Search NASA⌕ Search

SEARCH · Search NASA

Results for “Network performance forecasting”

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 217 records · Page 12

Prediction of the SYM-H Index Using a Bayesian Deep Learning Method With Uncertainty Quantification

We propose a novel deep learning framework, named SYMHnet, which employs a graph neural network and a bidirectional long short-term memory network to cooperatively learn patterns from solar wind and interplanetary magnetic field parameters for short-term forecasts of the SYM-H index based on 1- and 5-min resolution data. SYMHnet takes, as input, the time series of the parameters' values provided by NASA's Space Science Data Coordinated Archive and predicts, as output, the SYM-H index value at time point t + w hours for a given time point t where w is 1 or 2. By incorporating Bayesian inference into the learning framework, SYMHnet can quantify both aleatoric (data) uncertainty and epistemic (model) uncertainty when predicting future SYM-H indices. Experimental results show that SYMHnet works well at quiet time and storm time, for both 1- and 5-min resolution data. The results also show that SYMHnet generally performs better than related machine learning methods. For example, SYMHnet achieves a forecast skill score (FSS) of 0.343 compared to the FSS of 0.074 of a recent gradient boosting machine (GBM) method when predicting SYM-H indices (1 hr in advance) in a large storm (SYM-H = -393 nT) using 5-min resolution data. When predicting the SYM-H indices (2 hr in advance) in the large storm, SYMHnet achieves an FSS of 0.553 compared to the FSS of 0.087 of the GBM method. In addition, SYMHnet can provide results for both data and model uncertainty quantification, whereas the related methods cannot.

79 ASTRONOMY AND ASTROPHYSICS↗

Sequential Precipitation Input Tagging (SPIT) to Estimate Water Transit Times and Hydrologic Tracer Dynamics Within Water‐Tagging Enabled Hydrologic Models

Determining the age distribution of water exiting a catchment is important for understanding groundwater storage and mixing. New water-tagging capabilities within models track precipitation events as they move through simulated storages, yet forward modeling of individual events may not systematically capture the full transit time distribution (TTD). Here, we present a “sequential precipitation input tagging” (SPIT) framework to tag all input precipitation at regular intervals during extended model simulations. Monthly tags over 7 years were applied at six National Ecological Observatory Network sites to calculate TTDs and derive mean virtual tracer age, $\overline{T_{V}}$, fractions of young water, F yw , and hydrologic tracer concentrations (water isotopes δ 18 O and δ 2 H) within a tagging enabled version of the Weather Research and Forecast hydrologic model (WRF-Hydro). Throughout seven simulation years, the fraction of simulated discharge derived from tagged events, F tag , increased each year, with the final year's F tag ranging from 66% to 100% and highlights the need to apply SPIT over many years to understand TTDs. When the F tag was >75%, simulated $\overline{T_{V}}$ ranged 179–923 days and F yw 0.6%–23.9%, with daily values exhibiting a power-law relationship with precipitation, discharge, and groundwater. Through implementation of SPIT, we find this hydrologic model configuration performs poorly in estimation of $\overline{T_{V}}$ and F yw (root mean squared error of 469 days and 14.4% respectively), suggesting it misrepresents subsurface mixing. Thus, the SPIT framework provides a reproducible approach to calculate watershed transit times within tagging enabled models and thereby assess and improve representation of hydrologic processes.

fraction of young water↗

A Fusion of Geothermal and InSAR Data with Machine Learning for Enhanced Deformation Forecasting at the Geysers

The Geysers geothermal field in California is experiencing land subsidence due to the seismic and geothermal activities taking place. This poses a risk not only to the underlying infrastructure but also to the groundwater level which would reduce the water availability for the local community. Because of this, it is crucial to monitor and assess the surface deformation occurring and adjust geothermal operations accordingly. In this study, we examine the correlation between the geothermal injection and production rates as well as the seismic activity in the area, and we show the high correlation between the injection rate and the number of earthquakes. This motivates the use of this data in a machine learning model that would predict future deformation maps. First, we build a model that uses interferometric synthetic aperture radar (InSAR) images that have been processed and turned into a deformation time series using LiCSBAS, an open-source InSAR time series package, and evaluate the performance against a linear baseline model. The model includes both convolutional neural network (CNN) layers as well as long short-term memory (LSTM) layers and is able to improve upon the baseline model based on a mean squared error metric. Then, after getting preprocessed, we incorporate the geothermal data by adding them as additional inputs to the model. This new model was able to outperform both the baseline and the previous version of the model that uses only InSAR data, motivating the use of machine learning models as well as geothermal data in assessing and predicting future deformation at The Geysers as part of hazard mitigation models which would then be used as fundamental tools for informed decision making when it comes to adjusting geothermal operations.

