Search NASA⌕ Search

SEARCH · Search NASA

Results for “Subsurface characterization”

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 397 records · Page 22
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Gas diffusion through variably-water-saturated zeolitic tuff: Implications for transport following a subsurface nuclear event

Noble gas transport through geologic media has important applications in the characterization of underground nuclear explosions (UNEs). Without accurate transport models, it is nearly impossible to distinguish between xenon signatures originating from civilian nuclear facilities and UNEs. Understanding xenon transport time through the earth is a key parameter for interpreting measured xenon isotopic ratios. One of the most challenging aspects of modeling gas transport time is accounting for the effect of variable water saturation of geological media. In this study, we utilize bench-scale laboratory experiments to characterize the diffusion of krypton, xenon, and sulfur hexafluoride (SF6) through intact zeolitic tuff under different saturations. Here, we demonstrate that the water in rock cores with low partial saturation dramatically affects xenon transport time compared to that of krypton and SF6 by blocking sites in zeolitic tuff that preferentially adsorb xenon. This leads to breakthrough trends that are strongly influenced by the degree of the rock saturation. Xenon is especially susceptible to this phenomenon, a finding that is crucial to incorporate in subsurface gas transport models used for nuclear event identification. We also find that the breakthrough of SF6 diverges significantly from that of noble gases within our system. When developing field scale models, it is important to understand how the behavior of xenon deviates from chemical tracers used in the field, such as SF6 (Carrigan et al., 1996). These new insights demonstrate the critical need to consider the interplay between rock saturation and fission product sorption during transport modeling, and the importance of evaluating specific interactions between geomedia and gases of interest, which may differ from geomedia interactions with chemical tracers.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗
DOE OSTI2016

Canada's Underground Research Laboratory (1980-2014) - 16274

In 1980, Canada initiated a long-term, comprehensive research and development program of geologic characterization and large-scale in situ geotechnical experiments in a subsurface facility, the Underground Research Laboratory (URL). The URL was excavated specifically to evaluate the concept of permanent disposal of used CANDU{sup R} nuclear reactor fuel in a deep geological repository (DGR) constructed in Canadian Shield granite. Key phases of the URL included siting, detailed site characterization and monitoring, construction, operation, decommissioning, and closure. Much collaborative research with international organizations was undertaken at the URL and continues, even after the closure of the facility, through the ongoing Enhanced Sealing Project (ESP). The ESP monitors the engineered performance of a full-scale composite shaft seal that was emplaced during URL closure to permanently seal the main URL access shaft across a sub-horizontal hydraulically active fracture zone. Currently the ESP is jointly funded by Canadian Nuclear Laboratories (CNL) (Canada), Posiva (Finland), and Andra (France), and monitoring is planned to continue until the end of 2016. Decades of programmatic experience at the URL have shown that knowledge continuity is integral to supporting the development of a safe geological disposal facility for high level nuclear waste. This paper outlines the history of the URL and describes some technical lessons learned from its 30-plus years of construction, operation, and closure activities. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2024

Stochastic reconstruction of fracture network pattern using spatial point processes

Fracture spatial patterns can strongly affect fluid flow in the subsurface. Proximity and distribution of fractures control reservoir flow behavior over various length scales. In many studies, however, simplified geometrical patterns are generated for fractures which may lead to unrealistic subsurface models. Here we introduce a new method for characterization and modeling of fracture spatial patterns based on outcrop observations. We use Ripley's K-function to characterize the arrangement of fracture barycenters and intersection points over various length scales. In addition, we use semivariograms to quantify spatial correlation in fracture intensity maps. Using this information, we develop a stochastic algorithm that generates two-dimensional fracture network realizations with spatial properties similar to those of a real fracture network measured in the field. Numerical simulation models indicate that the generated fracture realizations exhibit similar flow behaviors as that of the original fracture network. Furthermore, such modeling tools expand and improve our capability in building representative fracture models and in quantification of uncertainty in naturally and hydraulically fractured reservoirs.

