Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “ITERATIVE NETWORK”

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 541 records · Page 30

General Line Ampacity State Solver

INL’s General Line Ampacity State Solver (GLASS) is a high performance software platform that enables advanced grid system operations and planning efforts to be computed in real time. It enables the combination of advanced sensors and sensor networks, communication and data transfer channels, secure and reliable data management, and real-time processing through a flexible system architecture, with the ability to support and integrate with any utility enterprise system. GLASS is designed to help resolve the problems associated with thermally limited transmission and distribution networks, namely bare overhead power transmission lines. GLASS enables the methodology of Concurrent Cooling (Dynamic Line Rating) by reducing the need of using a cost-prohibited number of weather stations (reduce the number of weather stations per mile of overhead power line) to achieve the resolution necessary for individual electric transmission line span capacity calculations. GLASS achieves this by coupling real-time grid operations data and infrastructure architecture information with real-time field deployed weather station measurements and pre-computed Computational Fluid Dynamics (CFD)-based wind modeling interpolation parameters (look-up tables). GLASS uses mathematical models based on the IEEE 738 standard and builds off of the provided iterative solution using modern software paradigms and programming styles to provide a more robust and modular software system. GLASS also interfaces in real-time with utility companies using known transmission line spans and structures. GLASS provides real-time ampacity calculations based on recent weather station data as well as transmission line specifications. GLASS also allows the user to interact graphically with internal databases, which include information such as transmission line conductor types and thermal rating limits, weather station names and locations, specific transmission line span and structure locations, as well as recent ampacity calculations.

Gentle, Jake↗

LLNL-G3Dv3: Global P wave tomography model for improved regional and teleseismic travel time prediction

[1] We develop a global-scale P wave velocity model (LLNL-G3Dv3) designed to accurately predict seismic travel times at regional and teleseismic distances simultaneously. The model provides a new image of Earth's interior, but the underlying practical purpose of the model is to provide enhanced seismic event location capabilities. The LLNL-G3Dv3 model is based on ~2.8 millionP and Pnarrivals that are re-processed using our global multiple-event locator called Bayesloc. We construct LLNL-G3Dv3 within a spherical tessellation based framework, allowing for explicit representation of undulating and discontinuous layers including the crust and transition zone layers. Using a multiscale inversion technique, regional trends as well as fine details are captured where the data allow. LLNL-G3Dv3 exhibits large-scale structures including cratons and superplumes as well numerous complex details in the upper mantle including within the transition zone. Particularly, the model reveals new details of a vast network of subducted slabs trapped within the transition beneath much of Eurasia, including beneath the Tibetan Plateau. We demonstrate the impact of Bayesloc multiple-event location on the resulting tomographic images through comparison with images produced without the benefit of multiple-event constraints (single-event locations). We find that the multiple-event locations allow for better reconciliation of the large set of direct P phases recorded at 0–97° distance and yield a smoother and more continuous image relative to the single-event locations. Travel times predicted from a 3-D model are also found to be strongly influenced by the initial locations of the input data, even when an iterative inversion/relocation technique is employed.

58 GEOSCIENCES↗

Constrained approximation of effective generators for multiscale stochastic reaction networks and application to conditioned path sampling

Efficient analysis and simulation of multiscale stochastic systems of chemical kinetics is an ongoing area for research, and is the source of many theoretical and computational challenges. In this paper, we present a significant improvement to the constrained approach, which is a method for computing effective dynamics of slowly changing quantities in these systems, but which does not rely on the quasi-steady-state assumption (QSSA). The QSSA can cause errors in the estimation of effective dynamics for systems where the difference in timescales between the “fast” and “slow” variables is not so pronounced. This new application of the constrained approach allows us to compute the effective generator of the slow variables, without the need for expensive stochastic simulations. This is achieved by finding the null space of the generator of the constrained system. For complex systems where this is not possible, or where the constrained subsystem is itself multiscale, the constrained approach can then be applied iteratively. This results in breaking the problem down into finding the solutions to many small eigenvalue problems, which can be efficiently solved using standard methods. Since this methodology does not rely on the quasi steady-state assumption, the effective dynamics that are approximated are highly accurate, and in the case of systems with only monomolecular reactions, are exact. We will demonstrate this with some numerics, and also use the effective generators to sample paths of the slow variables which are conditioned on their endpoints, a task which would be computationally intractable for the generator of the full system.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Active deep kernel learning of molecular properties from structural embeddings

As vast databases of chemical identities become increasingly available, the challenge shifts to how we effectively explore and leverage these resources to study molecular properties. This paper presents an active learning approach for molecular discovery using deep kernel learning (DKL), demonstrated on the QM9 dataset. DKL links structural embeddings directly to properties, creating organized latent spaces that prioritize relevant property information. By iteratively recalculating embedding vectors in alignment with target properties, DKL uncovers concentrated maxima representing key molecular properties and reveals unexplored regions with potential for innovation. This approach underscores DKL’s potential in advancing molecular research and discovery.

