Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “data driven optimization”

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 757 records · Page 42
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

Athena - Shuttle Bus Optimization and Event Driven Simulator [SWR 20-105]

The primary purpose of the software is to optimize shuttle routes to and from terminals, given some airport campus location like the rental car center. This code was developed as part of the DOE funded Athena project. One aspect of the Athena project was to investigate if the optimization of shuttle routes that move passengers to and from the five DFW terminals and the rental car center could result in lower energy consumption from the shuttle fleet while still maintaining an acceptable level of service to passengers. Hence, the optimization done by the code determines a set of shuttle routes, the type of shuttle on each route, and the number of shuttles servicing each route. The optimization aims to make these choices in a way that the energy consumption per hour by the fleet is minimized. Additionally, the software can fast simulate airport shuttle operations with collected data. The simulation inputs are passenger arrival rates, shuttle routes and frequencies, and simulation configurations. The outputs include: time dependent shuttle energy level, time dependent charging station usage, time dependent number of passengers on each bus, time dependent number of passengers at each bus stop, history of number of passengers left at each bus stop, route history of fleet shuttle buses and on-demand buses, and time dependent bus distance traveled.

Kelly, Kenneth↗
DOE OSTI2020

Simulation-driven optimization of high-order meshes in ALE hydrodynamics

Here we propose tools for high-order mesh optimization and demonstrate their benefits in the context of multi-material Arbitrary Lagrangian-Eulerian (ALE) compressible shock hydrodynamic applications. The mesh optimization process is driven by information provided by the simulation which uses the optimized mesh, such as shock positions, material regions, known error estimates, etc. These simulation features are usually represented discretely, for instance, as finite element functions on the Lagrangian mesh. The discrete nature of the input is critical for the practical applicability of the algorithms we propose and distinguishes this work from approaches that strictly require analytical information. Our methods are based on node movement through a high-order extension of the Target-Matrix Optimization Paradigm (TMOP). The proposed formulation is fully algebraic and relies only on local Jacobian matrices, so it is applicable to all types of mesh elements, in 2D and 3D, and any order of the mesh. We discuss the notions of constructing adaptive target matrices and obtaining their derivatives, reconstructing discrete data in intermediate meshes, node limiting that enables improvement of global mesh quality while preserving space-dependent local mesh features, and appropriate normalization of the objective function. The adaptivity methods are combined with automatic ALE triggers that can provide robustness of the mesh evolution and avoid excessive remap procedures. The benefits of the new high-order TMOP technology are illustrated on several simulations performed in the high-order ALE application BLAST.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Explainable Graph Learning for Particle Accelerator Operations

Particle accelerators are vital tools in physics, medicine, and industry, requiring precise tuning to ensure optimal beam performance. However, real-world deviations from idealized simulations make beam tuning a time-consuming and error-prone process. In this work, we propose an explanation-driven framework for providing actionable insight into beamline operations, with a focus on the injector beamline at the Continuous Electron Beam Accelerator Facility (CEBAF). We represent beamline configurations as heterogeneous graphs, where setting nodes represent elements that human operators can actively adjust during beam tuning, and reading nodes passively provide diagnostic feedback. To identify the most influential setting nodes responsible for differences between any two beamline configurations, our approach first predicts the resulting changes in reading nodes caused by variations in settings, and then learns importance scores that capture the joint influence of multiple setting nodes. Experimental results on real-world CEBAF injector data demonstrate the framework’s ability to generate interpretable insights that can assist human operators in beamline tuning and reduce operational overhead.

Wang, Song [Univ. of Virginia, Charlottesville, VA↗
DOE OSTI2021

Incorporating Valve Switching Losses Into a Static Optimal Control Algorithm for the Hybrid Hydraulic-Electric Architecture (HHEA)

