Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “SLICING”

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 1,261 records · Page 70

Load-balanced sparse MTTKRP on GPUs

Sparse matricized tensor times Khatri-Rao product (MT- TKRP) is one of the most computationally expensive kernels in sparse tensor computations. This work focuses on optimizing the MTTKRP for floating point operations, storage, and scalability. We begin by identifying the performance bottlenecks in directly extending the state-of-the-art CSF (compressed sparse fiber) formats from CPUs to GPUs. Our detailed analysis over the recently proposed formats shows that the lower bounds on storage and flop counts can vary significantly depending on the structure of the sparse tensor. To address this, we propose a load balanced, computation and storage-efficient scheme, HYB, which combines the best of COO (coordinate), CSF and CSL (compressed slice). With these enhancements, our GPU framework significantly out- performs the current formats on both CPU and GPU platforms.

Nisa, Israt↗

SeekAView: An Intelligent Dimensionality Reduction Strategy for Navigating High-Dimensional Data Spaces

Dealing with the curse of dimensionality is a key challenge in high-dimensional data visualization. We present SeekAView to address three main gaps in the existing research literature. First, automated methods like dimensionality reduction or clustering suffer from a lack of transparency in letting analysts interact with their outputs in real-time to suit their exploration strategies. The results often suffer from a lack of interpretability, especially for domain experts not trained in statistics and machine learning. Second, exploratory visualization techniques like scatter plots or parallel coordinates suffer from a lack of visual scalability: it is difficult to present a coherent overview of interesting combinations of dimensions. Third, the existing techniques do not provide a flexible workflow that allows for multiple perspectives into the analysis process by automatically detecting and suggesting potentially interesting subspaces. In SeekAView we address these issues using suggestion based visual exploration of interesting patterns for building and refining multidimensional subspaces. Compared to the state-of-the-art in subspace search and visualization methods, we achieve higher transparency in showing not only the results of the algorithms, but also interesting dimensions calibrated against different metrics. We integrate a visually scalable design space with an iterative workflow guiding the analysts by choosing the starting points and letting them slice and dice through the data to find interesting subspaces and detect correlations, clusters, and outliers. We present two usage scenarios for demonstrating how SeekAView can be applied in real-world data analysis scenarios.

Krause, Josua↗

Temporal Decompostion of a Distribution System Quasi-Static Time-Series Simulation

This paper documents the first phase of an investigation into reducing runtimes of complex OpenDSS models through parallelization. As the method seems promising, future work will quantify - and further mitigate - errors arising from this process. In this initial report, we demonstrate how, through the use of temporal decomposition, the run times of a complex distribution-system-level quasi-static time series simulation can be reduced roughly proportional to the level of parallelization. Using this method, the monolithic model runtime of 51 hours was reduced to a minimum of about 90 minutes. As expected, this comes at the expense of control- and voltage-errors at the time-slice boundaries. All evaluations were performed using a real distribution circuit model with the addition of 50 PV systems - representing a mock complex PV impact study. We are able to reduce induced transition errors through the addition of controls initialization, though small errors persist. The time savings with parallelization are so significant that we feel additional investigation to reduce control errors is warranted.

14 SOLAR ENERGY↗

Sonic Lift-off of GaAs-based Solar Cells with Reduced Surface Facets

Spalling has been proposed as a promising kerfless technique for slicing thin wafers and devices, enhancing the wafer yield from an ingot by reducing kerf-losses and enabling substrate reuse for the case of costly devices. The main challenge of spalling is to control the roughness and thickness variation of the spalled wafers that can be as high as 100% of the wafer thickness. Here, we have shown previously that the thickness variations are largely influenced by the interaction of the crack front with its own reflected waves. Herein we show that through Sonic Lift-off we can offer a modification to standard spalling technologies that minimizes the interaction of reflected waves with the crack front during lift-off of GaAs(100)-based solar cells. We show that the average surface facets amplitude can be reduced by 96% while obtaining a high quality bulk with no discernible residual strain as measured by PL. The resulting spalling depth can easily be tuned from 100 µm to less than 15µm. The electrical characterization of the Sonic lift-off cell shows no degradation of the VOC and a power conversion within 99% relative of a non-spalled baseline.

