Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Performance Tuning”

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 91 records · Page 5
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Performance Profile of Transformer Fine-Tuning in Multi-GPU Cloud Environments

The study presented here focuses on performance characteristics and trade-offs associated with running machine-learning tasks in multi-GPU environments on both on-site cloud computing resources and commercial cloud services (Azure). Specifically, this study examines these tradeoffs by examining the performance of training and fine-tuning of transformer-based deep-learning (DL) networks on clinical notes and data, a task of critical importance in the medical domain. To this end, we perform DL-related experiments on the widely deployed NVIDIA V100 GPUs and on the newer A100 GPUs connected via NVLink or PCIe. This study analyzes the execution time of major operations to train DL models and investigate popular options to optimize each of them. We examine and present the findings on the impacts that various operations (e.g. data loading into GPUs, training, fine-tuning), optimizations, and system configurations (single vs. multi-GPU, NVLink vs. PCIe) have on the overall training performance.

Begoli, Edmon↗
NASA NTRS2000

Autonomous Performance Monitoring System: Monitoring and Self-Tuning (MAST)

Maintaining the long-term performance of software onboard a spacecraft can be a major factor in the cost of operations. In particular, the task of controlling and maintaining a future mission of distributed spacecraft will undoubtedly pose a great challenge, since the complexity of multiple spacecraft flying in formation grows rapidly as the number of spacecraft in the formation increases. Eventually, new approaches will be required in developing viable control systems that can handle the complexity of the data and that are flexible, reliable and efficient. In this paper we propose a methodology that aims to maintain the accuracy of flight software, while reducing the computational complexity of software tuning tasks. The proposed Monitoring and Self-Tuning (MAST) method consists of two parts: a flight software monitoring algorithm and a tuning algorithm. The dependency on the software being monitored is mostly contained in the monitoring process, while the tuning process is a generic algorithm independent of the detailed knowledge on the software. This architecture will enable MAST to be applicable to different onboard software controlling various dynamics of the spacecraft, such as attitude self-calibration, and formation control. An advantage of MAST over conventional techniques such as filter or batch least square is that the tuning algorithm uses machine learning approach to handle uncertainty in the problem domain, resulting in reducing over all computational complexity. The underlying concept of this technique is a reinforcement learning scheme based on cumulative probability generated by the historical performance of the system. The success of MAST will depend heavily on the reinforcement scheme used in the tuning algorithm, which guarantees the tuning solutions exist.

Peterson, Chariya↗
DOE OSTI2022

Improvement of Oxygen Reduction Performance in Alkaline Media by Tuning Phase Structure of Pd–Bi Nanocatalysts

Tuning the crystal phase of bimetallic nanocrystals offers an alternative avenue to improving their electrocatalytic performance. Herein, we present a facile and one-pot synthesis approach that is used to enhance the catalytic activity and stability toward oxygen reduction reaction (ORR) in alkaline media via control of the crystal structure of Pd-Bi nanocrystals. By merely altering the types of Pd precursors under the same conditions, the monoclinic structured Pd 5 Bi 2 and conventional face-centered cubic (fcc) structured Pd 3 Bi nanocrystals with comparable size and morphology can be precisely synthesized, respectively. Interestingly, the carbon-supported monoclinic Pd 5 Bi 2 nanocrystals exhibit superior ORR activity in alkaline media, delivering a mass activity (MA) as high as 2.05 A/mg Pd . After 10,000 cycles of ORR durability test, the monoclinic structured Pd 5 Bi 2 /C nanocatalysts still remain a MA of 1.52 A/mg Pd , which is 3.6 times, 16.9 times, and 21.7 times as high as those of the fcc Pd 3 Bi/C counterpart, commercial Pd/C, and Pt/C electrocatalysts, respectively. Moreover, structural characterizations of the monoclinic Pd 5 Bi 2 /C nanocrystals after the durability test demonstrate the excellent retention of the original size, morphology, composition, and crystal phase, greatly alleviating the leaching of the Bi component. Overall, this work provides new insight for the synthesis of multimetallic catalysts with a metastable phase and demonstrates phase-dependent catalytic performance.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
NASA NTRSDate not supplied