The Hybrid Hydraulic-Electric Architecture (HHEA) has, in recent years, been proposed as an energy efficient alternative to conventional load-sensing architectures in mobile machines such as excavators and wheel-loaders. HHEA leverages the advantages of hydraulic power and electric power to eliminate throttling valves while also improving the energy and control performance of the system. The architecture utilizes a set of common pressure rails to provide a majority of power and and a small electric motor driven pump to modulate this power to meet the exact demand. Previous work has developed a computationally efficient Lagrange Multiplier approach for determining the optimal pressure rail selections that minimizes the energy losses in the system. The static model used considers only the energy use for each pressure rail selection but not the losses associated with the valves during the transition. This paper presents an approach to include the switching losses in the model and in the optimization procedure. To capture the switching losses, switching events between different rails and at various input and output flow rates were simulated with consideration of valve spool dynamics. A parameterized model that summarizes the losses is then obtained, allowing switching losses to be added to the previous energy analysis. The performance of the switching loss model was compared with reference data obtained from a high-fidelity simulation model. To incorporate the switching losses into optimal control algorithm, an efficient dynamic programming approach that prevents frequent switching is adopted in place of the Lagrange multiplier approach. The overall effect of switching losses on the energy consumption and optimal control decisions is presented. In general, switching losses contribute to about 9–10% of input energy.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2020

P-model v1.0: an optimality-based light use efficiency model for simulating ecosystem gross primary production

Terrestrial photosynthesis is the basis for vegetation growth and drives the land carbon cycle. Accurately simulating gross primary production (GPP, ecosystem-level apparent photosynthesis) is key for satellite monitoring and Earth system model predictions under climate change. While robust models exist for describing leaf-level photosynthesis, predictions diverge due to uncertain photosynthetic traits and parameters which vary on multiple spatial and temporal scales. Here, we describe and evaluate a GPP (photosynthesis per unit ground area) model, the P-model, that combines the Farquhar–von Caemmerer–Berry model for C 3 photosynthesis with an optimality principle for the carbon assimilation–transpiration trade-off, and predicts a multi-day average light use efficiency (LUE) for any climate and C 3 vegetation type. The model builds on the theory developed in Prentice et al. (2014) and Wang et al. (2017a) and is extended to include low temperature effects on the intrinsic quantum yield and an empirical soil moisture stress factor. The model is forced with site-level data of the fraction of absorbed photosynthetically active radiation (fAPAR) and meteorological data and is evaluated against GPP estimates from a globally distributed network of ecosystem flux measurements. Although the P-model requires relatively few inputs, the R 2 for predicted versus observed GPP based on the full model setup is 0.75 (8 d mean, 126 sites) – similar to comparable satellite-data-driven GPP models but without predefined vegetation-type-specific parameters. The R 2 is reduced to 0.70 when not accounting for the reduction in quantum yield at low temperatures and effects of low soil moisture on LUE. The R 2 for the P-model-predicted LUE is 0.32 (means by site) and 0.48 (means by vegetation type). Applying this model for global-scale simulations yields a total global GPP of 106–122 Pg C yr –1 (mean of 2001–2011), depending on the fAPAR forcing data. The P-model provides a simple but powerful method for predicting – rather than prescribing – light use efficiency and simulating terrestrial photosynthesis across a wide range of conditions. The model is available as an R package (rpmodel).

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

High entropy powering green energy: hydrogen, batteries, electronics, and catalysis

A reformation in energy is underway to replace fossil fuels with renewable sources, driven by the development of new, robust, and multi-functional materials. High-entropy materials (HEMs) have emerged as promising candidates for various green energy applications, having unusual chemistries that give rise to remarkable functionalities. This review examines recent innovations in HEMs, focusing on hydrogen generation/storage, fuel cells, batteries, semiconductors/electronics, and catalysis—where HEMs have demonstrated the ability to outperform state-of-the-art materials. We present new master plots that illustrate the superior performance of HEMs compared to conventional systems for hydrogen generation/storage and heat-to-electricity conversion. We highlight the role of computational methods, such as density functional theory and machine learning, in accelerating the discovery and optimization of HEMs. The review also presents current challenges and proposes future directions for the field. We emphasize the need for continued integration of modeling, data, and experiments to investigate and leverage the underlying mechanisms of the HEMs that are powering progress in sustainable energy.

batteries↗
DOE OSTI2023

Limits on the Range and Rate of Change in Power Take-Off Load in Ocean Wave Energy Conversion: A Study Using Model Predictive Control

Previous work comparing power take-off (PTO) architectures for ocean wave-powered reverse osmosis suggests that variable displacement in the wave energy converter (WEC)-driven pump does not offer a significant performance advantage. A limitation of that study is that the WEC was subject to a constant load within a given sea state (“Coulomb damping”) and did not account for controlled, moment-to-moment variation of the PTO load enabled by a variable displacement pump. This study explores the potential performance advantage of a variable PTO load over Coulomb damping. Model predictive control is used to provide optimal load control with constraints on the PTO load. The constraints include minimum and maximum loads and a limit on the rate of load adjustment. Parameter studies on these constraints enable conclusions about PTO design requirements in addition to providing an estimated performance advantage over Coulomb damping. Numerical simulation of the Oyster 1 WEC is carried out with performance weighted by historical sea state data from Humboldt Bay, CA. The results show a performance advantage of up to 20% higher yearly-average power absorption over Coulomb damping. Additionally, the parameter studies suggest that the PTO load should be adjustable down to at least 25% of the maximum load and should be adjustable between the minimum and maximum loads within a few seconds.

