Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Computational pipelines”

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 OSTI2023

Building a Computational and Experimental Rapid Response Pipeline to Counter the Coronavirus Disease 2019 Outbreak and Emerging Biothreats

The COVID-19 pandemic underscored the promise of monoclonal antibody-based prophylactic and therapeutic drugs, especially where protective candidates can be rapidly identified or developed for emerging biothreats and escape variants. Current cutting-edge technologies for this purpose still rely on pathogen-exposed convalescent volunteers and a large screening effort to find a proverbial needle in a haystack. Computational design of protective antibodies based on pre-existing templates skips those requirements and allows for greater control over the breadth and target epitope, while also co-optimizing for potency and developability or other biophysical characteristics. We approached this problem by building and expanding an in vitro experimental rapid antibody production and characterization pipeline to support development of an autonomous, closed loop, active learning software system based on structural simulation and ground truth experimental data to design and evaluate antibody antigen interactions. Starting from early in the pandemic, we targeted SARS-CoV-2, the causative agent of COVID-19, by re-purposing neutralizing antibodies against SARS-CoV-1 that had been identified in the wake of that outbreak in the early 2000’s. We successfully re-targeted three different anti-SARS-CoV-1 antibodies to neutralize SARS-CoV-2 in vitro, where the antibodies were generated externally and tested through conventional binding and neutralization assays internally or with collaborators. As antibodies were identified from the blood of humans infected with SARS-CoV-2, we shifted to collaborate with academic partners to develop improved versions of their human-derived antibodies. This work reached its most important stage in rapid response to the emergence of the Omicron variant of concern (VOC) in late 2021. In a matter of weeks, enabled by on demand innovation to our screening pipeline, we computationally designed and experimentally characterized derivative antibodies of COV2-2130, one of two antibodies from Vanderbilt that form the basis of the AstraZeneca Evusheld prophylactic drug product. This drug product suffers a serious loss of efficacy against Omicron BA.1 and BA.1.1, the first Omicron strains. Due to tight integration of computational design and experimental evaluation, we were able to identify a pair of designs with potent neutralization of the main targets Omicron BA.1 and BA.1.1; but also the earlier Delta variant, and subsequent Omicron strains including BA.2, BA.4, BA.5, and BA.2.75, demonstrating that our multi-target design process can, by its nature, produce robust antibody designs that strictly improve over the parental antibody. These results, recognized by a 2022 Director’s Science and Technology award, have enabled the follow-on GUIDE program, to commence in FY23. While earlier design campaigns were substantially outsourced, we have engineered better and faster processes internally to better compliment, calibrate, and speed computational designs. As part of the follow-on GUIDE program, we will stand up a rapid and high-throughput antibody production and characterization facility staffed with the expertise and capabilities to foster our current collaboration across PLS and ENG as well as other partnerships toward computational design of biologics.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

Automated Image Segmentation and Processing Pipeline Applied to X–Ray Computed Tomography Studies of Pitting Corrosion in Aluminum Wires

Understanding pitting corrosion is critical, yet its kinetics and morphology remain challenging to study from X-ray computed tomography (XCT) due to manual segmentation barriers. To address this, an automated pipeline leveraging deep learning for efficient large-scale XCT analysis is developed, revealing new corrosion insights. The pipeline enables pit segmentation, 3D reconstruction, statistical characterization, and a topological transformation for visualization. Here, the pipeline is applied to 87 648 XCT images capturing commercial purity aluminum (1100 Al) wire exposed to sodium chloride (NaCl) salt particles over a period of 122 h. The pipeline achieves complete feature extraction and statistical quantification across the entire XCT dataset, leveraging distributed computing environment for high efficiency. Global growth kinetics such as high-level stepwise sigmoidal volume loss patterns and granular individual pit developments are both captured for 36 detected pits. By combining automation, computer vision, and extensive XCT datasets, this research accelerates precise corrosion assessment to enable materials science discoveries at scale.

