Search NASA⌕ Search

SEARCH · Search NASA

Results for “Optimization Under Uncertainty”

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
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

A Behavioral Robotics Approach to Radiation Mapping Using Adaptive Sampling

Radiation mapping is a desirable task to automate because of the inherent risks involved and its tedious nature. A novel system was designed to address this by combining various existing technologies, utilizing behavior-based robotics and Bayesian optimization. The system uses a quadruped robot equipped with a manipulator and gamma detector to take measurements at locations that are selected based on the uncertainty of a surrogate model used to estimate the true radiation field. The robot uses input from the world with depth cameras to avoid collisions with the robot’s body, and unreachable points for the end effector are addressed by both allowing for a soft collision with the environment to occur, prompting the system to abandon that point, and varying the exploration tendency of the optimization based on consecutive collisions. This approach provides unique traversability and adaptability over other strategies in the literature. Experiments were performed by placing a Cesium-137 source on the ground and varying geometric setups and an optimization parameter demonstrating the adaptability to diverse environments and the increased robustness resulting from the designed behavior. The results additionally demonstrate that dynamically adjusting the optimization algorithm’s exploration tendency based on the arm’s collision history improves the system’s ability to navigate cluttered environments and construct accurate radiation maps without getting stuck in unreachable areas.

Adams, Joel↗
DOE OSTI2020

A Distributed Reinforcement Learning Yaw Control Approach for Wind Farm Energy Capture Maximization: Preprint

In this paper, we present a reinforcement-learning based distributed approach to wind farm energy capture maximization using yaw control, also known as wake steering. In order to maximize the power output of a wind farm, it is often necessary for individual turbines to decrease their own power output through yaw misalignment so as to deflect their wakes away from downstream turbines. Although using model-based methods to achieve yaw misalignment is one option, a model-free method might be better suited to incorporate factors that are difficult to model or changing conditions. We propose an algorithm that adapts concepts of temporal difference reinforcement learning distributed to a multi-agent environment that allows individual turbines to act so as to optimize overall wind farm output and react to unforeseen disturbances.

controls↗
DOE OSTI2023

Model to inform the expansion of hydrogen distribution infrastructure

A growing hydrogen economy requires new hydrogen distribution infrastructure to link geographically distributed hubs of supply and demand. The Hydrogen Optimization with Deployment of Infrastructure (HOwDI) Model helps meet this requirement. The model is a spatially resolved optimization framework that determines location-specific hydrogen production and distribution infrastructure to cost-optimally meet a specified location-based demand. While these results are useful in understanding hydrogen infrastructure development, there is uncertainty in some costs that the model uses for inputs. Thus, the project team took the modeling effort a step further and developed a Monte Carlo methodology to help manage uncertainties. Seven scenarios were run using existing infrastructure and new demand in Texas exploring different policy and tax approaches. The inclusion of tax credits increased the percentage of runs that could deliver hydrogen at <$\$4$ /kg from 31% to 77% and decreased the average dispensed cost from $\$4.35$ /kg to $\$3.55$ /kg. However, even with tax credits there are still some runs where unabated SMR is deployed to meet new demand as the low-carbon production options are not competitive. Every scenario, except for the zero-carbon scenario (without tax credits), resulted in at least 20% of the runs meeting the $\$4$ /kg dispensed fuel cost target. This indicates that multiple pathways exist to deliver $\$4$ /kg hydrogen.

08 HYDROGEN↗
DOE OSTI2024

Estimating Uncertainty in Simulated ENSO Statistics

Abstract Large ensembles of model simulations are frequently used to reduce the impact of internal variability when evaluating climate models and assessing climate change induced trends. However, the optimal number of ensemble members required to distinguish model biases and climate change signals from internal variability varies across models and metrics. Here we analyze the mean, variance and skewness of precipitation and sea surface temperature in the eastern equatorial Pacific region often used to describe the El Niño–Southern Oscillation (ENSO), obtained from large ensembles of Coupled model intercomparison project phase 6 climate simulations. Leveraging established statistical theory, we develop and assess equations to estimate, a priori, the ensemble size or simulation length required to limit sampling‐based uncertainties in ENSO statistics to within a desired tolerance. Our results confirm that the uncertainty of these statistics decreases with the square root of the time series length and/or ensemble size. Moreover, we demonstrate that uncertainties of these statistics are generally comparable when computed using either pre‐industrial control or historical runs. This suggests that pre‐industrial runs can sometimes be used to estimate the expected uncertainty of statistics computed from an existing historical member or ensemble, and the number of simulation years (run duration and/or ensemble size) required to adequately characterize the statistic. This advance allows us to use existing simulations (e.g., control runs that are performed during model development) to design ensembles that can sufficiently limit diagnostic uncertainties arising from simulated internal variability. These results may well be applicable to variables and regions beyond ENSO.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Enabling Hyper-Differential Sensitivity Analysis for Ill-Posed Inverse Problems