16 TIDAL AND WAVE POWER↗
DOE OSTI2026

Techno-Economic Assessment of Data Center Load Demand Powered by Small Modular Reactors and Distributed Energy Resources

The rapid increase in data center energy demand, driven by AI and large-scale data processing, poses significant challenges to global energy infrastructure. Data centers require substantial and reliable energy for continuous operations and high-performance computing. Current electrical grids face issues such as transmission bottlenecks and aging infrastructure, making it difficult to meet these demands. Integrating inverter-based-resources (IBRs) like solar and wind presents both opportunities and challenges due to their intermittent nature. Small Modular Reactors (SMRs) offer a promising solution with their enhanced safety, modularity, reliability, and scalability, providing consistent base load power ideal for data center operations. This study presents a comprehensive techno-economic assessment of powering data center load demand using a combination of SMRs and IBRs with grid-connected and islanded mode. This study utilized Idaho National Laboratory’s (INL) HPC data center hourly load profiles and Xendee microgrid optimization platform to conduct the analysis. In this configuration, SMRs serves as the primary base load power source, consistently providing a steady supply of electricity necessary to meet the minimum load demand of the data center with support from the IBRs. Key performance indicators such as Levelized Cost of Electricity (LCOE), Net Present Value (NPV) has been calculated to assess the economic feasibility. The findings from this research will underscore the strategic benefits of integrating SMR plant with DERs – particularly for critical infrastructure load such as data centers.

14 - SOLAR ENERGY↗
DOE OSTI2020

In vitro prototyping of limonene biosynthesis using cell-free protein synthesis

Metabolic engineering of microorganisms to produce sustainable chemicals has emerged as an important part of the global bioeconomy. Unfortunately, efforts to design and engineer microbial cell factories are challenging because design-build-test cycles, iterations of re-engineering organisms to test and optimize new sets of enzymes, are slow. To alleviate this challenge, we demonstrate a cell-free approach termed in vitro Prototyping and Rapid Optimization of Biosynthetic Enzymes (or iPROBE). In iPROBE, a large number of pathway combinations can be rapidly built and optimized. The key idea is to use cell-free protein synthesis (CFPS) to manufacture pathway enzymes in separate reactions that are then mixed to modularly assemble multiple, distinct biosynthetic pathways. As a model, we apply our approach to the 9-step heterologous enzyme pathway to limonene in extracts from Escherichia coli. In iterative cycles of design, we studied the impact of 54 enzyme homologs, multiple enzyme levels, and cofactor concentrations on pathway performance. In total, we screened over 150 unique sets of enzymes in 580 unique pathway conditions to increase limonene production in 24 h from 0.2 to 4.5 mM (23–610 mg/L). Finally, to demonstrate the modularity of this pathway, we also synthesized the biofuel precursors pinene and bisabolene. Finally, we anticipate that iPROBE will accelerate design-build-test cycles for metabolic engineering, enabling data-driven multiplexed cell-free methods for testing large combinations of biosynthetic enzymes to inform cellular design.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Enhanced descriptor identification and mechanism understanding for catalytic activity using a data-driven framework: revealing the importance of interactions between elementary steps

We report accurate identification of descriptors for catalytic activities has long been essential to the in-depth understanding of catalysis and recently to set the basis for catalyst screening. However, commonly used methods suffer from low accuracy in predictability. This study reports an enhanced approach to accurately identify the descriptors from a kinetic dataset using a machine learning (ML) surrogate model. CO hydrogenation to methanol over Cu-based catalysts was taken as a case study. Our model captures not only the contribution from individual elementary steps but also the interaction between relevant steps within a reaction network, which was found to be essential for high accuracy. As a result, six effective descriptors are identified, which are accurate enough to ensure the trained gradient boosted regression (GBR) model for good prediction of the methanol turnover frequency (TOF) over metal (M)-doped Cu(111) model surfaces (M = Au, Cu, Pd, Pt, Ni). More importantly, going beyond the purely mathematical ML model, the catalytic role of each identified descriptor can be revealed by using model-agnostic interpretation tools, which enhances the insight into the promoting effect of alloying. The trained GBR model outperforms the conventional derivative-based methods in terms of both the predictability and the mechanism understanding. It opens alternative possibilities toward accurate descriptor-based rational catalyst optimization.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Unlocking the potential: machine learning applications in electrocatalyst design for electrochemical hydrogen energy transformation