58 GEOSCIENCES↗
DOE OSTI2021

The First Detection of an Earthquake From a Balloon Using Its Acoustic Signature

Extreme temperature and pressure conditions on the surface of Venus present formidable technological challenges against performing ground-based seismology. Efficient coupling between the Venusian atmosphere and the solid planet theoretically allows the study of seismically generated acoustic waves using balloons in the upper atmosphere, where conditions are far more clement. However, earthquake detection from a balloon has never been demonstrated. We present the first detection of an earthquake from a balloon-borne microbarometer near Ridgecrest, CA in July 2019 and include a detailed analysis of the dependence of seismic infrasound, as measured from a balloon on earthquake source parameters, topography, and crustal and atmospheric structure. Our comprehensive analysis of seismo-acoustic phenomenology demonstrates that seismic activity is detectable from a high-altitude platform on Earth, and that Rayleigh wave-induced infrasound can be used to constrain subsurface velocities, paving the way for the detection and characterization of such signals on Venus.

58 GEOSCIENCES↗
DOE OSTI2013

RECONSTRUCTING THE SUBSURFACE THREE-DIMENSIONAL MAGNETIC STRUCTURE OF A SOLAR ACTIVE REGION USING SDO/HMI OBSERVATIONS

A solar active region (AR) is a three-dimensional (3D) magnetic structure formed in the convection zone, whose property is fundamentally important for determining the coronal structure and solar activity when emerged. However, our knowledge of the detailed 3D structure prior to its emergence is rather poor, largely limited by the low cadence and sensitivity of previous instruments. Here, using the 45 s high-cadence observations from the Helioseismic and Magnetic Imager on board the Solar Dynamics Observatory, we are able for the first time to reconstruct a 3D data cube and infer the detailed subsurface magnetic structure of NOAA AR 11158, and to characterize its magnetic connectivity and topology. This task is accomplished with the aid of the image-stacking method and advanced 3D visualization. We find that the AR consists of two major bipoles or four major polarities. Each polarity in 3D shows interesting tree-like structure, i.e., while the root of the polarity appears as a single tree-trunk-like tube, the top of the polarity has multiple branches consisting of smaller and thinner flux tubes which connect to the branches of the opposite polarity that is similarly fragmented. The roots of the four polarities align well along a straight line, while the top branches are slightly non-coplanar. Our observations suggest that an active region, even appearing highly complicated on the surface, may originate from a simple straight flux tube that undergoes both horizontal and vertical bifurcation processes during its rise through the convection zone.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2019

Phase, morphology, elemental composition, and formation mechanisms of biogenic and abiogenic Fe-Cu-sulfide nanoparticles: A comparative study on their occurrences under anoxic conditions

In this paper, we report on a systematic study on the physicochemical attributes of synthetic Fe-Cu-sulfide nanoparticles (NPs) precipitated under conditions similar to the anoxic, low temperature aqueous, sedimentary, soil and subsurface environments where these NPs have been repeatedly identified. Characterizing the basic attributes of these NPs is the first step in understanding their behaviors in various processes including in the bio-availability of essential and toxic metals, environmental remediation and resource recovery. All experiments are performed in the presence and absence of the sulfate-reducer Desulfovibrio vulgaris to elucidate biological controls on NP formation. First, the single-metal endmember NPs are determined by precipitation in solution containing either aqueous Fe(II) or Cu(II). Limited differences are observed between biogenic and abiogenic precipitates aged for up to one month; the Fe-only experiments resulted in 4-10 nm mackinawite (FeS) NPs that aggregate to form nanosheets up to ~1,000 nm in size, while the Cu-only experiments resulted in mixtures of covellite (CuS) NPs comprised of <10 nm fine nanocrystals, 20-40 x 6-9 nm nanorods and ~30 nm nanoplates. The crystal sizes of biogenic mackinawite and covellite are respectively larger and smaller than their abiogenic counterparts, indicating a mineral-specific response to biological presence. Structural defects are observable in the fine nanocrystals and nanorods of covellite in both biogenic and abiogenic experiments, indicative of intrinsic NP instability and formation mechanism via particle attachment. In contrast, covellite nanoplates are defect-free, indicating high stability and potentially rapid recrystallization following particle attachment. Next, mixed-metal sulfide NPs are precipitated at variable initial aqueous Fe-to-Cu ratios (2:1, 1:1 and 1:5). With increasing ratio of Fe-to-Cu, Fe-rich covellite, nukundamite (Cu 5.5 FeS 6.5 ), chalcopyrite (CuFeS 2 ), and Cu-rich mackinawite are formed. The Fe-rich covellite NPs are larger (100-200 nm) than covellite precipitated in the absence of Fe, indicating a role for Fe in promoting crystal growth. Chalcopyrite and nukundamite are formed through incorporation of Fe into precursor covellite NPs while retaining the original crystal morphology, as confirmed by doping a covellite suspension with aqueous Fe(II), resulting in the formation of chalcopyrite and nukundamite within days. Additionally, in the biological systems we observe the recrystallization of mackinawite to greigite (Fe 3 S 4 ) after six months of incubation in the absence of Cu, and the selective formation of chalcopyrite and nukundamite at lower initial Fe-to-Cu ratios compared to abiotic systems. These observations are consistent with NP precipitation that are influenced by the distinct (sub)micro-environments around bacterial cells compared to the bulk solution. Lastly, comparative TEM analyses indicate that the synthetic NPs are morphologically similar to NPs identified in natural environments, opening ways to studying behaviors of natural NPs using experimental approaches.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Distributed Temperature Sensing (DTS) to Monitor CO 2 Migration in an Enhanced Oil Recovery Field in Northern Michigan

