Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Data Reasoning”

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 469 records · Page 26
All sourcesNASA NTRSDOE OSTI
DOE OSTI2026

Summary of the Initial Post-Irradiation Characterization of HFIR-Irradiated Low-N and High-N HT-9 Steel

Reference cladding systems for sodium fast reactors are based on the historical steel, HT-9. HT-9 is a Fe12Cr ferritic/martensitic steel with additions of Mo, W, V, and other minor elements and demonstrates low irradiation swelling and adequate mechanical properties. Extensive irradiation literature exists on the use of HT-9 as cladding for metal fuel, primarily irradiation on the U-Zr/HT-9 system from the Experimental Breeder Reactor-II (EBR-II) and Fast Flux Test Facility (FFTF) sodium fast reactor, and as a structural material from experiments in the FFTF. The large amount of historical data makes the U-Zr/HT-9 system the reference fuel specification for many nuclear reactor vendors that seek to license modern sodium-cooled fast reactors in the United States. However, it is yet unclear how variations in impurity content within HT-9 fundamentally affect irradiation performance at various irradiation temperatures. Recent work suggests that impurity content may noticeably alter the production of helium through nuclear transmutation. For these reasons, High-Flux Isotope Reactor (HFIR) irradiation of HT-9 steels with known variations in the impurity content is particularly timely to generate data to enable more accurate refinement of the chemical specification for nuclear-grade HT-9 material. This report summarizes the initial transmission electron microscopy characterization of HFIR-irradiated HT-9 steels following mechanical property measurements by the Advanced Fuels Campaign (AFC). This report includes qualitative results of the cavity, dislocation loop and cluster/precipitate microstructures as well as radiation-induced segregation. Quantitative results are being shared with partner institutions and will be included in more detail in a future report in FY2026.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2020

Critical Infrastructure Classification via CNN-based Modeling and Image Analysis

With recent advances in the fields of satellite imagery and machine learning we now have the ability to develop explainable deep learning models that enhance critical infrastructure analysis. Funded through Idaho National Laboratory’s (INL) Laboratory Directed Research and Development (LDRD) office we are in the process of developing a deep learning model capable of identifying critical infrastructure facilities and embedded features within those facilities. Utilizing current limit of practice techniques in the machine learning areas of explainability and transfer learning our model, once complete, will have the capacity to be used on multiple different imagery data sets and produce results that not only classify critical infrastructure facilities, but also explain why a critical infrastructure facility was classified as a certain type of facility. These advancements eliminate the ‘black box’ approach deep learning models have had in the past, where a user will have to trust the conclusion of a model without understanding what reasoning when into the model’s classification process. They also expand a model’s usefulness, traditionally a deep learning model will have to use the same data set it was originally trained on. Given the long training times of deep learning models this is impractical in a number of scenarios. By utilizing transfer learning advancements, we are eliminating the need to train our model on the same data set it is then run on to classify critical infrastructure facilities. We are also enabling the analysis and classification of data sets that are potentially too small to be divided into a training and testing data set. Once completed this model can provide a foundation to enhanced critical infrastructure analysis, dependency analysis, and potential disaster relief efforts.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Data Privacy for the Grid: Toward a Data Privacy Standard for Inverter-Based and Distributed Energy Resources

The traditional approach to planning the distribution grid has focused on reliability in the context of gradual and reasonably predictable load growth. Forecasts of load growth, combined with asset management practices, were used by system planners to identify upgrades to the system to maintain or improve reliability. The decisions, typically based within load flow analysis tools, included considerations about contingency scenarios and corporate forecasts (i.e., top-down predictions at a summary level of what would happen in a particular area that could impact load growth and behavior). As a result, today, this traditional approach no longer fits all purposes.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2022