Artificial neural networks↗

Fast meta-solvers for 3D complex-shape scatterers using neural operators trained on a non-scattering problem

Three-dimensional target identification using scattering techniques requires high accuracy solutions and very fast computations for real-time predictions in some critical applications. We first train a deep neural operator (DeepONet) to solve wave propagation problems described by the Helmholtz equation in a domain without scatterers but at different wavenumbers and with a complex absorbing boundary condition. We then design two classes of fast meta-solvers by combining DeepONet with either relaxation methods, such as Jacobi and Gauss-Seidel, or with Krylov methods, such as GMRES and BiCGStab, using the trunk basis of DeepONet as a coarse-scale preconditioner. We leverage the spectral bias of neural networks to account for the lower part of the spectrum in the error distribution while the upper part is handled inexpensively using relaxation methods or fine-scale preconditioners. The meta-solvers are then applied to solve scattering problems with different shape of scatterers, at no extra training cost. We first demonstrate that the resulting meta-solvers are shape-agnostic, fast, and robust, whereas the standard standalone solvers may even fail to converge without the DeepONet. We then apply both classes of meta-solvers to scattering from a submarine, a complex three-dimensional problem. We achieve very fast solutions, especially with the DeepONet-Krylov methods, which require orders of magnitude fewer iterations than any of the standalone solvers.

97 MATHEMATICS AND COMPUTING↗

Identification of Faults Susceptible to Induced Seismicity (Final Report)

Central to the work documented in this report is the capability of geocellular models to represent the geologic conceptual model updated with fault identification from machine learning and joint inversion modeling of microseismic data measured and recorded as a consequence of CO 2 injection at a field demonstration site: the Illinois Basin - Decatur Project (IBDP). This work required seven unique geocellular models with 100s of simulated variations to gain a very high degree of confidence in the identification of geologic features present that contributed to induced microseismicity at IBDP. All forward modeling: pressure modeling, stress modeling, and seismic modeling used the same geologic conceptual model and representations of that model at different scales. The pressure modeling and poroelastic modeling created “snapshots” of pore pressure and stress field changes at different times during CO 2 injection, in which microseismic events were clustered (in time). These pressure and stress snapshots, within the framework and architecture of the geologic conceptual model via the geocellular model, informed the single fault and fault network models to ascertain the likelihood of fault movement (seismic or aseismic). The outcomes of the pressure, stress, and fault/fault network (seismic) modeling confirmed that the faults in the geologic conceptual model in Task 2 were likely the source of microseismic events measured at IBDP and acted as conduits for pressure to be transmitted from the injection interval into the Precambrian crystalline basement rock. This closely coordinated and integrated unique modeling approach was conducted to prove the viability of our proposed workflow 1) to better resolve crystalline basement faults, 2) detect subseismic faults that could be activated by injection, 3) increase the certainty in fault detection and their susceptibility to release seismic energy, and 4) understand transmission of pressure vertically from the well to the underlying fractured crystalline basement. The proposed methodology was effective in guiding an iterative process of calibrating forward modeling results based on similar geocellular models while honoring the geologic conceptual model (i.e., characterization data and knowledge of regional geology); this led to higher level of certainty in the identification of fault/faults zones to control seismicity and transmission of pressure to the regions of recorded and located injection induced seismicity.