Yazbeck, Joe (ORCID:0000000302235260)↗

Search for single top quark production in pp collisions at {radical}(s)=1.96 TeV in the missing transverse energy plus jets topology

We report a search for single top quark production with the CDF II detector using 2.1 fb{sup -1} of integrated luminosity of pp collisions at {radical}(s)=1.96 TeV. The data selected consist of events characterized by large energy imbalance in the transverse plane and hadronic jets, and no identified electrons and muons, so the sample is enriched in W{yields}{tau}{nu} decays. In order to suppress backgrounds, additional kinematic and topological requirements are imposed through a neural network, and at least one of the jets must be identified as a b quark jet. We measure an excess of signal-like events in agreement with the standard model prediction, but inconsistent with a model without single top quark production by 2.1 standard deviations ({sigma}), with a median expected sensitivity of 1.4{sigma}. Assuming a top quark mass of 175 GeV/c{sup 2} and ascribing the excess to single top quark production, the cross section is measured to be 4.9{sub -2.2}{sup +2.5}(stat+syst) pb, consistent with measurements performed in independent data sets and with the standard model prediction.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Field Experiments and Model Validation: The EGS Collab Project

Implementing enhanced geothermal systems (EGS) will require improvements in understanding stimulation of crystalline rock to create appropriate flow pathways, and the ability to effectively simulate both the stimulation and the flow and transport processes in the resulting fracture network. The US Department of Energy (DOE) is addressing these and other challenges at multiple scales. The EGS Collab project, addressed here, is performing tests and modeling at the 10 m scale. The FORGE project is performing tests at the full reservoir scale. The EGS Collab team created an underground testbed at the Sanford Underground Research Facility (SURF) in Lead, South Dakota, at a depth of approximately 1.5 km to examine hydraulic fracturing (Experiment 1). We are currently building a second testbed at SURF at a depth of about 1.25 km aimed at investigating shear stimulation (Experiment 2). In Experiment 1, we drilled eight boreholes in a well-characterized phyllite and installed geophysical sensors in six of them to create a well-instrumented testbed to allow careful monitoring of stimulation events and flow tests. Numerical simulation was used to answer key experimental design questions, to forecast fracture propagation trajectories and extents, and to analyze and evaluate results both in near-real-time and in detailed process studies. Stimulations performed in this testbed allowed quantification of processes occurring during stimulation and the examination of dynamic flow occurrences. Long-term ambient temperature and chilled water flow tests were performed in addition to many tracer tests to examine system behavior. Our second testbed, targeted at shear stimulation, is currently being built at the SURF Facility at a depth of about 1.25 km in amphibolite under a different set of stress and fracture conditions than Experiment 1.

artificial intelligence↗

Improving multiwell petrophysical interpretation from well logs via machine learning and statistical models

Well-log interpretation estimates in situ rock properties along well trajectory, such as porosity, water saturation, and permeability, to support reserve-volume estimation, production forecasts, and decision making in reservoir development. However, due to measurement errors, variability of well logs caused by multiple measurement vendors, different borehole tools, and nonuniform drilling/borehole conditions, estimations of rock properties with original well logs without proper preprocessing may not be accurate, especially in the context of multiwell estimation. Well-log normalization techniques such as two-point scaling and mean-variance normalization are commonly used to improve the robustness of multiwell rock-property estimation. However, these techniques do not consider the correlation between well logs and require subjective knowledge for their effective implementation. To reduce uncertainties and processing time associated with multiwell rock-property estimation from well logs, we develop discriminative adversarial (DA) and linear constraint models for well-log normalization and rock-property estimation. The DA neural network model developed for well-log normalization and interpretation can perform linear and nonlinear well-log normalization while considering the joint distribution of each well log and rock properties. However, the linear constraint model uses an ensemble of predictions from linear models to constrain well-log normalization and rock-property estimation. We also develop a divergence-based type well identification method to select type (training) wells for a test well based on the statistical similarity of associated well-log distributions instead of the interwell distance. We apply the DA model to perform well-log normalization and prediction of permeability for the Seminole San Andres Unit carbonate reservoir. Compared with the permeability predicted with the classical machine learning model without well-log normalization and models with two-point scaling normalization, the DA model yields the most accurate permeability prediction by decreasing the mean-squared error of permeability prediction by 20%–50%.

