Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Distributed Processing”

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 451 records · Page 25
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Siting bioenergy facilities in the United States: Measuring participation in decisions and distribution of effects

Scientists and stakeholders can inform the process of siting renewable energy facilities in ways that do not perpetuate socioeconomic disparities associated with fossil fuel industries or create new ones. Procedural justice indicators and distributional justice indicators that incorporate environmental, social, and economic objectives can be used to site energy facilities in ways that increase benefits and reduce negative impacts to disadvantaged and underserved populations. A generic list of potential energy justice indicators for siting bioenergy facilities was developed collaboratively between U.S. bioenergy researchers and diverse agriculture, energy, and energy and environmental justice stakeholders and experts. From this list smaller numbers of indicators can be selected or modified with communities for local siting of bioenergy facilities. Groups of indicators can be used to guide biorefinery or biopower siting and permitting decisions, e.g., to compare siting options, to draw early attention to key problems, or to track progress toward justice-related targets.

09 BIOMASS FUELS↗
DOE OSTI2024

Snow Distribution Patterns Revisited: A Physics-Based and Machine Learning Hybrid Approach to Snow Distribution Mapping in the Sub-Arctic

Snowpack distribution in Arctic and alpine landscapes often occurs in repeating, year-to-year patterns due to local topographic, weather, and vegetation characteristics. Previous studies have suggested that with years of observational data, these snow distribution patterns can be statistically integrated into a snow process modeling workflow. Recent advances in snow hydrology and machine learning (ML) have increased our ability to predict snowpack distribution using in-situ observations, remote sensing data sets, and simple landscape characteristics that can be easily obtained for most environments. Here, we propose a hybrid approach to couple a ML snow distribution pattern (MLSDP) map with a physics-based, snow process model. We trained a random forest ML algorithm on tens of thousands of snow survey observations from a subarctic study area on the Seward Peninsula, Alaska, collected during peak snow water equivalent (SWE). We validated hybrid model outputs using in-situ snow depth and SWE observations, as well as a light detection and ranging data set and a distributed temperature profiling sensor data set. When the hybrid results were compared with the physics-based method, the hybrid method more accurately depicted the spatial patterns of the snowpack, areas of drifting snow, and years when no in-situ observations were used in the random forest ML training data set. The hybrid method also showed improvements in root mean squared error at 61% of locations where time-series estimations of snow depth were observed. These results can be applied to any physics-based model to improve the snow distribution patterning to reflect observed conditions in high latitude and high elevation cold region environments.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Physics-Informed Gaussian Process Inference of Liquid Structure from Scattering Data

We present a nonparametric Bayesian framework to infer radial distribution functions from experimental scattering measurements with uncertainty quantification using nonstationary Gaussian processes. The Gaussian process prior mean and kernel functions are designed to mitigate well-known numerical challenges with the Fourier transform, including discrete measurement binning and detector windowing, while encoding fundamental yet minimal physical knowledge of the liquid structure. We demonstrate uncertainty propagation of the Gaussian process posterior to unmeasured quantities of interest. Experimental radial distribution functions of liquid argon and water with uncertainty quantification are provided as both a proof of principle for the method and a benchmark for molecular models.

Chemical structure↗
DOE OSTI2021

Solving inverse problems in stochastic models using deep neural networks and adversarial training

Inverse problems associated with stochastic models constitute a significant portion of scientific and engineering applications. In such cases the unknown quantities are distributions. The applicability of traditional methods is limited because of their demanding assumptions or prohibitive computational consumption; for example, maximum likelihood methods require closed-form density functions, and Markov Chain Monte Carlo needs a large number of simulations. We propose a new method that estimates the unknown distribution by matching the statistical properties between observed and simulated random processes. We leverage the expressive power of neural networks to approximate the unknown distribution and use a discriminative neural network for computing the statistical discrepancies between the observed and simulated random processes. Here we demonstrated numerically that the proposed methods can estimate both the model parameters and learn complicated unknown distributions.