58 GEOSCIENCES↗

A learning-based data-driven forecast approach for predicting future reservoir performance

Quantification of the predictive uncertainty of subsurface models has long been investigated. The traditional workflow is to calibrate prior models to match observed data, and then use the posterior models to simulate future system performance. Not only are these procedures computationally expensive, but they also have issues in maintaining geological model constraints during the calibration step. Data space inversion (DSI) was introduced recently to predict future system performance without the iterative history matching or model calibration step. In general, DSI approaches seek to establish a statistical relationship between the observed and forecast variables, as well as to quantify the predictive uncertainty of the forecast variables, by using an ensemble of uncalibrated prior models. Existing DSI approaches all require a number of complex transformation and mapping operations, which may deter their widespread use. Here in this study, we introduce a new and simpler DSI approach, the learning-based, data-driven forecast approach (LDFA), by combining dimension reduction and machine learning techniques to quickly provide accurate forecast results and reliably quantify corresponding uncertainty in the results. Our LDFA framework is demonstrated using two supervised learning algorithms, artificial neural network (ANN) and support vector regression (SVR), on two representative examples from reservoir engineering and geological carbon storage. Results suggest that our approach provides accurate forecast results (e.g., future oil production rate or cumulative injected CO 2 ) and reasonable predictive uncertainty intervals. Our framework is generic and may be applied to other surface and subsurface problems.

42 ENGINEERING↗

A data-driven approach to real-time vertical position estimation for NSTX-U vertical stability control

In this paper, a database of 77 996 plasma equilibrium reconstructions from 727 discharges during the initial operation of the NSTX-U spherical tokamak is analyzed to develop a statistically robust model of the plasma vertical position for real-time control. A variety of regression models are developed and tested, ranging in complexity from linear models to deep neural networks, and including input signals ranging from the four pairs of flux loops used historically on NSTX-U up to the full set of 389 real-time signals available to the plasma control system. A linear model based on 140 real-time magnetics signals is found to offer excellent accuracy, with a coefficient of determination R 2 = 0.906. The robustness of this model to limited training data, new operating scenarios, and signal errors is tested, and a procedure is demonstrated to tune the model parameters to optimize its robustness. A time-dependent plasma equilibrium solver, TokaMaker, is used to simulate vertical stability control in NSTX-U, demonstrating that it should be possible to iteratively tune the parameters of a linear vertical position model to stabilize both positive and negative triangularity plasmas in future experiments.

magnetic diagnostics↗

SODAs: sparse optimization for the discovery of differential and algebraic equations

Differential-algebraic equations (DAEs) integrate ordinary differential equations (ODEs) with algebraic constraints, providing a fundamental framework for developing models of dynamical systems characterized by time-scale separation, conservation laws and physical constraints. While sparse optimization has revolutionized model development by allowing data-driven discovery of parsimonious models from a library of possible equations, existing approaches for dynamical systems assume DAEs can be reduced to ODEs by eliminating variables before model discovery. This assumption limits the applicability of such methods for DAE systems with unknown constraints and time scales. We introduce sparse optimization for differential-algebraic systems (SODAs), a data-driven method for the identification of DAEs in their explicit form. By discovering the algebraic and dynamic components sequentially without prior identification of the algebraic variables, this approach leads to a sequence of convex optimization problems. It has the advantage of discovering interpretable models that preserve the structure of the underlying physical system. To this end, SODAs improves since SODAs is singular numerical stability when handling high correlations between library terms, caused by near-perfect algebraic relationships, by iteratively refining the conditioning of the candidate library. We demonstrate the performance of our method on biological, mechanical and electrical systems, showcasing its robustness to noise in both simulated time series and real-time experimental data.

DAE↗

Graph processing platforms at scale: practices and experiences

Graph analysis unveils hidden associations of data in many phenomena and artifacts, such as road network, social networks, genomic information, and scientific collaboration. Unfortunately, a wide diversity in the characteristics of graphs and graph operations make it challenging to find a right combination of tools and implementation of algorithms to discover desired knowledge from the target data set. This study presents an extensive empirical study of three representative graph processing platforms: Pegasus, GraphX, and Urika. Each system represents a combination of options in data model, processing paradigm, and infrastructure. We benchmarked each platform using three popular graph operations, degree distribution, connected components, and PageRank over a variety of real-world graphs. Our experiments show that each graph processing platform shows different strength, depending the type of graph operations. While Urika performs the best in non-iterative operations like degree distribution, GraphX outputforms iterative operations like connected components and PageRank. In addition, we discuss challenges to optimize the performance of each platform over large scale real world graphs.

