SEARCH · Engineering Papers

Results for “Point Optimization”

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 1,441 records · Page 80

All sources NASA NTRS DOE OSTI

NASA NTRSDate not supplied

Battery Evaluation Profiles for X-57 and Future Urban Electric Aircraft

Battery energy density is one of the most critical design parameters for sizing all-electric aircraft, however it’s easily overestimated. Establishing the effective, usable energy density is confused by varying degrees of margin needed to account for structural and thermal management between different cell chemistry and pack designs. Therefore, a better methodology is needed to fairly compare emerging battery technologies for electric aircraft. Currently, there is a loss of critical information when vehicle trade studies are performed using “nominal” published cell-level performance metrics. Aircraft power demands rarely match these nominal power profiles, and aircraft designers lack the ability to accurately simulate the battery performance and temperature off-nominally unless the battery chemistry is well established. Conversely, battery suppliers have no generalized reference cases to publish more realistic performance metrics. This can lead to poor assumptions, such as aircraft studies assuming a fixed discharge efficiency of a battery, when in reality the usable energy in a pack is dependent on the power and thermal profile. Information needed to properly assess weight penalties for thermal management is also typically poorly characterized when assessing candidate batteries. This paper serves to better inform battery development, and similarly, provide aircraft designers with more realistic assumptions for applying knockdown margins in their designs. Detailed power and thermal performance estimates are provided, which provide a starting point for sizing power and thermal budgets using experimentally derived battery models. Results show that the X-57 battery-to-shaft efficiency is 77.3% for a particular optimized mission. Considering a 25% reserve on the battery capacity, this means that only roughly half of the original 55.3kWh ‘nominal’ pack energy can be converted to useful work during a mission. Further estimates on a clean-sheet VTOL optimization show an average 82.7% battery-to-shaft efficiency, using 98% peak efficiency inverters and 97.4% peak efficiency motors. Although higher battery efficiencies are possible, the resulting weight penalty negates improvement in vehicle performance. These trade-offs and resulting power profiles are provided as a starting point to better assess future battery designs.

Battery Electric Aircraft↗

NASA NTRSDate not supplied

Battery Evaluation Profiles for X-57 and Future Urban Electric Aircraft

Battery↗

NASA NTRS1997

Preliminary Design of Low-Thrust Interplanetary Missions

For interplanetary missions, highly efficient electric propulsion systems can be used to increase the mass delivered to the destination and/or reduce the trip time over typical chemical propulsion systems. This technology is being demonstrated on the Deep Space 1 mission - part of NASA's New Millennium Program validating technologies which can lower the cost and risk and enhance the performance of future missions. With the successful demonstration on Deep Space 1, future missions can consider electric propulsion as a viable propulsion option. Electric propulsion systems, while highly efficient, produce only a small amount of thrust. As a result, the engines operate during a significant fraction of the trajectory. This characteristic makes it much more difficult to find optimal trajectories. The methods for optimizing low-thrust trajectories are typically categorized as either indirect, or direct. Indirect methods are based on calculus of variations, resulting in a two-point boundary value problem that is solved by satisfying terminal constraints and targeting conditions. These methods are subject to extreme sensitivity to the initial guess of the variables - some of which are not physically intuitive. Adding a gravity assist to the trajectory compounds the sensitivity. Direct methods parameterize the problem and use nonlinear programming techniques to optimize an objective function by adjusting a set of variables. A variety of methods of this type have been examined with varying results. These methods are subject to the limitations of the nonlinear programming techniques. In this paper we present a direct method intended to be used primarily for preliminary design of low-thrust interplanetary trajectories, including those with multiple gravity assists. Preliminary design implies a willingness to accept limited accuracy to achieve an efficient algorithm that executes quickly.

