Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “tensor networks”

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.

390 records · Page 22
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Microseismic Monitoring of Laboratory Hydraulic Fracturing Experiments in Granitic Rocks for Different Fracture Propagation Regimes

While hydraulic fracturing (HF) is a widely employed process, the underlying fracturing processes are still heavily contested. The attributes of the HF generated fracture network can exhibit substantial variation when dealing with specific HF propagation regimes encountered in the field. In this study, HF experiments were performed on true-triaxially loaded Barre granite cubes, with microseismic monitoring, to identify and characterize the fracturing mechanisms associated with different viscosity injection fluids. Utilizing fluids with high (oil/1450 cP) and low (water/1 cP) viscosity represented two key HF propagation regimes: viscosity- and toughness-dominated. The experiments conducted with oil involved higher breakdown pressures, larger fluid volumes, and slower fracture propagation speeds. Furthermore, the frequency–magnitude distribution (b value) for all experiments (1.9–2.3) was similar to those encountered for large-scale operations. Slightly larger b values were encountered during the initiation phase (2.4–2.7) relative to the fracture propagation and post-fracturing phases (1.9–2.2). Techniques such as polarity and moment-tensor inversion were utilized to characterize the source mechanisms. For the HF experiments with oil, tensile fractures were most dominant (92%) in the initiation phase compared to fracture propagation and post-fracturing phases (70–75%). Similar tensile fracturing dominancy was not observed with water, attributable to fluid permeation and leak-off. Regardless of the injection fluid or classification criteria employed, tensile fractures were the dominant type consistently, with fewer occurring in water experiments but the specific ratio of crack types varied with different source mechanism criteria employed.

58 GEOSCIENCES↗
DOE OSTI2025

Enhancing Short-Range Weather Forecasts through Temporal Variation Encoding: A Multiperiod Embedding Approach

Machine learning (ML) techniques have emerged as promising approaches to improve regional weather forecast accuracy and reliability through data-driven methods. We propose a novel ML-based weather forecasting model, the Multiperiod Embed Net (MPENet). A key distinguishing feature of MPENet is its explicit utilization of the inherent cyclic nature in weather dynamics, unlike the autoregressive strategies commonly used in other ML weather forecasting approaches. Critical cyclic structures are identified via Fourier analyses of dynamic time series. Cyclicity in the convolutional representation is achieved by transforming one-dimensional time series of meteorological variables into two-dimensional tensors based on identified periods. This approach enables the model to leverage intrinsic weather patterns, enhancing regional forecast performance. To demonstrate the effectiveness of MPENet, we conduct a comparative analysis with Nvidia’s FourCastNet. Both models are trained on High-Resolution Rapid Refresh (HRRR) data from 2015 to 2022, over a 192 km × 192 km region in Tennessee. The comparisons are performed locally at two specific locations known to have different weather dynamics due to orographic effects: Crossville, on the relatively flat Cumberland Plateau with fewer topographic airflow disruptions, and Oak Ridge, in the ridge-and-valley region, where airflow is heavily influenced by surrounding valleys and mountains. Our results indicate that FourCastNet achieves strong accuracy at very short lead times, while MPENet maintains competitive skill and shows advantages in capturing temporal evolution over longer periods. Cross-correlation analyses of MPENet and FourCastNet predictions with the HRRR data suggest that encoding critical cyclicity into the network architecture leads to improvements in the forecasting skill.

Artificial intelligence↗
DOE OSTI2022

H-GCN: A Graph Convolutional Network Accelerator on Versal ACAP Architecture

Recently Graph Neural Networks (GNNs) have drawn tremendous attentions due to their unique capability to extend the Machine Learning (ML) approaches to broadly defined applications with unstructured data, especially graphs. Comparing with other ML modalities, the acceleration of GNNs is as critical but even more challenging due to the irregularity and heterogeneity from graph typologies that together limit the performance. Existing efforts mainly focus on handling graphs’ irregularity, however, have not studied the heterogeneity. To this end, in this work, we propose H-GCN, a PL-AIE-based hybrid accelerator that leverages the emerging heterogeneity of Xilinx Versal ACAPs to achieve high-performance GNN inference. In particular, H-GCN partitions each graph into three subgraphs based on its inherent heterogeneity and processes them using PL and the newly emerged AIE respectively. To further improve the performance, we explore the sparsity support of AIE and develop an efficient density-aware method to map tiles of SpMM onto the systolic tensor array automatically. Compared with the current state-of-the-art GCN accelerator, HGCN achieves on average 1.5× speedups.