Inverse problems constrained by partial differential equations (PDEs) play a critical role in model development and calibration. In many applications, there are multiple uncertain parameters in a model that must be estimated. However, high dimensionality of the parameters and computational complexity of the PDE solves make such problems challenging. A common approach is to reduce the dimension by fixing some parameters (which we will call auxiliary parameters) to a best estimate and use techniques from PDE-constrained optimization to estimate the other parameters. In this article, hyper-differential sensitivity analysis (HDSA) is used to assess the sensitivity of the solution of the PDE-constrained optimization problem to changes in the auxiliary parameters. Foundational assumptions for HDSA require satisfaction of the optimality conditions which are not always practically feasible as a result of ill-posedness in the inverse problem. Here we introduce novel theoretical and computational approaches to justify and enable HDSA for ill-posed inverse problems by projecting the sensitivities on likelihood informed subspaces and defining a posteriori updates. Our proposed framework is demonstrated on a nonlinear multiphysics inverse problem motivated by estimation of spatially heterogeneous material properties in the presence of spatially distributed parametric modeling uncertainties.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Virtual to Physical: Reinforcement Learning to Optimize SNS Particle Accelerator Controls

Complex accelerators must have control systems that can handle dynamic nonlinear environments. This makes traditional control methods unsuitable as they can struggle to adapt to these uncertainties. This provides an ideal environment for reinforcement learning algorithms as they are adaptable and generalizable. We present a reinforcement learning pipeline that can effectively handle the dynamics of a complex accelerator. We test and prove our pipelines capabilities on multiple environments including the Spallation Neutron Source (SNS) and the Beam Test Facility (BTF) at Oakridge National Lab (ORNL). Due to the limited time available to train an online algorithm like reinforcement learning on a real accelerator, we utilize a virtual twin accelerator (VIRAC) developed by ORNL to pretrain the policy and show its ability to converge in the virtual environment. We then test the adaptability of the pretrained RL model by applying it on the real accelerator and comparing the results. Utilizing our Scientific Optimization and Controls Toolkit (SOCT) and open-source standards such as Gymnasium we create and solve for a MEBT orbit correction problem in the SNS and an emittance maximization problem in the BTF. We show how Twin Delayed Deep Deterministic Policy Gradient (TD3) can solve this optimization environment in the virtual accelerator and transfer this policy onto the real accelerator for inference and model retraining. We show how reinforcement learning can be utilized as a control system for complex accelerators and provide a model pipeline for how an implementation performs and can be adapted to new accelerator control problems.

Kasparian, Armen [Thomas Jefferson National Accele↗
DOE OSTI2020

Vehicle to Grid Frequency Regulation Capacity Optimal Scheduling for Battery Swapping Station Using Deep Q-Network

Battery swapping stations (BSSs) are ideal candidates for fast frequency regulation services (FFRS) due to their large battery stock capacity. In addition, BSSs can precharge batteries for customers and the batteries that are not in charging can provide a stable regulation capacity to the market. However, uncertainties, such as ACE signals and the EV per-hour visit counts, introduce stochastic nonlinear dynamics into the operation of a BSS-based FFRS. Currently, there is no quantification method to ensure its optimal economical operation. To close this gap, in this article, we propose a novel deep Q-learning-based FFRS capacity dynamic scheduling strategy. This method can autonomously schedule the hourly regulation capacity in real time to maximize the BSSx0027;s revenue for providing FFRS. Case studies using real-world data verify the efficacy of the proposed work.