42 ENGINEERING↗
DOE OSTI2021

The impact of r -process heating on the dynamics of neutron star merger accretion disc winds and their electromagnetic radiation

Neutron star merger accretion discs can launch neutron-rich winds of >10 -2 M⊙. This ejecta is a prime site for r-process nucleosynthesis, which will produce a range of radioactive heavy nuclei. The decay of these nuclei releases enough energy to accelerate portions of the wind by ~0.1c. Here, we investigate the effect of r-process heating on the dynamical evolution of disc winds. We extract the wind from a 3D general relativistic magnetohydrodynamic simulation of a disc from a post-merger system. This is used to create inner boundary conditions for 2D hydrodynamic simulations that continue the original 3D simulation. We perform two such simulations: one that includes the r-process heating, and another one that does not. We follow the hydrodynamic simulations until the winds reach homology (60 s). Using time-dependent multifrequency multidimensional Monte Carlo radiation transport simulations, we then calculate the kilonova light curves from the winds with and without dynamical r-process heating. We find that the r-process heating can substantially alter the velocity distribution of the wind, shifting the mass-weighted median velocity from 0.06c to 0.12c. The inclusion of the dynamical r-process heating makes the light curve brighter and bluer at $\sim 1\, \mathrm{d}$ post-merger. However, the high-velocity tail of the ejecta distribution and the early ($\lesssim 1\, \mathrm{d}$) light curves are largely unaffected.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2023

Reinforcement Learning for Load-balanced Parallel Particle Tracing

We explore an online reinforcement learning (RL) paradigm to dynamically optimize parallel particle tracing performance in distributed-memory systems. Our method combines three novel components: (1) a work donation algorithm, (2) a high-order workload estimation model, and (3) a communication cost model. First, we design an RL-based work donation algorithm. Our algorithm monitors workloads of processes and creates RL agents to donate data blocks and particles from high-workload processes to low-workload processes to minimize program execution time. The agents learn the donation strategy on the fly based on reward and cost functions designed to consider processes' workload changes and data transfer costs of donation actions. Second, we propose a workload estimation model, helping RL agents estimate the workload distribution of processes in future computations. Third, we design a communication cost model that considers both block and particle data exchange costs, helping RL agents make effective decisions with minimized communication costs. We demonstrate that our algorithm adapts to different flow behaviors in large-scale fluid dynamics, ocean, and weather simulation data. Our algorithm improves parallel particle tracing performance in terms of parallel efficiency, load balance, and costs of I/O and communication for evaluations with up to 16,384 processors.

Distributed and parallel particle tracing↗
DOE OSTI2021

Asynchronous and Load-Balanced Union-Find for Distributed and Parallel Scientific Data Visualization and Analysis

We present a novel distributed union-find algorithm that features asynchronous parallelism and k-d tree based load balancing for scalable visualization and analysis of scientific data. Applications of union-find include level set extraction and critical point tracking, but distributed union-find can suffer from high synchronization costs and imbalanced workloads across parallel processes. In this study, we prove that global synchronizations in existing distributed union-find can be eliminated without changing final results, allowing overlapped communications and computations for scalable processing. We also use a k-d tree decomposition to redistribute inputs, in order to improve workload balancing. We benchmark the scalability of our algorithm with up to 1,024 processes using both synthetic and application data. Here, we demonstrate the use of our algorithm in critical point tracking and super-level set extraction with high-speed imaging experiments and fusion plasma simulations, respectively.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Structural and mechanical properties of monolayer amorphous carbon and boron nitride

Amorphous materials exhibit various characteristics that are not featured by crystals and can sometimes be tuned by their degree of disorder (DOD). Here, we report results on the mechanical properties of monolayer amorphous carbon (MAC) and monolayer amorphous boron nitride (maBN) with different DOD. The pertinent structures are obtained by kinetic-Monte-Carlo (kMC) simulations using machine-learning potentials (MLP) with density-functional-theory (DFT)-level accuracy. An intuitive order parameter, namely the areal fraction F x occupied by crystallites within the continuous random network, is proposed to describe the DOD. We find that F x captures the essence of the DOD: Samples with the same F x but different sizes and arrangements of crystallites, obtained using two distinct kMC procedures, have virtually identical radial distributions functions as well as bond-length and bond-angle distributions. Furthermore, by simulating the fracture process with molecular dynamics, we found that the mechanical responses of MAC and maBN before fracture are mainly determined by F x and are insensitive to the sizes and specific arrangements and to some extent the numbers and area distributions of the crystallites. The behavior of cracks in the two materials is analyzed and found to mainly propagate in meandering paths in the CRN region and to be influenced by crystallites in distinct ways that toughen the material. Furthermore, the present results reveal the relation between structure and mechanical properties in amorphous monolayers and may provide a universal toughening strategy for 2D materials.