Conceptual Design of a Tiltduct Reference Vehicle for Urban Air Mobility

NASA is establishing a fleet of conceptual air vehicle designs to support research and development for Urban Air Mobility (UAM). This fleet of vehicles will enable examination of the sensitivity of UAM vehicle designs to technology assumptions, identify key research and development needs for UAM aircraft, and provide the UAM community with reference vehicles that are publicly available and based upon known assumptions. To date, five six-passenger reference vehicles have been published: a quadrotor, a side-by-side, a lift-plus-cruise, a single-main-rotor helicopter, and a tiltwing. To increase the breadth of vehicle technologies encapsulated in the fleet of NASA UAM reference vehicles, this paper establishes a tiltduct vehicle as an addition to the fleet. The fleet will continue to evolve as future analyses and trade studies are performed. The tiltduct reference vehicle has six tilting ducted proprotors. This paper describes the initial configuration downselection; discusses ducted proprotor design rules of thumb as they applied to the conceptual design of the reference vehicle; describes the vehicle sizing, trade studies, and tuning of models performed; and finally, compares the resulting tiltduct vehicle against the other six-passenger NASA UAM reference vehicles. The high-level analyses performed for this study did not indicate significant differences in performance between the tiltduct and tiltwing reference vehicles, and so vehicle performance alone may not be a key driver in the selection of a tiltduct vehicle over a tiltwing vehicle. However, if ducts are found to have significant acoustical benefits, then acoustical priorities may provide a compelling reason to incorporate ducted proprotors. One significant limitation of the design presented in this paper is that the ducted proprotor performance was tuned based upon performance characteristics observed during historical tests with disk loadings (defined as thrust divided by proprotor disk area) of 125-250 lb/ft^2. The tiltduct vehicle designed in this study has a disk loading of 30 lb/ft^2, to be more representative of UAM vehicles; further studies to understand performance of ducted proprotors at representative disk loadings are warranted.

Tiltduct↗
NASA NTRS2022

Conceptual Design of a Tiltduct Reference Vehicle for Urban Air Mobility

NASA is establishing a fleet of conceptual air vehicle designs to support research and development for Urban Air Mobility (UAM). This fleet of vehicles will enable examination of the sensitivity of UAM vehicle designs to technology assumptions, identify key research and development needs for UAM aircraft, and provide the UAM community with reference vehicles that are publicly available and based upon known assumptions. To date, five six-passenger UAM reference vehicles have been published: a single-main-rotor helicopter, a side-by-side helicopter, a quadrotor, a lift-plus-cruise, and a tiltwing. To increase the breadth of vehicle technologies encapsulated in the fleet of NASA UAM reference vehicles, this paper establishes a tiltduct vehicle as an addition to the fleet. The fleet will continue to evolve as future analyses and trade studies are performed. The tiltduct UAM reference vehicle has six tilting ducted proprotors. This paper describes the initial configuration downselection; discusses ducted proprotor design rules of thumb as they applied to the conceptual design of the reference vehicle; describes the vehicle sizing, trade studies, and tuning of models performed; and finally, compares the resulting tiltduct vehicle against the other six-passenger NASA UAM reference vehicles. The high-level analyses performed for this study did not indicate significant differences in performance between the tiltduct and tiltwing reference vehicles, and so vehicle performance alone may not be a key driver in the selection of a tiltduct vehicle over a tiltwing vehicle. However, if ducts are found to have significant acoustical benefits, then acoustical priorities may provide a compelling reason to incorporate ducted proprotors. One significant limitation of the design presented in this paper is that the ducted proprotor performance was tuned based upon performance characteristics observed during historical tests with disk loadings (defined as thrust divided by proprotor disk area) of 125-250 lb/ft$^2$. The tiltduct vehicle designed in this study has a disk loading of 30 lb/ft$^2$, to be more representative of UAM vehicles; further studies to understand performance of ducted proprotors at representative disk loadings are warranted.

Tiltduct↗
DOE OSTI2025

Simulation-Based Inference for Neutrino Interaction Model Parameter Tuning

High-energy physics experiments studying neutrinos rely heavily on simulations of their interactions with atomic nuclei. Limitations in the theoretical understanding of these interactions typically necessitate ad hoc tuning of simulation model parameters to data. Traditional tuning methods for neutrino experiments have largely relied on simple algorithms for numerical optimization. While adequate for the modest goals of initial efforts, the complexity of future neutrino tuning campaigns is expected to increase substantially, and new approaches will be needed to make progress. In this paper, we examine the application of simulation-based inference (SBI) to the neutrino interaction model tuning for the first time. Using a previous tuning study performed by the MicroBooNE experiment as a test case, we find that our SBI algorithm can correctly infer the tuned parameter values when confronted with a mock data set generated according to the MicroBooNE procedure. This initial proof-of-principle illustrates a promising new technique for next-generation simulation tuning campaigns for the neutrino experimental community.