Geochemistry & Geophysics↗

Continuous Time Group Discovery in Dynamic Graphs

With the rise in availability and importance of graphs and networks, it has become increasingly important to have good models to describe their behavior. While much work has focused on modeling static graphs, we focus on group discovery in dynamic graphs. We adapt a dynamic extension of Latent Dirichlet Allocation to this task and demonstrate good performance on two datasets. Modeling relational data has become increasingly important in recent years. Much work has focused on static graphs - that is fixed graphs at a single point in time. Here we focus on the problem of modeling dynamic (i.e. time-evolving) graphs. We propose a scalable Bayesian approach for community discovery in dynamic graphs. Our approach is based on extensions of Latent Dirichlet Allocation (LDA). LDA is a latent variable model for topic modeling in text corpora. It was extended to deal with topic changes in discrete time and later in continuous time. These models were referred to as the discrete Dynamic Topic Model (dDTM) and the continuous Dynamic Topic Model (cDTM), respectively. When adapting these models to graphs, we take our inspiration from LDA-G and SSN-LDA, applications of LDA to static graphs that have been shown to effectively factor out community structure to explain link patterns in graphs. In this paper, we demonstrate how to adapt and apply the cDTM to the task of finding communities in dynamic networks. We use link prediction to measure the quality of the discovered community structure and apply it to two different relational datasets - DBLP author-keyword and CAIDA autonomous systems relationships. We also discuss a parallel implementation of this approach using Hadoop. In Section 2, we review LDA and LDA-G. In Section 3, we review the cDTM and introduce cDTMG, its adaptation to modeling dynamic graphs. We discuss inference for the cDTM-G and details of our parallel implementation in Section 4 and present its performance on two datasets in Section 5 before concluding in Section 6.

97 MATHEMATICS AND COMPUTING↗

Integrating Reanalysis and Satellite Cloud Information to Estimate Surface Downward Long-Wave Radiation

The estimation of downward long-wave radiation (DLR) at the surface is very important for the understanding of the Earth’s radiative budget with implications in surface–atmosphere exchanges, climate variability, and global warming. Theoretical radiative transfer and observationally based studies identify the crucial role of clouds in modulating the temporal and spatial variability of DLR. In this study, a new machine learning algorithm that uses multivariate adaptive regression splines (MARS) and the combination of near-surface meteorological data with satellite cloud information is proposed. The new algorithm is compared with the current operational formulation used by the European Organization for the Exploitation of Meteorological Satellites (EUMETSAT) Satellite Application Facility on Land Surface Analysis (LSA-SAF). Both algorithms use near-surface temperature and dewpoint temperature along with total column water vapor from the latest European Centre for Medium-range Weather Forecasts (ECMWF) reanalysis ERA5 and satellite cloud information from the Meteosat Second Generation. The algorithms are trained and validated using both ECMWF-ERA5 and DLR acquired from 23 ground stations as part of the Baseline Surface Radiation Network (BSRN) and the Atmospheric Radiation Measurement (ARM) user facility. Results show that the MARS algorithm generally improves DLR estimation in comparison with other model estimates, particularly when trained with observations. When considering all the validation data, root mean square errors (RMSEs) of 18.76, 23.55, and 22.08 W·m –2 are obtained for MARS, operational LSA-SAF, and ERA5, respectively. The added value of using the satellite cloud information is accessed by comparing with estimates driven by ERA5 total cloud cover, showing an increase of 17% of the RMSE. The consistency of MARS estimate is also tested against an independent dataset of 52 ground stations (from FLUXNET2015), further supporting the good performance of the proposed model.