Zhang, Chengming↗
DOE OSTI2022

Effect of localization on photoluminescence and zero-field splitting of silicon color centers

The study of defect centers in silicon has been recently reinvigorated by their potential applications in optical quantum information processing. A number of silicon defect centers emit single photons in the telecommunication O-band, making them promising building blocks for quantum networks between computing nodes. The two-carbon G-center, self-interstitial W-center, and spin-1/2 T-center are the most intensively studied silicon defect centers, yet despite this, there is no consensus on the precise configurations of defect atoms in these centers, and their electronic structures remain ambiguous. Here, in this work, we employ ab initio density functional theory to characterize these defect centers, providing insight into the relaxed structures, band structures, and photoluminescence spectra, which are compared to experimental results. Motivation is provided for how these properties are intimately related to the localization of electronic states in the defect centers. In particular, we present the calculation of the zero-field splitting for the excited triplet state of the G-center defect as the structure is linearly interpolated from the A-configuration to the B-configuration, showing a sudden increase in the magnitude of the D zz component of the zero-field-splitting tensor. By performing projections onto the local orbital states of the defect, we analyze this transition in terms of the symmetry and bonding character of the G-center defect, which sheds light on its potential application as a spin-photon interface.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2024

The use of digital thread for reconstruction of local fiber orientation in a compression molded pin bracket via deep learning

A deep convolutional neural network (DCNN) was used for microstructure reconstruction using artificial intelligence (MR-AI) by predicting local average fiber orientation distributions (FOD) in a 3D prepreg platelet molded composite (PPMC) pin bracket. To train the MR-AI model, surface strain fields from residual stresses simulated in PPMC plates were used as the input to the DCNN. A training dataset included PPMC plates with various degrees of global fiber alignment, based on the information obtained from high-fidelity flow simulation of a pin bracket. Further, the MR-AI model was then deployed to analyze FOD in the 3D pin bracket by conducting thermo-elastic residual stress analysis. Initially, the MR-AI model was established entirely on the synthetic simulation data. Then, a μCT scan of a physically molded pin bracket was used to create a finite element model that provided data for additional validation of the DCNN model. For the μCT scan finite element pin bracket the MR-AI model predicted the distribution of fiber orientation tensor components with MAE of 0.10 indicating a global prediction error of 10%. For the flow simulated pin bracket, the MR-AI model predicted the distribution of fiber orientation tensor components with a global prediction error of 11%. The MR-AI model showed the ability to predict regions of varying alignment in the base and flange of the pin bracket. The proposed MR-AI methodology allows for rapid prediction of FOD in geometrically complex parts and offers a promising path to detecting unique fiber orientation states in molded components.

42 ENGINEERING↗
DOE OSTI2023

Analysis of the 2021 Milford, Utah earthquake swarm: Enhanced earthquake catalog and migration patterns

An earthquake swarm occurred in Spring 2021 in south-central Utah near the town of Milford. The University of Utah Seismograph Stations located 125 earthquakes between March 19 and May 10 with magnitudes ranging from 0.5 to 3.2. We implement a matched-filter technique in order to identify additional earthquakes that went undetected during the routine network location. The 125 network-located earthquakes are used as templates and are cross-correlated with continuous data for the dates Feb 17—June 10. This time period corresponds to approximately 1 month before the earthquake swarm began through 1 month after it ended. For the matched-filter analysis, we rely heavily on station FOR1, which is located within 5 km of most template events. Four other stations within 20–30 km of most template events provide a supplement to the closest station. The matched-filter implementation results in the detection of over 600 earthquakes in addition to the original 125 catalog events. This is one of the largest swarms ever recorded in Utah, and no previous large swarms have been recorded in this location. We use HypoDD to obtain relative double-difference locations of the catalog events. Both routine locations and HypoDD relocations of the catalog events suggest a fault dipping west beneath the mountain range, opposite of typical Basin and Range normal faults that dip beneath alluvial valleys. Moment tensors for the largest five events show normal faulting consistent with the west-dipping fault seen in the seismicity. Hydrothermal features in the area, including a geothermal power plant, suggest that fluids may be a contributing factor to the earthquake swarm triggering. We examine the role of fluids by exploring bounds on diffusion parameters and investigating spatial migration characteristics of the swarm seismicity. We conclude that this swarm is the result of heterogenous stress conditions in a prefractured region.