Synthesis, Structure, and Properties of Volatile Lanthanide Dialkyltriazenides

Several dialkyltriazenide complexes of the lanthanide elements neodymium, europium, and erbium have been prepared; these include the homoleptic complex Er(Bu t N 3 Bu t ) 3 , the tetrahydrofuran monoadducts Ln(Bu t N 3 Bu t ) 3 (THF) where Ln = Nd or Eu, and the lithium salts [Li(THF)][Ln(MeN 3 Bu t ) 4 ] where Ln = Eu or Er. Crystal structures, nuclear magnetic resonance data, and infrared data are reported for all complexes. Further, the di-tert-butyltriazenide complexes are thermally stable, sublime at reasonably low temperatures, and show smooth volatilization without decomposition, which make them potentially useful in lanthanide separation processes and as chemical vapor deposition precursors for lanthanide nitrides and other phases.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

Experimental distinction of the molecularly induced Balmer emission contribution and its application for inferring molecular divertor density with 2D filtered camera measurements during detachment in JET L-mode plasmas

A previously presented model for generating 2D estimates of the divertor plasma conditions at JET from deuterium Balmer line intensity ratios, obtained from tomographic reconstructions of divertor camera images, was amended to consider also the Balmer emission arising from molecular processes. Utilizing the AMJUEL and H2VIBR atomic and molecular databases of EIRENE enabled also inference of the molecular divertor density from the distinguished molecularly induced emission. Analysis of a JET L-mode density scan suggests the molecularly induced emission accounting for up to 60%–70% and 10%–20% of the Balmer Dα and Dγ intensities, respectively, at the onset of detachment, while electron-ion recombination becomes increasingly dominant with deepening detachment. Similar observations were made by post-processing EDGE2D-EIRENE simulations, which indicated significant roles of molecular D$^{+}_{2}$ ions and vibrational excitation of the D 2 molecules as precursors for the molecularly induced emission. The experimentally inferred molecular density at the outer strike point was found to increase monotonously with decreasing strike point temperature, reaching approximately 30%–50% of the local electron density at n mol, osp = 1–2 X 10 20 m -3 at T e,osp ≈ 0.7 eV. A further steep increase by a factor of 3–5 was observed with decrease of T e,osp to 0.5 eV. The observations are in qualitative and reasonable quantitative agreement with EDGE2D-EIRENE predictions of n mol, osp within the uncertainties of the experimental data.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2020

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

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

PDEs↗
DOE OSTI2021

Detection of significant antiviral drug effects on COVID-19 with reasonable sample sizes in randomized controlled trials: A modeling study

Development of an effective antiviral drug for Coronavirus Disease 2019 (COVID-19) is a global health priority. Although several candidate drugs have been identified through in vitro and in vivo models, consistent and compelling evidence from clinical studies is limited. The lack of evidence from clinical trials may stem in part from the imperfect design of the trials. We investigated how clinical trials for antivirals need to be designed, especially focusing on the sample size in randomized controlled trials. A modeling study was conducted to help understand the reasons behind inconsistent clinical trial findings and to design better clinical trials. We first analyzed longitudinal viral load data for Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) without antiviral treatment by use of a within-host virus dynamics model. The fitted viral load was categorized into 3 different groups by a clustering approach. Comparison of the estimated parameters showed that the 3 distinct groups were characterized by different virus decay rates (p-value < 0.001). The mean decay rates were 1.17 d -1 (95% CI: 1.06 to 1.27 d -1 ), 0.777 d -1 (0.716 to 0.838 d -1 ), and 0.450 d -1 (0.378 to 0.522 d -1 ) for the 3 groups, respectively. Such heterogeneity in virus dynamics could be a confounding variable if it is associated with treatment allocation in compassionate use programs (i.e., observational studies). Subsequently, we mimicked randomized controlled trials of antivirals by simulation. An antiviral effect causing a 95% to 99% reduction in viral replication was added to the model. To be realistic, we assumed that randomization and treatment are initiated with some time lag after symptom onset. Using the duration of virus shedding as an outcome, the sample size to detect a statistically significant mean difference between the treatment and placebo groups (1:1 allocation) was 13,603 and 11,670 (when the antiviral effect was 95% and 99%, respectively) per group if all patients are enrolled regardless of timing of randomization. The sample size was reduced to 584 and 458 (when the antiviral effect was 95% and 99%, respectively) if only patients who are treated within 1 day of symptom onset are enrolled. We confirmed the sample size was similarly reduced when using cumulative viral load in log scale as an outcome. We used a conventional virus dynamics model, which may not fully reflect the detailed mechanisms of viral dynamics of SARS-CoV-2. The model needs to be calibrated in terms of both parameter settings and model structure, which would yield more reliable sample size calculation. In this study, we found that estimated association in observational studies can be biased due to large heterogeneity in viral dynamics among infected individuals, and statistically significant effect in randomized controlled trials may be difficult to be detected due to small sample size. The sample size can be dramatically reduced by recruiting patients immediately after developing symptoms. We believe this is the first study investigated the study design of clinical trials for antiviral treatment using the viral dynamics model.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2024

