Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Expectation Maximization”

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 55 records · Page 3
All sourcesNASA NTRSDOE OSTI
NASA NTRS2014

Clustering Days with Similar Airport Weather Conditions

On any given day, traffic flow managers must often rely on past experience and intuition when developing traffic flow management initiatives that mitigate imbalances between the aircraft demand and the weather impacted airport capacity. The goal of this study was to build on recent efforts to apply data mining classification and clustering algorithms to vast archives of historical weather and air traffic data to identify patterns and past decisions that can ultimately inform day-of-operations decision-making. More specifically, this study identified similar weather impacted days at select U.S. airports, and analyzed the traffic management initiatives implemented on these representative days. The identification of the similar days was accomplished by applying a decision tree algorithm to the hourly Localized Aviation Model Output Statistics Program observations and the arrival delays for Newark Liberty International Airport. The branches from the trained decision tree were subsequently pruned to identify four weather conditions that resulted in medium to high delays for the arrivals scheduled to Newark in 2012. Using these weather conditions, four, daily airport-level Weather Impacted Traffic Index values were calculated using the Localized Aviation Model Output Statistics Program observations and the 2012 scheduled arrival counts from the FAAs Aviation System Performance Metric system. The four, daily Weather Impacted Traffic Index values for 2012 were subsequently clustered using an Expectation Maximization clustering algorithm, and nine unique types of weather days at Newark were identified. By far the most prominent type of day at Newark was a day associated with relatively good weather conditions, where there was little convective activity, winds were low, ceilings and visibility were high and there was little precipitation. Moderate levels of convective activity characterized the next most prominent type of day. Days with persistently high winds or low ceiling and visibility levels were relatively rare in 2012. Lastly, the frequency at which Ground Delay Programs, Ground Stops and Miles-in-Trail restrictions were implemented on each of the typical types of days at Newark were analyzed. Based on the results, it does appear as if the usage of Miles-in-Trail, Ground Delay Program and Ground Stop restrictions correlates well with the severity of the weather associated with each unique type of weather impacted day at Newark. Furthermore, the results demonstrate that it is feasible to use historical weather and air traffic archives to provide guidance on the types of traffic management restrictions to implement in response to the weather conditions impacting an airport.

weather↗
NASA NTRS2020

LHS 1815b: The First Thick-disk Planet Detected by TESS

We report the first discovery of a thick-disk planet, LHS 1815b (TOI-704b, TIC 260004324), detected in the Transiting Exoplanet Survey Satellite (TESS) survey. LHS 1815b transits a bright (V = 12.19 mag, K = 7.99 mag) and quiet M dwarf located 29.87 ± 0.02 pc away with a mass of 0.502 ± 0.015 M⊙ and a radius of 0.501 ± 0.030 R⊙. We validate the planet by combining space- and ground-based photometry, spectroscopy, and imaging. The planet has a radius of 1.088 ± 0.064 R⊕ with a 3σ mass upper limit of 8.7 M⊕. We analyze the galactic kinematics and orbit of the host star LHS 1815 and find that it has a large probability (Pthick/Pthin = 6482) to be in the thick disk with a much higher expected maximal height (Zmax = 1.8 kpc) above the Galactic plane compared with other TESS planet host stars. Future studies of the interior structure and atmospheric properties of planets in such systems using, for example, the upcoming James Webb Space Telescope, can investigate the differences in formation efficiency and evolution for planetary systems between different Galactic components (thick disks, thin disks, and halo).

LHS 1815b↗
NASA NTRSDate not supplied

Unsupervised Change Detection for Space Habitats Using 3D Point Clouds

This work presents an algorithm for scene change detection from point clouds to enable autonomous robotic caretaking in future space habitats. Autonomous robotic systems will help maintain future deep-space habitats, such as the Gateway space station, which will be uncrewed for extended periods. Existing scene analysis software used on the International Space Station (ISS) relies on manually-labeled images for detecting changes. In contrast, the algorithm presented in this work uses raw, unlabeled point clouds as inputs. The algorithm first applies modified Expectation-Maximization Gaussian Mixture Model (GMM) clustering to two input point clouds. It then performs change detection by comparing the GMMs using the Earth Mover’s Distance. The algorithm is validated quantitatively and qualitatively using a test dataset collected by an Astrobee robot in the NASA Ames Granite Lab comprising single frame depth images taken directly by Astrobee and full-scene reconstructed maps built with RGB-D and pose data from Astrobee. The runtimes of the approach are also analyzed in depth. The source code is publicly released to promote further development.