25 ENERGY STORAGE↗
DOE OSTI2023

Uncertain characterization of reservoir fluids due to brittleness of equation of state regression

Equations of state (EoS) play a central role in modeling the phase equilibrium of fluid mixtures. Their parameterization involves fitting a model to experimental data, i.e., solving a nonlinear, non-convex, multivariate optimization problem. The latter requires one to select design variables, domains of definition for each variable, and weights assigned to individual measurements. We demonstrate that subjective choices of an optimization algorithm and an initial guess also impact the regression process. Consequently, EoS predictions are fundamentally uncertain even after the EoS tuning to a limited set of experimental data points. We demonstrate this observation for two hydrocarbon reservoir fluids, in which five properties of the heaviest carbon fraction are treated as design variables. While all the optimization algorithms and initial guesses match experimental data for the gas and liquid properties, the resulting EoS parameterizations lead to dramatically different predictions of the fluid’s thermophysical behavior in the unsampled pressure and temperature regions. In conclusion, we propose the probabilistic treatment of design variables to quantify the predictive uncertainty of the resulting fluid models.

15 GEOTHERMAL ENERGY↗
DOE OSTI2020

An Uncertainty Management Framework for Integrated Gas-Electric Energy Systems

In many parts of the world, electric power systems have seen a significant shift toward generation from renewable energy and natural gas. Because of their ability to flexibly adjust power generation in real time, gas-fired power plants are frequently seen as the perfect partner for variable renewable generation. However, this reliance on gas generation increases interdependence and propagates uncertainty between power grids and gas pipelines and brings coordination and uncertainty management challenges. To address these issues, we propose an uncertainty management framework for uncertain, but bounded gas consumption by gas-fired power plants. The admissible ranges are computed based on a joint optimization problem for the combined gas and electricity networks, which involves chance-constrained scheduling for the electric grid and a novel robust optimization formulation for the natural-gas network. This formulation ensures feasibility of the integrated system with a high probability, while providing a tractable numerical formulation. A key advance with respect to existing methods is that our method is based on a physically accurate, validated model for transient gas pipeline flows. Our case study benchmarks our proposed formulation against methods that ignore how reserve activation impacts the fuel use of gas power plants and only consider predetermined gas consumption. Here, the results demonstrate the importance of considering uncertainty to avoid operating constraint violations and curtailment of gas to the generators.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2025

Accurate and uncertainty-aware multi-task prediction of HEA properties using prior-guided deep Gaussian processes

Surrogate modeling techniques have become indispensable in accelerating the discovery and optimization of high-entropy alloys (HEAs), especially when integrating computational predictions with sparse experimental observations. This study systematically evaluates the training and testing performance of four prominent surrogate models—conventional Gaussian processes (cGP), Deep Gaussian processes (DGP), encoder-decoder neural networks for multi-output regression and eXtreme Gradient Boosting (XGBoost)—applied to a hybrid dataset of experimental and computational properties of the 8-component HEA system Al-Co-Cr-Cu-Fe-Mn-Ni-V. We specifically assess their capabilities in predicting correlated material properties, including yield strength, hardness, modulus, ultimate tensile strength, elongation, and average hardness under dynamic/quasi-static conditions, alongside auxiliary computational properties. The comparison highlights the strengths of hierarchical deep modeling approaches in handling heteroscedastic, heterotopic, and incomplete data commonly encountered in materials science. Our findings illustrate that combined surrogate models such as DGPs infused with machine-learned priors outperform other surrogates by effectively capturing inter-property correlations and by assimilating prior knowledge. This enhanced predictive accuracy positions the combined surrogate models as powerful tools for robust and data-efficient materials design.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Towards real-time forecasting of natural gas production by harnessing graph theory for stochastic discrete fracture networks

In this work, we compare hydrocarbon production curves obtained from a graph-based reduced-order model with the high-fidelity Discrete Fracture Network (DFN) predictions for a fracture network created using data from a real shale site. We observe that the bounds for the high fidelity DFN model lie within the bounds for the reduced order model, implying that the reduced-order model provides a conservative estimate. Moreover, we found that except for first-passage times and late arriving mass, the production curves from the reduced-order model predict transport accurately. However, it is to be noted that the results are inspite of trading a three-dimensional geometry for a reduced system in the form of a graph, one that is 500–1000 times faster in terms of computational efficiency (for this particular application). In addition, we also compare the production curves for large drawdown and small drawdown using our graph approach. The reduced-order model is successful in showing that the long term productivity is higher in case of small drawdown although the initial productivity is higher for large drawdown. Thus, this reduced-order model offers great potential in uncertainty quantification for production, as well as in providing operators with information to make real-time decisions for optimal production.