54 ENVIRONMENTAL SCIENCES↗

Prediction and Annotation of Plant Protein Interaction Networks

Large-scale experimental studies of interactions between components of biological systems have been performed for a variety of eukaryotic organisms. However, there is a dearth of such data for plants. Computational methods for prediction of relationships between proteins, primarily based on comparative genomics, provide a useful systems-level view of cellular functioning and can be used to extend information about other eukaryotes to plants. We have predicted networks for Arabidopsis thaliana, Oryza sativa indica and japonica and several plant pathogens using the Bioverse (http://bioverse.compbio.washington.edu) and show that they are similar to experimentally-derived interaction networks. Predicted interaction networks for plants can be used to provide novel functional annotations and predictions about plant phenotypes and aid in rational engineering of biosynthesis pathways.

59 BASIC BIOLOGICAL SCIENCES↗

Knowledge-Informed Uncertainty-Aware Machine Learning for Time Series Forecasting of Dynamical Engineered Systems

The high complexity and multiscale nature of many engineered systems—such as those in nuclear power plants—make representing and forecasting their dynamic behavior challenging. Physics-based models can be overly complex and computationally intractable, whereas machine learning (ML) tools are often data-hungry and prone to unphysical solutions. This study proposes a knowledge-informed ML-aided hybrid residual modeling approach that offers accurate and efficient time series forecasting for the operation of dynamical engineered systems. Hybrid residual modeling entails a baseline solution from domain knowledge and known physics expressions about the system dynamics integrated with an ML model to capture undiscovered information from the mismatch (i.e., residuals) between true states from measurements and baseline-predicted outputs. This study further quantifies the ML model uncertainty to provide trustworthy solutions. Real-time operational data from thermal-hydraulic flow loops of the cryogenic moderator system in Oak Ridge National Laboratory’s Spallation Neutron Source facility were used to demonstrate the potential of knowledge-informed uncertainty-aware ML in real-world applications. The state variables of the cryogenic helium loop were modeled with (1) first principles–based system identification (sysID), (2) long short-term memory (LSTM) neural network, and (3) hybrid sysID (baseline) + LSTM (residual). The superior predictive capability of the sysID+LSTM model versus stand-alone sysID and LSTM is confirmed by average performance metrics and individual data points across different prediction horizons. By creating a robust representation of the underlying physical system, the widely applicable hybrid residual modeling approach will enable the future development of digital twins for performance prediction, prognostics, and operation control.

Zhao, Xingang↗

Time Series Foundation Models and Deep Learning Architectures for Earthquake Temporal and Spatial Nowcasting

Advancing the capabilities of earthquake nowcasting, the real-time forecasting of seismic activities, remains crucial for reducing casualties. This multifaceted challenge has recently gained attention within the deep learning domain, facilitated by the availability of extensive earthquake datasets. Despite significant advancements, the existing literature on earthquake nowcasting lacks comprehensive evaluations of pre-trained foundation models and modern deep learning architectures; each focuses on a different aspect of data, such as spatial relationships, temporal patterns, and multi-scale dependencies. This paper addresses the mentioned gap by analyzing different architectures and introducing two innovative approaches called Multi Foundation Quake and GNNCoder. We formulate earthquake nowcasting as a time series forecasting problem for the next 14 days within 0.1-degree spatial bins in Southern California. Earthquake time series are generated using the logarithm energy released by quakes, spanning 1986 to 2024. Our comprehensive evaluations demonstrate that our introduced models outperform other custom architectures by effectively capturing temporal-spatial relationships inherent in seismic data. The performance of existing foundation models varies significantly based on the pre-training datasets, emphasizing the need for careful dataset selection. However, we introduce a novel method, Multi Foundation Quake, that achieves the best overall performance by combining a bespoke pattern with Foundation model results handled as auxiliary streams.

