Search NASA⌕ Search

SEARCH · Search NASA

Results for “Reservoir Model”

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 865 records · Page 48

Reservoir Computing as a Tool for Climate Predictability Studies

Reduced-order dynamical models play a central role in developing our understanding of predictability of climate irrespective of whether we are dealing with the actual climate system or surrogate climate models. In this context, the linear inverse modeling (LIM) approach, by capturing a few essential interactions between dynamical components of the full system, has proven valuable in providing insights into predictability of the full system. We demonstrate that reservoir computing (RC), a form of learning suitable for systems with chaotic dynamics, provides an alternative nonlinear approach that improves on the predictive skill of the LIM approach. We do this in the example setting of predicting sea surface temperature in the North Atlantic in the preindustrial control simulation of a popular earth system model, the Community Earth System Model so that we can compare the performance of the new RC-based approach with the traditional LIM approach both when learning data are plentiful and when such data are more limited. The improved predictive skill of the RC approach over a wide range of conditions—larger number of retained EOF coefficients, extending well into the limited data regime, etc.—suggests that this machine-learning technique may have a use in climate predictability studies. While the possibility of developing a climate emulator—the ability to continue the evolution of the system on the attractor long after failing to be able to track the reference trajectory—is demonstrated in the Lorenz-63 system, it is suggested that further development of the RC approach may permit such uses of the new approach in more realistic predictability studies.

54 ENVIRONMENTAL SCIENCES↗

Numerical analysis of laser-driven reservoir dynamics for shockless loading

Laser-driven plasma loader for shockless compression provides a new approach to study the rapid compression response of materials not attainable in conventional shock experiments. In this method, the strain rate is varied from {approx}10{sup 6}/s to {approx}10{sup 8}/s, significantly higher than other shockless compression methods. Thus, this loading process is attractive in the research of solid material dynamics and astrophysics. The objective of the current study is to demonstrate the dynamic properties of the jet from the rear surface of the reservoir, and how important parameters such as peak load, rise time, shockless compression depth, and stagnating melt depth in the sample vary with laser intensity, laser pulse length, reservoir thickness, vacuum gap size, and even the sample material. Numerical simulations based on the space-time conservation element and solution element method, together with the bulk ablation model, were used. The dynamics of the reservoir depend on the laser intensity, pulse length, equation of state, as well as the molecular structure of the reservoir. The critical pressure condition at which the reservoir will unload, similar to a gas or weak plasma, is 40-80 GPa before expansion. The momentum distribution bulges downward near the front of the plasma jet, which is an important characteristic that determines shockless compression. The total energy density is the most important parameter, and has great influence on the jet characteristics, and consequently on the shockless compression characteristics. If the reservoir is of a single material irradiated at a given laser condition, the relation of peak load and shockless compression depth is in conflict, and the highest loads correspond to the smallest thickness of sample. The temperature of jet front runs up several electron volts after impacting on the sample, and the heat transfer between the stagnating plasma and the sample is sufficiently significant to induce the melting of the sample surface. However, this diffusion heat wave propagates much more slowly than the stress wave, and has minimal effect on the shockless compression progress at a deeper position.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

X-ray emission from M87 - A pressure confined cooling atmosphere surrounding a low mass galaxy

A conventional model for the mass of the galaxy M87 (500-billion solar masses) is used to show that a radiative cooling flow agrees well with spectroscopically derived mass flow rates (10 solar masses per year) and surface brightness profiles of X-ray emission around the galaxy. A temperature of 100-million K, and a density of a few x 10 to the -4th/cu cm are obtained without the use of a massive halo. The atmosphere contains 1-trillion solar masses of gas, with a temperature profile increasing outward and gas acting as a reservoir for the flow. The present models indicate that thermal conductivity in M87 is less than 0.002 of its classical value, and does not significantly affect properties of the surrounding gas.