Lim, Seung-Hwan↗

Precision measurement of relative permittivity of aluminum oxide for a high power resonant waveguide window with low return loss

An iterative method of estimating permittivity and window thickness with high precision is developed for half wavelength resonant waveguide windows for use in a lower hybrid launcher for DIII-D. The high Q factor of half-wavelength windows results in extreme sensitivity of return loss on small deviations in permittivity. Permittivity and loss tangent measurements of Morgan AL-995 99.5% aluminum oxide ceramic using a waveguide resonant cavity per ASTM D2520-13 determine variation of permittivity with frequency. Measurement of window resonant frequency in conjunction with the slope of permittivity from waveguide cavity measurements determine required thickness of a resonant window within 5 MHz of the desired resonant frequency. Methods to improve measurement accuracy of resonant peaks on a vector network analyzer are discussed. A sensitivity study to dimensional variations in window or sample rod geometry, and shift in resonant frequency due to RF heating of the window material is conducted.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

A thermodynamic model of integrated liquid-to-liquid thermoelectric heat pump systems

Thermoelectric (TE) heat pumps (TEHPs) are advantageous for heating and cooling in various applications because of their modularity and simple design. A TEHP system includes the TE modules with p- and n-type materials bonded to substrates, plus heat exchangers, thermal interfaces to the heat exchangers, and heat transfer fluids. Although modeling an individual TE module has been extensively studied, limited studies have reported performance at the larger system-level. Furthermore, no prior study has addressed the impact of temperature-dependent TE material properties (e.g., electric resistivity, thermal conductivity, and Seebeck coefficient) on overall heat-pump-system-level performance. Here, this work presents a mathematical model for TEHP system performance based on Goldsmid's approach for TE material performance, “effective” TE material properties, Gnielinski's correlation for convective heat transfer, and thermal balance theory for a heat exchange network. This combined approach provides an accurate model of the liquid-to-liquid TEHP system. Three different approaches—one empirical, one based on the manufacturer's specifications, and one drawn from the literature—were then used to determine values for TE material properties. The first two methods treated properties as constants, while the last approach treated properties as surface-temperature-based functions. Finally, experimental TEHP data was used to validate the models, all with relative absolute deviations of approximately 10% when predicting heating capacity and 10%–25% when forecasting cooling capacity up to a 30 K surface temperature lift. The results demonstrated that, at the TEHP system level, the TE material properties could be treated as constants, avoiding solver iterations and reducing the performance uncertainty by up to 95%.

42 ENGINEERING↗

The Transcription Unit Architecture of the Escherichia Coli Genome

Under EMSL User Proposal 25660, the authors reported that bacterial genomes are organized by structural and functional elements, including promoters, transcription start and termination sites, open reading frames, regulatory noncoding regions, untranslated regions and transcription units. Here, we iteratively integrate high-throughput, genome-wide measurements of RNA polymerase binding locations and mRNA transcript abundance, 5' sequences and translation into proteins to determine the organizational structure of the Escherichia coli K-12 MG1655 genome. Integration of the organizational elements provides an experimentally annotated transcription unit architecture, including alternative transcription start sites, 5' untranslated region, boundaries and open reading frames of each transcription unit. A total of 4,661 transcription units were identified, representing an increase of >530% over current knowledge. This comprehensive transcription unit architecture allows for the elucidation of condition-specific uses of alternative sigma factors at the genome scale. Furthermore, the transcription unit architecture provides a foundation on which to construct genome-scale transcriptional and translational regulatory networks.

59 BASIC BIOLOGICAL SCIENCES↗

Progressive transfer learning for low-frequency data prediction in full waveform inversion

To effectively overcome the cycle-skipping issue in full waveform inversion (FWI), we developed a deep neural network (DNN) approach to predict the absent low-frequency components by exploiting the hidden physical relation connecting the low- and the high-frequency data. To efficiently solve this challenging nonlinear regression problem, two novel strategies were proposed to design the DNN architecture and to optimize the learning process: (1) dual data feed structure; (2) progressive transfer learning. With the dual data feed structure, not only the high-frequency data, but also the corresponding beat tone data are fed into the DNN to relieve the burden of feature extraction. The second strategy, progressive transfer learning, enables us to train the DNN using a single evolving training dataset. Within the framework of the progressive transfer learning, the training dataset continuously evolves in an iterative manner by gradually retrieving the subsurface information through the physics-based inversion module, progressively enhancing the prediction accuracy of the DNN and propelling the inversion process out of the local minima. Here, the synthetic numerical experiments suggest that, without any a priori geological information, the low-frequency data predicted by the progressive transfer learning are sufficiently accurate for an FWI engine to produce reliable subsurface velocity models free of cycle-skipping artifacts.

