Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Traditional Machine Learning Models”

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 55 records · Page 3
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Machine Learning-Based Regression Models for Ironmaking Blast Furnace Automation

Computational fluid dynamics (CFD)-based simulation has been the traditional way to model complex industrial systems and processes. One very large and complex industrial system that has benefited from CFD-based simulations is the steel blast furnace system. The problem with the CFD-based simulation approach is that it tends to be very slow for generating data. The CFD-only approach may not be fast enough for use in real-time decisionmaking. To address this issue, in this work, the authors propose the use of machine learning techniques to train and test models based on data generated via CFD simulation. Regression models based on neural networks are compared with tree-boosting models. In particular, several areas (tuyere, raceway, and shaft) of the blast furnace are modeled using these approaches. The results of the model training and testing are presented and discussed. The obtained R 2 metrics are, in general, very high. The results appear promising and may help to improve the efficiency of operator and process engineer decisionmaking when running a blast furnace.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Machine learning models for rat multigeneration reproductive toxicity prediction

Reproductive toxicity is one of the prominent endpoints in the risk assessment of environmental and industrial chemicals. Due to the complexity of the reproductive system, traditional reproductive toxicity testing in animals, especially guideline multigeneration reproductive toxicity studies, take a long time and are expensive. Therefore, machine learning, as a promising alternative approach, should be considered when evaluating the reproductive toxicity of chemicals. We curated rat multigeneration reproductive toxicity testing data of 275 chemicals from ToxRefDB (Toxicity Reference Database) and developed predictive models using seven machine learning algorithms (decision tree, decision forest, random forest, k-nearest neighbors, support vector machine, linear discriminant analysis, and logistic regression). A consensus model was built based on the seven individual models. An external validation set was curated from the COSMOS database and the literature. The performances of individual and consensus models were evaluated using 500 iterations of 5-fold cross-validations and the external validation data set. The balanced accuracy of the models ranged from 58% to 65% in the 5-fold cross-validations and 45%–61% in the external validations. Prediction confidence analysis was conducted to provide additional information for more appropriate applications of the developed models. The impact of our findings is in increasing confidence in machine learning models. We demonstrate the importance of using consensus models for harnessing the benefits of multiple machine learning models (i.e., using redundant systems to check validity of outcomes). While we continue to build upon the models to better characterize weak toxicants, there is current utility in saving resources by being able to screen out strong reproductive toxicants before investing in vivo testing. The modeling approach (machine learning models) is offered for assessing the rat multigeneration reproductive toxicity of chemicals. Our results suggest that machine learning may be a promising alternative approach to evaluate the potential reproductive toxicity of chemicals.

consensus model↗
DOE OSTI2022

Supplementary Data for "Evaluation of the Economic Implications of Varied Pressure Drawdown Strategies Generated Using a Real-time, Rapid Predictive, Multi-fidelity Model for Unconventional Oil and Gas Wells" by Bello, K., Vikara, D., Sheriff, A., Viswanathan, H., Carr, T., Sweeney, M., O'Malley, D., Marquis, M., Vactor, R.T., and Cunha, L.

The Bello et al. study evaluates the impact of contrasting pressure drawdown on gas productivity and the resulting economics of a well in the Marcellus Shale of the Appalachian Basin. This research applies a techno-economic analysis approach to help identify potential ways pressure management strategies can be used to improve cumulative recovery of hydraulically fractured horizontal wells while maintaining project profitability. Gas production forecast outlook scenarios of the Marcellus Shale Energy and Environment Laboratory Laboratory's MIP-3H well were generated under varying pressure drawdowns using two approaches: 1) a novel physics-informed machine learning (PIML) workflow and 2) via traditional reservoir simulation in Computer Modeling Group’s (CMG) GEM Compositional & Unconventional Simulator. Cash flow and other economic metrics of interest were compiled on the production outlook using the U.S. Department of Energy's (DOE) National Energy Technology Laboratory (NETL) Unconventional Shale Well Economic Model (UShWEM).The sheets within this Microsoft ExcelTM workbook provide the economic metric outputs for the baseline condition and the one-at-a-time (OAT) sensitivity analysis of UShWEM's input parameters for each of the production scenarios evaluated.