Binney, J.↗

A cloud computing approach to massively-parallel distributed stochastic optimization with application to geomechanical reservoir characterization

Many search strategies used to optimize parameters during hydrologic model calibration are inherently limited in that simulations must be performed sequentially in order for the algorithm to converge. As a result, the overall time required to reach a solution increases linearly with the number of steps necessary. In contrast, sampling strategies that attempt to approximate the posterior distribution of the parameters do not require the model simulations to be run in sequence and therefore invite the use of a massively parallel approach across heterogeneous computing environments. We apply a cloud computing methodology to this problem where a head node running on Amazon Web Services continuously produces candidate models that are distributed to arbitrary compute environments to run the simulation, in our case Clemson’s Palmetto cluster computer. Thus the time required to obtain a sufficient number of samples to produce an adequate approximation of the posterior distribution of the parameters is limited by the number of available compute sites. Monte Carlo sampling of the posterior space would be duplicative and inefficient, thus we combine Latin hypercube methods with genetic algorithms to ensure that no large regions of the parameter space are left unexplored, and that high probability regions are investigated thoroughly. We independently assess the fit to different types of data sets without the need to construct a single objective function. Sets of Pareto optimal models are particularly valuable when the quality and information content of data is poorly known or conceptual model errors exist in our simulation. We illustrate our approach for a problem where in situ strain data are used to calibrate the parameters of an oil reservoir in the Avant oil field in Oklahoma. We use well injection and shut-in to stress a deep (500m) confined formation, and observe the geomechanical response using borehole strainmeters installed in the shallow subsurface (30m). These data were then used to calibrate the poroelastic properties of the reservoir and caprock along with the geometry of an elliptic channel associated with the formation. Using our approach we are able to identify issues with model error as well as substantial tradeoffs that occur within the parameters.

inversion, stochastic, geomechanics, poroelastic, ↗

CCUS 2024, Interpreting the strain tensor Larry Murdoch Interpreting strain tensor data to characterize and monitor reservoirs for CO2 storage and other applications

Recent advances in instrumentation have made it feasible to measure the transient strain tensor caused by small changes in fluid volume or pressure in the subsurface and this has opened the door to new opportunities for characterization and monitoring during CCUS. We have demonstrated this method by deploying strainmeters at shallow depths (30 to 40m) and then conducting injection well tests in an underlying reservoir at 530m depth. The resulting data indicated that the horizontal strain at shallow strainmeters was tensile and the vertical strain was compressive. The radial strain was less than the horizontal strain, and the strain rates decreased from 100 nanostrain/day to roughly 10 ne/d over a few days (1 nanostrain = 1 part per billion strain). We then used the strain data to estimate reservoir properties, geometry and pressure through inversion of poroelastic forward models using both numerical and novel analytical methods. The average horizontal strain in the caprock resembles the transient pressure in the underlying reservoir and classic type-curve methods from transient well testing can be used for preliminary interpretations of strain data. We have developed fast, closed-form analytical solutions to a pressurized poroelastic inclusion and inhomogeneity in a half-space. Numerical models developed using finite element methods allow more details of the subsurface to be included in the inversion, but they require much longer run times and this makes inversion cumbersome using standard methods. We have developed an inversion approach that uses a proxy model created using machine learning to do most of the forward calculations. This approach markedly reduces the computational requirements and makes it feasible to use Bayesian inversion with large numerical models. Bayesian inversion is important because it provides predictions with uncertainties, which makes the results useful for decision making. We have shown with field tests and simulations that the strain tensor in the caprock is sensitive to pressure in the reservoir, reservoir properties and boundaries, and pressure in the caprock caused by leaks. These results indicate that measuring and interpreting the shallow strain tensor could be a valuable tool for both initial reservoir characterization efforts and long-term monitoring during CCUS. Recent advances in instrumentation have made it feasible to measure the transient strain tensor caused by small changes in fluid volume or pressure in the subsurface and our objective was to evaluate opportunities for strain monitoring during characterization and monitoring for CCUS. Our approach was to deploy strainmeters at shallow depths (30 to 40m) and then conduct injection well tests in an underlying reservoir at 530m depth. The results indicate that the horizontal strain at shallow strainmeters was tensile and the vertical strain was compressive. The radial strain was less than the horizontal strain, and the strain rates decreased from 100 nanostrain/day to roughly 10 ne/d over a few days (1 nanostrain = 1 part per billion strain). We then used the strain data to estimate reservoir properties, geometry and pressure through inversion of poroelastic forward models using both numerical and novel analytical methods. The average horizontal strain in the caprock resembles the transient pressure in the underlying reservoir and classic type-curve methods from transient well testing can be used for preliminary interpretations of strain data. We have developed fast, closed-form analytical solutions to a pressurized poroelastic inclusion and inhomogeneity in a half-space. Numerical models developed using finite element methods allow more details of the subsurface to be included in the inversion, but they require much longer run times and this makes inversion cumbersome using standard methods. We have developed an inversion approach that uses a proxy model created using machine learning to do most of the forward calculations. This approach markedly reduces the computational requirements and makes it feasible to use Bayesian inversion with large numerical models. Bayesian inversion is important because it provides predictions with uncertainties, which makes the results useful for decision making. In conclusion, we have shown with field tests and simulations that the strain tensor in the caprock is sensitive to pressure in the reservoir, reservoir properties and boundaries, and pressure in the caprock caused by leaks. These results indicate that measuring and interpreting the shallow strain tensor could be a valuable tool for both initial reservoir characterization efforts and long-term monitoring during CCUS.