97 MATHEMATICS AND COMPUTING↗

Analyzing inference workloads for spatiotemporal modeling

Ensuring power grid resiliency, forecasting climate conditions, and optimization of transportation infrastructure are some of the many application areas where data is collected in both space and time. Spatiotemporal modeling is about modeling those patterns for forecasting future trends and carrying out critical decision-making by leveraging machine learning/deep learning. Once trained offline, field deployment of trained models for near real-time inference could be challenging because performance can vary significantly depending on the environment, available compute resources and tolerance to ambiguity in results. Users deploying spatiotemporal models for solving complex problems can benefit from analytical studies considering a plethora of system adaptations to understand the associated performance-quality trade-offs. To facilitate the co-design of next-generation hardware architectures for field deployment of trained models, it is critical to characterize the workloads of these deep learning (DL) applications during inference and assess their computational patterns at different levels of the execution stack. In this paper, we develop several variants of deep learning applications that use spatiotemporal data from dynamical systems. We study the associated computational patterns for inference workloads at different levels, considering relevant models (Long short-term Memory, Convolutional Neural Network and Spatio-Temporal Graph Convolution Network), DL frameworks (Tensorflow and PyTorch), precision (FP16, FP32, AMP, INT16 and INT8), inference runtime (ONNX and AI Template), post-training quantization (TensorRT) and platforms (Nvidia DGX A100 and Sambanova SN10 RDU). Overall, our findings indicate that although there is potential in mixed-precision models and post-training quantization for spatiotemporal modeling, extracting efficiency from contemporary GPU systems might be challenging. Instead, co-designing custom accelerators by leveraging optimized High Level Synthesis frameworks (such as SODA High-Level Synthesizer for customized FPGA/ASIC targets) can make workload-specific adjustments to enhance the efficiency.

97 MATHEMATICS AND COMPUTING↗

A species’ response to spatial climatic variation does not predict its response to climate change

The dominant paradigm for assessing ecological responses to climate change assumes that future states of individuals and populations can be predicted by current, species-wide performance variation across spatial climatic gradients. However, if the fates of ecological systems are better predicted by past responses to in situ climatic variation through time, this current analytical paradigm may be severely misleading. Empirically testing whether spatial or temporal climate responses better predict how species respond to climate change has been elusive, largely due to restrictive data requirements. Here, we leverage a newly collected network of ponderosa pine tree-ring time series to test whether statistically inferred responses to spatial versus temporal climatic variation better predict how trees have responded to recent climate change. When compared to observed tree growth responses to climate change since 1980, predictions derived from spatial climatic variation were wrong in both magnitude and direction. This was not the case for predictions derived from climatic variation through time, which were able to replicate observed responses well. Future climate scenarios through the end of the 21st century exacerbated these disparities. These results suggest that the currently dominant paradigm of forecasting the ecological impacts of climate change based on spatial climatic variation may be severely misleading over decadal to centennial timescales.

54 ENVIRONMENTAL SCIENCES↗

PERSIANN Dynamic Infrared–Rain Rate Model (PDIR) for High-Resolution, Real-Time Satellite Precipitation Estimation

Precipitation measurements with high spatiotemporal resolution are a vital input for hydrometeorological and water resources studies; decision-making in disaster management; and weather, climate, and hydrological forecasting. Moreover, real-time precipitation estimation with high precision is pivotal for the monitoring and managing of catastrophic hydroclimate disasters such as flash floods, which frequently transpire after extreme rainfall. While algorithms that exclusively use satellite infrared data as input are attractive owing to their rich spatiotemporal resolution and near-instantaneous availability, their sole reliance on cloud-top brightness temperature (T b ) readings causes underestimates in wet regions and overestimates in dry regions—this is especially evident over the western contiguous United States (CONUS). We introduce an algorithm, the Precipitation Estimations from Remotely Sensed Information Using Artificial Neural Networks (PERSIANN) Dynamic Infrared–Rain rate model (PDIR), which utilizes climatological data to construct a dynamic (i.e., laterally shifting) T b –rain rate relationship that has several notable advantages over other quantitative precipitation-estimation algorithms and noteworthy skill over the western CONUS. Validation of PDIR over the western CONUS shows a promising degree of skill, notably at the annual scale, where it performs well in comparison to other satellite-based products. Analysis of two extreme landfalling atmospheric rivers show that solely IR-based PDIR performs reasonably well compared to other IR- and PMW-based satellite rainfall products, marking its potential to be effective in real-time monitoring of extreme storms. This research suggests that IR-based algorithms that contain the spatiotemporal richness and near-instantaneous availability needed for rapid natural hazards response may soon contain the skill needed for hydrologic and water resource applications.

