Search NASA⌕ Search

SEARCH · Search NASA

Results for “optimal design parameters”

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 487 records · Page 27
All sourcesNASA NTRSDOE OSTI
DOE OSTI2026

Scaling considerations for supercritical carbon dioxide cycles including turbomachinery loss models

A modeling framework for the supercritical carbon dioxide recompressed closed Brayton cycle was developed. Unlike typical models, this effort incorporated generalized empirical turbomachinery loss models. Aerodynamic, windage, and leakage losses were considered in order to address the limitations of conventional constant-efficiency turbomachinery assumptions without relying on machine-specific or computationally expensive simulations. The model enables system-level exploration of optimal cycle design across a range of power scales, including smaller scales that are relevant to microreactors and extraterrestrial power applications. Parametric studies and multi-objective optimizations are used to evaluate the trade-offs between thermal efficiency and system compactness based on an analytical heat exchanger scaling model, yielding Pareto-optimal fronts across a range of operating pressures. Results reveal that at small power scales, the Pareto-optimal compressor inlet pressure becomes subcritical due to the increasing influence of density-dependent turbomachinery losses. Here, the relative contributions of each loss mechanism are quantified, and design recommendations are provided for key parameters such as recompression split ratio and generator cavity pressure across varying power scales.

Multi-objective optimization↗
DOE OSTI2022

PWR loading pattern optimization with reinforcement learning

The core loading pattern optimization problem belongs to the class of combinatorial optimization problem and has been studied since the dawn of commercial nuclear energy industry. It is characterized by multiple objectives and constraints, with a very high number of candidate patterns, which makes it impossible to solve explicitly. Stochastic optimization methodologies including Genetic Algorithms and Simulated Annealing are used by different nuclear utilities and vendors to perform fuel cycle reload design. Nevertheless, hand-designed solutions continue to be the prevalent method in the industry. To improve the state-of-the-art core reload patterns, we aim to create a method as scalable as possible, that agrees with the designer's goal of performance and safety. To help in this task Deep Reinforcement Learning (DRL), in particular Proximal Policy Optimization is leveraged. DRL has recently experienced a strong impetus from its successes applied to games, sometimes even reaching 'super-human' performances. This paper lays out the foundation of this method and proposes to study the behavior of several hyper-parameters that influence the DRL algorithm. The algorithm is highly dependent on multiple factors such as an exploration/exploitation trade-off that manifests through different parameters such as the number of loading patterns seen and the number of samples collected before a policy update, but also the shape of the objective function derived for the core design. Experimental results also demonstrate the effectiveness of the method in finding high-quality solutions from scratch within a reasonable amount of time. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2025

Virtual Engineering: Python framework for engineering process design

Virtual Engineering (VE) is a Python software framework designed to accelerate the research and development of engineering processes that are fundamentally defined by multiple unit operations executed in series. VE supports a wide variety of different multi-physics models and integrates them to simulate a complete end-to-end process. To automate the execution of this model sequence, VE provides (i) a robust method to communicate between models, (ii) a high-level, user-friendly interface to set model parameters and enable optimization, and (iii) an overall model-agnostic approach that allows new computational units to be swapped in and out of workflows. Although the VE framework was developed to support the biochemical conversion of biomass to fuel, we have designed each component to easily accommodate new domains and unit models.