Murdoch, Larry↗

Geological Thermal Energy Storage (GeoTES) Charged with Solar Thermal Technology Using Depleted Oil/Gas Reservoirs and Carnot-Battery Technique Using Shallow Reservoirs: Preprint

Geological thermal energy storage (GeoTES) utilizes the underground reservoirs to storage and dispatch energy per given demand schedule over a time scale up to seasons of a year. The energy input can be of various sources/forms; in this paper, we investigate 1) the GeoTES technology with solar thermal hybridization and using depleted oil/gas reservoirs; 2) the GeoTES technology with heat pumps charged by excess renewable electricity and using low-temperature shallow reservoirs. For each GeoTES technology, we carry out the suitability analysis of candidate reservoirs, develop initial techno-economic models, and validate the model with selected case study. The paper provides an overview of our technical progress on the topics of concern and aims to promote a wider acceptance of the GeoTES technologies in the future energy market.

Carnot battery↗

Heterogeneity, pore pressure, and injectate chemistry: Control measures for geologic carbon storage

Desirable outcomes for geologic carbon storage include maximizing storage efficiency, preserving injectivity, and avoiding unwanted consequences such as caprock or wellbore leakage or induced seismicity during and post injection. Here, to achieve these outcomes, three control measures are evident including pore pressure, injectate chemistry, and knowledge and prudent use of geologic heterogeneity. Field, experimental, and modeling examples are presented that demonstrate controllable GCS via these three measures. Observed changes in reservoir response accompanying CO 2 injection at the Cranfield (Mississippi, USA) site, along with lab testing, show potential for use of injectate chemistry as a means to alter fracture permeability (with concomitant improvements for sweep and storage efficiency). Further control of reservoir sweep attends brine extraction from reservoirs, with benefit for pressure control, mitigation of reservoir and wellbore damage, and water use. State-of-the-art validated models predict the extent of damage and deformation associated with pore pressure hazards in reservoirs, timing and location of networks of fractures, and development of localized leakage pathways. Experimentally validated geomechanics models show where wellbore failure is likely to occur during injection, and efficiency of repair methods. Use of heterogeneity as a control measure includes where best to inject, and where to avoid attempts at storage. Lastly, an example is use of waste zones or leaky seals to both reduce pore pressure hazards and enhance residual CO 2 trapping.

58 GEOSCIENCES↗

Time-lapse seismic data inversion for estimating reservoir parameters using deep learning

Geologic carbon sequestration involves the injection of captured carbon dioxide ([Formula: see text]) into subsurface formations for long-term storage. The movement and fate of the injected [Formula: see text] plume is of great concern to regulators because monitoring helps to identify potential leakage zones and determines the possibility of safe long-term storage. To address this concern, we design a deep-learning framework for [Formula: see text] saturation monitoring to determine the geologic controls on the storage of the injected [Formula: see text]. We use different combinations of porosities and permeabilities for a given reservoir to generate saturation and velocity models. We train the deep-learning model with a few time-lapse seismic images and their corresponding changes in saturation values for a particular [Formula: see text] injection site. The deep-learning model learns the mapping from the change in the time-lapse seismic response to the change in [Formula: see text] saturation during the training phase. We then apply the trained model to data sets comprising different time-lapse seismic image slices (corresponding to different time instances) generated using different porosity and permeability distributions that are not part of the training to estimate the [Formula: see text] saturation values along with the plume extent. Our algorithm provides a deep-learning assisted framework for the direct estimation of [Formula: see text] saturation values and plume migration in heterogeneous formations using the time-lapse seismic data. Our method improves the efficiency of time-lapse inversion by streamlining the large number of intermediate steps in the conventional time-lapse inversion workflow. This method also helps to incorporate the geologic uncertainty for a given reservoir by accounting for the statistical distribution of porosity and permeability during the training phase. Tests on different examples verify the effectiveness of our approach.