Fracture Network Model↗
DOE OSTI2023

Evaluation of the economic implications of varied pressure drawdown strategies generated using a real-time, rapid predictive, multi-fidelity model for unconventional oil and gas wells

Experience has suggested that pressure maintenance in hydraulically fractured reservoirs via lower, more sustained production drawdowns may offer improved cumulative recovery and overall resource extraction efficiency compared to more rapid drawdown approaches aimed at generating high initial production. However, given the inherent variability of oil and natural gas markets, operators pursue production strategies that maximize profitability over resource extraction efficiency. This study focuses on evaluating the implications of contrasting pressure drawdown strategies on the long-term production and resulting economics for a real, producing unconventional gas well in the Marcellus Shale of the Appalachian Basin using a techno-economic analysis approach. Our research combines elements of well-specific horizontal well design, production forecasting, equipment sizing and capital cost estimation, operating cost estimation, and revenue and tax calculations. Gas production forecast outlook scenarios were generated under varying pressure drawdowns using two approaches: 1) a novel physics-informed machine learning workflow and 2) traditional reservoir simulation. A discounted cash flow model was used to evaluate the resulting economic implications for each drawdown scenario—generating output for exploring the coupled effect of factors like the timing and volume of gas production, prevailing economic and market conditions for natural gas, and overall estimated ultimate recovery on profitability metrics such as internal rate of return and net present value. Results show that there is potential to maximize the cumulative gas produced in the specific case study well by employing a lower pressure drawdown. Conversely, the greatest profitability is achieved using rapid drawdown as signified by a small, specific subset of our outlook scenarios. On an averaging basis, we find that the combinations of highest cumulative producing and most profitable scenarios occur under lower drawdowns with long (>40 years) producing timeframes, but require higher relative gas price and lower discounting considerations. Further, the machine learning predictive outlooking capability proved effective for enabling rapid generation of a multitude of scenario forecasts. As a result, a variety of prominent example cases could be generated to strike the balance of greater productivity and economic return given their associated producing features and economic conditions when compared to similar producing scenarios—critical insight that offers improved decision support for unconventional oil and gas operations.

42 ENGINEERING↗
DOE OSTI2025

Leveraging large language models to automate the identification of healthcare access barriers for veterans

Objective: To develop and evaluate an automated system for identifying healthcare barriers focusing on transportation issues in veterans’ clinical notes using large language models (LLMs) and to assess the impact of different prompting strategies on classification performance and explanation consistency. Methods: We developed a hybrid system combining pattern matching for templated notes with LLM analysis for free-text notes. Using 2000 manually annotated clinical notes, we compared four prompting strategies (dual-role short, dual-role long, analysis-first, analysis-only) across Mistral-7B and Llama-3.1 models. We evaluated classification performance using standard metrics and assessed explanation consistency through embedding similarity analysis. Results: The analysis-first strategy achieved superior performance, with Mistral-7B reaching an F1 score of 0.914, outperforming traditional machine learning approaches (GBM: 0.786, BERT: 0.811). LLMs demonstrated higher explanation consistency within models (mean cosine similarity 0.887–0.908) compared to cross-model similarities (0.767–0.872). Pattern matching successfully handled 6.7% of templated notes deterministically. Mistral-7B showed greater internal consistency but higher abstention rates compared to Llama-3.1. Conclusion: Requiring LLMs to analyze evidence before classification improves both accuracy and explanation consistency for identifying transportation barriers in clinical notes. This approach enables automated barrier detection at scale while providing clinically relevant explanations, supporting both population-level healthcare planning and individual patient care decisions.

Healthcare access barriers↗
DOE OSTI2023

The entropy of galaxy spectra: how much information is encoded?

Abstract The inverse problem of extracting the stellar population content of galaxy spectra is analysed here from a basic standpoint based on information theory. By interpreting spectra as probability distribution functions, we find that galaxy spectra have high entropy, thus leading to a rather low effective information content. The highest variation in entropy is unsurprisingly found in regions that have been well studied for decades with the conventional approach. We target a set of six spectral regions that show the highest variation in entropy – the 4000 Å break being the most informative one. As a test case with real data, we measure the entropy of a set of high-quality spectra from the Sloan Digital Sky Survey, and contrast entropy-based results with the traditional method based on line strengths. The data are classified into star-forming (SF), quiescent (Q), and active galactic nucleus (AGN) galaxies, and show – independently of any physical model – that AGN spectra can be interpreted as a transition between SF and Q galaxies, with SF galaxies featuring a more diverse variation in entropy. The high level of entanglement complicates the determination of population parameters in a robust, unbiased way, and affects traditional methods that compare models with observations, as well as machine learning (especially deep learning) algorithms that rely on the statistical properties of the data to assess the variations among spectra. Entropy provides a new avenue to improve population synthesis models so that they give a more faithful representation of real galaxy spectra.