2-dimensional systems↗
DOE OSTI2024

An integrated atom array-nanophotonic chip platform with background-free imaging

Arrays of neutral atoms trapped in optical tweezers have emerged as a leading platform for quantum information processing and quantum simulation due to their scalability, reconfigurable connectivity, and high-fidelity operations. Individual atoms are promising candidates for quantum networking due to their capability to emit indistinguishable photons that are entangled with their internal atomic states. Integrating atom arrays with photonic interfaces would enable distributed architectures in which nodes hosting many processing qubits could be efficiently linked together via the distribution of remote entanglement. However, many atom array techniques cease to work in close proximity to photonic interfaces, with atom detection via standard fluorescence imaging presenting a major challenge due to scattering from nearby photonic devices. Here, we demonstrate an architecture that combines atom arrays with up to 64 optical tweezers and a millimeter-scale photonic chip hosting more than 100 nanophotonic cavities. We achieve high-fidelity ( ~ 99.2%), background-free imaging in close proximity to nanofabricated cavities using a multichromatic excitation and detection scheme. The atoms can be imaged while trapped a few hundred nanometers above the dielectric surface, which we verify using Stark shift measurements of the modified trapping potential. Finally, we rearrange atoms into defect-free arrays and load them simultaneously onto the same or multiple devices.

74 ATOMIC AND MOLECULAR PHYSICS↗
DOE OSTI2022

Three‐Dimensional Permeability Inversion Using Convolutional Neural Networks and Positron Emission Tomography

Abstract Quantification of heterogeneous multiscale permeability in geologic porous media is key for understanding and predicting flow and transport processes in the subsurface. Recent utilization of in situ imaging, specifically positron emission tomography (PET), enables the measurement of three‐dimensional (3‐D) time‐lapse radiotracer solute transport in geologic media. However, accurate and computationally efficient characterization of the permeability distribution that controls the solute transport process remains challenging. Leveraging the relationship between local permeability variation and solute advection behavior, an encoder‐decoder based convolutional neural network (CNN) is implemented as a permeability inversion scheme using a single PET scan of a radiotracer pulse injection experiment as input. The CNN can accurately capture the 3‐D spatial correlation between the permeability and the radiotracer solute arrival time difference maps in geologic cores. We first test the inversion accuracy using synthetic test datasets and then test the accuracy on a suite of experimental PET imaging datasets acquired on four different geologic cores. The network‐predicted permeability maps from the geologic cores are used to parameterize forward numerical models that are directly compared with the experimental PET imaging data. The results indicate that a single trained network can generate robust 3‐D permeability inversion maps in seconds. Numerical models parameterized with these permeability maps closely capture the experimentally observed solute arrival time behavior. This work provides an unprecedented approach for efficiently characterizing multiscale permeability heterogeneity in complex geologic samples.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Blockchain for Fault-Tolerant Grid Operations

Radial topology and vast geographic coverage make distribution systems prone to widespread power outages upon the failure of a single (or multiple) upstream component. Fault-handling algorithms depend heavily on correct estimations of the system’s state to effectively isolate the affected area and reduce the number of affected customers while maintaining operational safety. The work described here leverages the core features of distributed, consensus-based decision-making processes and the immutability of blockchain, and demonstrates their value in improving fault-tolerant grid operations. In this work, blockchain was used to create a trusted data-sharing platform that enables independent actors to reconstruct the system state; this enables distributed resources to make intelligent decisions with limited knowledge. Although the process requires data sharing, its algorithms have been designed to limit the amount of private information that is exchanged, which helps preserve business-sensitive data and maintain customer privacy. In addition, by reducing the information that must be shared, the communication requirements are also reduced; (however, an in-depth analysis of the communication requirements is beyond the scope of this project). The proposed use cases are intended to represent a foundational basis for third parties to develop functional solutions that can eventually be deployed in the field. To further provide guidance, the envisioned use cases have incorporated design requirements that consider the blockchain characteristics and a need to limit information from surrounding resources, which preserve the assumption and the possibility that such resources could belong to different entities. This report presents a detailed design of the three use cases with the tools needed to enable the analysis being tested. The implemented gross error detection method can detect mismatches when the error exceeds 3.8 times the sensor’s rated accuracy. Detection of the circuit breaker state successfully identified the correct states across all simulation tests. A distribution-system power-flow solution in the simulator OpenDSS generally possesses a convergency tolerance of 0.01% on the voltage magnitude. The evaluation of possible reconnection using voltage magnitude—preserving the data ownership—has a voltage magnitude difference smaller than 0.001% from the OpenDSS result. The results preserving data ownership have a difference within the expected power flow tolerance with full knowledge of the system, which surpasses expectations.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2022