02 PETROLEUM↗

Computational Workflow for Accelerated Molecular Design Using Quantum Chemical Simulations and Deep Learning Models

Efficient methods for searching the chemical space of molecular compounds are needed to automate and accelerate the design of new functional molecules such as pharmaceuticals. Given the high cost in both resources and time for experimental efforts, computational approaches play a key role in guiding the selection of promising molecules for further investigation. Here, we construct a workflow to accelerate design by combining approximate quantum chemical methods [i.e. density-functional tight-binding (DFTB)], a graph convolutional neural network (GCNN) surrogate model for chemical property prediction, and a masked language model (MLM) for molecule generation. Property data from the DFTB calculations are used to train the surrogate model; the surrogate model is used to score candidates generated by the MLM. The surrogate reduces computation time by orders of magnitude compared to the DFTB calculations, enabling an increased search of chemical space. Furthermore, the MLM generates a diverse set of chemical modifications based on pre-training from a large compound library. We utilize the workflow to search for near-infrared photoactive molecules by minimizing the predicted HOMO-LUMO gap as the target property. Our results show that the workflow can generate optimized molecules outside of the original training set, which suggests that iterations of the workflow could be useful for searching vast chemical spaces in a wide range of design problems.

Blanchard, Andrew↗

A genetically encoded, high-signal-to-noise maltose sensor

We describe the generation of a family of high-signal-to-noise single-wavelength genetically encoded indicators for maltose. This was achieved by insertion of circularly permuted fluorescent proteins into a bacterial periplasmic binding protein (PBP), Escherichia coli maltodextrin-binding protein, resulting in a four-color family of maltose indicators. The sensors were iteratively optimized to have sufficient brightness and maltose-dependent fluorescence increases for imaging, under both one- and two-photon illumination. We demonstrate that maltose affinity of the sensors can be tuned in a fashion largely independent of the fluorescent readout mechanism. Using literature mutations, the binding specificity could be altered to moderate sucrose preference, but with a significant loss of affinity. We use the soluble sensors in individual E. coli bacteria to observe rapid maltose transport across the plasma membrane, and membrane fusion versions of the sensors on mammalian cells to visualize the addition of maltose to extracellular media. The PBP superfamily includes scaffolds specific for a number of analytes whose visualization would be critical to the reverse engineering of complex systems such as neural networks, biosynthetic pathways, and signal transduction cascades. We expect the methodology outlined here to be useful in the development of indicators for many such analytes.

59 BASIC BIOLOGICAL SCIENCES↗

A Science Gateway for the Repeatable Analysis of Machine Learning Predicted Gravity Anomalies

In recent years, deep learning has become an increasingly popular alternative for modeling in geoscience applications due to its scalability and efficiency. However, the interpretability, compute, data volume, and hyperparameter tuning requirements of deep learning models make development and monitoring difficult. Furthermore, model explainability and communicating results obtained by these models to users or domain experts is a challenge, as domain experts in geoscience also need to have a deep understanding of how those models function in order to support their scientific works. Here, we describe a science gateway and machine learning pipeline for predicting gravity anomalies from geophysical data. The gateway, built on open-source technologies, provides a holistic view of the pipeline through interactive visualizations aimed at enabling efficient exploratory data analysis. The repeatability, reproducibility, and monitoring capabilities of this overall system allow us to iterate and analyze at scale. Using this pipeline and gateway, we can repeatedly produce accurate high-resolution gravity anomaly datasets. By describing the underlying technologies, implementation, and results, here we provide a foundation for the broader adoption of science gateways into cross-cutting geoscience and machine learning research projects as a means to improve the scientific discovery and collaboration in the geophysics and computational sciences community.

58 GEOSCIENCES↗

Event Classifications on DNE2 Main Experiment Data using a Convolutional Neural Network Ensemble

The Dynamic Networks (DN) Experiment for FY24 (DNE2) is an experiment within DN with the goal of quantitatively evaluating the effectiveness of solutions developed so far by various researchers under the Low Yield Nuclear Monitoring (LYNM) program using a shared set of metrics and datasets. A key component of this experiment is the mimicking of a signature processing pipeline, and comparing currently accepted and standard-use processing methods to more state-of-the-art processes developed under DN. In this work, we focus specifically on the Event Characterization (EC) Focus Area (FA) of the pipeline, where a seismic event’s magnitude, yield and class are identified. We use Deep Learning (DL) to classify the type of events being processed as either earthquakes (EQs) or explosions (EXs) for three iterations of experiment datasets. The model is noticeably more confident and accurate in classifying explosions than earthquakes, reflecting a known shortcoming of the model, that being of a bias towards predicting explosions over earthquakes in the west coast due to training data biases.

97 MATHEMATICS AND COMPUTING↗