58 GEOSCIENCES↗
DOE OSTI2025

Three-Receiver Quantum Broadcast Channels: Classical Communication with Quantum Non-unique Decoding

In network communication, it is common in broadcasting scenarios for there to exist a hierarchy among receivers based on information they decode due, for example, to different physical conditions or premium subscriptions. This hierarchy may result in varied information quality, such as higher-quality video for certain receivers. This is modeled mathematically as a degraded message set, indicating a hierarchy between messages to be decoded by different receivers, where the default quality corresponds to a common message intended for all receivers, a higher quality is represented by a message for a smaller subset of receivers, and so forth. We extend these considerations to quantum communication, exploring three-receiver quantum broadcast channels with two- and three-degraded message sets. Our technical tool involves employing quantum non-unique decoding, a technique we develop by utilizing the simultaneous pinching method. Here, we construct one-shot codes for various scenarios and find achievable rate regions relying on various quantum Rényi mutual information error exponents. Our investigation includes a comprehensive study of pinching across tensor product spaces, presenting our findings as the asymptotic counterpart to our one-shot codes. By employing the non-unique decoding, we also establish a simpler proof to Marton’s inner bound for two-receiver quantum broadcast channels without the need for more involved techniques. Additionally, we derive no-go results and demonstrate their tightness in special cases.

Salek, Farzin [Technical University of Munich (Ger↗
DOE OSTI2024

HydraGNN_Predictive_GFM_2024 - Ensemble of predictive graph foundation models for ground state atomistic materials modeling

We provide the ensemble of fifteen pre-trained graph foundation models (GFMs) for atomistic materials modeling applications. Each one of the fifteen GFMs has been trained on five open-source datasets that (once aggregated) amount to over 154 million atomistic structures, which cover over two-thirds of the natural elements of the periodic table and that comprises a broad set of organic and inorganic compounds. This vast set of atomistic structures comprises ground state configurations that are dynamically stable (i.e., equilibrated structures with atomic forces approximately close to zero values) as well as dynamically unstable structures (i.e., non-equilibrium structures with non-negligible non-zero values of atomic forces). The ensemble of datasets aggregated does NOT include excited states. The datasets have been curated to remove atomistic structures with spectral norm of the force tensor above 100 eV/angstrom. Moreover, a linear term of the energy was computed for each dataset using a linear regression model that uses the chemical concentration of each natural element as regressor. The linear term predicted by the linear regression model has been subtracted from each original energy value to perform a re-alignment of the energy values across different electronic structures approximation theories performed to generate the diverse multi-source, multi-fidelity datasets. The folder "ADIOS_files" contains the set of pre-processed datasets in Adaptable I/O System (ADIOS) format (https://www.exascaleproject.org/research-project/adios/) that have been used for the development and training of GFMs in this work. The "ADIOS_files" directory contains 6 sub-directories named as follows: - ANI1x-v3.bp - MPTrj-v3.bp - OC2020-20M-v3.bp - OC2020-v3.bp - OC2022-v3.bp - qm7x-v3.bp Each sub-directory contains the pre-processed datasets converted in Adaptable I/O System (ADIOS) format (https://www.exascaleproject.org/research-project/adios/) that have been used to the development, training, and performance testing of the ensemble go predictive graph foundation models. Each GFM was developed using HydraGNN (https://github.com/ORNL/HydraGNN) as underlying graph neural network (GNN) architecture. The multi-task learning (MTL) capability of HydraGNN was used to simultaneously train the GFMs on labeled values for direct predictions of energy (a total system property of an atomistic structure that measures the chemical stability) and atomic forces (an atomic level property of an atomistic structure that measures the dynamical stability). The hyper parameters of the GFM have been tuned using scalable hyperparameter optimization (HPO) algorithms implemented in the software DeepHyper (https://github.com/deephyper/deephyper). The pre-training of each HPO trial was performed using distributed data parallelism (DDP) to scale the training across 128 compute nodes of the exascale OLCF supercomputer Frontier. Each HPO trial was trained only for 10 epochs and an early stopping was performed to avoid wasting significant computational resources on GNN architectures that were clearly underperforming. For each HPO trial, the 'omnistat' tool developed by (AMD Research - Advanced Micro Device) was used to measure the total energy consumption in kWh. The ensemble of GFMs was obtained by selecting the fifteen best performing HPO trials. Four models have been selected for their clear advantage in accuracy, and these are the GFMs with IDs 229, 156, 147, 260. Additional eleven models have been selected based on judicious balance between accuracy and energy consumption needed for training, and these are the GFMs with IDs 165, 78, 137, 1, 175, 171, 181, 67, 179, 167, 351. Each selected GFM of the ensemble was continued to cumulate a total of at most 30 epochs. In some cases, the total number of epochs actually performed was les than 30 due to two combined factors: (1) the size of the GFM (i.e., the number of model parameters to train) and (2) the total wall-clock time for which the computational resources could be allocated on OLCF-Frontier. The "Ensemble_of_models" directory contains 15 sub-directories named as follows: - gfm_0.229 - gfm_0.156 - gfm_0.147 - gfm_0.260 - gfm_0.165 - gfm_0.78 - gfm_0.137 - gfm_0.1 - gfm_0.175 - gfm_0.171 - gfm_0.181 - gfm_0.67 - gfm_0.179 - gfm_0.167 - gfm_0.351 Each one of these sub-directories refers to one of the fifteen HPO trials that have been selected to continue the pre-training with at most 30 epochs. With each sub-directory associated with a specific HPO trial, the following files can be found: - config.json: file for argument parsing to develop and train an HydraGNN architecture - gfm_0.ID_epoch_N.pk: file with model parameters for HPO ID trial after N epochs of training The ensemble of fifteen GFM architectures was used for (1) ensemble averaging to stabilize the predictions of energy and atomic forces after pre-training for post-processing analysis and (2) ensemble uncertainty quantification (UQ). The code used to develop, pre-train, and load the pre-trained models for post-processing analysis is available on the ORNL-GitHub at the following link: https://github.com/ORNL/HydraGNN/tree/Predictive_GFM_2024