Sims, Jon A.↗

DOE OSTI2021

Instruction Roofline: An insightful visual performance model for GPUs

The Roofline performance model provides an intuitive approach to identify performance bottlenecks and guide performance optimization. However, the classic FLOP-centric approach is inappropriate for the emerging applications that perform more integer operations than floating point operations. In this article, we reintroduce our Instruction Roofline Model on NVIDIA GPUs and expand our evaluation of it. The Instruction Roofline incorporates instructions and memory transactions across all memory hierarchies together, and provides more performance insights than the FLOP-oriented Roofline Model, that is, instruction throughput, stride memory access patterns, bank conflicts, and thread predication. We use our Instruction Roofline methodology to analyze eight proxy applications: HPGMG from AMReX, Matrix Transpose benchmarks, ADEPT from MetaHipMer's sequence alignment phase, EXTENSION from MetaHipMer's local assembly phase, CUSP, cuSPARSE, cudaTensorCoreGemm, and cuBLAS. We demonstrate the ability of our methodology to understand various aspects of performance and performance bottlenecks on NVIDIA GPUs and motivate code optimizations.

Ding, N↗

NASA NTRS2013

Performance Gains of Propellant Management Devices for Liquid Hydrogen Depots

This paper presents background, experimental design, and preliminary experimental results for the liquid hydrogen bubble point tests conducted at the Cryogenic Components Cell 7 facility at the NASA Glenn Research Center in Cleveland, Ohio. The purpose of the test series was to investigate the parameters that affect liquid acquisition device (LAD) performance in a liquid hydrogen (LH2) propellant tank, to mitigate risk in the final design of the LAD for the Cryogenic Propellant Storage and Transfer Technology Demonstration Mission, and to provide insight into optimal LAD operation for future LH2 depots. Preliminary test results show an increase in performance and screen retention over the low reference LH2 bubble point value for a 325 2300 screen in three separate ways, thus improving fundamental LH2 LAD performance. By using a finer mesh screen, operating at a colder liquid temperature, and pressurizing with a noncondensible pressurant gas, a significant increase in margin is achieved in bubble point pressure for LH2 screen channel LADs.

Hartwig, Jason W.↗

NASA NTRSDate not supplied

A Nonlinear Schur Complement Solver for CFD-Based Multidisciplinary Models

CFD-based multidisciplinary models are the fundamental building blocks of multidis- ciplinary design optimization frameworks. Linear and nonlinear solutions of these coupled models are difficult, especially when the Jacobian matrices represent a saddle point problem, where a block-diagonal corresponding to a discipline is non-invertible. These scenarios necessitate the use of a coupled solver algorithm such as the Newton’s method instead of the popular block Gauss– Seidel-based methods because of this non-invertible block. To address this challenge, we introduce a nonlinear Schur complement solver suitable for CFD-based multidisciplinary models. The solver leverages the specialized linear and nonlinear solvers of the CFD code, and therefore, does not require the solution of a large coupled linear system as the coupled Newton’s method. Further- more, because the solver primarily uses the specialized linear and nonlinear solvers of the CFD code, it does not suffer from the same robustness limitations as the coupled Newton’s method. In this work, we will implement this solver in NASA’s OpenMDAO framework and demonstrate its effectiveness using a CFD-based aeropropulsive model. The solver will contribute to the develop- ment of aeropropulsive design optimization and CFD-based design optimization methods with the ultimate goal of accelerating the design and integration of advanced propulsion systems.