robotics↗
NASA NTRS2020

Shape Estimation for Elongated Deformable Object using B-spline Chained Multiple Random Matrices Model

In this paper, a B-spline chained multiple random matrix models (RMMs) representation is proposed to model geometric characteristics of an elongated deformable object. The hyper degrees of freedom structure of the elongated deformable object make its shape estimation challenging. Based on the likelihood function of the proposed B-spline chained multiple RMMs, an expectation-maximization (EM) method is derived to estimate the shape of the elongated deformable object. A split and merge method based on the Euclidean minimum spanning tree (EMST) is proposed to provide initialization for the EM algorithm. The proposed algorithm is evaluated for the shape estimation of the elongated deformable objects in scenarios, such as the static rope with various configurations (including configurations with intersection), the continuous manipulation of a rope and a plastic tube, and the assembly of two plastic tubes. The execution time is computed and the accuracy of the shape estimation results is evaluated based on the comparisons between the estimated width values and its ground-truth, and the intersection over union (IoU) metric.

Gang Yao↗
DOE OSTI2023

3D Source Reconstruction Using Coded Aperture Gamma-Ray Imaging

Recent measurements with coded-aperture imagers demonstrate material mass determination in a holdup setting to an accuracy within a few percent. This capability is of particular interest to the Surplus Plutonium Disposition (SPD) project, which aims to dilute and dispose of surplus plutonium oxide. Gamma-ray imagers can be used to determine holdup without interrupting normal operations. In this work, we examine techniques for 3D source localization and mass determination using gamma-ray imagers. Coded-aperture imagers provide excellent source localization within the 2D image plane; however, multiple imagers operating in tandem are necessary to identify source location in 3D space. A Maximum Likelihood Expectation-Maximization (MLEM) method for fitting detector mappings is a powerful tool for accomplishing this task. MLEM allows 3D source localization to be simultaneously constrained using multiple gamma-ray imagers by constructing the basis for the MLEM fit using detector mappings from different detector locations stitched together. Each of these basis points represents a singular response from a source in 3D space and is generated using Monte Carlo simulations of sources placed individually at different locations throughout the imager’s field of view. Additionally, implementing knowledge of the physical equipment in the simulations of the glovebox used for the SPD project incorporates attenuation effects that are needed to calculate material holdup.

Laminack, Alex↗
NASA NTRS2019

On Expected Value Strong Controllability

The Probabilistic Simple Temporal Network (PSTN) generalizes Simple Temporal Networks with Uncertainty (STNUs) by introducing probability distributions over the timing of uncontrollable timepoints. PSTNs are controllable if there is a strategy to execute the controllable timepoints while bounding the risk of violating any constraint to a small value. If this risk bound can't be satisfied, PSTNs are not considered controllable. We introduce the Expected Value Probabilistic SimpleTemporal Network (EPSTN), which extends PSTNs by including a benefit to the satisfaction of temporal constraints. We study the problem of Expected Value Strong Controllability (EvSC) of EPSTNs, which seeks a schedule maximizing the expected value of satisfied constraints. We solve the EvSC problem by extending a previously developed linear program, combined with search over constraints to violate at execution time. We describe conditions under which the solution to this linear program is the maximum expected value schedule. We then show how to search for constraints to discard, using the linear program at the core of the search. While the general problem is shown to be exponential, we conclude by providing several methods to bound the complexity of search.

Planning↗
DOE OSTI2025

Entanglement maximization and mirror symmetry in two-Higgs-doublet models

We consider 2-to-2 scatterings of Higgs bosons in a CP-conserving two-Higgs-doublet model (2HDM) and study the implication of maximizing the entanglement in the flavor space, where the two doublets Φ a , a = 1, 2, can be viewed as a qubit: Φ 1 = |0⟩ and Φ 2 = |1⟩. More specifically, we compute the scattering amplitudes for Φ a Φ b → Φ c Φ d and require the outgoing flavor entanglement to be maximal for a full product basis such as the computational basis, which consists of {|00⟩, |01⟩, |10⟩, |11⟩}. In the unbroken phase and turning off the gauge interactions, entanglement maximization results in the appearance of an U(2) × U(2) global symmetry among the quartic couplings, which in general is broken softly by the mass terms. Interestingly, once the Higgs bosons acquire vacuum expectation values, maximal entanglement enforces an exact U(2) × U(2) symmetry, which is spontaneously broken to U(1) × U(1). As a byproduct, this gives rise to Higgs alignment as well as to the existence of 6 massless Nambu-Goldstone bosons. The U(2) × U(2) symmetry can be gauged to lift the massless Goldstones, while maintaining maximal entanglement demands the presence of a discrete Z 2 symmetry interchanging the two gauge sectors. The model is custodially invariant in the scalar sector, and the inclusion of fermions requires a mirror dark sector, related to the standard one by the Z 2 symmetry.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
NASA NTRS2011