Ruddlesden–Popper Perovskites with Narrow Phase Distribution for Air-Stable Solar Cells

2D Ruddlesden-Popper perovskites have risen to prominence as stable and efficient photovoltaic materials because of their structural diversity, rich photophysics, and low moisture ingression. However, thin films processed from stoichiometric precursor solutions possess a broad phase distribution of different number of inorganic layers with random crystal orientation, crippling device performance. The effect of methylammonium chloride (MACI) and 3-amino4-phenolsulfonic acid (APSA) on the fabrication of perpendicularly oriented (PEA) 2 MA 4 Pb 5 I 16 films with narrow phase distribution using antisolvent and hot-casting processing techniques is investigated. MACI plays a critical role in suppressing parasitic n ≤ 2 and 3D-like phases. APSA performs the dual function of trap passivation and further narrowing phase polydispersity through strong coordination with Pb 2+ . Ex situ grazing-incident wide-angle X-Ray scattering (GIWAXS) and ultrafast spectroscopic characterization reveal uniformly mixed-phase distribution with disordered orientation in antisolvent treated films, while additive-assisted hot-casting treatment results in oriented, reverse-graded phase distribution, i.e., small-n on the film surface and large-n at the bottom. Arising thin films enable efficient p-i-n solar cells with an efficiency of 14.34%, and a V oc of 1.20 V, retaining 96% initial efficiency after 1440 h under ambient conditions (RH = 50-60%) without encapsulation.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Deep Generative Models that Solve PDEs: Distributed Computing for Training Large Data-Free Models

Recent progress in scientific machine learning (SciML) has opened up the possibility of training novel neural network architectures that solve complex partial differential equations (PDEs). Several (nearly data free) approaches have been recently reported that successfully solve PDEs, with examples including deep feed forward networks, generative networks, and deep encoder-decoder networks. However, practical adoption of these approaches is limited by the difficulty in training these models, especially to make predictions at large output resolutions (≥1024×1024). Here we report on a software framework for data parallel distributed deep learning that resolves the twin challenges of training these large SciML models - training in reasonable time as well as distributing the storage requirements. Our framework provides several out of the box functionality including (a) loss integrity independent of number of processes, (b) synchronized batch normalization, and (c) distributed higher-order optimization methods. We show excellent scalability of this framework on both cloud as well as HPC clusters, and report on the interplay between bandwidth, network topology and bare metal vs cloud. We deploy this approach to train generative models of sizes hitherto not possible, showing that neural PDE solvers can be viably trained for practical applications. We also demonstrate that distributed higher-order optimization methods are 2-3× faster than stochastic gradient-based methods and provide minimal convergence drift with higher batch-size.

PDEs↗
DOE OSTI2025

Inference of Multichannel r -process Element Enrichment in the Milky Way Using Binary Neutron Star Merger Observations

Observations of GW170817 strongly suggest that binary neutron star (BNS) mergers produce rapid neutron-capture nucleosynthesis ( r -process) elements. However, it remains an open question whether these mergers can account for all the r -process element enrichment in the Milky Way’s history. Here, we constrain the contributions of the BNS channel using astrophysical neutron star observations. The rate and mass distributions are constrained by LIGO/Virgo/Kagra through the latest catalog GWTC-3, the neutron star equation of state by gravitational-wave, radio, and X-ray observations, and the delay time distribution by short gamma-ray burst (GRB) host galaxy associations. We present a Bayesian framework to consistently combine these observations with abundance information to quantify the contribution and uncertainties of single and multiple astrophysical enrichment sources, and obtain a distribution of per-event BNS r -process element yields consistent with geophysical and astrophysical abundance constraints. We then adopt a Galactic chemical evolution model assuming an instantaneous and fixed amount of Fe enrichment from core-collapse supernovae, and show that BNS-only enrichment scenarios remain inconsistent with the observed r-process abundance trend of disk stars in the Galaxy even with the uncertainties in BNS merger observations. Using stellar abundance observations instead of the short GRB constraints, we can infer a shorter BNS delay time distribution with power-law index α ≤ −2.0 and minimum delay time ${t}_{{\rm{\min }}}\leqslant 40$ Myr at 90% confidence, consistent with detailed Galactic chemical evolution models. Such delay times are in tension with those predicted by standard BNS formation models. Alternatively, we confirm that a two-channel scenario, in which the second channel tracks the star formation history without significant delay, can account for both Galactic stellar and short GRB observations. We estimate that 45%–90% of the r -process abundance in the Milky Way today would have been produced by this star formation-tracking channel, rather than BNS mergers with significant delay times.