Ferreras, Ignacio (ORCID:0000000345843127)↗
DOE OSTI2022

A Comparative Study of Machine Learning Algorithms for Industry-Specific Freight Generation Model

According to Bureau of Transportation Statistics, the U.S. transportation system handled 14,329 million ton-miles of freight per day in 2020. Understanding the generation of these freight shipments is crucial for transportation researchers, planners, and policymakers to design and plan for a more efficient and connected freight transportation system. Traditionally, the freight generation modeling has been based on Ordinary Least Square (OLS) regression, although more advanced Machine Learning (ML) algorithms have been evaluated and proven to have excellent performance in various transportation applications in recent years. Furthermore, one modeling approach applied for one industry might not always be applicable for another as their freight generation logics can be quite different. The objective of this study is to apply and evaluate alternative ML algorithms in the estimation of freight generation for each of 45 industry types. Seven alternative ML algorithms, along with the base OLS regression, were evaluated and compared. In addition, the study considered different combinations of variables in both the original and logarithmic form as well as hyperparameters of those ML algorithms in the model selection for each industry type. The results showed statistically significant improvements in the root mean square error reduction by the alternative ML algorithms over the OLS for over 80% of cases. The study suggests utilizing the alternative ML algorithms can reduce the root mean square error by about 30%, depending on industry types.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Inferring Plant Acclimation and Improving Model Generalizability With Differentiable Physics‐Informed Machine Learning of Photosynthesis

Net photosynthesis (A N ) is a key component of the global carbon cycle influencing climate feedback over decadal scales. Although plant acclimation to environmental changes can modify A N , traditional vegetation models in Earth system models (ESMs) often rely on plant functional type (PFT)-specific parameterizations or simplified acclimation assumptions limiting generalizability across time, space, and PFTs. In this study, we developed a differentiable photosynthesis model to learn the environmental dependencies of V c,max25 (maximum carboxylation rate at 25°C, representing photosynthetic capacity), as this genre of hybrid physics-informed machine learning can seamlessly train neural networks and process-based equations together. Compared to PFT-specific parameterization of V c,max25 , learning the environment dependencies of key photosynthetic parameters improved model spatiotemporal generalizability. Applying environmental acclimation to V c,max25 led to substantial variations in global mean A N indicating the need to address acclimation in ESMs. The model effectively captured multivariate observations (V c,max25 , A N , and stomatal conductance (g s )) simultaneously with multivariate constraints, improving generalization across space and PFTs. It also learned sensible acclimation relationships of V c,max25 to different environmental conditions. The model explained more than 54%, 57%, and 62% of the variance of A N , g s , and V c,max25 , respectively, presenting a first global-scale spatial test benchmark of A N and g s . These results highlight the potential for differentiable modeling to enhance process-based modules in ESMs and effectively leverage information from large, multivariate data sets.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Huge ensembles – Part 1: Design of ensemble weather forecasts using spherical Fourier neural operators

Abstract. Simulating low-likelihood high-impact extreme weather events in a warming world is a significant and challenging task for current ensemble forecasting systems. While these systems presently use up to 100 members, larger ensembles could enrich the sampling of internal variability. They may capture the long tails associated with climate hazards better than traditional ensemble sizes. Due to computational constraints, it is infeasible to generate huge ensembles (comprised of 1000–10 000 members) with traditional, physics-based numerical models. In this two-part paper, we replace traditional numerical simulations with machine learning (ML) to generate hindcasts of huge ensembles. In Part 1, we construct an ensemble weather forecasting system based on spherical Fourier neural operators (SFNOs), and we discuss important design decisions for constructing such an ensemble. The ensemble represents model uncertainty through perturbed-parameter techniques, and it represents initial condition uncertainty through bred vectors, which sample the fastest-growing modes of the forecast. Using the European Centre for Medium-Range Weather Forecasts Integrated Forecasting System (IFS) as a baseline, we develop an evaluation pipeline composed of mean, spectral, and extreme diagnostics. With large-scale, distributed SFNOs with 1.1 billion learned parameters, we achieve calibrated probabilistic forecasts. As the trajectories of the individual members diverge, the ML ensemble mean spectra degrade with lead time, consistent with physical expectations. However, the individual ensemble members' spectra stay constant with lead time. Therefore, these members simulate realistic weather states during the rollout, and the ML ensemble passes a crucial spectral test in the literature. The IFS and ML ensembles have similar extreme forecast indices, and we show that the ML extreme weather forecasts are reliable and discriminating. These diagnostics ensure that the ensemble can reliably simulate the time evolution of the atmosphere, including low-likelihood high-impact extremes. In Part 2, we generate a huge ensemble initialized each day in summer 2023, and we characterize the simulations of extremes.