Planning the FUSE Mission Using the SOVA Algorithm

Three documents discuss the Sustainable Objective Valuation and Attainability (SOVA) algorithm and software as used to plan tasks (principally, scientific observations and associated maneuvers) for the Far Ultraviolet Spectroscopic Explorer (FUSE) satellite. SOVA is a means of managing risk in a complex system, based on a concept of computing the expected return value of a candidate ordered set of tasks as a product of pre-assigned task values and assessments of attainability made against qualitatively defined strategic objectives. For the FUSE mission, SOVA autonomously assembles a week-long schedule of target observations and associated maneuvers so as to maximize the expected scientific return value while keeping the satellite stable, managing the angular momentum of spacecraft attitude- control reaction wheels, and striving for other strategic objectives. A six-degree-of-freedom model of the spacecraft is used in simulating the tasks, and the attainability of a task is calculated at each step by use of strategic objectives as defined by use of fuzzy inference systems. SOVA utilizes a variant of a graph-search algorithm known as the A* search algorithm to assemble the tasks into a week-long target schedule, using the expected scientific return value to guide the search.

Lanzi, James↗
NASA NTRS2006

Generation of Data-Rate Profiles of Ka-Band Deep-Space Links

A short report discusses a methodology for designing Ka-band Deep-Space-to- Earth radio-communication links. This methodology is oriented toward minimizing the effects of weather on the Ka-band telecommunication link by maximizing the expected data return subject to minimum link availability and a limited number of data rates. This methodology differs from the current standard practices in which a link is designed according to a margin policy for a given link availability at 10 elevation. In this methodology, one chooses a data-rate profile that will maximize the average data return over a pass while satisfying a minimum-availability requirement for the pass, subject to mission operational limitations expressed in terms of the number of data rates used during the pass. The methodology is implemented in an intelligent search algorithm that first finds the allowable data-rate profiles from the mission constraints, spacecraft-to-Earth distance, spacecraft EIRP (effective isotropic radiated power), and the applicable zenith atmospheric noise temperature distribution, and then selects the best data rate in terms of maximum average data return from the set of allowable data-rate profiles.

Shambayati, Shervin↗
DOE OSTI2026

Supercharging simulation-based inference for Bayesian optimal experimental design

Abstract Bayesian optimal experimental design (BOED) seeks to maximize the expected information gain (EIG) of experiments. This requires a likelihood estimate, which in many settings is intractable. Simulation-based inference (SBI) provides powerful tools for this regime. However, existing work explicitly connecting SBI and BOED is restricted to a single contrastive EIG bound. We show that the EIG admits multiple formulations which can directly leverage modern SBI density estimators, encompassing neural posterior, likelihood, and ratio estimation. Building on this perspective, we define a novel EIG estimator using neural likelihood estimation. Further, we identify optimization as a key bottleneck of gradient based EIG maximization and show that a simple multi-start parallel gradient ascent procedure can substantially improve reliability and performance. With these innovations, our SBI-based BOED methods are able to match or outperform by up to 22% existing state-of-the-art approaches across standard BOED benchmarks.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Analysis and optimization of seismic monitoring networks with Bayesian optimal experimental design