gravitational wave astronomy↗
DOE OSTI2020

Inferring the Dynamics of the State Evolution During Quantum Annealing

To solve an optimization problem using a commercial quantum annealer, one has to represent the problem of interest as an Ising or a quadratic unconstrained binary optimization (QUBO) problem and submit its coefficients to the annealer, which then returns a user-specified number of low-energy solutions. It would be useful to know what happens in the quantum processor during the anneal process so that one could design better algorithms or suggest improvements to the hardware. However, existing quantum annealers are not able to directly extract such information from the processor. Hence, in this work we propose to use advanced features of D-Wave 2000Q to indirectly infer information about the dynamics of the state evolution during the anneal process. Specifically, D-Wave 2000Q allows the user to customize the anneal schedule, that is, the schedule with which the anneal fraction is changed from the start to the end of the anneal. Furthermore, using this feature, we design a set of modified anneal schedules whose outputs can be used to generate information about the states of the system at user-defined time points during a standard anneal. With this process, called "slicing", we obtain approximate distributions of lowest-energy anneal solutions as the anneal time evolves. We use our technique to obtain a variety of insights into the annealer, such as the state evolution during annealing, when individual bits in an evolving solution flip during the anneal process and when they stabilize, and we introduce a technique to estimate the freeze-out point of both the system as well as of individual qubits.

42 ENGINEERING↗
DOE OSTI2024

Simulant Development of Potential 200 West Area Waste Feeds

Preliminary planning for retrieval, qualification, and pretreatment of waste in Hanford’s 200 West Area (200W) has begun as part of the West Area Risk Management project. Experimental studies to technically mature pretreatment process operations will likely be needed because of the uniqueness of 200W waste. Pacific Northwest National Laboratory formulated five simulants to represent 200W-qualified feed based on the preliminary flowsheet provided by Washington River Protection Solutions, LLC. The simulant recipes were devised using applicable historical information as a reference point to support the use of the flowsheet waste vectors, which were combined into five distinct groups. These five groups formed the basis for the liquid composition targets that were adapted into recipes using charged-balanced salt species. The liquid phase recipes were batched in 1-L quantities and analyzed at Pacific Northwest National Laboratory. Once confirmed to be stable, the liquid solutions were tested for compatibility with candidate solid components. Specific solid components were recommended based on cross-examining the proposed solid phases in the flowsheet with relevant data from the literature. Mixtures of solid components were added to aliquots of the liquid batches and sub-sampled to measure particle size distribution. The measured distribution was compared to independently created benchmark distributions appropriate for each simulant. This process was iterated until a solid phase composition that resulted in a representative particle size distribution was found. After the final compositions were confirmed, a suite of chemical and physical characterization data was collected. This report describes the simulant basis, formulation methodology, laboratory measurements, and data collected for the recipes recommended to represent 200W waste feeds.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2021

Coupled Microbial-Conversion and Computational-Fluid-Dynamics (CFD) Models for Butanediol Production in Micro-Aerated Reactors