42 ENGINEERING↗

A Metallurgical Inspection Method to Assess the Damage in Performance-Limiting Nb 3 Sn Accelerator Magnet Coils

The design and production of Nb 3 Sn-based dipole and quadrupole magnets is critical for the realization of the High-Luminosity Large Hadron Collider (HL-LHC) at the European Organization for Nuclear Research (CERN). Nb 3 Sn superconducting coils are aimed at enhancing the bending and focusing strengths of accelerator magnets for HL-LHC and beyond. Due to the brittle nature of Nb 3 Sn, the coil fabrication steps are very challenging and require very careful QA/QC. Flaws in the Nb 3 Sn filaments may lead to quenches, and eventually, performance limitation below nominal during magnet testing. A novel inspection method, including advanced non-destructive and destructive techniques, was developed to explore the root-causes of quenches occurring in performance-limiting coils. The most relevant results obtained for MQXF coils through this innovative inspection method are presented. This approach allows for precise assessment of the physical events associated to the quenches experienced b y magnet coils, mainly occurring under the form of damaged strands with transversely broken sub-elements. Coil-slice preparation, micro-optical observations of transverse and longitudinal cross-sections, and a deep etching technique of copper will be illustrated in the present work, with a focus on the results achieved for a CERN coil from a non-conforming quadrupole magnet prototype, and two coils fabricated in the US, in the framework of the Accelerator Upgrade Project (AUP) collaboration, from two different non-conforming quadrupole magnets, respectively. The results obtained through the proposed inspection method will be illustrated.

43 PARTICLE ACCELERATORS↗

An Accurate, Error-Tolerant, and Energy-Efficient Neural Network Inference Engine Based on SONOS Analog Memory

In this work, we demonstrate SONOS (silicon-oxide-nitrideoxide- silicon) analog memory arrays that are optimized for neural network inference. The devices are fabricated in a 40nm process and operated in the subthreshold regime for in-memory matrix multiplication. Subthreshold operation enables low conductances to be implemented with low error, which matches the typical weight distribution of neural networks, which is heavily skewed toward near-zero values. This leads to high accuracy in the presence of programming errors and process variations. We simulate the end-to-end neural network inference accuracy, accounting for the measured programming error, read noise, and retention loss in a fabricated SONOS array. Evaluated on the ImageNet dataset using ResNet50, the accuracy using a SONOS system is within 2.16% of floating-point accuracy without any retraining. The unique error properties and high On/Off ratio of the SONOS device allow scaling to large arrays without bit slicing, and enable an inference architecture that achieves 20 TOPS/W on ResNet50, a >10× gain in energy efficiency over state-of-the-art digital and analog inference accelerators.

97 MATHEMATICS AND COMPUTING↗

Diagnosis of Performance Faults in LargeScale MPI Applications via Probabilistic Progress-Dependence Inference

Debugging large-scale parallel applications is challenging. Most existing techniques provide little information about failure root causes. Further, most debuggers significantly slow down program execution, and run sluggishly with massively parallel applications. This paper presents a novel technique that scalably infers the tasks in a parallel program on which a failure occurred, as well as the code in which it originated. Our technique combines scalable runtime analysis with static analysis to determine the least-progressed task(s) and to identify the code lines at which the failure arose. We present a novel algorithm that infers probabilistically progress dependence among MPI tasks using a globally constructed Markov model that represents tasks' control-flow behavior. In comparison to previous work, our algorithm infers more precisely the least-progressed task. Further, we combine this technique with static backward slicing analysis, further isolating the code responsible for the current state. A blind study demonstrates that our technique isolates the root cause of a concurrency bug in a molecular dynamics simulation, which only manifests itself at 7,996 tasks or more. We extensively evaluate fault coverage of our technique via fault injections in 10 HPC benchmarks and show that our analysis takes less than a few seconds on thousands of parallel tasks.

97 MATHEMATICS AND COMPUTING↗

The combined and separate impacts of climate extremes on the current and future US rainfed maize and soybean production under elevated CO 2

Heat and drought stresses are two emerging climatic threats to the US maize and soybean production, yet their impacts on yields are collectively determined by the magnitude of climate change and rising atmospheric CO2 concentration. Here we present a study that quantified the current and future yield responses of US rainfed maize and soybean to climate extremes, and for the first time characterized spatial shifts in the relative importance of temperature, heat and drought stress. Crop yields are simulated using the Agricultural Production Systems sIMulator (APSIM), driven by the high-resolution (12 km) Weather Research and Forecasting (WRF) Model downscaled future climate scenarios at two time slices (1995-2005 and 2085-2094). Our results show that climatic yield gaps and interannual variability are greater in the core production area than in the remaining US by the late 21st century under both Representative Concentration Pathway (RCP) 4.5 and RCP8.5 scenarios, and the magnitude of change is highly dependent on the current climate sensitivity and vulnerability. Elevated CO2 partially offsets the climatic yield gaps and reduces interannual yield variability, and effect is more prominent in soybean than in maize. We demonstrate that drought will continue to be the largest threat to US rainfed maize and soybean production, although its dominant role gradually gives way to other impacts of heat extremes. We also reveal that shifts in the geographic distributions of dominant stressors are characterized by increases in the concurrent stress, especially for the US Midwest. These findings imply the importance of considering drought and extreme heat simultaneously for future agronomic adaptation and mitigation strategies, particularly for breeding programs and crop management.