SUMMARY Monitoring networks increasingly aim to assimilate data from a large number of diverse sensors covering many sensing modalities. Bayesian optimal experimental design (OED) seeks to identify data, sensor configurations or experiments which can optimally reduce uncertainty and hence increase the performance of a monitoring network. Information theory guides OED by formulating the choice of experiment or sensor placement as an optimization problem that maximizes the expected information gain (EIG) about quantities of interest given prior knowledge and models of expected observation data. Therefore, within the context of seismo-acoustic monitoring, we can use Bayesian OED to configure sensor networks by choosing sensor locations, types and fidelity in order to improve our ability to identify and locate seismic sources. In this work, we develop the framework necessary to use Bayesian OED to optimize a sensor network’s ability to locate seismic events from arrival time data of detected seismic phases at the regional-scale. This framework requires five elements: (i) A likelihood function that describes the distribution of detection and traveltime data from the sensor network, (ii) A prior distribution that describes a priori belief about seismic events, (iii) A Bayesian solver that uses a prior and likelihood to identify the posterior distribution of seismic events given the data, (iv) An algorithm to compute EIG about seismic events over a data set of hypothetical prior events, (v) An optimizer that finds a sensor network which maximizes EIG. Once we have developed this framework, we explore many relevant questions to monitoring such as: how to trade off sensor fidelity and earth model uncertainty; how sensor types, number and locations influence uncertainty; and how prior models and constraints influence sensor placement.

58 GEOSCIENCES↗
DOE OSTI2023

Off-policy deep reinforcement learning with automatic entropy adjustment for adaptive online grid emergency control

Electric overloading conditions and contingencies put modern power systems at risk of voltage collapse and blackouts. Load shedding is crucial to maintain voltage stability for grid emergency control. However, the rule- or model-based schemes rely on accurate dynamic system models and face considerable challenges in adapting to various operating conditions and uncertain event occurrences. Here, to address these issues, this paper proposes a novel deep reinforcement learning (DRL)-based voltage stability control algorithm with automatic entropy adjustment (AEA) for grid emergency control. Various dynamic network components for complex system operations are modeled to construct the DRL environment. An off-policy soft actor-critic architecture is developed to maximize the expected reward and policy entropy simultaneously. The AEA mechanism is proposed to facilitate the policy maximum entropy procedure, and the proposed method can automatically provide effective discrete and continuous actions against various fault scenarios. Our approach accomplishes high sampling efficiency, scalability, and auto-adaptivity of the control policies under high uncertainties. Comparative studies with the existing DRL-based control methods in IEEE benchmarks indicate salient performance improvement of the proposed method for dynamic system emergency control.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2025

Co-optimization of fuel properties, combustion system geometry, and injection strategy for conventional diesel fuel

Here, studies have shown that fuel properties can impact an engine’s operation in several ways, including ignition delay, sooting tendency, mixture formation, and combustion temperature. In mixing-controlled compression ignition (MCCI) engines, the fuel system design and piston bowl geometry significantly affect combustion performance and emissions. Based on current information, it is difficult to draw conclusions about fuel property effects and sensitivities. The central fuel hypothesis approach used in the US Department of Energy Co-Optima program has worked well for spark ignition fuels: identifying critical fuel property ranges is sufficient to screen fuel blends that are expected to maximize efficiency and reduce pollutant emissions. However, for MCCI-relevant fuels, the information gained from past studies is not sufficient to build such a merit function or to allow for performing a similar screening of fuel blends. It is hypothesized that a co-optimization of a fuel’s physical and chemical properties, combustion system geometry, and injection strategy could leverage synergies between the effects of the fuel properties and geometries, resulting in improved performance over state-of-the-art. A machine learning–assisted unconstrained global optimization algorithm was used to explore a design space comprising 23 independent variables. The results show that physical property effects were minimal even for large variations in fuel properties, and the only interaction effect that was observed was the effect of varied fuel density parameters on fuel/air mixture formation. Nevertheless, these interactions were not sufficient in magnitude to significantly affect optimization results. Therefore, analysis of the results suggests that fuel physical properties cannot be leveraged in a co-optimization context to increase engine efficiency.

33 ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2023

The impact of market design and clean energy incentives on strategic generation investments and resource adequacy in low-carbon electricity markets

Well-designed electricity markets play a crucial role in maintaining reliable electric power systems, which are critical in modern society. Here, this study examines the impact of different electricity market designs and clean energy incentive schemes on supporting renewable energy integration and achieving clean energy goals. To this end, we utilize a game-theoretical generation expansion planning model where generation companies make investment and retirement decisions to maximize their expected profit. The model is structured as an equilibrium problem with equilibrium constraints (EPEC) and solved using a diagonalization approach combined with progressive hedging. We analyze three types of electricity market designs: an energy-only market, a capacity market, and a clean energy market, and consider a wide range of market parameters resulting in 14 total scenarios. Wind and solar capacity comprise the majority of new investments in all considered scenarios, but the resultant system planning reserve margin (PRM) can differ significantly depending on market parameters. We also find that profit-driven investments lead to lower PRMs than a traditional system cost minimization approach. These individual scenario results further demonstrate how different market designs and clean energy incentive schemes may influence investor decision-making and impact resource adequacy throughout the clean energy transition.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2022