Microbial conversion of substrates to macromolecules has been widely used in the synthesis of value-added products in pharmaceutical and biotechnology industries. These bioreactions are also being investigated in the production of low-value commodities such as biofuels [1]. Gas-liquid mass-transfer and transport-reaction coupling are important challenges when designing and scaling up these reactor systems. Experiments in wellmixed small-scale reactors have enabled characterization of microbial reactivity, while their coupling with macroscale transport remains relatively unexplored. In this work, we use a coupled metabolic-CFD model to study the action of a genetically engineered microbe Zymomonas mobilis [2] on sugars to produce 2,3-Butanediol (BDO). BDO is an important hydrocarbon intermediate that can be catalytically upgraded to several fuels and chemicals [3]. An important aspect to this particular microbial conversion is the need for micro-aerated environments as opposed to traditional aerobic fermentation. Slight variations in oxygen concentration can result in competing reaction pathways that disable BDO production. Hence, gas-liquid mass transfer and transport need to be optimized in large-scale reactors to maximize BDO production, for which CFD is a valuable tool. The aerobic-fermentation CFD model previously developed by the authors [5] for simulating bubble-column and airlift reactors at scale was used in this study. The Reynolds-averaged mass, momentum and energy transport equations for interpenetrating gas and liquid phase are solved in this model along with the transport and interphase mass transfer of oxygen. Our previous work used a phenomenological model for microbial oxygen uptake that neglected microbial growth and other reaction pathways. In this work, a detailed metabolic model enabled prediction of product formation and inhibition pathways. In order to manage computational cost, we used a subcycling technique [6] that takes advantage of the clear separation in transport (~ 200 sec) and reaction (~ 2-3 hours) timescales. The CFD model is first solved to steady state, after which the metabolic model is advanced at every cell in the computational domain using the local oxygen concentration. The CFD model is then run to achieve a new steady state that provides a new oxygen distribution for the metabolic model. This process, where reaction and fluid updates are interleaved together, is iterated until reactants are completely exhausted. This work will examine the performance of different reactor designs such as bubble column and airlift reactors at scale (250-500 m3). Oxygen mass-transfer coefficient and distribution are critically analyzed among reactors, and optimization studies pertaining to aeration is presented. Furthermore, it has been observed in experiments that high BDO production may be achieved by manipulating the aerobic environment over the course of reaction, such that oxygen concentration is high during the growth phase, and very low as sugar is depleted. This characteristic will be addressed by our simulations for which a timedependent scheduling strategy for aeration is presented that maximizes BDO production. [1] Humbird, D., Davis, R., and McMillan, J., Aeration costs in stirred-tank and bubble column bioreactors, Biochemical Engineering Journal, 127, 161—166, 2017 [2] Yang, S., Mohagheghi, A., Franden, M. A., Chou, Y.-C., Chen, X., Dowe, N., Himmel, M. E., and Zhang, M., Metabolic engineering of zymomonas mobilis for 2, 3-butanediol production from lignocellulosic biomass sugars. Biotechnology for biofuels, 9(1):189, 2016 [3] Kim, S. J., Sim, H. J., Kim, J. W., Lee, Y. G., Park, Y. C., and Seo, J. H., Enhanced production of 2,3-butanediol from xylose by combinatorial engineering of xylose metabolic pathway and cofactor regeneration in pyruvate decarboxylase-deficient Saccharomyces cerevisiae. Bioresource Technology, 245:1551–1557, 2017 [4] Weller, H., Tabor, G., Jasak, H. and Fureby, C., A tensorial approach to computational continuum mechanics using object-oriented techniques, Computers in physics, 12, 6, 620--631, 1998

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2021

Deep Cellular Recurrent Network for Efficient Analysis of Time-Series Data With Spatial Information

Efficient processing of large-scale time series data is an intricate problem in machine learning. Conventional sensor signal processing pipelines with hand engineered feature extraction often involve huge computational cost with high dimensional data. Deep recurrent neural networks have shown promise in automated feature learning for improved time-series processing. However, generic deep recurrent models grow in scale and depth with increased complexity of the data. This is particularly challenging in presence of high dimensional data with temporal and spatial characteristics. Consequently, this work proposes a novel deep cellular recurrent neural network (DCRNN) architecture to efficiently process complex multi-dimensional time series data with spatial information. Here, the cellular recurrent architecture in the proposed model allows for location-aware synchronous processing of time series data from spatially distributed sensor signal sources. Extensive trainable parameter sharing due to cellularity in the proposed architecture ensures efficiency in the use of recurrent processing units with high-dimensional inputs. This study also investigates the versatility of the proposed DCRNN model for classification of multi-class time series data from different application domains. Consequently, the proposed DCRNN architecture is evaluated using two time-series datasets: a multichannel scalp EEG dataset for seizure detection, and a machine fault detection dataset obtained in-house. The results suggest that the proposed architecture achieves state-of-the-art performance while utilizing substantially less trainable parameters when compared to comparable methods in the literature.

60 APPLIED LIFE SCIENCES↗