36 MATERIALS SCIENCE↗
NASA NTRS1990

Efficient algorithms for dilated mappings of binary trees

The problem is addressed to find a 1-1 mapping of the vertices of a binary tree onto those of a target binary tree such that the son of a node on the first binary tree is mapped onto a descendent of the image of that node in the second binary tree. There are two natural measures of the cost of this mapping, namely the dilation cost, i.e., the maximum distance in the target binary tree between the images of vertices that are adjacent in the original tree. The other measure, expansion cost, is defined as the number of extra nodes/edges to be added to the target binary tree in order to ensure a 1-1 mapping. An efficient algorithm to find a mapping of one binary tree onto another is described. It is shown that it is possible to minimize one cost of mapping at the expense of the other. This problem arises when designing pipelined arithmetic logic units (ALU) for special purpose computers. The pipeline is composed of ALU chips connected in the form of a binary tree. The operands to the pipeline can be supplied to the leaf nodes of the binary tree which then process and pass the results up to their parents. The final result is available at the root. As each new application may require a distinct nesting of operations, it is useful to be able to find a good mapping of a new binary tree over existing ALU tree. Another problem arises if every distinct required binary tree is known beforehand. Here it is useful to hardwire the pipeline in the form of a minimal supertree that contains all required binary trees.

Iqbal, M. Ashraf↗
NASA NTRS1990

Optimal pipelining

An effort is made to characterize the tradeoffs and overheads limiting the speedup potential theoretically projected for pipeline-incorporating computer architectures, using a mathematical model of the roles played by the various parameters. Pipeline optimization proceeds by a partitioning of the pipeline into an optimum number of segments so that maximization of throughput is obtained. Inferences are drawn from the model, and potential improvements to it are identified. Substantial agreement is obtained with Kunkel and Smith's (1986) CRAY-1S simulations of pipelining.

Dubey, Pradeep K.↗
DOE OSTI2020

Distributed Many-to-Many Protein Sequence Alignment using Sparse Matrices

Identifying similar protein sequences is a core step in many computational biology pipelines such as detection of homologous protein sequences, generation of similarity protein graphs for downstream analysis, functional annotation, and gene location. Performance and scalability of protein similarity search have proven to be a bottleneck in many bioinformatics pipelines due to increase in cheap and abundant sequencing data. This work presents a new distributed-memory software PASTIS. PASTIS relies on sparse matrix computations for efficient identification of possibly similar proteins. We use distributed sparse matrices for scalability and show that the sparse matrix infrastructure is a great fit for protein similarity search when coupled with a fully-distributed dictionary of sequences that allow remote sequence requests to be fulfilled. Our algorithm incorporates the unique bias in amino acid sequence substitution in search without altering basic sparse matrix model, and in turn, achieves ideal scaling up to millions of protein sequences.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1984

JPL Robotics Laboratory computer vision software library

The past ten years of research on computer vision have matured into a powerful real time system comprised of standardized commercial hardware, computers, and pipeline processing laboratory prototypes, supported by anextensive set of image processing algorithms. The software system was constructed to be transportable via the choice of a popular high level language (PASCAL) and a widely used computer (VAX-11/750), it comprises a whole realm of low level and high level processing software that has proven to be versatile for applications ranging from factory automation to space satellite tracking and grappling.

Cunningham, R.↗
DOE OSTI2021

IGRINS RV: A Precision Radial Velocity Pipeline for IGRINS Using Modified Forward Modeling in the Near-infrared

Application of the radial velocity (RV) technique in the near-infrared is valuable because of the diminished impact of stellar activity at longer wavelengths, making it particularly advantageous for the study of late-type stars but also for solar-type objects. In this paper, we present the IGRINS RV open-source python pipeline for computing infrared RV measurements from reduced spectra taken with IGRINS, an R ≡ λ/Δλ ∼ 45,000 spectrograph with simultaneous coverage of the H band (1.49–1.80 μm) and K band (1.96–2.46 μm). Using a modified forward-modeling technique, we construct high-resolution telluric templates from A0 standard observations on a nightly basis to provide a source of common-path wavelength calibration while mitigating the need to mask or correct for telluric absorption. Telluric standard observations are also used to model the variations in instrumental resolution across the detector, including a yearlong period when the K band was defocused. Without any additional instrument hardware, such as a gas cell or laser frequency comb, we are able to achieve precisions of 26.8 m s{sup −1} in the K band and 31.1 m s{sup −1} in the H band for narrow-line hosts. These precisions are empirically determined by a monitoring campaign of two RV standard stars, as well as the successful retrieval of planet-induced RV signals for both HD 189733 and τ Boo A; furthermore, our results affirm the presence of the Rossiter–McLaughlin effect for HD 189733. The IGRINS RV pipeline extends another important science capability to IGRINS, with publicly available software designed for widespread use.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2022