The Midwest Regional Carbon Sequestration Partnership (MRCSP) was founded in 2003 as part of the U.S. Department of Energy’s (DOE’s) Regional Carbon Sequestration Partnership initiative. Since its founding, MRCSP has made significant strides toward making CCUS a viable option for states in the region. The public/private consortium, funded through the DOE Regional Carbon Sequestration Initiative, brings together nearly 40 industry partners and 10 states. Battelle, as the project lead, oversees research, development and operations and coordinates activities among the partners. The incremental, phased approach has built a valuable knowledge base for the industry and paved the way for commercial-scale adoption of CCUS technologies. From 2008 to 2020, MRCSP Phase III focused on the development of large-scale injection projects. This report is part of a series of reports prepared under the Midwestern Regional Carbon Sequestration Partnership (MRCSP) Phase III (Development Phase). These reports summarize and detail the findings of the work conducted under the Phase III project. MRCSP conducted analysis of CO 2 flow patterns using distributed temperature sensing (DTS) data acquired from the Chester 16 pinnacle reef located in Otsego County, Michigan. The reef has one injection well Chester 6-16 and one monitoring well Chester 8-16 where DTS is installed. DTS provides a means to track temperature within the wellbores in real time and characterize the migration patterns of CO 2 in the subsurface when DTS data is combined with other operational data such as the injection rates, bottomhole pressures and temperature. The primary objective of this study was to determine the distribution of CO 2 inflow zones in the A1 Carbonate and Brown Niagaran formations.

01 COAL, LIGNITE, AND PEAT↗
NASA NTRS2006

Miniature Ground Penetrating Radar, CRUX GPR

Under NASA instrument development programs (PIDDP 2000-2002, MIPD 2003-2005, ESR and T, 2005) we have been developing miniature ground penetrating radars (GPR) for use in mapping subsurface stratigraphy from planetary rovers for Mars and lunar applications. The Mars GPR is for deeper penetration (up to 50 m depth) into the Martian subsurface at moderate resolution (0.5 m) for a geological characterization. As a part of the CRUX (Construction and Resource Utilization Explorer) instrument suite, the CRUX GPR is optimized for a lunar prospecting application. It will have shallower penetration (5 m depth) with higher resolution (10 cm) for construction operations including ISRU (in-situ resource utilization).

ground penetrating radars (GPR↗
NASA NTRS2008

Propulsive Maneuver Design for the 2007 Mars Phoenix Lander Mission

On May 25, 2008, the Mars Phoenix Lander (PHX) successfully landed in the northern planes of Mars in order to continue and complement NASA's "follow the water" theme as its predecessor Mars missions, such as Mars Odyssey (ODY) and Mars Exploration Rovers, have done in recent years. Instruments on the lander, through a robotic arm able to deliver soil samples to the deck, will perform in-situ and remote-sensing investigations to characterize the chemistry of materials at the local surface, subsurface, and atmosphere. Lander instruments will also identify the potential history of key indicator elements of significance to the biological potential of Mars, including potential organics within any accessible water ice. Precise trajectory control and targeting were necessary in order to achieve the accurate atmospheric entry conditions required for arriving at the desired landing site. The challenge for the trajectory control maneuver design was to meet or exceed these requirements in the presence of spacecraft limitations as well as other mission constraints. This paper describes the strategies used, including the specialized targeting specifically developed for PHX, in order to design and successfully execute the propulsive maneuvers that delivered the spacecraft to its targeted landing site while satisfying the planetary protection requirements in the presence of flight system constraints.