Geochemistry & Geophysics↗

Probabilistic Geomechanical Analysis of Compartmentalization at the Snøhvit CO 2 Sequestration Project

Pressure build-up caused by large-scale CO 2 injection is one of the key concerns during a carbon sequestration project. Large overpressures can potentially compromise seal integrity, reactivate faults, and induce concerning seismicity, among other hazards. Furthermore, pressure build-up is directly related with storage capacity. In this work we study the geomechanical response to CO 2 injection in the Tubåen Fm at Snøhvit, with the objective of understanding the potential for fault reactivation, leakage, and contamination of the producing interval in an adjacent block through the bounding faults. Furthermore, we evaluate the potential contribution of a structural component to reservoir compartmentalization. We combine simplified analytical models, based on critically stressed fracture theory and a Mohr-Coulomb failure criterion, with a quantitative risk assessment methodology to perform a rigorous uncertainty quantification. Large stress uncertainties are present and are reflected in the modeling results. It was found that under the “most likely” stress state (Strike-Slip and North-South maximum horizontal stress (S Hmax ) azimuth) the faults are fairly stable, and tensile hydrofracturing of the caprock would be expected before fault reactivation. In most of the analyzed cases, the critical pressure perturbation needed for reactivation is above 13 MPa, which was the limiting pressure increase before reaching the fracture pressure. Faults were found to be ~ 20% less stable when considering plausible variations in the orientation of S Hmax . In those cases, fault reactivation could be expected before caprock failure if injection continued, posing risks of induced seismicity, CO 2 leakage and/or contamination of the gas producing reservoir. Finally, a potential structural component in the compartmentalization and fluid migration is difficult to assess due to stress orientation uncertainty.

54 ENVIRONMENTAL SCIENCES↗

Wabash CarbonSAFE. Geologic Analysis of the Potosi Dolomite Reservoir Interval and Potential Confining Units, at the Wabash #1 Well, Vigo Co., Indiana, Subtask 7.2, Technical Report

The Wabash CarbonSAFE project drilled the Wabash #1 stratigraphic test well (ID# 168045) at the Wabash Valley Resources (WVR) IGCC facility in Vigo County, Indiana, to evaluate the feasibility of commercial-scale CO 2 storage near the site. The Mt. Simon Sandstone was the initial target for storage evaluation and was found to have generally poor reservoir qualities in the Wabash #1 well. Focus was then placed on the secondary target, the Potosi Dolomite. For storage and containment evaluation an extensive suite of geophysical logs was collected throughout the borehole, and whole cores were obtained from the Mt. Simon Sandstone, Eau Claire Formation, and Maquoketa Group. The Maquoketa Group is identified as a regionally extensive secondary sealing interval above the Mt. Simon Sandstone and as a primary seal for the Potosi Dolomite storage complex. Due to concerns over potential lost circulation zones in the Potosi Dolomite during drilling, no core samples or Formation Micro Imager (FMI) logs were acquired in the Potosi Dolomite. However, cased hole well testing was performed over a 20 feet (6 m) interval within the Potosi Dolomite during which a fluid swab sample was obtained. Approximately 35 miles of 2D seismic information was acquired in the project area near Wabash #1 to aid in reservoir and caprock characterization. The highly porous and permeable zones in the Potosi Dolomite and lower Oneota Dolomite observed within the Wabash #1 well are, for the purpose of this report, considered the same reservoir package and referred to as the Potosi reservoir interval. This report presents a regional depositional context for the Potosi reservoir interval, the overlying confining zones in the Knox Group, and Maquoketa Group primary seal. This context is developed through detailed information obtained from the Wabash #1 well such as lithologic data collected from cuttings and core, geophysical logging, geomechanical analysis of core samples, and well testing and fluid sampling within the Potosi reservoir interval. Information discussed in this report was compiled as a basis for the development of static and dynamic geologic models of the Potosi Dolomite storage complex. The static and dynamic modeling of CO 2 injection in the Potosi reservoir interval are discussed in a separate report (Dessenberger et al., 2022) under the Wabash CarbonSAFE project.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Rates of mineral dissolution under CO 2 storage conditions