Nonlinear Solvers↗

DOE OSTI2025

Optimizing the optimizer for physics-informed neural networks and Kolmogorov-Arnold networks

Physics-Informed Neural Networks (PINNs) have revolutionized the computation of PDE solutions by integrating partial differential equations (PDEs) into the neural network’s training process as soft constraints, becoming an important component of the scientific machine learning (SciML) ecosystem. More recently, physics-informed Kolmogorv-Arnold networks (PIKANs) have also shown to be effective and comparable in accuracy with PINNs. In their current implementation, both PINNs and PIKANs are mainly optimized using first-order methods like Adam, as well as quasi-Newton methods such as BFGS and its low-memory variant, L-BFGS. However, these optimizers often struggle with highly nonlinear and non-convex loss landscapes, leading to challenges such as slow convergence, local minima entrapment, and (non)degenerate saddle points. In this study, we investigate the performance of Self- Scaled BFGS (SSBFGS), Self-Scaled Broyden (SSBroyden) methods and other advanced quasi-Newton schemes, including BFGS and L-BFGS with different line search strategies. These methods dynamically rescale updates based on historical gradient information, thus enhancing training efficiency and accuracy. We systematically compare these optimizers – using both PINNs and PIKANs – on key challenging PDEs, including the Burgers, Allen-Cahn, Kuramoto-Sivashinsky, Ginzburg-Landau, and Stokes equations. Additionally, we evaluate the performance of SSBFGS and SSBroyden for Deep Operator Network (DeepONet) architectures, demonstrating their effectiveness for data-driven operator learning. Our findings provide state-of-the-art results with orders-of-magnitude accuracy improvements without the use of adaptive weights or any other enhancements typically employed in PINNs. More broadly, our work reveal insights into the effectiveness of quasi-Newton optimization strategies in significantly improving the convergence and accurate generalization of PINNs and PIKANs.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2020

Multi-Agent Control of Lane-Switching Automated Vehicles for Energy Efficiency

The proliferation of automatic control systems and their connectivity accents the performance of their interactions. In particular, connected and automated vehicles could become high-impact examples thanks to their energy use and productivity effects. Ideally, a controller might collectively optimize all agents' control moves for a given objective. However, limits on computational complexity, incomplete knowledge of the central planner, and risks of a single point of failure make distributed control attractive as well. This paper proposes a collaborative heuristic to approach the performance of centralized control with decentralized-like computational effort. The related centralized controller is also described in detail and evaluated as a baseline. A collaboration-intensive obstacle avoidance scenario involving electric vehicles is simulated to demonstrate benefits over fully decentralized control. While centralized optimization performed best with an 8.6 % energy reduction, the collaborative decentralized scheme reached a favorable computation-performance tradeoff with a 6.7 % energy reduction.

Dollar, R. Austin↗

DOE OSTI2024

Building Automation

In this chapter, we focus on both need-based and vision-based control innovation for building automation for HVAC systems. After providing a brief background that describes different HVAC architectures and products, we describe the control challenges and opportunities from a needs-based point of view, at the equipment level, the systems level, and the building level. At the equipment level, increased use of continuously variable actuation such as variable speed compressors and fans, requires application of robust multivariable control that must consider various types of nonlinear behavior across increasingly larger operating envelopes. The objective at this level is to achieve robust and energy efficient operation. At the system level, coupled and interactive dynamics among subsystems are increasingly important to consider, and functional integration of these subsystems, along with set points and operating schedules are designed to minimize energy consumption and improve occupant comfort. At the building level, issues such as integration among other building systems, such as the building envelope and the electric grid, offer opportunities to develop new types of demand response and grid-interactive behaviors and technologies. At the higher levels, system dynamics and robust stability are less of an issue, similar to process control applications, while optimization and integration are more important, although dynamic stability is increasing an issue at the higher levels too. From a vision-driven innovation point of view, we describe three areas that will require sustained research efforts and will eventually impact building automation: Digital Twins, Model Predictive Control, and Grid Interactive Buildings.