03 NATURAL GAS↗
DOE OSTI2022

Modeling Global Carbon Costs of Plant Nitrogen and Phosphorus Acquisition

Abstract Most Earth system models (ESMs) do not explicitly represent the carbon (C) costs of plant nutrient acquisition, which leads to uncertainty in predictions of the current and future constraints to the land C sink. We integrate a plant productivity‐optimizing nitrogen (N) and phosphorus (P) acquisition model (fixation & uptake of nutrients, FUN) into the energy exascale Earth system (E3SM) land model (ELM). Global plant N and P uptake are dynamically simulated by ELM‐FUN based on the C costs of nutrient acquisition from mycorrhizae, direct root uptake, retranslocation from senescing leaves, and biological N fixation. We benchmarked ELM‐FUN with three classes of products: ILAMB, a remotely sensed nutrient limitation product, and CMIP6 models; we found significant improvements in C cycle variables, although the lack of more observed nutrient data prevents a comprehensive level of benchmarking. Overall, we found N and P co‐limitation for 80% of land area, with the remaining 20% being either predominantly N or P limited. Globally, the new model predicts that plants invested 4.1 Pg C yr −1 to acquire 841.8 Tg N yr −1 and 48.1 Tg P yr −1 (1994–2005), leading to significant downregulation of global net primary production (NPP). Global NPP is reduced by 20% with C costs of N and 50% with C costs of NP. Modeled and observed nutrient limitation agreement increases when N and P are considered together ( r 2 from 0.73 to 0.83).

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

Absolute calibration of the Lyman-α measurement apparatus at DIII-D

The LLAMA (Lyman-Alpha Measurement Apparatus) diagnostic was recently installed on the DIII-D tokamak [Rosenthal et al., Rev. Sci. Instrum. (submitted) (2020)]. LLAMA is a pinhole camera system with a narrow band Bragg mirror, a bandpass interference filter, and an absolute extreme ultraviolet photodiode detector array, which measures the Ly-α brightness in the toroidal direction on the inboard, high field side (HFS) and outboard, low field side (LFS). This contribution presents a setup and a procedure for an absolute calibration near the Ly-α line at 121.6 nm. The LLAMA in-vacuum components are designed as a compact, transferable setup that can be mounted in an ex situ vacuum enclosure that is equipped with an absolutely calibrated Ly-α source. The spectral purity and stability of the Ly-α source are characterized using a vacuum ultraviolet spectrometer, while the Ly-α source brightness is measured by a NIST-calibrated photodiode. The non-uniform nature of the Ly-α source emission was overcome by performing a calibration procedure that scans the Ly-α source position and employs a numerical optimization to determine the emission pattern. Nominal and measured calibration factors are determined and compared, showing agreement within their uncertainties. A first conversion of the measured signal obtained from DIII-D indicates that the Ly-α brightness on the HFS and LFS is on the order of 1020 Ph sr-1 m-2 s-1. The established calibration setup and procedure will be regularly used to re-calibrate the LLAMA during DIII-D vents to monitor possible degradation of optical components and detectors.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2025

𝜂 and 𝜂′ Production in 𝐽/𝜓 Radiative Decays from Quantum Chromodynamics

We present a first principles calculation within quantum chromodynamics (QCD) of the radiative decays of the 𝐽/𝜓 into the light pseudoscalar mesons 𝜂 and 𝜂′. Within a lattice computation we obtain the transition form factors as a function of photon virtuality from the timelike region, accessible experimentally via the “Dalitz” decay 𝐽/𝜓 →𝑒 + ⁢𝑒 − ⁢𝜂 (′) , through to the real photon point corresponding to 𝐽/𝜓 →𝛾⁢𝜂 (′) . This is the first calculation in lattice QCD with two (heavier than physical) degenerate flavors of light quark and a heavier strange quark, in which the 𝜂′ appears as the first excited state with pseudoscalar isoscalar quantum numbers. We access it reliably by using variationally optimized operators , the use of which also improves the purity of the 𝐽/𝜓 and 𝜂 signals, reducing systematic uncertainties. High quality results at a large number of kinematic points are obtained in a typically noisy disconnected process by using a novel correlator averaging procedure. Our results show the expected enhanced production of the 𝜂′ over the 𝜂 in this process, and suggest that the demonstrated lattice technology is suitable for future calculations considering processes in which light meson resonances are produced.