Mahesh, Ankur↗
DOE OSTI2021

Machine Learning Tools Set for Natural Gas Fuel Cell System Design

This study is focusing on leveraging the system design tools set for the next-generation solid oxide fuel cell (SOFC) based natural gas fuel cell (NGFC) system. Conventionally, system design and optimization of NGFC systems rely heavily on traditional reduced order model (ROM) techniques and designers’ experience level. For overcoming the technical barriers of system design, multiple multi-physics models and machine learning (ML) tools have been utilized to automate the conceptual design process and enhance the reliability of solutions for the NGFC system. The proposed tools set includes a physics-informed ML tool for automated ROM construction that leverages advances in deep neural networks to significantly reduce ROM prediction error for the NGFC power island compared to traditional approaches. The constructed physics-informed ML ROM can be used in system design, and optimization tools set Institute for the Design of Advanced Energy Systems (IDAES) Process Systems Engineering (PSE) framework. The tools set also provides a user-friendly graphic user interface built within Jupyter Notebooks, and the complete tools set is open-source public available.

Wang, Dewei↗
DOE OSTI2025

Machine learning models for segmentation and classification of cyanobacterial cells

Abstract Timelapse microscopy has recently been employed to study the metabolism and physiology of cyanobacteria at the single-cell level. However, the identification of individual cells in brightfield images remains a significant challenge. Traditional intensity-based segmentation algorithms perform poorly when identifying individual cells in dense colonies due to a lack of contrast between neighboring cells. Here, we describe a newly developed software package called Cypose which uses machine learning (ML) models to solve two specific tasks: segmentation of individual cyanobacterial cells, and classification of cellular phenotypes. The segmentation models are based on the Cellpose framework, while classification is performed using a convolutional neural network named Cyclass. To our knowledge, these are the first developed ML-based models for cyanobacteria segmentation and classification. When compared to other methods, our segmentation models showed improved performance and were able to segment cells with varied morphological phenotypes, as well as differentiate between live and lysed cells. We also found that our models were robust to imaging artifacts, such as dust and cell debris. Additionally, the classification model was able to identify different cellular phenotypes using only images as input. Together, these models improve cell segmentation accuracy and enable high-throughput analysis of dense cyanobacterial colonies and filamentous cyanobacteria.

Huffine, Clair A.↗
DOE OSTI2024

Power System Frequency Dynamics Modeling, State Estimation, and Control using Neural Ordinary Differential Equations (NODEs) and Soft Actor-Critic (SAC) Machine Learning Approaches

With the global energy transition of the electric power system, grid control, supervision, and protection is becoming more challenging. With the increasing integration of renewable energy sources (RES), the system dynamics are changing, causing traditional power system dynamic modeling with swing equation-based modeling approaches to fail. Additionally, the converter-dominated power grid is decreasing the system inertia, making the power system more fragile to the frequency swings. This paper first investigates and compares the application of a model-based Kalman filter state estimation approach with (i) a model-free machine learning approach --- neural ordinary differential equations (NODEs) --- and (ii) a data-driven system identification (SysId) approach to model and infer critical state values of the power system frequency dynamics. Then a model predictive control (MPC) framework is compared to a model-free Soft Actor-Critic (SAC) reinforcement learning (RL) control algorithm in providing efficient fast frequency response (FFR) to the power system frequency dynamics. The approaches are compared in terms of their performance goals as well as their per-timestep computational efficiency. Furthermore, the comparative study for state estimation shows that for the model-free requirement, both NODEs and SysId can provide accurate state estimates; however, with increasing model complexity, NODEs can be a better choice for model identification. Similarly, the results from the FFR comparative study show that the SAC RL-based FFR, once trained, outperforms MPC with better control signals and faster computation time, making the SAC RL-based FFR better option for providing FFR to the power system.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