Bortoff, Scott↗

DOE OSTI2023

Certifiably Correct Range-Aided SLAM

We present the first algorithm capable of efficiently computing certifiably optimal solutions to range-aided simultaneous localization and mapping (RA-SLAM) problems. Robotic navigation systems are increasingly incorporating point-to-point ranging sensors, leading state estimation which takes the form of RA-SLAM. However, the RA-SLAM problem is more difficult to solve than traditional pose-graph SLAM; ranging sensor models introduce additional non-convexity, unlike pose-pose or pose-landmark measurements, a single range measurement does not uniquely determine the relative transform between the involved sensors, and RA-SLAM inference is highly sensitive to initial estimates. Our approach relaxes the RA-SLAM problem to a semidefinite program (SDP), which we show how to solve efficiently using the Riemannian staircase methodology. The solution of this SDP provides a high-quality initialization for our original RA-SLAM problem, which is subsequently refined via local optimization, as well as a lower-bound on the RA-SLAM problem's optimal value. Our algorithm, named certifiably correct RA-SLAM (CORA), applies to problems comprised of arbitrary pose-pose, pose-landmark, and ranging measurements. Evaluation on simulated and real-world marine examples shows that our algorithm frequently produces certifiably optimal RA-SLAM solutions; moreover, even suboptimal estimates are typically within 1-2\% of the optimal value.

Papalia, Alan↗

DOE OSTI2022

Secure mmWave Spectrum Sharing with Autonomous Beam Scheduling for 5G and Beyond

Spectrum Sharing (SS) has seen a renewed set of initiatives in 5G with the availability of shared and unlicensed spectrum bands that can be used by multiple cellular service providers and private cellular networks. Beam based transmission, instead of the traditional sector based transmission in conjunction with the spectrum agility of the 5G New Radio (NR) has brought new opportunities to optimized sharing of spectrum. Currently in the U.S., a centralized Spectrum Access Server (SAS) is used to co-ordinate spectrum sharing among networks sharing the same spectrum band. However, SAS becomes a focal point for security attacks and a performance bottleneck. In addition, SAS relies on an Environmental Sensor Network (ESN), separate from the 5G network. Without trusted spectral occupancy information, false reporting of spectrum sensing data can create sub-optimal and unfair spectrum usage. This paper summarizes our recent research findings in using a decentralized scheme for multiple networks to securely share spectrum with autonomous beam scheduling : 1) A new stochastic network framework based on Lyapunov Optimization approach is developed to optimize scheduling at the base stations; 2) Game theoretic (GT) approach is used to formulate the distributed scheduler; 3) Another distributed scheduler with Q-learning is presented that utilizes the Reinforcement Learning (RL) approach; 4) The performance and convergence rate of these distributed solutions to use shared and unlicensed spectrum are compared with existing solutions. Conditions under which the performance of these schedulers approach the theoretical upper bound, which is the performance possible with no interference among the operators sharing the spectrum, are presented; 5) The ability of a base station to use its own user equipment as sensors, for optimal spectrum sharing with base stations in other operator networks, is demonstrated to be an effective approach.

5G↗

DOE OSTI2024

Tailored Micromagnet Sorting Gate for Simultaneous Multiple Cell Screening in Portable Magnetophoretic Cell‐On‐Chip Platforms

Abstract Conventional magnetophoresis techniques for manipulating biocarriers and cells predominantly rely on large‐scale electromagnetic systems, which is a major obstacle to the development of portable and miniaturized cell‐on‐chip platforms. Herein, a novel magnetic engineering approach by tailoring a nanoscale notch on a disk micromagnet using two‐step optical and thermal lithography is developed. Versatile manipulations are demonstrated, such as separation and trapping, of carriers and cells by mediating changes in the magnetic domain structure and discontinuous movement of magnetic energy wells around the circumferential edge of the micromagnet caused by a locally fabricated nano‐notch in a low magnetic field system. The motion of the magnetic energy well is regulated by the configuration of the nanoscale notch and the strength and frequency of the magnetic field, accompanying the jump motion of the carriers. The proposed concepts demonstrate that multiple carriers and cells can be manipulated and sorted using optimized nanoscale multi‐notch gates for a portable magnetophoretic system. This highlights the potential for developing cost‐effective point‐of‐care testing and lab‐on‐chip systems for various single‐cell‐level diagnoses and analyses.