09 BIOMASS FUELS↗
DOE OSTI2020

CONTROLLED END GAS AUTO IGNITION WITH EXHAUST GAS RECIRCULATION ON A STOICHIOMETRIC, SPARK IGNITED, NATURAL GAS ENGINE

The goal of this study is to address fundamental limitations to achieving diesel-like efficiencies in heavy duty on-highway natural gas (NG) engines. Engine knock and misfire are barriers to pathways leading to higher efficiency engines. This study explores enabling technologies for development of high efficiency stoichiometric, spark ignited, natural gas engines. These include design strategies for fast and stable combustion and higher dilution tolerance. Additionally, advanced control methodologies are implemented to maintain stable operation between knock and misfire limits. To implement controlled end- gas autoignition (C-EGAI) strategies a Combustion Intensity Metric (CIM) is used for ignition control with the use of a Woodward large engine control module (LECM). Tests were conducted using a single cylinder, variable compression ratio, cooperative fuel research (CFR) engine with baseline conditions of 900 RPM, engine load of 800 kPa indicated mean effective pressure (IMEP), and stoichiometric air/fuel ratio. Exhaust gas recirculation (EGR) tests were performed using a custom EGR system that simulates a high pressure EGR loop and can provide a range of EGR rates from 0 to 40%. The experimental measurements included the variance of EGR rate, compression ratio, engine speed, IMEP, and CIM. These five variables were optimized through a Modified Box- Benken design Surface Response Method (RSM), with brake efficiency as the merit function. A positive linear correlation between CIM and f-EGAI was identified. Consequently, CIM was used as the feedback control parameter for C-EGAI. As such, implementation of C-EGAI effectively allowed for the utilization of high EGR rates and CRs, controlling combustion between a narrower gap between knock and lean limits. The change from fixed to parametric ignition timing with CIM targeted select values of f-EGAI with an average coefficient of variance (COV) of peak pressure of 5.4. The RSM efficiency optimization concluded with operational conditions of 1080 RPM, 1150 kPa IMEP, 10.55:1 compression ratio, and 17.8% EGR rate with a brake efficiency of 21.3%. At this optimized point of peak performance, a f-EGAI for C-EGAI was observed at 34.1% heat release due to auto ignition, a knock onset crank angle value of 10.3° aTDC and ignition timing of - 24.7° aTDC. This work has demonstrated that combustion at a fixed f-EGAI can be maintained through advanced ignition control of CIM without experiencing heavy knocking events.

Cooperative Fuel Research Engine, Auto Ignition, E↗
DOE OSTI2022

Schema Elements for Granta Annual Report: FY2022

Granta: Materials Intelligence, also known as Granta:MI or Granta, is a commercial database software by Ansys, Inc. that is utilized by the Nuclear Security Enterprise (NSE) to organize and store relevant materials data. For a complete discussion of the use of Granta:MI at NSE sites, see the FY21 annual report. Granta:MI is used by five NSE sites locally, and all NSE sites have access to an enterprise instance on the Enterprise Secure Network (ESN) as well as an unclassified development instance. It has been recognized by NNSA management that a shared repository for additive manufacturing (AM) data would not only ensure data and knowledge is not lost but would provide a pool of information relating AM inputs (build parameters, raw materials properties, post-processing information) to the as-built properties of AM parts. Such a pool of data would enable optimization of AM build design and help NNSA achieve the goals of shortening fielding times for new components.

36 MATERIALS SCIENCE↗
DOE OSTI2021

EUV and Hard X-ray Hartmann Wavefront Sensing for Optical Metrology, Alignment and Phase Imaging

For more than 15 years, Imagine Optic have developed Extreme Ultra Violet (EUV) and X-ray Hartmann wavefront sensors for metrology and imaging applications. These sensors are compatible with a wide range of X-ray sources: from synchrotrons, Free Electron Lasers, laser-driven betatron and plasma-based EUV lasers to High Harmonic Generation. In this paper, we first describe the principle of a Hartmann sensor and give some key parameters to design a high-performance sensor. We also present different applications from metrology (for manual or automatic alignment of optics), to soft X-ray source optimization and X-ray imaging.