Pion Nucleon Scattering in BCHPT Combined with 1/Nc Expansion

Pion nucleon scattering has played a crucial role in advancing our understanding of the low-energy regime of strong interactions. This dissertation presents the development of a new theoretical framework that combines Baryon Chiral Perturbation Theory (BChPT) with the 1/Nc Expansion to analyze pion nucleon scattering. The resulting effective Lagrangian incorporates both chiral and spin-flavor symmetries. Scattering amplitudes are calculated up to the one-loop level (NNLO), which includes four tree-level Feynman diagrams and 45 non-zero loop diagrams. The effective Lagrangian is renormalized at the one-loop level. With spin-3/2 baryons naturally included, the framework demonstrates strong convergence at higher energies. The NNLO pion nucleon scattering amplitude shows good agreement with experimental data from the SAID database. This novel approach, BChPT combined with 1/Nc Expansion, demonstrates substantial predictive power at higher energies where other low-energy theories fall short, achieving the goals of this combined framework. The contribution from loop diagrams is essential for achieving good agreement with the data, highlighting the importance of conducting the calculation up to NNLO. Extracted values for physically measurable quantities agree reasonably well with independently extracted values, underscoring the robustness of the theory.

Jayakodige, Dulitha [Hampton Univ., Hampton, VA (U↗
DOE OSTI2025

TASTI-GRID: Overview of the State of Oregon’s Resilience and Reliability [Slides]

This report is intended to help your state identify the most effective investments to improve grid resilience and reliability, based on analysis of outage data, weather events, and the current state of the electric grid. The information presented is derived from the best available data. The discussions and recommendations included provide context and valuable insights into potential high returns on investment (ROI). Specifically, the report offers an overview of electric grid resilience in Oregon, covering outage information, reasons for outages, calculated "resilience scores" across the state, and recommendations for both immediate and long-term investments in grid infrastructure. The document provides a concise overview of the analytical capabilities of the TASTI-GRID team. By utilizing additional data and resources, it synthesizes insights from the platform to present a distinct perspective on investing in grid resilience throughout Oregon. This document is not intended to influence or direct state-level decisions related to investments in grid resilience and reliability. TASTI-GRID should not be used for real-time monitoring, supporting the ESF #12 emergency response functions, or predicting the restoration times for outages.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2023

Array-Based Machine Learning for Functional Group Detection in Electron Ionization Mass Spectrometry

Mass spectrometry is a ubiquitous technique capable of complex chemical analysis. The fragmentation patterns that appear in mass spectrometry are an excellent target for artificial intelligence methods to automate and expedite the analysis of data to identify targets such as functional groups. To develop this approach, we trained models on electron ionization (a reproducible hard fragmentation technique) mass spectra so that not only the final model accuracies but also the reasoning behind model assignments could be evaluated. The convolutional neural network (CNN) models were trained on 2D images of the spectra using transfer learning of Inception V3, and the logistic regression models were trained using array-based data and Scikit Learn implementation in Python. Our training dataset consisted of 21,166 mass spectra from the United States’ National Institute of Standards and Technology (NIST) Webbook. The data was used to train models to identify functional groups, both specific (e.g., amines, esters) and generalized classifications (aromatics, oxygen-containing functional groups, and nitrogen-containing functional groups). We found that the highest final accuracies on identifying new data were observed using logistic regression rather than transfer learning on CNN models. It was also determined that the mass range most beneficial for functional group analysis is 0–100 m/z. We also found success in correctly identifying functional groups of example molecules selected from both the NIST database and experimental data. Beyond functional group analysis, we also have developed a methodology to identify impactful fragments for the accurate detection of the models’ targets. The results demonstrate a potential pathway for analyzing and screening substantial amounts of mass spectral data.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