Tame-Narvaez, Karla Maria [Fermilab]↗
NASA NTRS1992

A generalized reusable guidance algorithm for optimal aerobraking

A practical real-time guidance algorithm was developed for guiding aerobraking vehicles in such a way that the maximum heating rate, the maximum structural loads, and the post-aeropass delta-V requirements (for post-aeropass orbit insertion) are all minimized. The algorithm is general and reusable in the sense that a minimum of assumptions are made, thus minimizing the number of gains and mission-dependent parameters that must be laboriously determined prior to a particular mission. A particularly interesting feature is that inplane guidance performance is tuned by simply adjusting one mission-dependent parameter, the bank margin; similarly, the out-of-plane guidance performance is turned by simply adjusting a plane controller time constant. Other objectives in the algorithm development are simplicity, efficiency, and ease of use. The algorithm is developed for, but not necessarily restricted to, a single pass mission and a trimmed vehicle with a bank angle modulation as the method of trajectory control. Guidance performance is demonstrated via results obtained using this algorithm integrated into an aerobraking test-bed program. Comparisons are made with numerical results from a version of the aerobraking guidance algorithm that was to be flown onboard NASA's aeroassist flight experiment (AFE) vehicle. Promising results are obtained with a minimum of development effort.

Dukeman, G. A.↗
DOE OSTI2025

Simulation-based inference for neutrino interaction model parameter tuning

High-energy physics experiments studying neutrinos rely heavily on simulations of their interactions with atomic nuclei. Limitations in the theoretical understanding of these interactions typically necessitate ad hoc tuning of simulation model parameters to data. Traditional tuning methods for neutrino experiments have largely relied on simple algorithms for numerical optimization. While adequate for the modest goals of initial efforts, the complexity of future neutrino tuning campaigns is expected to increase substantially, and new approaches will be needed to make progress. In this paper, we examine the application of simulation-based inference (SBI) to the neutrino interaction model tuning for the first time. Using a previous tuning study performed by the MicroBooNE experiment as a test case, we find that our SBI algorithm can correctly infer the tuned parameter values when confronted with a mock data set generated according to the MicroBooNE procedure. This initial proof-of-principle illustrates a promising new technique for next-generation simulation tuning campaigns for the neutrino experimental community.

Tame-Narvaez, Karla [Fermilab] (ORCID:000000022249↗
NASA NTRS2025

A Comparison of Bank Control and Direct Force Control for Aerocapture at Uranus

We compare bank angle control and direct force control for aerocapture at Uranus. We utilize the Fully Numerical Predictor-corrector Aerocapture Guidance for both configurations and discuss modifications to the direct force control corrector to handle multiple roots. Using the Genesis flight mechanics simulation, we analyze both nominal and dispersed three-degree-of-freedom trajectories with perfect navigation and control. We present Monte Carlo results and discuss differences between the performance of the two control approaches. We find that increased bank angle rate and acceleration limits do not increase the performance of the bank control configuration. We analyze a variation with tighter angle of attack bounds and find that the tuning and performance of the two configurations converges, because direct force control is less able to modulate drag.

Sergio A Sandoval↗
NASA NTRS2011

Unsteady Aerodynamic Model Tuning for Precise Flutter Prediction

A simple method for an unsteady aerodynamic model tuning is proposed in this study. This method is based on the direct modification of the aerodynamic influence coefficient matrices. The aerostructures test wing 2 flight-test data is used to demonstrate the proposed model tuning method. The flutter speed margin computed using only the test validated structural dynamic model can be improved using the additional unsteady aerodynamic model tuning, and then the flutter speed margin requirement of 15 % in military specifications can apply towards the test validated aeroelastic model. In this study, unsteady aerodynamic model tunings are performed at two time invariant flight conditions, at Mach numbers of 0.390 and 0.456. When the Mach number for the unsteady model tuning approaches to the measured fluttering Mach number, 0.502, at the flight altitude of 9,837 ft, the estimated flutter speed is approached to the measured flutter speed at this altitude. The minimum flutter speed difference between the estimated and measured flutter speed is -.14 %.

Pak, Chan-Gi↗
NASA NTRS2011

Unsteady Aerodynamic Model Tuning for Precise Flutter Prediction

A simple method for an unsteady aerodynamic model tuning is proposed in this study. This method is based on the direct modification of the aerodynamic influence coefficient matrices. The aerostructures test wing 2 flight-test data is used to demonstrate the proposed model tuning method. The flutter speed margin computed using only the test validated structural dynamic model can be improved using the additional unsteady aerodynamic model tuning, and then the flutter speed margin requirement of 15 percent in military specifications can apply towards the test validated aeroelastic model. In this study, unsteady aerodynamic model tunings are performed at two time invariant flight conditions, at Mach numbers of 0.390 and 0.456. When the Mach number for the unsteady aerodynamic model tuning approaches to the measured fluttering Mach number, 0.502, at the flight altitude of 9,837 ft, the estimated flutter speed is approached to the measured flutter speed at this altitude. The minimum flutter speed difference between the estimated and measured flutter speed is -0.14 percent.