20 metrology↗
DOE OSTI2023

Deep Learning for Modeling Enhanced Geothermal Systems

Enhanced Geothermal Systems (EGS) offer a vast potential to expand the use of geothermal energy. Heat is extracted from this engineered system by injecting cold water into a subsurface fractures, which are in contact with the hot dry rock, and pulled through the production wells. Creating EGS requires improving the natural permeability of hot crystalline rocks. To develop economically viable EGS reservoirs, significant technical barriers (e.g., better stimulation technologies without adequate water and/or permeability) and non-technical barriers (e.g., land access and permitting) must be overcome. In this short conference paper, we present a workflow to address a part of this challenge – “How to develop economically viable EGS using existing technologies?”. Our workflow called the GeoThermalCloud (GTC) for EGS, leverages recent advances in machine learning, deep learning, and cloud computing. This GTC framework is open-source and available at https://github.com/SmartTensors/GeoThermalCloud.jl. The GTC framework provides trained deep learning (DL) models to estimate the net present value of a given EGS design scenario. The Geothermal Design Tool (https://github.com/GeoDesignTool/GeoDT.git), a fast and simplified multi-physics solver, is used to develop a database for training DL models. The database consists of EGS design parameters (inputs to DL model) and their net present value (output of DL model) in uncertain geologic systems. The EGS design parameters for constructing this training database are based on Utah FORGE but include the options of more wells and deeper depths. The DL models are trained by ingesting the EGS design parameters and estimating the corresponding net present value. Such an emulation allows us to screen various EGS designs quickly and identify good development strategies by coupling them with optimization techniques. Our preliminary results show promise in DL emulation of net present value. However, a lot more work is needed to improve the predictive capability of DL models (i.e., extensive hyperparameter tuning is necessary). This will be the primary focus of our future work.

artificial neural networks, geothermal↗
DOE OSTI2026

Kinetics Modeling and Reactor Design Study of Glucose-to-Terpenes Cell-Free Conversion

Cell-free systems offer many advantages over traditional biological conversion by eliminating biological growth constraints. It also offers easy manipulation and finetuning of the reaction conditions for each individual enzyme. The conversion of cellulosic glucose to Limonene, a terpene, is a promising pathway for producing fuels and chemicals. Recent advances in developing cell-free systems focuses on bench scale optimization of terpene yield and to demonstrate its feasibility towards commercialization [1,2]. There is significant knowledge gap regarding reaction kinetics of these cell-free systems to further study how it will perform at larger scale. We present here, our studies on reaction kinetics and reactor design implications of cell-free glucose to Limonene conversion to facilitate the further development and commercialization of this process. We developed a novel kinetic model based on the metabolic-network structure of the cell-free system with multi-substrate reversible Michaelis-Menten rate law. To estimate kinetic parameters for this system of rate equations, we employed Bayesian optimization to perform global search with the assistance of gaussian processes to balance exploration and exploitation. The model parameters estimated showed good results compared with experimental data. The estimated parameters were used to perform sensitivity analysis. We found that Hexokinase is one of the most critical enzymes that affect the conversion of the glucose. We also observed that abundance of co-factors is also critical to the conversion of glucose to limonene. We investigated packed bed reactors with enzymes immobilized on the surface of particles to convert glucose stream into Limonene for larger scale production. The reactor design such as particle size, enzyme loading, and flow rate are found to be critical for improving yields. [1] Dudley, Q.M., Nash, C.J. and Jewett, M.C., 2019. Synthetic Biology, 4(1), p.ysz003. [2] Korman, T.P., Opgenorth, P.H. and Bowie, J.U., 2017. Nature communications, 8(1), p.15526.

09 BIOMASS FUELS↗
DOE OSTI2025

Fast-RF-Shimming: Accelerate RF shimming in 7T MRI using deep learning

Ultrahigh field (UHF) Magnetic Resonance Imaging (MRI) offers an elevated signal-to-noise ratio (SNR), enabling exceptionally high spatial resolution that benefits both clinical diagnostics and advanced research. However, the jump to higher fields introduces complications, particularly transmit radiofrequency (RF) field ($B^{+}_{1}$) inhomogeneities, manifesting as uneven flip angles and image intensity irregularities. These artifacts can degrade image quality and impede broader clinical adoption. Traditional RF shimming methods, such as Magnitude Least Squares (MLS) optimization, effectively mitigate $B^{+}_{1}$ inhomogeneity, but remain time-consuming. Recent machine learning approaches, including RF Shim Prediction by Iteratively Projected Ridge Regression and other deep learning architectures, suggest alternative pathways. Although these approaches show promise, challenges such as extensive training periods, limited network complexity, and practical data requirements persist. In this paper, we introduce a holistic learning-based framework called Fast-RF-Shimming, which achieves a 5000 ​× ​speed-up compared to the traditional MLS method. In the initial phase, we employ random-initialized Adaptive Moment Estimation (Adam) to derive the desired reference shimming weights from multi-channel $B^{+}_{1}$ fields. Next, we train a Residual Network (ResNet) to map $B^{+}_{1}$ fields directly to the ultimate RF shimming outputs, incorporating the confidence parameter into its loss function. Finally, we design Non-uniformity Field Detector (NFD), an optional post-processing step, to ensure the extreme non-uniform outcomes are identified. Comparative evaluations with standard MLS optimization underscore notable gains in both processing speed and predictive accuracy, which indicates that our technique shows a promising solution for addressing persistent inhomogeneity challenges.