form factors↗
DOE OSTI2025

Electrifying Airport GSE: Monte Carlo Grid Impacts

Airports globally are shifting from ICE-powered to electric Ground Support Equipment (eGSE) to enhance efficiency, reduce operational costs, and improve operator health. Leveraging predictable routes, flat terrain, and low operational speeds, airports provide ideal conditions for electrification. This study evaluates freight GSE electrification at Dallas-Fort Worth International Airport (DFW), USA, using the Agile@ platform, which integrates three analytical methods: Freight Facility Model (FFM), Activity-Structure-Intensity-Fuel (ASIF), and Monte Carlo simulations. Results from 10,000 simulations indicate modest but critical increases in electricity demand and significant variability in GSE energy consumption. These insights emphasize the importance of data-driven scheduling, targeted maintenance, and strategic infrastructure planning. For high-uncertainty scenarios, airports are advised to deploy buffer energy storage systems (battery banks), implement demand-response charging strategies, schedule flexible workforce shifts, and prioritize proactive maintenance-particularly for equipment with higher operational uncertainty, such as tug tractors with trailers. Agile@ thus offers a robust, scalable, and data-driven framework to optimize long-term GSE planning and enhance reliability across diverse airport environments.

Bose, Ranjan [ORNL] (ORCID:0009000791026327)↗
DOE OSTI2022

Revisiting metastable immiscibility in SiO 2 –Al 2 O 3 : Structure and phase separation of supercooled liquids and glasses

Understanding and controlling liquid–liquid phase separation in aluminosilicates is crucial for optimizing glass properties. However, the metastable nature of aluminosilicates’ phase separation has made it difficult to study experimentally, and uncertainty persists regarding the compositional and temperature extents of the miscibility gap. Here, we present new experimental evidence that suggests a consolute temperature between 1440 and 1590°C and endmember compositions of 7 and 62 mol.% Al 2 O 3 for the phase-separated glasses. Using containerless melt processing, deeply supercooled liquids over the 0–60 mol.% Al 2 O 3 range are probed with in situ small- and wide-angle X-ray scattering, which simultaneously reveals changes in nanoscale density heterogeneity and atomic structure. Correlations between phase separation and atomic coordination environments are compared for liquids and glasses. Pair distribution function analysis shows mean O–(Si + Al) coordination increases with Al 2 O 3 content and decreases with temperature.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Constraining the nuclear gluon PDF with inclusive hadron production data

The nuclear parton distribution functions (nPDFs) of gluons are known to be difficult to determine with fits of deep inelastic scattering (DIS) and Drell-Yan (DY) data alone. Therefore, the nCTEQ15 analysis of nuclear PDFs added inclusive neutral pion production data from RHIC to help in constraining the gluon. In this analysis, we present a new global analysis of nuclear PDFs based on a much larger set of single inclusive light hadron data from RHIC and the LHC. Using our new nCTEQ code (nCTEQ++) with an optimized version of INCNLO we study systematically the limitations of the theory and the impact of the fragmentation function uncertainty.

Duwentäster, Pit↗
DOE OSTI2024

Ducted Fuel Injection and Cooled Spray Technologies for Particulate Control in Heavy-Duty Diesel Engines

Heavy-duty diesel engine manufacturers are continually seeking simple, low-cost technologies to reduce emissions. Ducted Fuel Injection (DFI) and Cooled Spray (CS) technologies have emerged as promising solutions, offering the potential for nearly sootless operation. These innovations could significantly decrease harmful Particulate Matter (PM) emissions while enabling further optimization to reduce Nitrogen Oxides (NOx) and increase efficiency. While initial experiments across various engine types have shown promise, uncertainties remain regarding the ideal geometry, scaling properties, and effectiveness of these technologies across different operating conditions. This project aims to address these knowledge gaps through experiments in both optical and metal single-cylinder engines.

33 ADVANCED PROPULSION SYSTEMS↗