Joint Inversion of Surface Electrical Resistivity Tomography and Seismic Refraction Data between the 200 Areas

Geologic stratigraphy on the Hanford Site influences groundwater and contaminant migration through the aquifer system and the vadose zone. The current geologic framework model (GFM) relies heavily on a sparse distribution of borehole data in some locations to map geologic contacts and hydrologic properties in the subsurface. Non-invasive geophysical methods such as electrical resistivity tomography (ERT), transient electromagnetic surveying, and seismic imaging are being used at Hanford to map subsurface structure in areas with limited well observations. This is to develop and mature the capability of geophysical methods to aid in GFM refinement, to identify regions of subsurface complexity, and for optimal well siting. A joint inversion of co-located seismic refraction and ERT data was carried out for data collected on a ~2.3-km profile between the 200 Areas on the Hanford Site. While ERT and seismic refraction images have sensitivity to overlapping physical properties (porosity, moisture content, lithology), the resolution and physics used to acquire each of these datasets are different and therefore information can be different or mutually complementary. Performing a joint inversion provides a reasonable option for a coherent, coupled interpretation for mutually complementary datasets. Between the 200 Areas, there are few boreholes to interpret the geologic framework model, and these data sets were obtained to provide a first line of evidence toward identifying stratigraphic structure. The seismic refraction and ERT data were independently inverted during fiscal year 2022 and broadly showed a two-layer structure with a trough-like feature that is ~1 km wide and upwards of 150 m deep. The depth of the trough feature was greater in the ERT image compared to the seismic image, which indicated a maximum depth of approximately 110 m. The objective of the joint inversion described in this report was to invert the seismic refraction and ERT data together while constraining the ERT image to be structurally similar to the seismic refraction image. The approach was applied using the geophysical inverse modeling program E4D, which has the capability to invert first-arrival times from seismic refraction data and ERT resistances using a “cross-gradient” constraint. The application of cross-gradient constraints with different weights produces ERT models that show a high degree of similarity within the upper 100 m (above ~120 m elevation). None of the ERT models show an improved structural similarity to the seismic result; therefore, it is recommended that further attempts to jointly interpret these models focus on petrophysics and image resolution. Petrophysical measurements of core samples would improve knowledge of what drives the ERT response in this region and, along with downhole geophysical measurements, could be used to “ground truth” the surface-based geophysical results. Image resolution studies would provide insight into which regions of the inverted images are reliable and which regions are poorly constrained.

58 GEOSCIENCES↗
DOE OSTI2021

An optimized FM-index library for nucleotide and amino acid search

Abstract Background Pattern matching is a key step in a variety of biological sequence analysis pipelines. The FM-index is a compressed data structure for pattern matching, with search run time that is independent of the length of the database text. Implementation of the FM-index is reasonably complicated, so that increased adoption will be aided by the availability of a fast and flexible FM-index library. Results We present AvxWindowedFMindex (AWFM-index), a lightweight, open-source, thread-parallel FM-index library written in C that is optimized for indexing nucleotide and amino acid sequences. AWFM-index introduces a new approach to storing FM-index data in a strided bit-vector format that enables extremely efficient computation of the FM-index occurrence function via AVX2 bitwise instructions, and combines this with optional on-disk storage of the index’s suffix array and a cache-efficient lookup table for partial k-mer searches. The AWFM-index performs exact match count and locate queries faster than SeqAn3’s FM-index implementation across a range of comparable memory footprints. When optimized for speed, AWFM-index is $$\sim $$ ∼ 2–4x faster than SeqAn3 for nucleotide search, and $$\sim $$ ∼ 2–6x faster for amino acid search; it is also $$\sim $$ ∼ 4x faster with similar memory footprint when storing the suffix array in on-disk SSD storage. Conclusions AWFM-index is easy to incorporate into bioinformatics software, offers run-time performance parameterization, and provides clients with FM-index functionality at both a high-level (count or locate all instances of a query string) and low-level (step-wise control of the FM-index backward-search process). The open-source library is available for download at https://github.com/TravisWheelerLab/AvxWindowFmIndex.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Dynamic logistic regression and variable selection: Forecasting and contextualizing civil unrest