FPC↗
DOE OSTI2013

Characterization of Persistent Volatile Contaminant Sources in the Vadose Zone

Remediation activities over time will alter the subsurface distribution of contaminants and likely create significant changes in the source-zone architecture. A field method was demonstrated for use of data collected from multiple individual soil vapor extraction (SVE) system well tests to locate and characterize the distribution of persistent VOC sources in the vadose zone. Operational data collected at the Department of Energy’s Hanford site were used to examine source zone alteration over time due to SVE operation and to illustrate the source-zone characterization approach. Individual well test results confirmed a heterogeneous distribution of permeability and contaminant mass discharge throughout the vadose zone. The trends in mass discharge and concentration were analyzed to determine the location and extent of the primary source zone within a lower-permeability unit at the site. This information is useful to evaluate the performance of SVE operations, and support decisions concerning system alteration or closure based on risk assessments of the impact of vadose-zone sources on groundwater contamination or vapor intrusion.

carbon tetrachloride↗
DOE OSTI2021

Bedrock architecture, soil texture, and hyporheic zone characterization combining electrical resistivity and induced polarization imaging

The shallow (<15 m depth) subsurface environment of a short reach along a bedrock stream was investigated with electrical resistivity and induced polarization (IP) to map details of bedrock and soil sediments. The bedrock is mostly comprised of limestone, and is generally resistive compared to the overlying soil. The soil-bedrock interface was determined through a trial and error approach using a sharp boundary feature in the inversion model. The inferred bedrock surface determined from the inversions exhibited undulating patterns with troughs and ridges. A near continuous trough ran alongside the stream within the floodplain, and is suggestive of a paleochannel among other interpretations of this feature. The structure of the electrical resistivity above the bedrock showed small-scale elongated features. The chargeability from the IP method showed larger scale features. High values of chargeability were associated with the sediments in the floodplain, and low values were associated with bedrock, stream, and soil on the elevated banks above the stream. If the chargeability is associated with membrane polarization characteristic of clayey soils, then IP seems to highlight the mere existence of clay, while the resistivity may be more discernable of the relative proportion of clay. Vegetation differences may also explain the chargeability distribution, where parts of the survey with high chargeability had dense pine with no understory making the soils more organically rich.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Estimating biofuel contaminant concentration from 4D ERT with mixing models

Here, we present the results of a lab-scaled feasibility study to assess the performance of electrical resistivity tomography for detection, characterization, and monitoring of fuel grade ethanol releases to the subsurface. Further, we attempt to determine the concentration distribution of the ethanol from the electrical resistivity tomography data using mixing-models. Ethanol is a renewable fuel source as well as an oxygenate fuel additive currently used to replace the known carcinogen methyl tert-butyl ether; however, ethanol is preferentially biodegraded and a cosolvent. When introduced to areas previously impacted by nonethanol-based fuels, it will facilitate the persistence of carcinogenic fuel compounds like benzene and ethylbenzene, as well as remobilize them to the ground water. These compounds would otherwise be retained in the soil column undergoing active or passive remediation processes such as soil vapor extraction or natural attenuation. Here, we introduce ethanol to a saturated Ottawa sand in a tank instrumented for four-dimensional geoelectrical measurements. Forward model results suggest pure phase ethanol released into a water saturated silica sand should present a detectable target for electrical resistivity tomography relative to a saturated silica sand only. We observe the introduction of ethanol to the closed hydraulic system and subsequent migration over the duration of the experiment. One-dimensional and three–dimensional temporal data are assessed for the detection, characterization, and monitoring of the ethanol release. Results suggest one-dimensional geoelectrical measurements may be useful for monitoring a release, while three-dimensional geoelectrical field imaging would be useful to characterize, monitor, and design effective remediation approaches for an ethanol release, assuming field conditions do not preclude the application of geoelectrical methods. We then attempt to use predictive mixing models to calculate the distribution of ethanol concentration within the measurement domain. For this study we examine four different models: a nested parallel mixing model, a nested cubic mixing model, the complex refractive index model (CRIM), and the Lichtenecker-Rother (L-R) model. The L-R model, modified to include an electrical formation factor geometry term, provided the best agreement with expected EtOH concentrations.