Deep learning↗
DOE OSTI2024

How Does Escherichia coli Allocate Proteome?

Microorganisms are shown to actively partition their intracellular resources, such as pro- teins, for growth optimization. Recent experiments have begun to reveal molecular com- ponents unpinning the partition; however, it remains unclear quantitatively how individual parts orchestrate to yield precise resource allocation that is both robust and dynamic. Here we developed a coarse-grained mathematical framework that centers on guanosine pentaphosphate (ppGpp)-mediated regulation, and used it to systematically uncover the design principles of proteome allocation in Escherichia coli. Our results showed that cellular ability of resource partition lies in an ultrasensitive, negative feedback control- ling topology with the ultrasensitivity arising from zero-order amino acid kinetics and the negative feedback from ppGpp-controlled ribosome synthesis. In addition, together with the time-scale separation between slow ribosome kinetics and fast turnovers of ppGpp and amino acids, the network topology confers the organism an optimization mechanism which mimics sliding mode control, a nonlinear optimization strategy that is widely used in man-made systems. We further showed that such a controlling mechanism is robust against parameter variations and molecular fluctuations, and is also efficient for biomass production over time. Furthermore, this work elucidates the fundamental controlling mechanism of E. coli proteome allocation, thereby providing insights into quantitative microbial physiology as well as the design of synthetic gene networks.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Active and Durable PGM-free Cathodic Electrocatalysts for Fuel Cell Application

Platinum Group Metal-free (PGM-free) Oxygen Reduction Reaction (ORR) electrocatalysts possess high intrinsic activity measured by a Rotating Disk Electrode (RDE) method. However, under fuel cell operating conditions, PGM-free ORR electrocatalysts have underperformed compared to platinum catalysts. PGM-free ORR electrocatalyst’s fuel cell performance can be improved by designing and optimizing the cathodic catalyst layer (CL) and Membrane Electrode Assembly (MEA) construction such that: 1) it efficiently provides oxygen access to ORR active sites (through catalyst morphology control); 2) it removes water from the CL (by tuning the hydrophobicity of the PGM-free catalysts and the catalyst layer structure); and 3) it increases proton conductivity (by homogeneous mixing of catalysts and ionomer). Studying the CL is complex due to the absence of well-established protocols of MEA activation, especially compared to platinum-containing catalysts. PGM-free fuel cell testing protocols would need to optimize potentiostatic vs galvanostatic measurements, scan rates, parameters of Electrochemical Impedance Spectroscopy (EIS) and Beginning of Experiment (BOE) criteria. To make the PGM-free catalyst’s fuel cell performance comparable to platinum, the synergistic effort of materials design, fine tuning of the catalyst layer and comprehensive electrochemical analysis is required.

08 HYDROGEN↗
DOE OSTI2023

Optimal PMU design based on sampling model and sensitivity analysis

The precise measurements of the synchrophasor and frequency from phasor measurement units (PMUs) are widely used in power grid applications. With the improvement of the technique, applications always require stable dynamic performance and higher accuracy for the synchrophasor and frequency measurements, which is challenging for PMU development. To evaluate the contribution of PMU hardware to measurement accuracy, this paper proposes a general-purpose sampling model to analyze the measurement error. In the proposed sampling model, a strict mathematical derivation is derived, where its error is purely determined by the parameters of the PMU hardware. The sensitivity analysis is carried out by three methods, including mathematical analysis, computer simulation, and variance-based sensitivity analysis. Through the sensitivity analysis, this paper establishes the systematic formulation and the inclusion of synchrophasor, frequency, and ROCOF. Experimental results based on the real-world testbench involving distribution-level PMUs match the mathematical analysis conclusion, which verifies the correctness of the general-purpose sampling model. Furthermore, a strategy for the optimal PMU design is proposed, which could guide PMU design in the future.