36 MATERIALS SCIENCE↗

DOE OSTI2024

Efficient estimation of the modified Gromov–Hausdorff distance between unweighted graphs

Abstract Gromov–Hausdorff distances measure shape difference between the objects representable as compact metric spaces, e.g. point clouds, manifolds, or graphs. Computing any Gromov–Hausdorff distance is equivalent to solving an NP-hard optimization problem, deeming the notion impractical for applications. In this paper we propose a polynomial algorithm for estimating the so-called modified Gromov–Hausdorff (mGH) distance, a relaxation of the standard Gromov–Hausdorff (GH) distance with similar topological properties. We implement the algorithm for the case of compact metric spaces induced by unweighted graphs as part of Python library , and demonstrate its performance on real-world and synthetic networks. The algorithm finds the mGH distances exactly on most graphs with the scale-free property. We use the computed mGH distances to successfully detect outliers in real-world social and computer networks.

Oles, Vladyslav (ORCID:0000000188727463)↗

DOE OSTI2022

A systematic method for selecting molecular descriptors as features when training models for predicting physiochemical properties

Machine learning has proven to be a powerful tool for accelerating biofuel development. Although numerous models are available to predict a range of properties using chemical descriptors, there is a trade-off between interpretability and performance. Neural networks provide predictive models with high accuracy at the expense of some interpretability, while simpler models such as linear regression often lack in accuracy. In addition to model architecture, feature selection is also critical for developing interpretable and accurate predictive models. We present a method for systematically selecting molecular descriptor features and developing interpretable machine learning models without sacrificing accuracy. Our method simplifies the process of selecting features by reducing feature multicollinearity and enables discoveries of new relationships between global properties and molecular descriptors. To demonstrate our approach, we developed models for predicting melting point, boiling point, flash point, yield sooting index, and net heat of combustion with the help of the Tree-based Pipeline Optimization Tool (TPOT). For training, we used publicly available experimental data for up to 8351 molecules. Our models accurately predict various molecular properties for organic molecules (mean absolute percent error (MAPE) ranges from 3.3% to 10.5%) and provide a set of features that are well-correlated to the property. This method enables researchers to explore sets of features that significantly contribute to the prediction of the property, offering new scientific insights. To help accelerate early stage biofuel research and development, we also integrated the data and models into a open-source, interactive web tool.

09 BIOMASS FUELS↗

DOE OSTI2022

Ligand Control of Structural Diversity in Luminescent Hybrid Copper(I) Iodides

Copper(I) iodide hybrids are of interest for next-generation lighting technologies because of their efficient luminescence in the absence of rare-earth elements. Here, we report 10 structurally diverse hybrid copper(I) iodides that emit in the green–red region with quantum yields reaching 67%. The compounds display a diversity of structures including ones with one-dimensional (1D) Cu–1 chains, Cu 2 I 2 rhomboid dimers, and structures with two different arrangements of Cu 4 I 4 tetramers. The compounds with Cu 2 I 2 rhomboid dimers or Cu 4 I 4 cubane tetramers have higher photoluminescence quantum yields than those with Cu–I 1D chains and octahedral Cu 4 I 4 tetramers, owing to the optimal degree of condensation of the inorganic motifs, which suppresses nonradiative processes. Electronic structure calculations on these compounds point out the critical influence of the inorganic motif and organic ligand on the nature of the band gaps and thus the excitation characteristics. Temperature-dependent photoluminescence spectra are presented to better understand the nature of luminescence in compounds with different inorganic motifs. Here. the emerging understanding of composition–structure–property correlations in this family provides inspiration for the rational design of hybrid phosphors for general lighting applications.