APSIM↗

Developments in TEM Nanotomography of Calcium Silicate Hydrate

This investigation was designed to explore the possibility of using transmission electron microscope ( TEM ) tomography on cement‐based systems gain a greater understanding of their nanostructure and pore network. The preliminary results show a clearly a well‐defined pore network at the nanoscale, with pore size approximately 1.7–2.4 nm in diameter and spaced around 5–8 nm apart. A comparison of small angle X‐ray scattering data with 2‐D TEM images analyzed with the Fourier slice theorem documents an excellent structural correlation.

Taylor, Rae↗

Measured Biaxial Residual Stress Maps in a Stainless Steel Weld

Here, this paper describes a sequence of residual stress measurements made to determine a two-dimensional map of biaxial residual stress in a stainless steel weld. A long stainless steel (316L) plate with an eight-pass groove weld (308L filler) was used. The biaxial stress measurements follow a recently developed approach, comprising a combination of contour method and slitting measurements, with a computation to determine the effects of out-of-plane stress on a thin slice. The measured longitudinal stress is highly tensile in the weld- and heat-affected zone, with a maximum around 450 MPa, and compressive stress toward the transverse edges around ₋250 MPa. The total transverse stress has a banded profile in the weld with highly tensile stress at the bottom of the plate (y = 0) of 400 MPa, rapidly changing to compressive stress (at y = 5 mm) of ₋200 MPa, then tensile stress at the weld root (y = 17 mm) and in the weld around 200 MPa, followed by compressive stress at the top of the weld at around ₋150 MPa. Finally, the results of the biaxial map compare well with the results of neutron diffraction measurements and output from a computational weld simulation.

36 MATERIALS SCIENCE↗

A Perspective of Energy Codes and Regulations for the Buildings of the Future

Today’s building energy codes focus on prescriptive requirements for features of buildings that are directly controlled by the design and construction teams and verifiable by municipal inspectors. Although these code requirements have had a significant impact, they fail to influence a large slice of the building energy use pie – including not only miscellaneous plug loads, cooking equipment and commercial/industrial processes, but the maintenance and optimization of the code-mandated systems as well. Currently, code compliance is verified only through the end of construction, and there are no limits or consequences for the actual energy use in an occupied building. In the future, our suite of energy regulations will likely expand to include building efficiency, energy use or carbon emission budgets over their full life cycle. Intelligent building systems, extensive renewable energy, and a transition from fossil fuel to electric heating systems will likely be required to meet ultra-low-energy targets. This paper lays out the authors’ perspectives on how buildings may evolve over the course of the 21st century and the roles that codes and regulations will play in shaping those buildings of the future.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

A Critical Review of Heliostat Design for Concentrating Solar Thermal Technologies

Heliostat designs have undergone a widespread and eclectic development process, with many unique designs demonstrated. However, recent developments do not show the overall cohort converging toward a globally accepted universal design. Here, this study characterizes heliostats by breaking down and evaluating design traits based on emergent patterns from a comprehensive compilation of known heliostat designs spanning several decades. Four main categories for evaluation emerged: heliostat base, heliostat primary axis, heliostat drive, and facet support. Each of these four categories is further defined by four subtypes so that all heliostats fall into a single subtype within each category. The classified heliostats are ranked, yielding several view slices into the heliostat compilation or a breakdown of heliostats by type. An analysis of the breakdown shows several trends: a scale-up of established designs, new approaches at small to medium scales, and a movement toward greater adoption of linear drives. These trends reflect the most meaningful contributing factors to a proposed trajectory for a new era of heliostat designs striving to meet widely considered cost targets of $\$$50/m 2 or $\$$75/m 2 .