Automated defect identification in electroluminescence images of solar modules

Solar photovoltaic (PV) modules are susceptible to manufacturing defects, mishandling problems or extreme weather events that can limit energy production or cause early device failure. Trained professionals use electroluminescence (EL) images to identify defects in modules, however, field surveys or inline image acquisition can generate millions of EL images, which are infeasible to analyze by rote inspection. Here, we develop a rapid automatic computer vision pipeline (~0.5 seconds/module) to analyze EL images and identify defects including cracks, intra-cell defects, oxygen-induced defects, and solder disconnections. Defect identification is achieved with a machine learning model (Random Forest, ResNet models and YOLO) trained on 762 manually-labeled EL images of PV modules. We compare model performance on an imbalanced real-world validation set containing 134 EL images and determine that ResNet18 and YOLO are the optimal models; we next evaluated these models on a dedicated testing set (129 module images) with resulting macro F1 scores of 0.83 (ResNet18) and 0.78 (YOLO). Using a field EL survey of a PV power plant damaged in a vegetation fire, we analyze 18,954 EL images (2.4 million cells) and inspect the spatial distribution of defects on the solar modules. The results find increased frequency of ‘crack’, ‘solder’ and ‘intra-cell’ defects on the edges of the solar module closest to the ground after fire. We also find an abnormal increase of striation rings on cells which were assumed to be caused mainly in fabrication process. Our methods are published as open-source software. It can also be used to identify other kinds of defects or process different types of solar cells with minor modification on models by transfer learning.

14 SOLAR ENERGY↗
DOE OSTI2025

Structure and Synthesizability of Iron–Sulfur Metal–Organic Frameworks

Sulfur-based metal–organic frameworks (MOFs) and coordination polymers (CPs) are an emerging class of hybrid materials that have received growing attention due to their magnetic, conductive, and catalytic properties with potential applications in electrocatalysis and energy storage. In this work, we report a high-throughput virtual screening protocol to predict the synthesizability of candidate metal–sulfur MOFs/CPs by computing the thermodynamically stable structures resulting from a particular combination of metal cluster, linker, cation, and synthetic conditions. Free energies are computed by using all-atom classical mechanical thermodynamic integration. Low-free-energy structures are refined using ab initio density functional theory, and pair distribution functions and powder X-ray diffraction patterns are calculated to complement and guide experimental structure determination. We validate the computational approach by retrospective predictions of the stable structure produced by experimental syntheses, and a subsequent screen predicts Fe 4 S 4 -BDT–TPP as a new thermodynamically stable one-dimensional (1D) CP comprising a redox-active Fe 4 S 4 cluster, a 1,4-benzenedithiolate (BDT) linker, and a tetraphenylphosphonium (TPP) countercation. Furthermore, this material is experimentally synthesized, and the 1D chain structure of the crystal is confirmed using microcrystal electron diffraction. The computational screening pipeline is generically transferable to neutral and ionic MOFs/CPs comprising arbitrary metal clusters, linkers, cations, and synthetic conditions, and we make it freely available as an open source tool to guide and accelerate the discovery and engineering of novel porous materials.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

A high-throughput framework for lattice dynamics

We develop an automated high-throughput workflow for calculating lattice dynamical properties from first principles including those dictated by anharmonicity. The pipeline automatically computes interatomic force constants (IFCs) up to 4th order from perturbed training supercells, and uses the IFCs to calculate lattice thermal conductivity, coefficient of thermal expansion, and vibrational free energy and entropy. It performs phonon renormalization for dynamically unstable compounds to obtain real effective phonon spectra at finite temperatures and calculates the associated free energy corrections. The methods and parameters are chosen to balance computational efficiency and result accuracy, assessed through convergence testing and comparisons with experimental measurements. Deployment of this workflow at a large scale would facilitate materials discovery efforts toward functionalities including thermoelectrics, contact materials, ferroelectrics, aerospace components, as well as general phase diagram construction.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2021