42 ENGINEERING↗
DOE OSTI2023

Sensitivity Analysis, Reduced-order Modeling, and Optimization of a Gas-Cooled Pebble Bed Reactor using Equilibrium-Core and DLOFC Performance

This work presents and applies a workflow for performing design optimization on gas-cooled pebble-bed reactors. Based on previous research, a representative equilibrium core of a pebble-bed reactor and a depressurized loss-of-forced-cooling model are created. These applications are built using the Multiphysics Object-Oriented Simulation Environment (MOOSE), specifically utilizing Griffin, Pronghorn, and Bison. After defining design-related parameters and quantities of interest regarding reactor safety and efficiency, this multiphysics model is sampled using the MOOSE stochastic tools module. The result is a comprehensive dataset of configurations, enabling sensitivity analysis and the generation of reduced-order models. Subsequently, the dataset and reduced-order models are employed in an optimization study aimed at maximizing fuel utilization while adhering to safety and operational constraints. The optimization process leads to an improvement of fuel utilization by approximately 10\%, compared to engineering-judgment-based nominal conditions.

97 - MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Optimal Management of Grid-Interactive Efficient Buildings via Safe Reinforcement Learning

Reinforcement learning (RL)-based methods have achieved significant success in managing grid-interactive efficient buildings (GEBs). However, RL does not carry intrinsic guarantees of constraint satisfaction, which may lead to severe safety consequences. Besides, in GEB control applications, most existing safe RL approaches rely only on the regularisation parameters in neural networks or penalty of rewards, which often encounter challenges with parameter tuning and lead to catastrophic constraint violations. To provide enforced safety guarantees in controlling GEBs, this paper designs a physics-inspired safe RL method whose decision-making is enhanced through safe interaction with the environment. Different energy resources in GEBs are optimally managed to minimize energy costs and maximize customer comfort. The proposed approach can achieve strict constraint guarantees based on prior knowledge of a set of developed hard steady-state rules. Simulations on the optimal management of GEBs, including heating, ventilation, and air conditioning (HVAC), solar photovoltaics, and energy storage systems, demonstrate the effectiveness of the proposed approach.

Huo, Xiang↗
DOE OSTI2024

A Review of Ring Motors with Integrated Loads

Ring motors are electric machines that are typically characterized by having a hollow rotor / stator, a small difference between the inner and outer radii, and a large outer diameter relative to the axial length. The hollow portion of the ring motor allows integrating loads, such as an aerial or marine propeller, enabling power-dense systems. This paper reviews integrated ring motor designs from literature across different applications. Based on this review, first, design trends and performance parameters are identified and compared with conventional radial flux machines. Next, the bearing challenges posed by the unique form-factors of these machines are identified and approaches to realize bearings are presented. Finally, a research outlook is presented that identifies the benefits of applying multi-physics optimization, additive manufacturing, and bearingless machine technology to realize improved integrated ring motor designs.

Asgodom, Adonay↗
DOE OSTI2022

Neural Architecture and Feature Search for Predicting the Ridership of Public Transportation Routes

Accurately predicting the ridership of public-transit routes provides substantial benefits to both transit agencies, who can dispatch additional vehicles proactively before the vehicles that serve a route become crowded, and to passengers, who can avoid crowded vehicles based on publicly available predictions. The spread of the coronavirus disease has further elevated the importance of ridership prediction as crowded vehicles now present not only an inconvenience but also a public-health risk. At the same time, accurately predicting ridership has become more challenging due to evolving ridership patterns, which may make all data except for the most recent records stale. One promising approach for improving prediction accuracy is to fine-tune the hyper-parameters of machine-learning models for each transit route based on the characteristics of the particular route, such as the number of records. However, manually designing a machine-learning model for each route is a labor-intensive process, which may require experts to spend a significant amount of their valuable time. To help experts with designing machine-learning models, we propose a neural-architecture and feature search approach, which optimizes the architecture and features of a deep neural network for predicting the ridership of a public-transit route. Our approach is based on a randomized local hyper-parameter search, which minimizes both prediction error as well as the complexity of the model. We evaluate our approach on real-world ridership data provided by the public transit agency of Chattanooga, TN, and we demonstrate that training neural networks whose architectures and features are optimized for each route provides significantly better performance than training neural networks whose architectures and features are generic.