14 SOLAR ENERGY↗

Study of stress in tensile nitrogen-plasma-treated multilayer silicon nitride films

The authors conducted a physico-chemical analysis of tensile sequential-nitrogen-plasma-treated silicon nitride films, which function as stressor liners in complementary metal oxide semiconductor (CMOS) technologies. These films are made of stacked nanometer-thick, plasma-enhanced, chemical vapor-deposited layers which were individually treated with N{sub 2}-plasma, to increase stress. This study allowed us to monitor the evolution of the films' chemical composition and stress as a function of process parameters such as deposition and post-N{sub 2}-plasma duration. Consistent with secondary ion mass spectroscopy (SIMS), transmission electron microscopy (TEM) and other physico-chemical analysis results, it was shown that the elementary component of the films can be modeled with a bi-layer consisting of an untreated slice at the bottom that is covered by a more tensile post-treated film. In addition, we observed that longer plasma treatments increase residual stress, SiN bond concentration and layer density, while reducing hydrogen content. The stress increase induced by the plasma treatment was shown to correlate with the increase in SiN bonds following a percolation mechanism that is linked to hydrogen dissociation. Kinetics laws describing both SiN bond generation and stress increase are proposed and it is demonstrated that stress increase follows first-order kinetics.

36 MATERIALS SCIENCE↗

Black-box optimization of CT acquisition and reconstruction parameters: a reinforcement learning approach

Protocol optimization is critical in Computed Tomography (CT) for achieving desired diagnostic image quality while minimizing radiation dose. Due to the inter-effect of influencing CT parameters, traditional optimization methods rely on the testing of exhaustive combinations of these parameters. This poses a notable limitation due to the impracticality of exhaustive parameter testing. This study introduces a novel methodology leveraging Virtual Imaging Trials (VITs) and reinforcement learning to more efficiently optimize CT protocols. Computational phantoms with liver lesions were imaged using a validated CT simulator and reconstructed with a novel CT reconstruction Toolkit. The optimization parameter space included tube voltage, tube current, reconstruction kernel, slice thickness, and pixel size. The optimization process was done using a Proximal Policy Optimization (PPO) agent which was trained to maximize the Detectability Index (d’) of the liver lesion for each reconstructed image. Results showed that our reinforcement learning approach found the absolute maximum d’ across the test cases while requiring 79.7% fewer steps compared to an exhaustive search, demonstrating both accuracy and computational efficiency, offering a efficient and robust framework for CT protocol optimization. The flexibility of the proposed technique allows for use of varying image quality metrics as the objective metric to maximize for. Our findings highlight the advantages of combining VIT and reinforcement learning for CT protocol management.

Fenwick, David [Duke University Medical Center]↗

Sparse measurement medical CT reconstruction using multi-fused block matching denoising priors

A major challenge for medical X-ray CT imaging is reducing the number of X-ray projections to lower radiation dosage and reduce scan times without compromising image quality. However these under-determined inverse imaging problems rely on the formulation of an expressive prior model to constrain the solution space while remaining computationally tractable. Traditional analytical reconstruction methods like Filtered Back Projection (FBP) often fail with sparse measurements, producing artifacts due to their reliance on the Shannon-Nyquist Sampling Theorem. Consensus Equilibrium, which is a generalization of Plug and Play, is a recent advancement in Model-Based Iterative Reconstruction (MBIR), has facilitated the use of multiple denoisers are prior models in an optimization free framework to capture complex, non-linear prior information. However, 3D prior modelling in a Plug and Play approach for volumetric image reconstruction requires long processing time due to high computing requirement. Instead of directly using a 3D prior, this work proposes a BM3D Multi Slice Fusion (BM3D-MSF) prior that uses multiple 2D image denoisers fused to act as a fully 3D prior model in Plug and Play reconstruction approach. Our approach does not require training and are thus able to circumvent ethical issues related with patient training data and are readily deployable in varying noise and measurement sparsity levels. In addition, reconstruction with the BM3D-MSF prior achieves similar reconstruction image quality as fully 3D image priors, but with significantly reduced computational complexity. We test our method on clinical CT data and demonstrate that our approach improves reconstructed image quality.

Hossain, Maliha [ORNL]↗

Ultrasonic Phased Array Sound Field Mapping Through Large-Bore Coarse Grained Cast Austenitic Stainless Steel (CASS) Piping Materials

A sound field beam mapping exercise was conducted to further understand the effects of coarse grained microstructures found in CASS materials on phased array ultrasonic wave propagation. Laboratory measurements were made on three CASS specimens with different microstructures; the specimens were polished and etched to reveal measurable grain sizes, shapes and orientations. Three longitudinal, phased array probes were fixed on a specimen's outside diameter with the sound field directed toward one end (face) of the pipe segment over a fixed range of angles. A point receiver was raster scanned over the surface of the specimen face generating a sound field image. A slice of CASS material was then removed from the specimen end and the beam mapping exercise repeated. The sound fields acquired were analyzed for spot size, coherency, and beam redirection. Analyses were conducted between the resulting sound fields and the microstructural characteristics of each specimen.