Machine learning (ML) is rapidly emerging as a pivotal tool in the hydrogen energy industry for the creation and optimization of electrocatalysts, which enhance key electrochemical reactions like the hydrogen evolution reaction (HER), the oxygen evolution reaction (OER), the hydrogen oxidation reaction (HOR), and the oxygen reduction reaction (ORR). This comprehensive review demonstrates how cutting-edge ML techniques are being leveraged in electrocatalyst design to overcome the time-consuming limitations of traditional approaches. ML methods, using experimental data from high-throughput experiments and computational data from simulations such as density functional theory (DFT), readily identify complex correlations between electrocatalyst performance and key material descriptors. Leveraging its unparalleled speed and accuracy, ML has facilitated the discovery of novel candidates and the improvement of known products through its pattern recognition capabilities. This review aims to provide a tailored breakdown of ML applications in a format that is readily accessible to materials scientists. Hence, we comprehensively organize ML-driven research by commonly studied material types for different electrochemical reactions to illustrate how ML adeptly navigates the complex landscape of descriptors for these scenarios. We further highlight ML's critical role in the future discovery and development of electrocatalysts for hydrogen energy transformation. Potential challenges and gaps to fill within this focused domain are also discussed. As a practical guide, we hope this work will bridge the gap between communities and encourage novel paradigms in electrocatalysis research, aiming for more effective and sustainable energy solutions.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

ParaStell: parametric modeling and neutronics support for stellarator fusion power plants

The three-dimensional variation inherent to stellarator geometries and fusion sources motivates three-dimensional modeling to obtain accurate results from computational modeling in support of design and analysis of first wall, blanket, and shield (FWBS) systems. Manually constructing stellarator fusion power plant geometries in computer-aided design (CAD) and defining the corresponding fusion source can be cumbersome and challenging. The open-source parametric modeling toolset ParaStell has been developed to automate construction of such geometries in low-fidelity. Low-fidelity modeling is useful during the conceptual phase of engineering design as a means of rapidly exploring the design space of a given device. The modeling capability of ParaStell includes in-vessel components and magnets, for any given stellarator configuration, using a parametric definition and plasma equilibrium data. Furthermore, the toolset automates the generation of detailed, tetrahedral neutron source definitions and DAGMC geometries for use in neutronics modeling. ParaStell assists rapid design iteration, parametric study, and design optimization of stellarator fusion cores. As a demonstration of the design iteration capability, the effect of the three-dimensional parameter space on tritium breeding and magnet shielding is investigated, using the WISTELL-D configuration as a design basis. Blanket and shield thicknesses are varied in three dimensions, using the space available between the plasma edge and magnet coils as a constraint. The corresponding effects on tritium breeding ratio and magnet heating are tallied using the open-source Monte Carlo particle transport code OpenMC. The inclusion of additional and higher-fidelity modeling capabilities is planned for ParaStell’s future, as well as its implementation in machine-driven optimization.