Ayman, Afiya↗
DOE OSTI2026

HARMONY: Large-Scale Architecture Search for Efficient Hybrid Language Models

As large language models scale to trillions of parameters, their computational and memory requirements present critical challenges for efficient training and deployment. While Mixture of Experts (MoE) architectures enable efficient scaling through sparse parameter activation, and state-space models like Mamba offer linear-time complexity, principled methods for combining these paradigms remain undeveloped. We introduce HARMONY (Hybrid Architecture Research for Mamba, Optimized with Neural efficiencY), a multi-objective evolutionary neural architecture search framework for discovering efficient hybrid language models that integrate Transformer attention mechanisms, Mixture-of-Experts routing, and Mamba state-space components. Through large-scale distributed search using 16,384 MI250X GPUs on the Frontier supercomputer, HARMONY explores a comprehensive design space encompassing six attention variants (MHA, MQA, GQA, MLA, SWA, and Mamba-2), variable MoE configurations with both routed and shared experts, and extensive Mamba hyperparameters. Our framework discovers heterogeneous architectures that balance training performance with computational efficiency through multi-objective optimization incorporating latency penalties and fitness-based selection. Analysis of discovered architectures reveals that optimal hybrid designs favor heterogeneous component mixing rather than homogeneous patterns, with Mamba-2 and Multi-Head Latent Attention (MLA) emerging as preferred mechanisms. Discovered architectures demonstrate superior training efficiency: our best configuration achieves a final perplexity of 1.0874 with 2.38B parameters while processing 4,320 tokens/second, outperforming significantly larger manually designed models. Full-scale evaluation shows HARMONY's top architectures achieve better loss trajectories than equivalently-sized models using state-of-the-art configurations including Mixtral, Jamba, and Samba. Additionally, we demonstrate 91% weak scaling efficiency when training discovered 36B-parameter models across 1,024 GPUs. HARMONY is released as an open framework with comprehensive tools for building and training hybrid models using expert-data-pipeline parallelism, democratizing access to automated architecture design for next-generation language models.

Herron, Emily [ORNL] (ORCID:0000000273008172)↗
DOE OSTI2021

A fast reduced model for a shell-and-tube based latent heat thermal energy storage heat exchanger and its application for cost optimal design by nonlinear programming

Numerical simulation of latent heat thermal energy storage (LHTES) systems plays a fundamental role in studying the physical process and guiding the engineering design. Discretization of the PDEs describing the nonlinear solidification/melting process of phase change materials (PCMs) leads to a large-scale complex dynamics system, where the system behavior depends on a set of parameters. In a design setting, repeated model evaluations are required over the set of parameters results in significant computational burden. In this paper, an explicit analytic solution was built for the propagation of the solidification front in a cylindrical coordinate. The analytic solution approach is further employed to develop a low computational reduced model (RM) as a module for a shell-and-tube based LHTES heat exchanger. The levelized Cost of Energy (LCOE) is used as a design metric and the RM model is used to apply system-level constraints in the nonlinear programming formulation that facilitates efficient global optimal design of the PCM properties, flow conditions and tube geometries. The use of LCOE as the design metric prevents over design of the heat transfer rate and also establishes a fair ground for evaluation of different thermal storage technologies and their integrated applications with other systems. Optimal results showed that a higher effectiveness results in a higher LCOE; the velocity of the HTF and the length of the channel are highly correlated with each other; both larger PCM latent energy and conductivity result in lower LCOE.

25 ENERGY STORAGE↗