Ultrasonic, Phased Array, NDE, Austenitic Steel, B↗

Assessment of three dead detector correction methods for cone-beam computed tomography

Purpose: Dead detectors due to manufacturing defects or radiation damage in the electronic portal imaging devices (EPIDs) used for cone-beam computed tomography (CBCT) can lead to image degradation and ring artifacts. In this work three dead detector correction methods were assessed using megavoltage CBCT (MVCBCT) as a test system, with the goals of assessing the relative effectiveness of the three methods and establishing the conditions for which they fail. Methods: MVCBCT projections acquired with four linacs at 8 and 60 MU (monitor units) were degraded with varying percentages (2%-95%) of randomly distributed dead single detectors (RDSs), randomly distributed dead detector clusters (RDCs) of 2 mm diameter, and nonrandomly distributed dead detector disks (NRDDs) of varying diameter (4-16 mm). Correction algorithms were bidirectional linear interpolation (BLI), quad-directional linear interpolation (QLI), and a Laplacian solution (LS) method. Correction method failure was defined to occur if ring artifacts were present in the reconstructed phantom images from any linac or if the modulation transfer function (MTF) for any linac dropped below baseline with a p value, calculated with the two sample t test, of less than 0.01. Results: All correction methods failed at the same or lower RDC/RDS percentages and NRDD diameters for the 60 MU as for the 8 MU cases. The LS method tended to outperform or match the BLI and QLI methods. If ring artifacts anywhere in the images were considered unacceptable, the LS method failed for 60 MU at >33% RDS, >2% RDC, and >4 mm NRDD. If ring artifacts within 4 mm longitudinally of the phantom section interfaces were considered acceptable, the LS method failed for 60 MU at >90% RDS, >80% RDC, and >4 mm NRDD. LS failed due to MTF drop for 60 MU at >50% RDS, >25% RDC, and >4 mm NRDD. Conclusions: The LS method is superior to the BLI and QLI methods, and correction algorithm effectiveness decreases as imaging dose increases. All correction methods failed first due to ring artifacts and second due to MTF drop. If ring artifacts in axial slices within a 4 mm longitudinal distance from phantom section interfaces are acceptable, statistically significant loss in spatial resolution does not occur until over 25% of the EPID is covered in randomly distributed dead detectors, or NRDDs of 4 mm diameter are present.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Coronary centerline extraction from CT coronary angiography images using a minimum cost path approach

Purpose: The application and large-scale evaluation of minimum cost path approaches for coronary centerline extraction from computed tomography coronary angiography (CTCA) data and the development and evaluation of a novel method to reduce the user-interaction time. Methods: A semiautomatic method based on a minimum cost path approach is evaluated for two different cost functions. The first cost function is based on a frequently used vesselness measure and intensity information, and the second is a recently proposed cost function based on region statistics. User interaction is minimized to one or two mouse clicks distally in the coronary artery. The starting point for the minimum cost path search is automatically determined using a newly developed method that finds a point in the center of the aorta in one of the axial slices. This step ensures that all computationally expensive parts of the algorithm can be precomputed. Results: The performance of the aorta localization procedure was demonstrated by a success rate of 100% in 75 images. The success rate and accuracy of centerline extraction was quantitatively evaluated on 48 coronary arteries in 12 images by comparing extracted centerlines with a manually annotated reference standard. The method was able to extract 88% and 47% of the vessel centerlines correctly using the vesselness/intensity and region statistics cost function, respectively. For only the proximal part of the vessels these values were 97% and 86%, respectively. Accuracy of centerline extraction, defined as the average distance from correctly automatically extracted parts of the centerline to the reference standard, was 0.64 mm for the vesselness/intensity and 0.51 mm for the region statistics cost function. The interobserver variability was 99% for the success rate measure and 0.42 mm for the accuracy measure. Qualitative evaluation using the best performing cost function resulted in successful centerline extraction for 233 out of the 252 coronaries (92%) in 63 additional CTCA images. Conclusions: The presented results, in combination with minimal user interaction and low computation time, show that minimum cost path approaches can effectively be applied as a preprocessing step for subsequent analysis in clinical practice and biomedical research.

62 RADIOLOGY AND NUCLEAR MEDICINE↗