Evaluating the potential of a sedimentary basin reservoir to securely store CO 2 benefits from a comprehensive understanding of the geochemical reactions that take place once CO 2 is injected into a formation. In particular, models that predict the transport and reaction of CO 2 within a reservoir require a definition of the types of reactions affected by enhanced levels of CO 2 and how the kinetics of these reactions will affect a heterogeneous mineralogy and formation waters within a reservoir over time. In this review we evaluate rate models used to describe mineral dissolution kinetics and compare the range in values reported for the kinetic parameters used to describe the reactivity of various minerals relevant to mainly siliciclastic reservoirs. Parameters that have a significant impact on model results include the reactive surface area of a mineral, the apparent activation energies used to extrapolate reaction rates to the temperatures of potential storage reservoirs (c. 50–125 °C) and the in-situ pH of formation waters with elevated concentrations of dissolved CO 2 . The variation in reported values for these parameters can lead to predicted rates that span many orders of magnitude for a given mineral. Despite these uncertainties recent success with geochemical models has been made by applying a Monte Carlo approach to optimise the kinetic parameters for minerals where robust thermodynamic and kinetic data do not exist.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Proppant embedment in coal and shale: Impacts of stress hardening and sorption

During methane production in CBM reservoirs, the influence of proppant embedment and permeability damage cannot be neglected – especially where the wall-rock is soft. Effective stresses are elevated during methane recovery, increasing both normal loading stress and confinement and simultaneously overprinting sorption-induced volumetric strains. Experiments and analytic modeling are conducted to define key mechanisms controlling these competitive effects. We independently measure overall sample compaction (external LVDT) and local strain (strain gauge) in the matrix to deconvolve proppant embedment in a propped fracture for different conditions of confining stress. The results show symptomatic behaviors of elastic (shale) and elastoplastic (coal) responses of embedment. Different from shale, the evolution of embedment is convex upwards with increased stress where indented depth increases more rapidly as loading stress increases under constant confinement. In addition, a stress-hardening effect is found to play a pivotal role in determining the characteristics of indentation, which are examined in terms of evolution profiles, deformation regimes, embedment slopes, curvatures, yield points and irreversible indentations. Based on the experimental observations a semianalytical model predicts indentation and the evolution of propped permeability under recreated in-situ stress conditions. A simplified case study is conducted to further illustrate the evolution of aperture and permeability of a propped fracture in CBM reservoirs. The modeling results suggest that proppant embedment is significantly overestimated if the variable stress-hardening (VSH) effect is neglected, especially when effective stress is large. Moreover, a decrease in indentation depth possibly occurs during late stage methane production, resulting in a reversal/recovery in fracture closure. This is because desorption-induced shrinkage becomes the predominant effect, causing an increase in aperture and a reduction in the indented volume of proppant. The resulting recovery in permeability implies that the propped coal fracture has the potential to optimally facilitate methane production as a pathway, even at high closure stresses generated by methane drainage.

01 COAL, LIGNITE, AND PEAT↗

A multi-scale experimental investigation of flow properties in coarse-grained hydrate reservoirs during production (Final Scientific/Technical Report)

In this project, “A multi-scale experimental investigation of flow properties in coarse-grained hydrate reservoirs during production” we explored the relative permeability and dissipation behavior of hydrate-bearing coarse-grained sediment at the core scale and we observed the formation and dissociation of these materials at the pore scale with micro-CT and Raman experiments. The 3-year project results inform reservoir simulation efforts, critical to determining the viability of the coarse-grained hydrate reservoir as an energy resource. We performed the first 3 phase relative permeability experiments in a hydrate-bearing medium using the steady state method (gas and water flow in the presence of hydrate). Results support a model where water is the most wetting phase, hydrate is the intermediate wetting phase, and gas is the least wetting phase. This model can be used to predict relative permeability of gas and water in the presence of hydrate. We showed that when hydrate dissociates, it does so at a pressure and temperature predicted for fresh water (no salinity) conditions. At the scale of a grid block, reservoir simulation models should assume that hydrate dissociates at the freshwater phase boundary. We observed that the chemistry and the pore habit of methane hydrates change dramatically over the timescales of experiments (hours to weeks). It takes weeks to months for hydrates to evolve from a non-stoichiometric to stoichiometric compound with 3:1 large cage vs small cage occupancy. We documented hydrates forming initially in small clayey silt surfaces but gradually concentrated in large pores in sand-sized sediment.