09 BIOMASS FUELS↗
DOE OSTI2021

Spectral Induced Polarization-Biogeochemical Relationships for Remediation Amendment Monitoring

Geophysical tools such as electrical resistivity (ER) can indirectly monitor subsurface changes in response to remedial injections. These methods exhibit relatively low spatial resolution compared to sediment core characterization but are advantageous due to the ability to collect measurements non-intrusively over time across large volumes of the subsurface. Moreover, along with confirmatory groundwater or core sampling, geophysical methods can be used during active biogeochemical remedies to monitor short-term contaminant transformations and mobility, as well as part of an overall strategy for long-term monitoring of subsurface contaminated sites. The use of alternating current spectral induced polarization (SIP) provides significantly more information than conventional geophysical methods like direct current ER. SIP allows for monitoring changes in both solution and surface conductivity by separation of real and imaginary conductivity, respectively, as well as surface capacitance. In principle, SIP can measure indicators of remedy progression such as precipitation reactions that sequester contaminants, potentially providing a better indication of amendment delivery and reactivity as compared to conventional ER methods. However, multiple processes and material properties have overlapping (interacting) electrical responses across a range of frequencies. Hence, the purpose of this scoping study was to evaluate if SIP can be used to monitor (a) amendment delivery and (b) precipitation and reactivity of amendments under consideration at Hanford. SIP measurements were collected in fully saturated columns packed with sand and Hanford formation sediments containing (a) amendments that were highly conductive with significant capacitance (zero valent iron – ZVI, sulfur modified iron – SMI) and (b) amendments that exhibited low electrical conductivity with a small capacitance (calcite, apatite, bismuth). The sand was a quartz material with homogenous particle size that exhibited a relatively low surface conductivity. It was used as a control for comparison with the sediments from the Hanford Site, which have a greater surface conductivity because of their complex mineralogy and heterogeneous particle size distribution and may have complex interactions with amendments. The amendment mass fraction was varied to represent the different stages and subsurface locations associated with the delivery of a remedy. The primary objective was to identify the solution and solid surface changes associated with the delivery amendments and their secondary reactions within the subsurface. The figure below summarizes results for the amendments tested through this project. The SIP phase shift or imaginary conductivity change for the high conductivity amendments was more than 10 times that of the low conductivity amendments, highlighting the relative ease of detection of ZVI and SMI independent of the background signal from sand or Hanford sediments. The ZVI phase shift and imaginary conductivity changes occur primarily at high frequency (> 100 Hz) whereas SMI changes were at low frequency (0.01 to 10 Hz). SIP signals of SMI also increased over time and the maximum shifted to lower frequencies. The low conductivity amendments (calcite, abiotic and biotic apatite, bismuth subnitrate) exhibited relatively small phase and imaginary conductivity changes when added to sediments. The changes were above minimum detection limits (0.5 mrad for phase shift, 0.03 µS/cm for imaginary conductivity) for the highest concentration except for the commercial bismuth material. The lowest amendment concentration that can be detected is likely sediment specific, as minerals in sediments (clays, magnetite, Fe-oxides) have some capacitance and, therefore, exhibit a variable phase shift.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
NASA NTRS2003

On the Determination of Local Sites for Shallow Ground Water Prospection Using Low Frequency Sounding Radars on Mars

Low frequency sounding radars should be able to probe the Martian subsurface layers down to varying depths, depending on the geo-electrical properties of the sounded sites. We present in this work four frequency dependent geo-electrical models of the Martian subsurface in the 1-20 MHz frequency band, based on laboratory electromagnetic characterization of Martian soil analogues. Those models correspond to local Martian sites, where we considered particular interest for the search of water using mainly the Ground Penetrating Radar (GPR) instrument of the Netlander mission. Results and discussion are also valid for both sounding experiments MARSIS and SHARAD. The four models of the Martian subsurface are designed to represent terrains where recent fluvial like features suggest the presence of near subsurface ground ice and probably liquid water. We performed measurements on volcanic and sedimentary materials that may be present on these sites under the appropriate geophysical conditions that may exist in those terrains. We then simulated the backscattered radar echo arising from each site in the 2 MHz frequency band, using the Finite Difference Time Domain (FDTD) algorithm, in order to evaluate the instrument performances to probe the subsurface stratigraphy of each site. Our results confirm that the near subsurface rich iron oxide mineralogy controls the instrument performances in terms of penetration depth and signal to noise ratio in the 2 MHz frequency band. We also discuss the geophysical and geo-electrical sounding conditions that could lead to an ambiguous detection of shallow subsurface water on Mars for the Netlander GPR. Finally we hope to present by the conference time a terrestrial test analogue site selected according to those creteria where deep sub-surface water have been detected 600 m deep using a prototype of the 2 MHz Netlander GPR in a field survey performed last febrary in the eastern part of the Egyptian desert.