Reinforcement learning for adaptive maintenance policy optimization under imperfect knowledge of the system degradation model and partial observability of system states

Maintenance policy optimization usually is faced with challenges that arise from an imperfect knowledge of system degradation models and from the partial observability of system degradation states. Here, this paper proposes a reinforcement learning method to address these two challenges for a class of maintenance problems with Markov degradation processes. The reinforcement learning approach consists of a learning component and a planning component. Using sequentially collected observations, at each step of decision-making the learning component improves the knowledge of system degradation in terms of the probability distributions of the transition rates based on sequential Bayesian inference. Using the updated transition rates, at each step of decision-making the maintenance policy optimization problem is then formulated as a partially observable Markov decision problem, and the planning component computes the optimal maintenance policy that maximizes the expected cumulative reward. The proposed method is illustrated using a numerical example with repair and inspection maintenance actions. The result shows that as more observations are collected, the learning component progressively learns the true system degradation process, and the planning component adjusts the optimal maintenance policy accordingly as well, which leads to increased reward.

42 ENGINEERING↗
DOE OSTI2026

Probabilistic inference in very large universes

Our current favored cosmological theories allow for the striking and controversial possibility that the observable universe is just a small part of a much larger universe in which parameters that describe the effective, low-energy laws of physics vary from one region to another. The controversy is largely driven by the fact that such a “very large universe” is mostly observationally inaccessible to us, so the issue arises of how we can reasonably assess a theory that describes such a universe. In this paper, we propose a Bayesian method for theory assessment based on theory-generated probability distributions for our observations. We focus on the principles that define this method, leaving aside concerns about how, in practice, one would carry out the required calculations. (One important issue that we set aside is the measure problem.) We argue that cosmological theories can be tested by the standard method of Bayesian updating, but we need to use theoretical predictions for “first-person” probabilities—that is, probabilities that we should use for our observations, taking into account all relevant selection effects. These selection effects can vary from one observer to another and can vary with time, so, in principle, first-person probabilities are defined for each observer instant—an observer at a specific instant of time. Calculations of first-person probabilities should take into account everything that the observer believes about herself and her surroundings, which we refer to as her subjective state. If the universe is very large, a theory might predict that there are many observer instants in the same subjective state; we argue that first-person probabilities should be calculated using a principle of self-locating indifference (PSLI), the assumption that any real observer should make predictions for her future as if she were chosen randomly and uniformly from the theoretically predicted observer instants that share her subjective state. We believe the PSLI is intuitively very reasonable, but we also argue that, if the theory is correct, the use of this principle maximizes the expected fraction of observers who will make correct predictions. A further complication is that cosmological theories are not expected to fully predict the detailed properties of the universe, but rather will predict a set of possible universes, each with a probability. Different possible universes will generically have different numbers of observers. We argue that, in the calculation of first-person probabilities, the probability for each possible universe should be weighted by the number of observer instants in the specified subjective state that it contains. These issues have been controversial in the literature, so we also provide a rebuttal to the claim that principles like the PSLI involve a “selection fallacy”; a rebuttal to what we dub the principle of required certainty; an argument rejecting theories that predict a preponderance of Boltzmann brains; a rebuttal to a parable about humans and Jovians used by Hartle and Srednicki to argue that assumptions of typicality can lead to absurd consequences; and, finally, a discussion about how the use of “old evidence” can be fit into a Bayesian mold.

Azhar, Feraz [University of Notre Dame, IN (United↗
DOE OSTI2025

Stochastic Model Predictive Control With Gaussian Wind Direction Preview for Wake Steering

This article addresses the problem of wake steering control for wind farms that explicitly consider the tradeoff between farm-level power generation and yaw duty cycle under variable and uncertain wind conditions. A novel stochastic model predictive control (MPC) algorithm is presented, which utilizes a stochastic model of the freestream wind field components in a receding horizon framework to compute optimal yaw set points that maximize the expected value of the farm power while constraining the yaw actuation. Different configurations of the algorithm are evaluated using a steady-state wind farm simulator. The proposed stochastic MPC algorithm can plan control actions over a future prediction horizon based on probabilistic estimates of the incoming wind magnitude and direction.

17 WIND ENERGY↗