54 ENVIRONMENTAL SCIENCES↗

A Data-Driven Multi-Period Importance Sampling Strategy for Stochastic Economic Dispatch

Power systems with high penetrations of renewable energy (e.g., wind power) require sophisticated approaches to optimize system performance due to uncertainty in short-term system generation capacity. In this paper, we combine a data-driven analog scenario selection method with importance sampling to create a novel scenario construction approach for two-stage stochastic economic dispatch problems with a large number of wind farms on a network. The proposed method produces scenarios with realistic physics by finding high-fidelity analogs that can describe future states of the system. We show how to extend this method to multi-period operations and demonstrate the effectiveness of this technique by simulating economic dispatch operations on a synthetic test system over the course of a week.

data-driven forecasting↗

PMU Placement for Enhancing Dynamic Observability of a Power Grid

Power grids are operated in an increasingly complicated environment. However, operators lack effective and accurate tools for real-time monitoring and control of power systems. The U.S. Department of Energy, along with several utilities and system operators, is making a major $108 million investment in the Western Interconnection for phasor measurement unit (PMU) installation and phasor application development. This phasor measurement network opens up many opportunities for the estimation and prediction of power system states in real time, which enable operators to evaluate the system dynamic security in advance and allow them more time to respond to disturbances. Kalman filter based dynamic state estimation offers a solution suitable for this purpose. Our work indicates that the performance of Kalman filters in dynamic state estimation would degrade if PMU measurements cannot adequately capture the system dynamics. This paper develops a framework to identify how to place PMUs to improve dynamic observability of the power grid. Simulation results validate the concept, and the guidelines for PMU placement are derived.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Catalytic Resonance Theory: Forecasting the Flow of Programmable Catalytic Loops

Chemical transformations on catalyst surfaces occur through series and parallel reaction pathways. These complex networks and their behavior can be most simply evaluated through a three-species surface reaction loop (A* to B* to C* to A*) that is internal to the overall chemical reaction. Application of an oscillating dynamic catalyst to this reactive loop has been shown to exhibit one of three types of behavior: (1) a positive net flux of molecules about the loop in the clockwise direction, (2) a negative net flux of molecules about the loop in the counterclockwise direction, or (3) negligible flux of molecules about the loop at the limit cycle of reaction. Three-species surface loops were simulated with microkinetic modeling to assess the reaction loop behavior resulting from a catalytic surface oscillating between two or more catalyst surface energy states. Selected input parameters for the simulations spanned an 11-dimensional parameter space using 127 688 different parameter combinations. Their converged limit cycle solutions were analyzed for their loop turnover frequencies, the majority of which were found to be approximately zero. Classification and regression machine learning models were trained to predict the sign and magnitude of the loop turnover frequency and successfully performed above accessible baselines. Notably, the classification models exhibited a baseline weighted F1 score of 0.49, whereas trained models achieved weighted F1 scores of 0.94 and 0.96 when trained on the parameters used to define the simulations and derived rate constants, respectively. The trained models successfully predicted catalytic loop behavior, and interpretation of these models revealed all input parameters to be important for the prediction and performance of each model.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Modeling the 4D discharge of lithium-ion batteries with a multiscale time-dependent deep learning framework

The lithium-ion battery (LIB) field is moving towards the direction of investigating spatially resolved physical phenomena in the 3D porous microstructure of electrodes. These pore-scale simulations give new insights into the local dynamics of lithiation/de-lithiation and charge transport, Nevertheless, the computational time of these simulations limits the integration of these models in optimization workflows of cycling conditions or electrode manufacturing processes. Machine learning models present a way of assessing in real-time the performance of materials. While several successful techniques for replicating simulations with machine learning have been proposed, this case study presents a more demanding problem, due to the necessity of understanding the behavior of heterogeneous 3D local data, as it evolves in time: this poses both a scientific and a technical challenge. To this end, we propose an autoregressive multiscale convolutional neural network model to predict relevant quantities at the pore-scale in the solid phase: the lithium concentration (in the active material) and potential (in the active material and carbon binder). Here, these are ultimately used to reconstruct the battery discharge curve. 3D images of the electrode microstructures are the input to the network, trained with a dataset of finite element method simulations to predict the discharge behavior of the cathode side in lithium ion batteries. We propose this machine learning model as a proof-of-concept of the applicability of multiscale networks for time-dependent physics problems. The trained model exhibits very high accuracy (with errors lower than 2 %) in forecasting the discharge behavior of new unseen cathodes.

25 ENERGY STORAGE↗