Moreno, Connor↗
DOE OSTI2021

Machine Learning Assisted Safety Modeling and Analysis of Advanced Reactors

With the advances in computational power and numerical methods, analysts can now rely on first-principle simulations to predict ultra-fine details in a variety of applications. Advances in machine learning (ML) have produced algorithms that can now learn high-level abstractions via hierarchical models. This project aims to leverage advances in ML techniques and the available high-resolution simulation data to develop a novel modeling and simulation (M\&S) methodology for reactor safety analysis. While application-agnostic ML techniques are available, complex physics constraints need to be incorporated into ML techniques to build ML-based closures for computationally efficient predictive simulations. This project intends to develop a physics-guided data-driven multi-scale methodology for M\&S of advanced reactors. The project focuses on thermal fluid (T/F) phenomena, which play major roles in advanced reactor safety. Specifically, we propose a data-driven coarse-mesh turbulence model based on local flow features for the transient analysis of thermal mixing and stratification in a sodium-cooled fast reactor (SFR). The model has a coarse-mesh setup to ensure computational efficiency, while it is trained by fine-mesh computational fluid dynamics (CFD) data with Reynolds-averaged Navier-Stokes (RANS) turbulence model to ensure accuracy. Three different neural networks are developed and tested for loss-of-flow transients in the hot pool of SFR, i.e. the densely connected convolutional neural network (DCNN), long-short-term-memory network based on proper orthogonal decomposition (POD-LSTM), and the DCNN informed by LSTM (DCNN-LSTM). The performances of these three neural networks are evaluated based on baseline models. The DCNN-LSTM model has been chosen for further hyperparameter optimization. Furthermore, based on a simplified two-dimensional case, uncertainty quantification (UQ) of the developed ML-based closure are investigated with three methods, i.e. Monte Carlo dropout, deep ensemble, and Bayesian neural network. The developed ML-based turbulent viscosity closure relation based on deep ensemble is then integrated into the system analysis module SAM and serves as a term in the conservation equations. Such a SAM-ML based procedure guarantees that the obtained results are consistent with the physical constraints of the thermal-fluid system. The SAM-ML simulation on the same loss-of-flow transient showed comparable accuracy with the CFD simulation but with a much coarser mesh setup. Last but not least, the ML-based closure improvement with the support of higher-fidelity data from large eddy simulation (LES) is discussed. As a first step towards this direction, a baseline LES simulation is performed to obtain comparable data with RANS results. Based on the early results, future investigation on further improving the ML-based closure is discussed. We believe the developed approach that combines scientific machine learning with nuclear system analysis code can benefit the advanced reactor community as more accurate safety analyses will better characterize reactor safety margins and reduce licensing efforts.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2024

Next-Generation Materials Design: Quantum Mechanics and Data-Driven Modeling

The future of materials design is rapidly advancing through the combination of quantum mechanics and data-driven modeling. These approaches integrate quantum principles with advanced data analysis, enabling precise insights into material behavior. This talk will highlight recent progress in using these methods for computational design, particularly in high-entropy alloy catalysts, emphasizing the role of hierarchical machine-learning architectures for accurate predictions. Additionally, I will discuss our work on developing machine learning interatomic potentials (MLPs) for single-element metals, metal oxides, and alloys under extreme conditions, focusing on melting behavior and phase properties at high temperatures and pressures. We have also refined our MLP models to capture dynamic surface interactions, such as CO2 and CO adsorption on MgO, using both static and molecular dynamics simulations. These models maintain high accuracy while significantly reducing computational costs compared to first-principles calculations. By enabling efficient and accurate simulations, this work supports broader community adoption, optimizes datasets for materials discovery, and extends the accessible time, size, and environmental conditions beyond the limits of experiments and traditional simulations.

machine learning↗
DOE OSTI2025

A physics-constrained neural ordinary differential equations approach for robust learning of stiff chemical kinetics

The high computational cost associated with solving for detailed chemistry poses a significant challenge for predictive computational fluid dynamics (CFD) simulations of turbulent reacting flows. While deep learning techniques have been explored to develop faster surrogate models, they often fail to integrate reliably with CFD solvers. This instability arises because traditional deep learning approaches optimize for training error without ensuring compatibility with ordinary differential equation (ODE) solvers, resulting in accumulation of errors over time. Recently, neuralODE (NODE) based approaches have been shown to be a promising technique to emulate and accelerate detailed chemistry computations. Here, in the present work, we extend this NODE framework for stiff chemical kinetics by incorporating mass conservation constraints directly into the loss function during training. This ensures that the total mass as well as the individual elemental species masses are conserved in an a-posteriori manner. Proof-of-concept studies are performed with the novel physics-constrained NODE (PC-NODE) approach for homogeneous autoignition of hydrogen-air mixture over a range of composition and thermodynamic conditions. It is demonstrated that the PC-NODE framework not only improves the physical consistency of the resulting data-driven model with respect to mass conservation criteria, but also improves training efficiency. PC-NODE is shown to achieve 2–100× speedup relative to the hydrogen-air detailed chemical mechanism depending on the type of the ODE solver (implicit or explicit) used during autoregressive inference tests. Lastly, a-posteriori studies are performed wherein the trained PC-NODE model is coupled with a CFD solver. It is shown that higher accuracy is achieved with PC-NODE relative to the purely data-driven NODE approach. Moreover, PC-NODE also exhibits robustness and generalizability to unseen initial conditions from within (interpolative capability) as well as outside (extrapolative capability) the training regime.

