the Royal Society of London, A 454, pp 903-995) consisted of two steps: First, the data has to be reduced into a finite number of Intrinsic Mode Function by the Empirical Mode Decomposition method, then the resulting Intrinsic Mode Functions are converted to time-frequency-energy distribution through Hilbert transform. In this approach, the Empirical Mode Functions served as the basis functions with which the data is expanded. This basis function is adaptive, and the decomposition is nonlinear. Furthermore, as the Hilbert transform is a singular transform, it retains a high degree of local information. The instantaneous frequency is determined by differentiation of the phase function; therefore, there is no restriction of the 'uncertainty principle' for all the time-frequency analysis resulting from a priori basis approach. With the adaptive basis and the instantaneous frequency, the Hilbert Spectral analysis can represent data from nonlinear and nonstationary processes without resorting to the harmonics. Another advantage of using instantaneous frequency is the ability to find out frequency from limited length of data, which is a critical problem in climate studies. As the processes driving the climate changes could be both nonlinear and nonstationary, the Hilbert Spectral Analysis could be of great use in examining the underlying mechanisms. A preliminary study based on the length of day data will be presented as example for the application of the Hilbert Spectral Analysis for climate study.

Huang, Norden E.↗

DOE OSTI2023

Driving mode analysis—How uncertain functional inputs propagate to an output

Abstract Driving mode analysis elucidates how correlated features of uncertain functional inputs jointly propagate to produce uncertainty in the output of a computation. Uncertain input functions are decomposed into three terms: the mean functions, a zero‐mean driving mode, and zero‐mean residual. The random driving mode varies along a single direction, having fixed functional shape and random scale. It is uncorrelated with the residual, and under linear error propagation, it produces an output variance equal to that of the full input uncertainty. Finally, the driving mode best represents how input uncertainties propagate to the output because it minimizes expected squared Mahalanobis distance amongst competitors. These characteristics recommend interpretation of the driving mode as the single‐degree‐of‐freedom component of input uncertainty that drives output uncertainty. We derive the functional driving mode, show its superiority to other seemingly sensible definitions, and demonstrate the utility of driving mode analysis in an application. The application is the simulation of neutron transport in criticality experiments. The uncertain input functions are nuclear data that describe how Pu reacts to bombardment by neutrons. Visualization of the driving mode helps scientists understand what aspects of correlated functional uncertainty have effects that either reinforce or cancel one another in propagating to the output of the simulation.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

On the Need to Align Intent and Implementation in Uncertainty Quantification for Machine Learning

Quantifying uncertainties for machine learning (ML) models is a foundational challenge in modern data analysis. This challenge is compounded by at least two key aspects of the field: (a) inconsistent terminology surrounding uncertainty and estimation across disciplines, and (b) the varying technical requirements for establishing trustworthy uncertainties in diverse problem contexts. In this position paper, we aim to clarify the depth of these challenges by identifying these inconsistencies and articulating how different contexts impose distinct epistemic demands. We examine the current landscape of estimation targets (e.g., prediction, inference, simulation-based inference), uncertainty constructs (e.g., frequentist, Bayesian, fiducial), and the approaches used to map between them. Drawing on the literature, we highlight and explain examples of problematic mappings. To help address these issues, we advocate for standards that promote alignment between the \textit{intent} and \textit{implementation} of uncertainty quantification (UQ) approaches. We discuss several axes of trustworthiness that are necessary (if not sufficient) for reliable UQ in ML models, and show how these axes can inform the design and evaluation of uncertainty-aware ML systems. Our practical recommendations focus on scientific ML, offering illustrative cases and use scenarios, particularly in the context of simulation-based inference (SBI).

Trivedi, Shubhendu [MIT] (ORCID:0000000312374301)↗

DOE OSTI2024

Measurement of the muon spin precession frequency using the straw tracking detectors at the Fermilab Muon g-2 experiment

The measurement of the anomalous magnetic dipole moment of the muon ($a_{\mu}$) has long stood as an excellent precision test of the Standard Model (SM). The Fermilab Muon g-2 experiment has recently finished data-taking and in July 2023 published its latest determination of $a_\mu$ with a world-leading precision of 0.2\,ppm. In this publication, it surpassed the systematic uncertainty goal defined in the TDR. The analyses of a dataset approximately four times larger than this recent publication is now underway. The principle measurement of the Muon g-2 experiment measures $a_{\mu}$ by taking the ratio of two frequencies; the anomalous precession frequency ($\omega_a$) and the muon-weighted magnetic field of the experiment's storage ring measured from the precession frequency of protons in water using nuclear magnetic resonance (NMR) probes. In all publications to date, $\omega_a$ has been determined using energy deposits in the 24 calorimeters. However, the Fermilab experiment has t wo straw tracker detectors measuring the time and momentum of charged particles which can in principle also be used to to measure $\omega_a$ and such a measurement can provide an invaluable cross-check of the calorimeter result with different, and reduced, systematic uncertainties. This thesis presents the first (blinded) determination of $\omega_a$ using just charged tracks from the straw tracking detectors as opposed to calorimeter energy deposits. This analysis was undertaken using the Run-2/3 dataset which represents approximately 25\% of the final dataset. A total uncertainty of 2.19\,ppm on $\omega_a$ was obtained which is dominated by the statistical uncertainty of 2.16\,ppm. Additionally two new methodologies important to the analysis of the straw tracking data have been developed: one to better determine the track arrival time ($t_0$) and one to determine the level of pileup in the tracking detectors. The new $t_0$ algorithm which incorporates angular information improves t he resolution on the determination of the $t_0$ by a factor of two and results in 19\% more tracks being successfully reconstructed. The data from the trackers is also used to determine the beam profile that weights the magnetic field in the determination of $a_\mu$ and in determining several of the systematic uncertainties in the calorimeter-based $\omega_a$ analysis. A detailed study of the impact of the internal alignment of the tracker, the $t_0$ and pileup on the determination of the beam position was undertaken and propagated through to an uncertainty in the $\omega_a$ determination. These uncertainties were used in the Fermilab Muon g-2 experiment's recent publication in Phys. Rev. Lett.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