Heggy, E.↗
DOE OSTI2017

Characterization of natural organic matter in low-carbon sediments: Extraction and analytical approaches

Sediment-associated natural organic matter (NOM) is a complex assemblage of organic molecules with a wide range of sizes, functional groups, and structures, which is intricately associated with mineral particles. Organic carbon (OC) concentrations in subsurface sediments are typically 10 to 200 times lower than in surface soils, posing a distinct challenge for characterization. A range of chemical extractions were evaluated for extraction of NOM, and a NOM extraction scheme was developed using a combination of sequential extraction with water (MQ) and sodium pyrophosphate at pH 10 (PP), and purification by dialysis and solid phase extraction in order to isolate different fractions of sediment-associated NOM. Analysis of these different NOM fractions was then carried out by Fourier transform infrared spectroscopy (FTIR), ultraviolet-visible (UV-Vis) spectroscopy, and electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry (ESI-FTICR-MS). The OC extraction efficiency of the tested extractions followed the order PP>NaOH>HCl=hydroxylamine hydrochloride>dithionite. Greater concentrations of OC in PP vs NaOH extracts suggest that metal complexation and/or ligand exchange plays an important role in OC stabilization. Characterization of different pools of extraction NOM by FITR shows that the water soluble fraction has a higher fraction of aliphatic and carboxylic groups, while the PP fractions have higher C=C groups. This trend from aliphatic to more aromatic is also supported by the UV-Vis and ESI-FTICR-MS data.

Environmental Molecular Sciences Laboratory↗
DOE OSTI2018

Defining Modeling and Analysis Complexity Needs for Remediation - 18425

Many subsurface remedy decisions rely on model predictions or other types of data analyses to support site characterization and remedy selection, design, implementation, performance assessment, and closure. Thus, determining the appropriate level of complexity for modeling and data analysis is important to facilitate subsurface remedy decisions. Multiple factors are important in selecting the appropriate level of complexity in modeling and data analysis. For instance, the type and number of inputs required typically increases with increased complexity of analysis, and the ability to obtain or verify those inputs needs to be considered. The approach needs to establish the relationship between parameters or processes and contaminant transport and the detail of the description needed to meet the analysis objectives. Assessing the modeling complexity associated with the appropriate level of detail can be facilitated by considering complexity of the 1) modeling approach, 2) description of process, and 3) description of heterogeneity. As demonstrated by case studies, using a framework to guide determination of modeling and analysis complexity facilitates selection of the appropriate approach for the site situation. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2022

Machine-Learning of Nonlocal Kernels for Anomalous Subsurface Transport from Breakthrough Curves

Anomalous behavior is ubiquitous in subsurface solute transport due to the presence of high degrees of heterogeneity at different scales in the media. Although fractional models have been extensively used to describe the anomalous transport in various subsurface applications, their application is hindered by computational challenges. Simpler nonlocal models characterized by integrable kernels and finite interaction length represent a computationally feasible alternative to fractional models; yet, the informed choice of their kernel functions still remains an open problem. We propose a general data-driven framework for the discovery of optimal kernels on the basis of very small and sparse data sets in the context of anomalous subsurface transport. Using spatially sparse breakthrough curves recovered from fine-scale particle-density simulations, we learn the best coarse-scale nonlocal model using a nonlocal operator regression technique. Predictions of the breakthrough curves obtained using the optimal nonlocal model show good agreement with fine-scale simulation results even at locations and time intervals different from the ones used to train the kernel, confirming the excellent generalization properties of the proposed algorithm. A comparison with trained classical models and with black-box deep neural networks confirms the superiority of the predictive capability of the proposed model.

97 MATHEMATICS AND COMPUTING↗