Pak, Chan-gi↗
DOE OSTI2024

Improved Subseasonal Forecasting of Extreme Polar Vortices Using Machine Learning

Our research was focused on forecasting the position and shape of the winter stratospheric polar vortex at a subseasonal timescale of 15 days in advance. To achieve this, we employed both statistical and neural network machine learning techniques. The analysis was performed on 42 winter seasons of reanalysis data provided by NASA giving us a total of 6,342 days of data. The state of the polar vortex for determined by using geometric moments to calculate the centroid latitude and the aspect ratio of an ellipse fit onto the vortex. Timeseries for thirty additional precursors were calculated to help improve the predictive capabilities of the algorithm. Feature importance of these precursors was performed using random forest to measure the predictive importance and the ideal number of precursors. Then, using the precursors identified as important, various statistical methods were tested for predictive accuracy with random forest and nearest neighbor performing the best. An echo state network, a type of recurrent neural network that features sparsely connected hidden layer and a reduced number of trainable parameters that allows for rapid training and testing, was also implemented for the forecasting problem. Hyperparameter tuning was performed for each methods using a subset of the training data. The algorithms were trained and tuned on the first 41 years of data, then tested for accuracy on the final year. In general, the centroid latitude of the polar vortex proved easier to predict than the aspect ratio across all algorithms. Random forest outperformed other statistical forecasting algorithms overall but struggled to predict extreme values. Forecasting from echo state network suggested a strong predictive capability past 15 days, but further work is required to fully realize the potential of recurrent neural network approaches.

54 ENVIRONMENTAL SCIENCES↗
NASA NTRS2001

Structural Acoustic Prediction and Interior Noise Control Technology

This report documents the results of Task 14, "Structural Acoustic Prediction and Interior Noise Control Technology". The task was to evaluate the performance of tuned foam elements (termed Smart Foam) both analytically and experimentally. Results taken from a three-dimensional finite element model of an active, tuned foam element are presented. Measurements of sound absorption and sound transmission loss were taken using the model. These results agree well with published data. Experimental performance data were taken in Boeing's Interior Noise Test Facility where 12 smart foam elements were applied to a 757 sidewall. Several configurations were tested. Noise reductions of 5-10 dB were achieved over the 200-800 Hz bandwidth of the controller. Accelerometers mounted on the panel provided a good reference for the controller. Configurations with far-field error microphones outperformed near-field cases.

Mathur, G. P.↗
DOE OSTI2021

Spray: Sparse Reductions of Arrays in OPEN MP

We present SPRAY, an open-source header-only C++ library for sparse reductions of arrays. SPRAY is meant for applications in which a large array is collaboratively updated by multiple threads using an associative and commutative operation such as +=. Especially when each thread accesses only parts of the array, SPRAY can perform significantly better than OPENMP's built-in reduction clause or atomic updates, while also using less memory than the former. SPRAY provides both an easy-to-use interface that can serve as a drop-in replacement for OPENMP reductions and a selection of reducer objects that accumulate the final result in different thread-safe ways. We demonstrate SPRAY through multiple test cases including the LULESH shock hydrodynamics code and a transpose-matrix-vector multiplication for sparse matrices stored in CSR format. SPRAY reductions outperform built-in OPENMP reductions consistently, in some cases improving run time and memory overhead by 20X, and even beating domain-specific approaches such as Intel MKI, by over 2X in some cases. Furthermore, SPRAY reductions have a minimal impact on the code base, requiring only a few lines of source code changes. Once in place, SPRAY reduction schemes can be switched easily, allowing performance portability and tuning opportunities by separating performance-critical implementation details from application code.