03 NATURAL GAS↗

Hydrate Evolution in Response to Ongoing Environmental Shifts

Natural gas hydrates have the potential to become a vital domestic clean-burning energy source. However, past changes in environmental conditions have caused hydrates to become unstable and trigger both massive submarine landslides and the development of crater-like pockmarks, thereby releasing methane into the overlying seawater and atmosphere, where it acts as a powerful greenhouse gas. This project was designed to fill critical gaps in our understanding of domestic hydrate resources and improve forecasts for their response to environmental shifts. Project work can be separated into three interrelated components, each involving the development of predictive mathematical models. The first project component concerns the role of sediment properties on the development and dissociation of concentrated hydrate anomalies. To this end, we developed numerical models to predict equilibrium solubility of methane in twophase equilibrium with hydrate as a function of measureable porous medium characteristics. The second project component concerned the evolution of hydrate distribution in heterogeneous reservoirs. To this end, we developed numerical models to predict the growth and decay of anomalies in representative physical environments. The third project component concerned the stability of hydrate-bearing slopes under changing environmental conditions. To this end, we developed numerical treatments of pore pressure evolution and consolidation, then used "infinite-slope" analysis to approximate the landslide potential in representative physical environments, and developed a "rate-and-state" frictional formulation to assess the stability of finite slip patches that are hypothesized to develop in response to the dissociation of hydrate anomalies. The increased predictive capabilities that result from this work provide a framework for interpreting field observations of hydrate anomalies in terms of the history of environmental forcing that led to their development. Moreover, by taking explicit account of anomaly dissociation, project results are designed to help improve forecasts for changes in slope stability that could pose significant threats to energy infrastructure, disrupt hydrate reserves, and pollute the atmosphere with vast quantities of methane. This report presents the details of our work and outlines some of the highlights from our findings.

03 NATURAL GAS↗

Predicting the potential for mineral scale precipitation in unconventional reservoirs due to fluid-rock and fluid mixing geochemical reactions

Mineral precipitation within hydraulically fractured shale may affect fluid flow pathways and impact longterm hydrocarbon production. The ability to predict geochemical reactions that lead to problematic mineral precipitation will lead to active reservoir management strategies to improve overall production. Using the Marcellus Shale as a case study, a combination of laboratory experiments and reaction path modeling was applied to determine which reactions are likely to occur upon introduction of hydraulic fracturing fluid into the shale reservoir. Experimental results indicate that contact between fracturing fluid and shale will result in dissolution of primary minerals (quartz, feldspars, kaolinite, chlorite, pyrite) and secondary mineral precipitation over time periods of less than one week. Precipitation of barite, Fe-oxides, feldspars, amorphous silica and clay is likely to occur within the reservoir during shut in and early flowback due to mixing between fracturing fluid and reservoir brine as based on modeling saturation indices using experimental fluid data. Reaction path modeling of the experimental scenarios corroborates the specific dissolution and precipitation reactions observed experimentally. Comparison of the results to injected and produced waters from a Marcellus Shale well pad in Greene County, PA, USA, shows that the mineral reactions occur during the hydraulic fracturing, shut in, and early flowback periods. The results presented in this paper demonstrate the value in applying experimental approaches to identify mineral precipitation/dissolution reactions that may significantly impact reservoir performance. The good agreement between geochemical models and experimental results provides confidence that numerical models can be applied to screen the potential fluid-mineral and fluid-mixing reactions in unconventional reservoirs that result in undesired mineral scale precipitation.