36 MATERIALS SCIENCE↗
DOE OSTI2024

Birefringence tests of gravity with multimessenger binaries

Extensions to General Relativity (GR) allow the polarization of gravitational waves (GW) from astrophysical sources to suffer from amplitude and velocity birefringence, which respectively induce changes in the ellipticity and orientation of the polarization tensor. Here, we introduce a multi-messenger approach to test this polarization behavior of GWs during their cosmological propagation using binary sources, for which the initial polarization is determined by the inclination and orientation angles of the orbital angular momentum vector with respect to the line of sight. In particular, we use spatially resolved radio imaging of the jet from a binary neutron star (BNS) merger to constrain the orientation angle and hence the emitted polarization orientation of the GW signal at the site of the merger, and compare to that observed on Earth by GW detectors. For GW170817, using past measurements of the inclination angle, we constrain the deviation from GR due to amplitude birefringence to κ A = $-0.12^{+0.60}_ {-0.61}$, while the velocity birefringence parameter κ V remains unconstrained. The inability to constrain κ V is due to the low amplitude of GW170817 in the Virgo detector, and measurements of the polarization orientation require information from a combination of multiple detectors with different alignments. For this reason, we also mock future BNS mergers with resolved afterglow proper motion and project that κ V could be constrained to a precision of 5 rad (corresponding to an angular shift of the GW polarization of δ φV ≈ 0.2 rad for a BNS at 100 Mpc) by a future network of third-generation ground-based GW detectors such as Cosmic Explorer and the radio High Sensitivity Array. Crucially, this velocity birefringence effect cannot be constrained with dark binary mergers as it requires polarization information at the emission time, which can be provided only by electromagnetic emission.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS1995

Nearly sonic and transsonic convective motions in the solar atmosphere related to the solar wind origin

MHD equations are considered for the solar atmosphere. 15 different simplest MHD regimes are indicated for the momentum transport equation depending on the mutual binary interplay between 6 possible and locally dominant terms: non-stationarity and inhomogeneity of the flow, gas pressure, magnetic tensions, viscous and gravity forces. These regimes are delimited by five physically independent dimensionless parameters, for example, Strouhal, sonic Mach, alfvenic Mach, Reynolds and Froude numbers or their combinations. Another partially overlapping classification of the simplest regimes may be introduced based on the energy conservation equation. There are also 15 independent binary combinations between nonstationary and inhomogeneous convective terms, dissipative energy sinks and sources (viscous, heat-conductive, Joule and radiative ones) in the energy equation. More complicated regimes are considered with multiple dominated terms. All these MHD regimes play their important role somewhere in the solar atmosphere complicated by the tensor transport coefficients in the magnetically dominated regions of the upper atmosphere. Nearly sonic and transsonic nonstationary convective motions with ascending and descending flows are observed in the solar chromosphere. the transition region and the lower pans of the solar corona together with related horizontal velocity components. This convection represents a kind of the 'cocoonery' manufacturing nonstationary vortices generated here and partially connected to the photosphere and to the solar wind. The solar wind originates from this powerful transsonic muddle in the solar atmosphere as a tiny fraction of the streamlines which are temporarily getting detached from the 'cocoons; and going to the infinity. The topologically complicated instantaneous 'runaway surface' around the Sun, i.e., the surface which separates outgoing to the infinity streams from other finite flows in the solar atmosphere was not described in the literature and needs additional investigation. We conclude that a simple one-connected smooth and quasistationary 'critical surface' which is often supposed to be placed somewhere in the solar corona (say, at 4 solar radii) in reality does not exist. Observational and theoretical arguments favor instead of this a highly structured, disordered, patched and time-dependent sonic transition surface in the solar atmosphere permanently fluctuating at positional dependent heights starting sometimes from photospheric (and maybe subphotospheric) levels and more frequently from the chromosphenc network up to several solar radii in the solar corona.