computational combustion↗
DOE OSTI2024

Datasets for Widespread Residential Space Heating Electrification in Texas

In this experiment, we explore long term patterns in electricity demand driven by the dual effects of full electrification of space heating in Texas (by adoption of electric heat pumps), and climate change. We use a predictive model of electricity demand, climate projections, and an open source nodal power system (DC Optimal Power Flow) model of the Electric Reliability Council of Texas (ERCOT) system. Heat pumps are a more energy efficient way of providing space heating and cooling in homes. We attempt to exhaustively investigate the impacts of full residential space heating electrification by adoption of heat pumps for the segment of Texas households that currently rely on fossil fuels (about 40%), while simultaneously incorporating climate change meteorological variables. We explore a range of scenarios of heat pump efficiency and climate uncertainty over a long period of future years (2020-2099). In total, the simulation experiment generates 1,280 simulation years of hourly data. We report and analyze results in form of impacts on residential load, total load, peak load, seasonality of peaking, and reliability measured by occurrence and frequency of loss of load events. While the experiment is for ERCOT, the insights and approach can be applied to other regions. The results from the analysis can inform system planners on a range of potential capacity requirements/ reliability implications and/or risks of full space heating electrification via the adoption of electric heat pumps, given the uncertainty in the scenarios/ climate futures. The dataset includes model output for residential, non residential and total load, and the results from the GO ERCOT model runs for 4 RCP Scenarios (RCP 4.5 Cooler, RCP 4.5 Hotter, RCP 8.5 Cooler, RCP 8.5 Hotter), 4 Heating electrification Scenarios (Base , Standard Efficiency HP, High Efficiency HP, Ultra-High Efficiency HP) over 80 years (2020-2099). The metrological variables at BA scale were weighted weighted using population projections consistent with the SSP3 scenario.

Climate Change↗
DOE OSTI2024

Machine Learning-Enabled Image Classification for Automated Electron Microscopy

Abstract Traditionally, materials discovery has been driven more by evidence and intuition than by systematic design. However, the advent of “big data” and an exponential increase in computational power have reshaped the landscape. Today, we use simulations, artificial intelligence (AI), and machine learning (ML) to predict materials characteristics, which dramatically accelerates the discovery of novel materials. For instance, combinatorial megalibraries, where millions of distinct nanoparticles are created on a single chip, have spurred the need for automated characterization tools. This paper presents an ML model specifically developed to perform real-time binary classification of grayscale high-angle annular dark-field images of nanoparticles sourced from these megalibraries. Given the high costs associated with downstream processing errors, a primary requirement for our model was to minimize false positives while maintaining efficacy on unseen images. We elaborate on the computational challenges and our solutions, including managing memory constraints, optimizing training time, and utilizing Neural Architecture Search tools. The final model outperformed our expectations, achieving over 95% precision and a weighted F-score of more than 90% on our test data set. This paper discusses the development, challenges, and successful outcomes of this significant advancement in the application of AI and ML to materials discovery.

Materials Science↗
DOE OSTI2026

Predicting initial trans-membrane pressure across cycles in the ultrafiltration process using random forest

With growing freshwater scarcity, direct potable reuse (DPR) systems that reclaim wastewater for drinking are becoming increasingly important for sustainable water supply. Reliable operation requires minimizing downtime in ultrafiltration (UF) units, where membrane fouling leads to elevated trans-membrane pressure (TMP). This study develops data-driven regression models based on random forest (RF) and autoregressive (AR) approaches to forecast the initial TMP at the start of each UF filtration cycle in a pilot-scale DPR system. The RF model consistently outperforms baseline methods, including historical mean, last observation carried forward, and AR models, across multiple forecast horizons, achieving the lowest root mean square error. To evaluate how different classes of process variables contribute to TMP dynamics over time, we examine the feature importance of independent input variables across multiple forecast horizons. This analysis provides insight into the temporal relevance of operational and sensor-derived features, guiding control and monitoring strategies. Additionally, the impact of hyperparameter tuning on TMP prediction performance is assessed for both direct and recursive RF modelling approaches. The proposed RF framework establishes a robust foundation for predictive monitoring and real-time optimization of UF operations, supporting sustainable and reliable water reuse.

direct potable reuse↗