Civil unrest can range from peaceful protest to violent furor, and researchers are working to monitor, forecast, and assess such events to allocate resources better. Twitter has become a real-time data source for forecasting civil unrest because millions of people use the platform as a social outlet. Additionally, daily word counts are used as model features, and predictive terms contextualize the reasons for the protest. To forecast civil unrest and infer the reasons for the protest, we consider the problem of Bayesian variable selection for the dynamic logistic regression model and propose using penalized credible regions to select parameters of the updated state vector. This method avoids the need for shrinkage priors, is scalable to high-dimensional dynamic data, and allows the importance of variables to vary in time as new information becomes available. A substantial improvement in both precision and F1-score using this approach is demonstrated through simulation. Finally, we apply the proposed model fitting and variable selection methodology to the problem of forecasting civil unrest in Latin America. Our dynamic logistic regression approach shows improved accuracy compared to the static approach currently used in event prediction and feature selection.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

PVAnalytics [SWR-20-33]

PVAnalytics inputs long term (multiple years) photovoltaic (PV) outdoor sensor data and completes autonomous quality assurance (QA) analysis on the data. Time series data from individual sensors such as electric meters, inverter outputs, irradiance, and temperature sensors are compared to physically expected outputs of these sensors under clear sky conditions based on the latitude, longitude, and orientation of the PV system. PVAnalytics outputs as QA pass or fail for each sensor, reasons determined for failure, time series flags identifying QA thresholds have been exceeded, flags for sunny days, the time zone found for the given sensor, sensor drift, time periods for passing data, and results of orientation checks.

Muller, Matthew↗
DOE OSTI2023

Human and Technology Integration Evaluation of Advanced Automation and Data Visualization

While the existing United States (U.S.) light water reactors are highly reliable, safe, and provide a significant proportion of carbon-free electricity, the cost of operating and maintaining them has become less competitive compared to other electricity generating sources. The reason for the gap in operating and maintenance (O&M) costs can be at least in part attributed to the advent of new digital technologies that other electricity generating industries are currently using. Advanced capabilities including digital instrumentation and control (I&C) systems, advanced automation and analytics, and greater span of data integration (i.e., connectedness) across these non-nuclear plants has transformed the way work is performed and ultimately given them a competitive advantage in terms of the cost required for operating, maintaining, and supporting them. To reduce O&M cost and address obsolescence of the aging I&C infrastructure of the existing U.S. light water reactors, the U.S. Department of Energy (DOE) Light Water Reactor Sustainability (LWRS) Program Plant Modernization Pathway is conducting targeting multidisciplinary research that 1) delivers a sustainable business model to enable a cost-competitive U.S. nuclear industry and 2) is developing technology modernization solutions that address aging and obsolescence challenges. The work described in this report supports these two objectives and describes the demonstration of human and technology integration across recent industry collaborations to support their large-scale digital I&C modifications. This technical report describes the demonstration of the human and technology integration methodology in performing full-scale performance-based human-in-the-loop tests to evaluate plant-specific advanced automation and data visualization applications within these collaborators’ digital modifications. This technical report also documents future applications of human and technology integration that expand beyond main control room modernization and digital I&C upgrades, which have been a central focus to date. Thus, this technical report discusses how to implement human and technology integration across new business opportunities and how to develop an evaluation plan that defines measures and criteria, and documents key assumptions to support full plant modernization.