Veselovsky, I. S.↗
DOE OSTI2021

Toward Robust and Routine Determination of Mw for Small Earthquakes: Application to the 2020 Mw 5.7 Magna, Utah, Seismic Sequence

To better characterize seismic hazard, particularly, for induced seismicity, there is an increasing interest in methods to estimate moment magnitude (M w ) for small earthquakes. M w is generally preferred over other magnitude types, but, it is difficult to estimate M w for earthquakes with local magnitude (M L ) <3-3.5, using conventional moment tensor (MT) inversion. The 2020 M ww 5.7 Magna, Utah, seismic sequence provides an opportunity to illustrate and evaluate the value of spectral methods for this purpose. Starting with a high-quality seismic catalog of 2103 earthquakes (M L <5.6), we estimate M w using two independent spectral methods—one based on direct waves, yielding M w,direct , and the other based on coda waves, yielding M w,coda . For the direct-wave method, we present a non-parametric (NP) inversion scheme that solves for apparent geometrical spreading, G(R), and site effects (S), similar to other NP procedures that have been used to calibrate regional M L scales. The NP inversion is constrained using M ws derived from MTs for nine events in the Magna sequence. We recover statistically robust and physically reasonable G(R) and S and compute M w,direct for 635 Magna earthquakes down to M L 0.7. For the coda-wave method, we consider two separate calibration schemes involving previous MT solutions and compute M w,coda for 311 earthquakes down to M L 1.0. For 280 of the events that were processed with both methods—M w,direct and M w,coda —are strongly correlated (r = 0.98), with a mean difference of only 0.05. We compare M w,direct and M w,coda with M L and find reasonably good agreement for M L <3.6 with the theoretically predicted relationship of M w =(2/3)M L +C, in which C is a regional constant. Our results imply that seismic network operators can use spectral-based M w estimates to replace M L estimates for events with M L ≥1.0, and possibly smaller. The main requirement is the existence of a small number of MT solutions for calibration purposes.

58 GEOSCIENCES↗
DOE OSTI2022

Automatic Generation of High-Performance Convolution Kernels on ARM CPUs for Deep Learning

In this work, we present FastConv, a template-based code auto-generation open source library that can automatically generate high-performance deep learning convolution kernels of arbitrary matrices/tensors shapes. FastConv is based on the Winograd algorithm, which is reportedly the highest performing algorithm for the time-consuming convolution layers of convolutional neural networks. ARM CPUs cover a wide range designs and specifications, from embedded devices to HPC-grade CPUs. The leads to the dilemma of how to consistently optimize Winograd-based convolution solvers for convolution layers of different shapes. FastConv addresses this problem by using templates to auto-generate multiple shapes of tuned kernels variants suitable for skinny tall matrices. As a performance portable library, FastConv transparently searches for the best combination of kernel shapes, cache tiles, scheduling of loop orders, packing strategies, access patterns, and online/offline computations. Auto-tuning is used to search the parameter configuration space for the best performance for a given target architecture and problem size. The experiments with layer-wise evaluation on the VGG--16 model confirms a 1.25x performance gains is got by tuning the Winograd library. Integrated comparison results shows 1.02x to 1.40x, 1.14x to 2.17x, and 1.22x and 2.48x speedup is achieved over NNPACK, Arm NN, and FeatherCNN on the Kunpeng 920 beside few cases. Furthermore, problem size performance portability experiments with various convolution shapes shows that FastConv achieves 1.2x to 1.7x speedup and 2x to 22x speedup over NNPACK and ARM NN inference engine using Winograd on Kunpeng 920 . CPU performance portability evaluation on the VGG--16 show an average speedup over NNPACK of 1.42x, 1.21x, 1.26x, 1.37x, 2.26x, and 11.02x is observed on Kunpeng 920, Snapdragon 835, 855, 888, Apple M1, and AWS Graviton2, respectively.

97 MATHEMATICS AND COMPUTING↗