58 GEOSCIENCES↗

A Distributed Modeling System for Short-Term to Seasonal Ensemble Streamflow Forecasting in Snowmelt Dominated Basins

This paper describes a distributed modeling system for short-term to seasonal water supply forecasts with the ability to utilize remotely-sensed snow cover products and real-time streamflow measurements. Spatial variability in basin characteristics and meteorology is represented using a raster-based computational grid. Canopy interception, snow accumulation and melt, and simplified soil water movement are simulated in each computational unit. The model is run at a daily time step with surface runoff and subsurface flow aggregated at the basin scale. This approach allows the model to be updated with spatial snow cover and measured streamflow using an Ensemble Kalman-based data assimilation strategy that accounts for uncertainty in weather forecasts, model parameters, and observations used for updating. Model inflow forecasts for the Dworshak Reservoir in northern Idaho are compared to observations and to April-July volumetric forecasts issued by the Natural Resource Conservation Service (NRCS) for Water Years 2000 – 2006. October 1 volumetric forecasts are superior to those issued by the NRCS, while March 1 forecasts are comparable. The ensemble spread brackets the observed April-July volumetric inflows in all years. Short-term (one and three day) forecasts also show excellent agreement with observations.

54 ENVIRONMENTAL SCIENCES↗

Scaling Behavior of Thermally Driven Fractures in Deep Low‐Permeability Formations: A Plane Strain Model With 1‐D Heat Conduction

Abstract Injection of cold fluids through/into deep formations may cause significant cooling, thermal stress, and possible thermal fracturing. In this study, the thermal fracturing of low‐permeability formations under one‐dimensional heat conduction was investigated using a plane strain model. Dimensionless governing equations, with dimensionless fracture length , aperture , spacing , time , and effective confining stress , were derived. Solution of single thermal fracture was derived analytically, while solution of multiple fractures with constant (or dynamic) spacing were obtained using the displacement discontinuity method (and stability analysis). For single fracture, increases nonlinearly with and then transitions to scaling law , indicating that late‐time fracture length increases linearly with the square root of cooling time. For constantly spaced fractures, deviates from the single‐fracture solution at a later for a larger , showing slower propagation under inter‐fracture stress interaction . For dynamically spaced fractures, fracture arrest induced by stress interaction was determined by the stability analysis; the fully transient solution provides evolution of dimensionless fracture length, spacing, aperture, and pattern; a similar scaling law, with , obtained shows the effect of both stress interaction and fracture arrest. The solution and scaling law provide fast predictions for all reservoir and cooling conditions using (single) model parameter . Application to a geothermal site with demonstrates that thermal fractures reach 0.67, 6.25, and 78.00 m in length, 0.49, 2.30, and 13.00 m in spacing, and 0.43, 2.09, and 12.19 mm in aperture at 1, 100, and 10,000 days.

Chen, Bin↗

Application of machine learning to characterize gas hydrate reservoirs in Mackenzie Delta (Canada) and on the Alaska north slope (USA)

Here, artificial neural network-trained models were used to predict gas hydrate saturation distributions in permafrost-associated deposits in the Eileen Gas Hydrate Trend on the Alaska North Slope (ANS), USA and at the Mallik research site in the Beaufort-Mackenzie Basin, Northwest Territories, Canada. The database of Logging-While-Drilling (LWD) and wireline logs collected at five wells (Mount Elbert, Ignik Sikumi, and Kuparuk 7–11–12 wells at ANS, plus 2L-38 and 5L-38 wells at the Mallik research site) includes more than 10,000 depth points, which were used for training, validation, and testing the machine learning (ML) models. Data used in training the ML models include the well logs of density, porosity, electrical resistivity, gamma radiation, and acoustic wave velocity measurements. Combinations of two or three out of these five well logs were found to reliably predict the gas hydrate saturation with accuracy varying between 80 and 90% when compared to the gas hydrate saturations derived from Nuclear Magnetic Resonance (NMR)-based technique. The ML models trained on data from three ANS wells achieved high fidelity predictions of gas hydrate saturation at the Mallik site. The results obtained in this study indicate that ML models trained on data from one geological basin can successfully predict key reservoir parameters for permafrost-associated gas hydrate accumulations within another basin. A generalized approach for selecting a well log combination that can improve model accuracy is discussed. Overall, the study outcome supports earlier work demonstrating that ML models trained on non-NMR well logs are a viable alternative to physics-driven methods for predicting gas hydrate saturations.

58 GEOSCIENCES↗