99 GENERAL AND MISCELLANEOUS↗
DOE OSTI2020

Model Driven Deception for Defense of Operational Technology Environments

Due to the strong integration of real-world physics, OT deception platforms must operate differently than traditional IT deceptions. For instance, turning off a valve will be detected downstream by other sensors because the flow will reduce and stop. Additionally, controllers and applications leverage data from sensors to send control commands to each other. A believable deception must be integrated with the system to project the effects of events. An attack will likely attempt to control the physical process in a negative manner. To make the attacker believe they are achieving their objective, it must predict the effects of these actions, to a reasonable degree. Our approach to simulating a model to generate realistic decoy behavior is explored including description of two approaches: a physics model-based approach and a data driven approach. The performance of two machine learning techniques are investigated in their ability to learn a good enough model of the physics of the system.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Evaluating the limitations of Bayesian metabolic control analysis

Bayesian Metabolic Control Analysis (BMCA) is a promising framework for inferring metabolic control coefficients in data-limited scenarios, combining Bayesian inference with linear-logarithmic (lin-log) rate laws. These metabolic control coefficients quantify how changes in enzyme activities affect steady-state fluxes and metabolite concentrations across a metabolic network. However, its predictive accuracy and limitations remain underexplored. This study systematically evaluates BMCA’s ability to infer elasticity values, flux control coefficients (FCC), and concentration control coefficients (CCC) under varying data availability conditions using three synthetic metabolic network models. We demonstrate that BMCA predictions are highly dependent on the inclusion of flux and enzyme concentration data, with the omission of these datasets leading to severe inaccuracies. In our synthetic, enzyme-perturbation datasets, external metabolite concentrations had minimal impact and, in some cases, their exclusion improved predictions; when external-nutrient perturbations were introduced and those concentrations were observed, gains were at most modest. Additionally, we find that posterior estimation with both ADVI and HMC can underestimate large-magnitude elasticities in our synthetic settings, with ADVI showing somewhat higher variance under strong up-regulation; thus, recovering |elasticity| ≳ 1.5 remains challenging regardless of the inference engine. ADVI also fails to accurately infer allosteric interactions, even when regulatory effects are strong. While BMCA maintains reasonable accuracy in partially recovering the rankings of the highest FCC values, its estimates of absolute values remain constrained by prior assumptions and data limitations. Our findings reveal the BMCA algorithm’s strengths and weaknesses, providing guidance on its application in metabolic engineering, and highlighting the need for methodological refinements to enhance its predictive capabilities.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2023

Predicting future well performance for environmental remediation design using deep learning

Here in this study, we developed a deep learning (DL) framework with a multi-channel three-dimensional convolutional neural network (MC3D-CNN) to predict well performance and thereby assist future environmental remediation design. Such prediction of extraction well performance at designated locations is critical for configuring pump-and-treat (P&T) well network design and operation, setting reasonable target closure dates for overall remedying, and estimating remedy costs. The framework is developed with operational and monitoring data routinely collected during P&T remedy operations, including well extraction and injection rates as well as in situ contaminant concentrations. Traditionally, the collected data were rarely used for purposes other than assessing past well performance and the accuracy of the conceptual site model. However, recent advances in data-driven computational approaches enable better use of the large datasets to inform future well performance, enhance site characterization, and improve remediation planning. In this study, we established a DL framework to integrate transient three-dimensional contaminant plumes and multiple aquifer properties (e.g., hydraulic conductivity and hydrostratigraphic maps) to identify characteristic patterns controlling and representing extraction well mass recovery, aiming at providing future mass recovery estimates for existing wells and candidate wells at any proposed locations. We evaluated our framework by using a realistic synthetic dataset generated from a well-calibrated flow and transport model used in the 200 West Area of the U.S. Department of Energy’s Hanford Site in southeastern Washington state. The multi-channel feature in our framework allows integration of various types and temporal densities of training datasets for DL model development. Overall, we found that the trained DL model achieved an accuracy of over 90% in ranking extraction well performance in validation datasets, and over 80% in predicting high-performance-ranking well locations. This data-informed approach provides a flexible tool to support adaptive site management, streamline decision-making, and potentially reduce remediation time and costs. Our DL framework can be used as a filtering tool to improve the current P&T network optimization design by reducing the number of candidate well locations.

54 ENVIRONMENTAL SCIENCES↗