Planck constraints on the tensor-to-scalar ratio

In this work, we present constraints on the tensor-to-scalar ratio r using Planck data. We use the latest release of Planck maps (PR4), processed with the NPIPE code, which produces calibrated frequency maps in temperature and polarization for all Planck channels from 30 GHz to 857 GHz using the same pipeline. We computed constraints on r using the BB angular power spectrum, and we also discuss constraints coming from the TT spectrum. Given Planck's noise level, the TT spectrum gives constraints on r that are cosmic-variance limited (with $\sigma$ r = 0.093), but we show that the marginalized posterior peaks towards negative values of r at about the 1.2$\sigma$ level. We derived Planck constraints using the BB power spectrum at both large angular scales (the 'reionization bump') and intermediate angular scales (the 'recombination bump') from $\ell$=2 to 150, and find a stronger constraint than that from TT, with $\sigma$ r = 0.069. The Planck BB spectrum shows no systematic bias, and is compatible with zero, given both the statistical noise and the systematic uncertainties. The likelihood analysis using B modes yields the constraint r < 0.158 at 95% confidence using more than 50% of the sky. This upper limit tightens to r < 0.069 when Planck EE, BB, and EB power spectra are combined consistently, and it tightens further to r < 0.056 when the Planck TT power spectrum is included in the combination. Finally, combining Planck with BICEP2/Keck 2015 data yields an upper limit of r < 0.044.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

The miniJPAS survey quasar selection – I. Mock catalogues for classification

In this series of papers, we employ several machine learning (ML) methods to classify the point-like sources from the miniJPAS catalogue, and identify quasar candidates. Since no representative sample of spectroscopically confirmed sources exists at present to train these ML algorithms, we rely on mock catalogues. In this first paper, we develop a pipeline to compute synthetic photometry of quasars, galaxies, and stars using spectra of objects targeted as quasars in the Sloan Digital Sky Survey . To match the same depths and signal-to-noise ratio distributions in all bands expected for miniJPAS point sources in the range 17.5 ≤ r < 24, we augment our sample of available spectra by shifting the original r-band magnitude distributions towards the faint end, ensure that the relative incidence rates of the different objects are distributed according to their respective luminosity functions, and perform a thorough modelling of the noise distribution in each filter, by sampling the flux variance either from Gaussian realizations with given widths, or from combinations of Gaussian functions. Finally, we also add in the mocks the patterns of non-detections which are present in all real observations. Although the mock catalogues presented in this work are a first step towards simulated data sets that match the properties of the miniJPAS observations, these mocks can be adapted to serve the purposes of other photometric surveys.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

Towards Low-Overhead Resilience for Data Parallel Deep Learning

Data parallel techniques have been widely adopted both in academia and industry as a tool to enable scalable training of deep learning models. At scale, DL training jobs can fail due to software or hardware bugs, may need to be preempted or terminated due to unexpected events, or may perform suboptimally because they were misconfigured. Under such circumstances, there is a need to recover and/or reconfigure data-parallel DL training jobs on-the-fly, while minimizing the impact on the accuracy of the DNN model and the runtime overhead. In this regard, state-of-art techniques adopted by the HPC community mostly rely on checkpoint-restart, which inevitably leads to loss of progress, thus increasing the runtime overhead. In this paper we explore alternative techniques that exploit the properties of modern deep learning frameworks (overlapping of gradient averaging and weight updates with local gradient computations through pipeline parallelism) to reduce the overhead of resilience/elasticity. To this end we introduce a failure simulation framework and two resilience strategies (immediate mini-batch rollback and lossy forward recovery), which we study compared with checkpoint-restart approaches in a variety of settings in order to understand the trade-offs between the accuracy loss of the DNN model and the runtime overhead.

data-parallel training↗
DOE OSTI2025

Catalyst-Vision (PEM Catalyst Layer Image Analysis Tool) [SWR-25-100]

Catalyst-Vision (PEM Catalyst Layer Image Analysis Tool) provides an advanced Python-based tool, primarily designed for use in a Jupyter/Colab notebook, for the quantitative morphological analysis of pre-segmented shapes. While developed for analyzing PEM catalyst layers from microscopy, its methodology is suitable for characterizing any grayscale object provided on a uniform white background. The tool uses a robust computer vision pipeline based on the Euclidean Distance Transform and skeletonization to accurately measure local thickness and tortuosity, providing a comprehensive characterization of an object's geometry and internal texture. If you find this code useful, please cite our preprint as: Chan, Ai-Lin and Hayden, Steven and Harvey, Steven P. and Smeaton, Michelle and Okrucky, Caleb and Watt, John and Ulična, Soňa and Spurgeon, Steven and Jungjohann, Katherine and Alia, Shaun, Mechanism-informed breakdown: understanding degradation by controlling voltage hold patterns in PEM water electrolyzers. Preprint (2025).