sPHENIX heavy flavor jet tagging studies in p+p at $\sqrt{s_{NN}}=200~GeV$

Heavy-flavor jets, which are initiated from heavy quarks, are ideal probes for studying flavor dependent parton energy loss. We report on the performance of jet flavor tagging using two Neural Network Machine Learning (ML) models: the Long Short-Term Memory (LSTM) model and an Attention-based Neural Network, in simulations of 200 GeV p + p collisions. The tagging performance of bottom quark initiated jets with both ML models surpasses that of the traditional cut-based method. Technical details, including sample and kinematic variable selections, the machine learning training and testing setup with parameter tuning, and outcome comparisons, will be discussed.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2023

Temporal Subsampling Diminishes Small Spatial Scales in Recurrent Neural Network Emulators of Geophysical Turbulence

The immense computational cost of traditional numerical weather and climate models has sparked the development of machine learning (ML) based emulators. Because ML methods benefit from long records of training data, it is common to use data sets that are temporally subsampled relative to the time steps required for the numerical integration of differential equations. Here, we investigate how this often overlooked processing step affects the quality of an emulator's predictions. We implement two ML architectures from a class of methods called reservoir computing: (a) a form of Nonlinear Vector Autoregression (NVAR), and (b) an Echo State Network (ESN). Despite their simplicity, it is well documented that these architectures excel at predicting low dimensional chaotic dynamics. We are therefore motivated to test these architectures in an idealized setting of predicting high dimensional geophysical turbulence as represented by Surface Quasi-Geostrophic dynamics. In all cases, subsampling the training data consistently leads to an increased bias at small spatial scales that resembles numerical diffusion. Interestingly, the NVAR architecture becomes unstable when the temporal resolution is increased, indicating that the polynomial based interactions are insufficient at capturing the detailed nonlinearities of the turbulent flow. The ESN architecture is found to be more robust, suggesting a benefit to the more expensive but more general structure. Spectral errors are reduced by including a penalty on the kinetic energy density spectrum during training, although the subsampling related errors persist. Future work is warranted to understand how the temporal resolution of training data affects other ML architectures.

58 GEOSCIENCES↗
DOE OSTI2024

Geographical Insights into Suicide Mortality Through Spatial Machine Learning

Suicide mortality is a leading cause of death in the United States, with an upward trend that emphasizes its significance as a public health issue. Previous research has employed global models like ordinary least squares (OLS) regression and local models such as geographically weighted regression (GWR). While local models are useful for analyzing spatial variations in suicide mortality, they share limitations with traditional global models, particularly about their inability to handle multi-collinearity and non-linear relationships. Machine learning approaches, like random forests (RF), can address some of these limitations but often fail to account for spatial variability. This gap highlights the need for spatial ML models specifically designed to tackle suicide mortality. This research seeks to fill this void by using a geographically weighted random forest model (GWRF) to examine the associations between county-level suicide mortality in the U.S. from 2010 to 2020 and various social and environmental determinants of health. A key aspect of our methodology is disciplined feature selection, which reduces the pool of explanatory variables by about 90%. This refinement enhances the explanatory power of both global (R2 improved from 0.59 to 0.67) and local (R2 improved from 0.64 to 0.67) RF models while reducing their run times. An analysis of the importance scores for these selected features reveals that the drivers of suicide mortality vary by context. Thus, to effectively address regional disparities and inform targeted public health interventions, a holistic approach that incorporates multiple county-level characteristics is essential.

Lebakula, Viswadeep [ORNL] (ORCID:0000000152935914↗
DOE OSTI2024

Generative models for simulation of KamLAND-Zen

Abstract The next generation of searches for neutrinoless double beta decay ($$0 \nu \beta \beta $$ 0 ν β β ) are poised to answer deep questions on the nature of neutrinos and the source of the Universe’s matter–antimatter asymmetry. They will be looking for event rates of less than one event per ton of instrumented isotope per year. To claim discovery, accurate and efficient simulations of detector events that mimic$$0 \nu \beta \beta $$ 0 ν β β is critical. Traditional Monte Carlo (MC) simulations can be supplemented by machine-learning-based generative models. This work describes the performance of generative models that we designed for monolithic liquid scintillator detectors like KamLAND to produce accurate simulation data without a predefined physics model. We present their current ability to recover low-level features and perform interpolation. In the future, the results of these generative models can be used to improve event classification and background rejection by providing high-quality abundant generated data.