OpenMP↗
NASA NTRS2000

Eighth Goddard Conference on Mass Storage Systems and Technologies in Cooperation with the Seventeenth IEEE Symposium on Mass Storage Systems

This document contains copies of those technical papers received in time for publication prior to the Eighth Goddard Conference on Mass Storage Systems and Technologies which is being held in cooperation with the Seventeenth IEEE Symposium on Mass Storage Systems at the University of Maryland University College Inn and Conference Center March 27-30, 2000. As one of an ongoing series, this Conference continues to provide a forum for discussion of issues relevant to the management of large volumes of data. The Conference encourages all interested organizations to discuss long term mass storage requirements and experiences in fielding solutions. Emphasis is on current and future practical solutions addressing issues in data management, storage systems and media, data acquisition, long term retention of data, and data distribution. This year's discussion topics include architecture, future of current technology, new technology with a special emphasis on holographic storage, performance, standards, site reports, vendor solutions. Tutorials will be available on stability of optical media, disk subsystem performance evaluation, I/O and storage tuning, functionality and performance evaluation of file systems for storage area networks.

Kobler, Benjamin↗
DOE OSTI2021

Correlation of the Structure with Performance in MEH-PPV/dPS Thin Films Illuminated during Processing

Thin films containing conjugated polymers have been used in a wide array of optoelectronic devices, and much research has focused on the conformation of the conjugated polymer as a key aspect to tuning the performance of the resulting devices. White light exposure has been studied as a post-processing method to alter the film’s morphology and photoluminescence (PL) in systems composed of poly[2-methoxy-5-(2-ethylhexyloxy)-1,4-phenylenevinylene] (MEH-PPV) and polystyrene (PS). To date, there have been few studies that correlate the changes in the depth profile with the PL performance of MEH-PPV/PS films. This study is designed to address this gap in knowledge. Thin films of MEH-PPV/dPS annealed above the glass transition temperatures (T g ) in an illuminated environment led to lower PL than that found in annealed samples that are not illuminated. However, at higher MEH-PPV loading in the blend film, the extent of variation of PL with illumination diminished. Additionally, time-of-flight secondary ion mass spectrometry and small-angle neutron scattering document the three-dimensional morphology of the films under illumination conditions, which are correlated with the changes in PL. Illumination during annealing compressed MEH-PPV-rich layers in the thin film blends at low loadings of MEH-PPV (below 10 wt %) but resulted in significant in-plane phase separation at higher loadings (above 15 wt %). Both changes increase interchain interactions and lower the PL of the illuminated samples. The changes in the depth profile significantly alter the PL of the films, while the in-plane phase separation affected the optoelectronic properties to a lesser degree. This work, therefore, provides insights into how illumination and film composition can be utilized to predictably alter the structure and optoelectronic performance of conjugated polymer blend films.

36 MATERIALS SCIENCE↗
DOE OSTI2024

$\overline{TKE}$ Parameterization and $\bar{v}$ Uncertainty Analysis for CGMF

Previous work was performed on tuning CGMF parameters for 235 U, 238 U, and Plutonium isotopes. Now work is being done to tune minor uranium isotopes. However, uranium isotopes like 232 U and 236 U have almost no experimental data. We are applying cross-isotope models to extrapolate and tune CGMF on isotopes that lack experimental data. There exist several internal CGMF physics quantities that affect the output of CGMF—multi-chance fission probability, excitation energy sharing, spin-cutoff factor, spin scaling, and fragment total kinetic energy to name a few. The mean fragment total kinetic energy, $\overline{TKE}$, is particularly interesting because of its strong anti-correlation with $\bar{v}$. We are most interested in the mean fragment total kinetic energy before neutron emissions. $\overline{TKE}$ is assumed to be pre-neutron emission unless otherwise stated. Currently in CGMF, the $\overline{TKE}$ model for 233,234,235,238 U are tuned independently to reproduce ν for the associated isotopes. In this report, we will tune a cross-isotope $\overline{TKE}$ model to experimental $\overline{TKE}$ data for 232,233,234,235,236,238 U. Because of the unreliable and sparse nature of $\overline{TKE}$ experimental data, future work will use more reliable experimental $\bar{v}$ data to infer the $\overline{TKE}$ model (and likely other internal CGMF parameters) for uranium isotopes. Such work has been performed previously using a sensitivity analysis and Kalman filter methods.

07 ISOTOPE AND RADIATION SOURCES↗
NASA NTRS1992

RAID-2: Design and implementation of a large scale disk array controller

We describe the implementation of a large scale disk array controller and subsystem incorporating over 100 high performance 3.5 inch disk drives. It is designed to provide 40 MB/s sustained performance and 40 GB capacity in three 19 inch racks. The array controller forms an integral part of a file server that attaches to a Gb/s local area network. The controller implements a high bandwidth interconnect between an interleaved memory, an XOR calculation engine, the network interface (HIPPI), and the disk interfaces (SCSI). The system is now functionally operational, and we are tuning its performance. We review the design decisions, history, and lessons learned from this three year university implementation effort to construct a truly large scale system assembly.

Katz, R. H.↗