Spurgeon, Steven [National Laboratory of the Rocki↗
DOE OSTI2025

Replacing non-biomedical concepts improves embedding of biomedical concepts

Embeddings are semantically meaningful representations of words in a vector space, commonly used to enhance downstream machine learning applications. Traditional biomedical embedding techniques often replace all synonymous words representing biological or medical concepts with a unique token, ensuring consistent representation and improving embedding quality. However, the potential impact of replacing non-biomedical concept synonyms has received less attention. Embedding approaches often employ concept replacement to replace concepts that span multiple words, such as non-small-cell lung carcinoma, with a single concept identifier (e.g., D002289). Also, all synonyms of each concept are merged into the same identifier. Here, we additionally leveraged WordNet to identify and replace sets of non-biomedical synonyms with their most common representatives. This combined approach aimed to reduce embedding noise from non-biomedical terms while preserving the integrity of biomedical concept representations. We applied this method to 1,055 biomedical concept sets representing molecular signatures or medical categories and assessed the mean pairwise distance of embeddings with and without non-biomedical synonym replacement. A smaller mean pairwise distance was interpreted as greater intra-cluster coherence and higher embedding quality. Embeddings were generated using the Word2Vec algorithm applied to a corpus of 10 million PubMed abstracts. Our results demonstrate that the addition of non-biomedical synonym replacement reduced the mean intra-cluster distance by an average of 8%, suggesting that this complementary approach enhances embedding quality. Future work will assess its applicability to other embedding techniques and downstream tasks. Python code implementing this method is provided under an open-source license.

algorithms↗
NASA NTRSDate not supplied

AAM NC Tech Talk: Intro to Amazon Web Services

The purpose of this Tech Talk is to engage with the Advance Air Mobility National Campaign community on the use of a cloud provider to promote public confidence and accelerate the realization of emerging aviation markets for passenger and cargo transportation in urban, suburban, rural, and regional environments.

AWS↗
DOE OSTI2020

Scalable All-pairs Shortest Paths for Huge Graphs on Multi-GPU Clusters

We present an optimized Floyd-Warshall (Floyd-Warshall) algorithm that computes the All-pairs shortest path (APSP) for GPU accelerated clusters. The Floyd-Warshall algorithm due to its structural similarities to matrix-multiplication is well suited for highly parallel GPU architectures. To achieve high parallel efficiency, we address two key algorithmic challenges: reducing high communication overhead and addressing limited GPU memory. To reduce high communication costs, we redesign the parallel (a) to expose more parallelism, (b) aggressively overlap communication and computation with pipelined and asynchronous scheduling of operations, and (c) tailored MPI-collective. To cope with limited GPU memory, we employ an offload model, where the data resides on the host and is transferred to GPU on-demand. The proposed optimizations are supported with detailed performance models for tuning. Our optimized parallel Floyd-Warshall implementation is up to 5x faster than a strong baseline and achieves 8.1 PetaFLOPS/sec on 256~nodes of the Summit supercomputer at Oak Ridge National Laboratory. This performance represents 70% of the theoretical peak and 80% parallel efficiency. The offload algorithm can handle 2.5x larger graphs with a 20% increase in overall running time.

Sao, Piyush↗
NASA NTRS2008

High speed three-dimensional laser scanner with real time processing

A laser scanner computes a range from a laser line to an imaging sensor. The laser line illuminates a detail within an area covered by the imaging sensor, the area having a first dimension and a second dimension. The detail has a dimension perpendicular to the area. A traverse moves a laser emitter coupled to the imaging sensor, at a height above the area. The laser emitter is positioned at an offset along the scan direction with respect to the imaging sensor, and is oriented at a depression angle with respect to the area. The laser emitter projects the laser line along the second dimension of the area at a position where a image frame is acquired. The imaging sensor is sensitive to laser reflections from the detail produced by the laser line. The imaging sensor images the laser reflections from the detail to generate the image frame. A computer having a pipeline structure is connected to the imaging sensor for reception of the image frame, and for computing the range to the detail using height, depression angle and/or offset. The computer displays the range to the area and detail thereon covered by the image frame.

Lavelle, Joseph P.↗