36 MATERIALS SCIENCE↗

DOE OSTI2021

Gamma-ray burst detection prospects for next generation ground-based VHE facilities

ABSTRACT Gamma-ray Bursts (GRB) were discovered by satellite-based detectors as powerful sources of transient γ-ray emission. The Fermi satellite detected an increasing number of these events with its dedicated Gamma-ray Burst Monitor (GBM), some of which were associated with high energy photons ($E \gt 10$ GeV), by the Large Area Telescope (LAT). More recently, follow-up observations by Cherenkov telescopes detected very high energy emission ($E \gt 100$ GeV) from GRBs, opening up a new observational window with implications on the interpretation of their central engines and on the propagation of very energetic photons across the Universe. Here, we use the data published in the 2nd Fermi-LAT Gamma Ray Burst Catalogue to characterize the duration, luminosity, redshift, and light curve of the high energy GRB emission. We extrapolate these properties to the very high energy domain, comparing the results with available observations and with the potential of future instruments. We use observed and simulated GRB populations to estimate the chances of detection with wide-field ground-based γ-ray instruments. Our analysis aims to evaluate the opportunities of the Southern Wide-field-of-view Gamma-ray Observatory (SWGO), to be installed in the Southern Hemisphere, to complement CTA. We show that a low-energy observing threshold ($E_{low} \lt 200$ GeV), with good point source sensitivity ($F_{lim} \approx 10^{-11} \,\mathrm{erg\, cm^{-2}\, s^{-1}}$ in $1$ yr), are optimal requirements to work as a GRB trigger facility and to probe the burst spectral properties down to time-scales as short as $10$ s, accessing a time domain that will not be available to Imaging Atmospheric Cherenkov Telescopes instruments.

La Mura, G. (ORCID:000000018553499X)↗

DOE OSTI2026

Simulation of Divertor Performance in ST40 Under Dynamic Double-Null Plasmas

A power fraction model was implemented for the simultaneous prediction of 3-D surface temperature evolution at all four divertor targets in near-double-null (DN) tokamak configurations, which is especially important for compact high-field devices that may not have the ability to dissipate large amounts of power on the high-field side. Evaluating the power-sharing between the four divertor strike points in a disconnected DN configuration is important for understanding the overall power balance, as well as for optimizing the power exhaust performance and prolonging the survivability of the plasma facing components (PFCs). This power-sharing is typically evaluated in terms of the separation between the primary and secondary separatrices at the outboard midplane, $\textit {dR}_{\text {sep}}$. The Heat flux Engineering Analysis Toolkit (HEAT) is coupled with Brunner’s power fraction model to simulate the deposited heat flux and resultant temperature change on 3-D divertor targets in a dynamic DN (DDN) pulse operation in ST40, a high-field spherical tokamak. The simulation results showed that with DDN operation, the operation time has significantly increased compared with single-null geometry configurations.

ST40↗

DOE OSTI2023

Final Cooling with Thick Wedges for a Muon Collider

A muon collider is a particle accelerator that collides muons rather than protons or electrons. An important component of such a collider is cooling, in which the emittance of the muon beam is reduced to make it suitable for use in a collider. In the last stage of this process, called final 4D cooling, emittance in the transverse axes is reduced while emittance in the longitudinal axis is allowed to grow. In previous studies, Final 4-D cooling used absorbers within very high field solenoids to cool low-momentum muons. Simulations of the systems did not reach the desired cooling design goals. In this study, we develop and optimize a different conceptual design for the final 4D cooling channel, which is based on single thick wedges. We used G4Beamline to simulate the channel and Python to generate and analyze particle distributions. We characterized the effects of various parameters on the performance of this system, including the initial emittances and Twiss parameters, the length and angle of the wedges, the length of the drift channel, and the length, frequency, gradient, and phase of the RF cavities. We optimized the design parameters of the cooling channel and produced two conceptual designs (corresponding to two possible starting points for the input beam) which achieve transverse cooling in both x and y by a factor of ≈ 3.5. These channels achieve a lower transverse and longitudinal emittance than the best design previously published.

43 PARTICLE ACCELERATORS↗