Physics↗
DOE OSTI2022

Neural simulation-based inference approach for characterizing the Galactic Center γ-ray excess

The nature of the Fermi γ-ray Galactic Center Excess (GCE) has remained a persistent mystery for over a decade. Although the excess is broadly compatible with emission expected due to dark matter annihilation, an explanation in terms of a population of unresolved astrophysical point sources e.g., millisecond pulsars, remains viable. The effort to uncover the origin of the GCE is hampered in particular by an incomplete understanding of diffuse emission of Galactic origin. This can lead to spurious features that make it difficult to robustly differentiate smooth emission, as expected for a dark matter origin, from more “clumpy” emission expected from a population of relatively bright, unresolved point sources. Here, we use recent advancements in the field of simulation-based inference, in particular density estimation techniques using normalizing flows, in order to characterize the contribution of modeled components, including unresolved point source populations, to the GCE. Compared to traditional techniques based on the statistical distribution of photon counts, our machine learning-based method is able to utilize more of the information contained in a given model of the Galactic Center emission, and in particular can perform posterior parameter estimation while accounting for pixel-to-pixel spatial correlations in the γ-ray map. This makes the method demonstrably more resilient to certain forms of model misspecification. On application to Fermi data, the method generically attributes a smaller fraction of the GCE flux to unresolved point sources when compared to traditional approaches. We nevertheless infer such a contribution to make up a non-negligible fraction of the GCE across all analysis variations considered, with at least $38^{+9}_{-19}$% of the excess attributed to unresolved point sources in our baseline analysis.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Application of Modified Meshgraphnets for Subsurface Prediction during CO2 Sequestration

In the face of the increasingly dire consequences of anthropogenic climate change, capturing and storing carbon dioxide is paramount. However, several impediments exist to the safe and effective subsurface storage of CO2, such as cost of transport, identification of suitable sites for subsurface storage, and assessment of long-term risk from storage in subsurface aquifers. Accurate subsurface modeling is necessary to ensure that CO2 storage is both safe and effective. Still, such modeling has traditionally required either substantial time and computational power (numerical simulation) or a substantial amount of pre-existing data for training (machine learning models). Additionally, these models lack flexibility in dealing with both changes in discretization of the input data and generalizability beyond the data on which they are trained. In order to address these issues, this research applies graph neural networks (GNNs) to predict subsurface saturation and pressure during CO₂ injection in a model of the Illinois Basin-Decatur Project (IBDP). GNNs provide a flexible, intuitive method for representing and manipulating complex unstructured data, which is often found in many practical domain problems such as fluid flow and subsurface characterization. These unstructured grids are easily represented in GNNs by representing spatially-localized features such as permeability, porosity, saturation, and pressure as nodes in a graph and relationships between these properties as edges connecting these nodes. This research applies a specific GNN model called MeshGraphNets (MGN) to model the change in CO2 saturation and pressure over a 50-month time period (36 months of injection, 14 months post-injection). The MGN model leverages a message passing process that allows the network to learn both the spatial and temporal dynamics of this system simultaneously. Additionally, training on a limited dataset (64 realizations, 20 time points each) resulted in a high degree of accuracy in saturation prediction both within the same timeframe as the training (20 months, 0.039 average RMSE) and when projecting out to the end of injection (36 months, 0.053 average RMSE). Temporal predictions such as those generated by MGNs and other similar models are prone to accumulated error over time; in order to address this, a multi-step rollout (MSR) training process was applied to calculate training loss. This method mimics the forward prediction during inference by “rolling out” multiple time points in a single training step using the previous prediction as input to the MGN model. By calculating the loss several time steps forward from the current prediction, the model is forced to find a more stable state over time. Application of MSR to the MGN model resulted in an average 15% reduction in inference error over time during forward prediction. This study showcases the immense potential of GNNs as a game-changing methodology for predicting pressure and saturation evolution in CCS projects, ultimately paving the way for more sustainable and effective carbon storage solutions. Presentation prepared for the 2024 AiChE Annual Meeting, October 27 to November 1 2024, San Diego, CA.

Holcomb, Paul↗